initn betas 9B meeting mur berionera at rit os = : on ae thy aes een S Spree » a NN TRANSACTIONS OF THE ZOOLOGICAL SOCIETY OF LONDON. VOLUME XV. LONDON: PRINTED FOR THE SOCIETY : SOLD AT THEIR HOUSE IN HANOVER-SQUARE ; AND BY MESSRS. LONGMANS, GREEN, AND CO., PATERNOSTER-ROW. 1901. PRINTED BY TAYLOR AND FRANCIS. RED LION COURT, FLEET STRERT. CONTENTS. I. Report on the Collection of Fishes made by Mr. J. E. 8. Moore in Lake Tanganyika during his Expedition, 1895-96. By G. A. BouLencsr, F.RS., F.Z.S. With an Appendix by J. E. 8. Moorz, A.R.C.S. (Plates I.-VIII.) . . . . page 1 Il. On the Marine Copepoda of New Zealand. By G. Srnwarpson Brapy, M.D., D.Sc., PIGS) habeas nC MVMEZiS cau (lates) Xe Xels) se an) ee seme oll III. On the Extinct Birds of Patagonia.—1. The Skull and Skeleton of Phororhacos inflatus Ameghino. By C. W. Anpruws, B.Sc., F.Z.S., Assistant in the British Museum (Natural History). (Plates XIV.-XVII.) ...... =. =. 98d IV. Second Contribution to the Ichthyology of Lake Tanganyika.—On the Fishes obtained by the Congo Free State Expedition under Ineut. Lemaire in 1898. By G. A. Boutencmr, /.#.S., F.Z.S. (Plates XVITI-XX.) . ..... 8 V. On a Specimen of Cervus beigrandi Lart. (C. verticornis Dawk.) from the Forest- Bed of East Anglia. By Stoney F. Harmer, Sc.D., FRS., Superintendent of the University Museum of Zoology, Cambridge. (Plate XXI.) . . . . 97 VI. A Monograph of the Genus Casuarius. By the Hon. Waurer Roruscuip, Ph.D., F.ZS. Witha Dissertation onthe Morphology and Phylogeny of the Paleognathe (Ratitee and Crypturi) and Neognathe (Carinate). By W.P. Pycrarr. (Plates ONDE OUNTS)) Gy Boe tee ne cee eee eee 5 1(OKY VII. Notes on the Transformations of some South-African Lepidoptera. By Lneut.- Colonel J. Maucotm Fawcett. (Plates XLVI-XLIX.) .... . . 291 VILL. On some Points in the Anatomy of Polypterus. By J. S. Buperrr, WA., F.Z.S., Trinity College, Cambridge. (Plates L.-LIL) . ....... . . 828 Histroimcthesbapersicontained sim Vole XV.) 9. ai 2) ee ener ooo ING EXSOTASPECLES te CCOMM ne ae olen Mira, 6. 4 Gul le Uae ey sete: sa Caen 4 TRANSACTIONS OF THE ZOOLOGICAL SOCIETY OF LONDON. Vout. XV.—Parr 1. LONDON: | PRINTED FOR THE SOCIETY, SOLD AT THEIR HOUSE IN HANOVER-SQUARE ; AND BY MESSRS. LONGMANS, GREEN, AND CO., PATERNOSTER-ROW. Devember 1898. Price 21s. Taylor and Francis, Printers, | [Red Lion Court, Fleet Street. TRANSACTIONS OF THE ZOOLOGICAL SOCIETY OF LONDON. To Fellows. To the Public. HS. 1d. Be Se Oe VOLUME J. (1888-1835, containing 59 Plates). . Price 313 6 . . . 418 O* VOLUME II. (1835-1841, containing 71 Plates). . ,, 4 0 O Raye oe VOLUME III. (1842-1849, containing 63 ave Stl pe Ore Ow Oakey seman eetemlal OE VOLUME IY. (1851-1862, containing 77 Plates). . ,, 6 2 0 822) 464, VOLUME _ V. (1862-1866, containing 67 Plates). . , 5 4 8 .. . 619 OF VOLUME VI. (1866-1869, containing 92 Plates). . ,, 11 5 O 15 0 OF VOLUME VII. (1869-1872, containg 73 Plates). . ,, 10 4 0 . . . 13812 OF VOLUME VIII. (1872-1874, containing 82 Plates). . , 9 8 3... .1211 OF VOLUME IX. (1875-1877, containing 99 Plates). . ,, 12 16... .16 2 OF VOLUME X. (1877-1879, containing 95 Plates). . ,, 10 0 3 . . .18 7 OF GENERAL INDEX, Vols. I. to X. (1833-1879) .,, We) ices Taiememe sts (0) NOY 0) VOLUME XI. (1880-1885, containing 97 Plates). . , 912 0 .. .1216 O VOLUME XII. (1886-1890, containing 65 Plates). . ,, 5 Gas O Mes Deer eA * No copies of these volumes in stock. tT Only complete copies of these volumes left in stock. Continued on page 3 of Wrapper. TRANSACTIONS OF THE ZOOLOGICAL SOCIETY OF LONDON. I. Report on the Collection of Fishes made by Mr. J. E. S. Moore in Lake Tanganyika during his Expedition, 1895-96. By G. A. Bounencnr, F.R.S., ZS. With an Appendix by J. E. S. Moors, A.R.C.S. Received and read June 21, 1898, [Puates I.-VIIT.] THE examination of the collection of fishes brought home from Lake Tanganyika by Mr. Moore, and with which I have been entrusted by him, has not yielded any such startling results as have already been announced after study of the Invertebrates! Of the two series in the fauna of Tanganyika, distinguished by its explorer as the normal freshwater and the halolimnic groups?, the latter is unrepresented in the collection. This may be due either to the origin of the present fish-fauna not reaching so far back in time as that of the Mollusca and other Invertebrates, or to the incompleteness of the series brought home. ‘The latter explanation may be the correct one, since, owing to the lack of suitable dredging-apparatus, and to the difficulties of preserving, to which Mr. Moore has alluded ?, only a small proportion of the fishes of the Lake could be collected, mostly littoral forms found about the rocks; of the deeper-water fishes, which were observed to be abundant both in species and in individuals, we are still almost entirely ignorant; it is therefore to be expected that further collections, 1 J, E.S. Moore, “On the Zoological Evidence for the Connection of Lake Tanganyika with the Sea,” Proc. R. Soe. lxii. 1898, pp. 451-458. 2 J. E. S. Moore, ‘“‘ The Molluses of the Great African Lakes,” Q. Journ. Micr, Sci. 1898, pp. 159-204, 3 L.c. See also‘‘The Physiographical Features of the Nyasa and Tanganyika Districts,’ Geogr. Journ 1897. VoL. XV.—PART I. No. 1.—Deceniber, 1898. B 2 MR. G. A. BOULENGER ON A COLLECTION made under more favourable circumstances, would considerably modify our views on the general character of this most important section of the fauna. How incomplete the collection is may be partly realized from the fact that, of the six species previously described from Tanganyika, only one was rediscovered by Mr. Moore. His series of sketches executed from fresh specimens, free use of which he has kindly given me, thus enabling me to represent some of the new species in their natural colours, also indicate the existence of several fishes which are unrepresented in the collection. Large fishes, owing to the impossibility of preserving them, had to be left behind, and the difficulties of transport by carriers resulted in the loss of several jars containing spirit-specimens. Nevertheless, small as it is, and though deficient in any but typically African freshwater forms, the collection is of great interest, and Mr. Moore well deserves the thanks of all zoologists for the manner in which, amid so many difficulties, he has succeeded in affording us a first glimpse at a fish-fauna which has so long remained a mystery. A study of the freshwater fishes of Africa has hitherto led to the assumption that the bulk of the fauna originated in the region of the great lakes, whence they have radiated towards the Mediterranean and the Atlantic and Indian Oceans—a view based on the close affinity, often amounting to specific identity, of the fishes of the Nile, the Niger, the Congo, and the Zambesi. The homogeneity of the fauna is absolutely opposed to the conception of the great watersheds having been stocked from the sea, within the geological period of which these fishes are representative, this being evident even in the case of such forms as Zates and Tilapia, which are known to enter salt water. Nothing contrary to this theory is brought to light by an examination of the fishes obtained by Mr. Moore in Lake Tanganyika. The striking character of the Tanganyika fish-fauna, as revealed by Mr. Moore’s collection, is the extraordinary variety of the Cichliidw). This is a natural group distributed all over Africa, including Madagascar, but, although rich in species, nowhere else showing within a limited area anything like the modifications of structure described in this report, which have necessitated the establishment of nearly as many, new genera as were previously known from the whole of Africa. The generalized characters of some of the Cichlidw occurring in Lake Tanganyika, regarding as indicative of generalization the greater development of the anal fin, consistently with the system followed in classifying their more primitive allies the Centrarchidw?, and the extent of the lateral lines, both of which are complete in some of the newly-discovered forms®, afford further support to the 1 See P. Z. 8. 1898, p. 182. * Cat. Fish. 2nd ed. 1. p. 2 (1895). $ The lateral line has usually been described as “ interrupted” in the Cichlidw and other groups in which it has a similar disposition. ‘This expression conveys a serious misconception, and I have replaced it in the diagnoses of the Cichlide and the Serranide allied to Plesiops by “ lateral lines two,” the necessity for this change being particularly obvious in the case of some of the species described in this report. Fishes had no doubt originally a greater number of lines of sense-organs along the body, as we may still observe in many of OF FISHES FROM LAKE TANGANYIKA. 8 proposition enunciated above. Lake Tanganyika might be regarded as the centre of origin of all the African genera of Cichlide. Unfortunately, nothing reliable is yet known of this family in a fossil condition, with the exception of a few American Tertiary types, which do not differ materially from those inhabiting the same part of the world at the present time. There is no reason for referring to it the Cretaceous fishes from Mount Lebanon described as “ Chromides”’ by Heckel! and by Davis. All we can gather from Heckel’s description is that Pycnosterynz is based on some ‘ Pharyngognath Acanthopterygian,” and that it does not belong to the Cichlide is conclusively proved by his statement: “ Rippen kurz, dtinn, die hinteren auf langen Querfortsitzen ansitzend,’ a sentence that has been translated by Davis as “ Ribs short and slender, the posterior ones supporting long transverse apophyses”*. An examination of some of the specimens of Pycnosteryna’ exhibited in the British Museum suggests to me special relation to the Berycide, with which they are provisionally associated by Mr. Smith Woodward, and certainly no affinity to the Cichlide. Considering that coalesced lower pharyngeal bones occur at the present day in such widely different groups as the Serramde, Gerride, Scienide, Cichlide, Pomacentride, Embiotocide, Labride, Scaride, Pleuronectide, and Scombre- socidc, this character, if correctly ascertained by Heckel, would in itself be no serious objection to the allocation of Pycnosterynz among the Berycide. Giinther* also doubts the correctness of Heckel’s determination when he states that “the position of Pycnosterynx is uncertain,’ but we are not enlightened as to its affinities by the further remark that “it approaches certain Pharyngognaths.” It appears certain that true Perciform fishes, to which large group the Cichlidw belong, have not yet been described from pre-Tertiary beds. In order to show what is now known of the fish-faunas of the great lakes, lists are appended of the fishes of Lakes Nyassa, Tanganyika, Victoria Nyanza, and Rudolf. These lists must of course be taken as giving a very inadequate idea of the fauna, owing to the incompleteness of the collections on which they are based; but, imperfect as they are, they nevertheless will prove useful as a basis from which to judge of the relation existing between the piscine inhabitants of these lakes. Of the other lakes, unfortunately, nothing can be said at present, the occurrence of an undetermined species of Haplochilus in the Albert Nyanza being all the information we possess. the lower forms. In the Teleosteans, as a general rule, the lateral lines are reduced to one on each side, extending uninterrupted from the head to the root of the caudal fin, although among this order there is hardly a single large group that does not offer exceptions such as are shown in the Cichlidw. In these, both upper and lower lateral lines may be complete ; or the upper may be complete and the lower reduced, or the reverse; or the upper may be incomplete posteriorly and the lower anteriorly, which latter disposition has led to the misconception involved in the term “ interrupted.” 1 Abbild. u. Beschr. n. Thier. Syrien, p. 235 (1843). * Tr. R. Dublin Soe. (2) iii. 1888, p. 531. 2 1D, @ i BBL * «Study of Fishes,’ p. 200 (1880). BZ 4 MR. G. A. BOULENGER ON A COLLECTION I. Lake Nyassa}, CicHLIDz. SILURIDA. 1. Paratilapia robusta Gthy. 24. Bagrus meridionalis Gthr. 2. fs afra Gthr. 25. Anoplopterus platychir Gthr. 3. ie modesta Gthr. 26. Synodontis zambesensis Gthyr. 4., 5) livingstonit Gthr, De BS intermedia Gthr, Coro 6. a dimidiata Gthr. 7. x longiceps Gthy. 27, Labeo mesops Gthr. 8. Corematodus shiranus Bley. 28. Barbus trimaculatus Ptrs. 9. Tilapia shirana Blgr. 29. Barilius guentheri Blgr. TO, 45 qoseomlibon ins. 30. Engraulicypris pinguis Gthy, 11. ,, kirkii Gthr. 31. Pelotrophus microlepis Gthr. 12. 4, == squamipinnis Gthr, 82. 2 microcephalus Gthr. 13. 4, vrendalli Bigr. 14. 4, ~~‘ dateristriga Gthr. @HARACINIDE: 15. 4, subocularis Gthr. 16. 4, johnstont Gthr. 17. ~=4, ~——‘lethrinus Gthr. 18. ,, — ¢tetrastigma Gthr. CyPRINoDONTIDs. 19. 4, ~ callipterus Gthr. 34. Haplochilus johnstoni Gthr. 20. 4, williamsi Gthr. 21. 4, aurata Bigr. 22. Docimodus johnstoni Blgr. 33. Alestes imberi Ptrs. MorMyRID&. 35. Mormyrus discorhynchus Ptyrs. MAstTAacEMBELIDA. 36. 35 catostoma Gthr. 23. Mastacembelus shiranus Gthr. 37. Mormyrops zambanenje Ptrs. Il. Lake Tanganyika ?. SERRANIDZ. 9. Telmatochromis temporalis Blgr. 1, Lates microlepis Blgy. 10. Julidochromis ornatus Blgy. ll. Paratilapia pfefferi Blgr. CicHLIps. 12. sy macrops Blgr. 2. Lamprologus fasciatus Blgr. 13. aa ventralis Blgr. 3. $ compressiceps Blgy. 14, 53 furcifer Blgr. 4. i moori Blgr. Nie i leptosoma Blgy. 5. 90 modestus Blgr. 16. Ectodus descampsii Blgr. 6 = elongatus Blgr. 17. 4, + melanogenys Bigr. 7 os furcifer Blgr. 18. Bathybates ferox Blgr. 8. Telmatochromis vittatus Blgr. 19. Eretmodus cyanostictus Blgr. ’ Of. Giinther, P. Z. 8. 1864, p. 307, 1898, p. 616, and Ann, & Mag. N. H. (6) xvii. 1896, p.397; Boulenger P. Z.8. 1896, p. 915, and Ann. & Mag. N. H. (6) xix. 1897, p. 155, and (7) i. 1898, p. 254. * Cf. Giinther, P. Z. 8. 1893, p. 628. OF FISHES FROM LAKE TANGANYIKA, Cicuiip# (continued). 20. Tilapia tanganice Gthr. al, » Ourtoni Gthr. 22. a5 labiata Blgr. 23. Tropheus moorii Bley. 24, Simochromis diagramma Gthy. 25. Petrochromis polyodon Blgr. 26. Perissodus microlepis Bley. 27. Plecodus paradozus Bley. MasraceMBrLipa. 28. Mastacembelus moorii Blgr. 29. B tanganice Gthr, 30. 5 ophidium Gthr. SILuRIDA, 31. Clarias anguillaris L. 32. ,, ~—_ liocephalus Blgr. 33. Anoplopterus platychir Gthr. 34, Auchenaspis biscutata Geoftr. 35. Synodontis multipunctatus Blgr. 36. Malapterurus electricus Gm, CyPRINID&. 37. Labeo, sp. CHARACINIDA, 38. Alestes macrolepidotus C. & V. 39. ,, = macrophthalmus Gthr. 40. Hydrocyon forskalit Cuv. CyPRINODONTID&. 41. Haplochilus tanganicanus Blgr. PoLyPrERID&m. 42. Polypterus bichir Geoffr, (?). LEPIDOSIRENIDA, 43. Protopterus annectens Ow.?’. Ill. Lake Victoria Nyanza?. Cicuiipe. - Paratilapia longirostris Hilg. 99 cavifrons Hilg. % retrodens Hilg. . Tilapia nilotica Cuv. » nuchisquamulata Hilg. » Ssaduvagii Pfeff,' » obliquidens Hilg. ow GD) £8 E> $9 8S IS MastaceMBELIDA, 8. Mastacembelus, sp. SILURIDZ. 9. Clarias, sp. 10. Synodontis afrofischeri Hilg. CyprinIp@. ll. Labeo forskalii Riipp. 12. Labeo rueppellit Pfeff. 13. Barbus pagenstecheri Fisch. 14, 4, trimaculatus Ptrs. CHARACINIDA, 15. Alestes rueppellit Gthr. CyYPRINODONTIDA. 16. Fundulus teniopygus Hilg. Mormynripa&, 17. Mormyrus oxyrhynchus Geoftr. 18, fe longibarbis Hilg. LeEpIDUSIRENIDE, 19. Protopterus annectens Ow. ' Reported from Lake Tanganyika by Sir H. H. Johnston, Brit. C. Afr. p. 362 (1897). 2 Hilgendorf, Sitzb. Ges. naturf. Fr. Berl. 1888, p. 75; Pfeffer, Thierw. O.-Afr., Fische (1896); Giinther, Ann. & Mag. N. H. (6) xvii. 1896, p. 397. 6 MR. G. A. BOULENGER ON A COLLECTION IV. Lake Rudolf. CicH1iip2. CHARACINIDA. 1. Tilapia DURE Cae 6. Citharinus geoffroyi Cuv. 2. 5, tristrami Gthr. 7. Alestes rueppellii Gthr. SILURIDE. 8. Distichodus rudolphi Gthy. 3. Synodontis schal Bl. Schn. : 4, ” smithit Gthr. PouyPrEeRIp”®. Cyprinip&. 9. Polypterus senegalus Cuy. 5. Barbus, sp. The fishes of Lake Nyassa are, with two exceptions, specifically distinct from those of the Nile, as pointed out by Dr. Giinther ?, while seven out of forty-three species represented in Lake Tanganyika occur both in the Nile and in the rivers flowing into the Atlantic. And as the Mormyride, which furnish the two species common to Nyassa and the Nile®, have not yet been recorded from Tanganyika, while, with the exception of a small stream-Siluroid, not one of the species described from the former lake has been rediscovered in the latter, it follows that, although similar in general character, the fish-fauna of the two lakes shows no trace of community so far as specific forms are concerned, as might have been expected from the absence of direct communication between them. Before concluding these prefatory remarks, I wish to express my thanks to Messrs. J. H. Gardiner and J. Green for their kind assistance in supplying me with sciagraphs of the new fishes, which, supplementing the skeletons prepared by Mr. Groenvold, have enabled me to add some notes on the osteological characters of the genera described as new. SERRANIDA. 1. Laves MICROLEPIS, sp. n. (Plate I. fig. 1.) Body elongate, its depth 33 times in the total length. Length of head 3 times in total length; upper profile nearly straight ; diameter of eye equal to length of snout, 33 times in length of head; lower jaw projecting ; maxillary extending to below centre of eye, the width of its distal extremity not quite half diameter of eye; pree- and suborbitals finely serrated; cheeks, opercles, -and occiput covered with small scales ; preopercular border forming nearly a right angle, finely toothed on its vertical limb, with two or three widely-separated spines on its lower limb, and with one or two very 1 Giinther, P. Z. 8. 1896, p. 217. > Study of Fishes,’ p. 230. 3 Tf, as seems probable, the distinction between Mormyrops zambanenje and M. anguilloides should not be maintained. On the other hand, the Nilotic specimens referred to Mormyrus discorhynchus may prove to be specifically separable. Lake Tanganyika might thus ultimately possess no species of fish in common with the Nile. OF FISHES FROM LAKE TANGANYIKA. U strong spines at the angle; opercular spine as much developed as the latter ; clavicle with a group of 3 or 4 strong spines. 16 gill-rakers on lower part of anterior arch. Dorsal VII, II 11; first and second spines short, third very strong, longest, ? length of head; longest soft rays 3 length of head. Pectoral } length of head. Anal III 8; spines short, second and third equal. Caudal truncate. Caudal peduncle twice as long as deep. Scales 100-110 ee Body silvery, spotted and marbled with brown ; dorsals spotted with brown ; caudal with irregular brown bars. Total length 155 millim. Two young specimens from Kinyamkolo. This species is closely allied to the widely-distributed Lates niloticus Hasselq. Compared to young specimens of the latter, it differs in the smaller scales, the higher spinous dorsal, the longer caudal peduncle, aad the shape of the caudal fin, which is truncate instead of rounded. CICHLIDA. 2. LAMPROLOGUS FASCIATUS, sp. n. (Plate I. fig. 2.) A few moderately large curved canine teeth in front of each jaw, followed by a narrow band of minute teeth; lateral teeth very small. Depth of body 4 times in total length, length of head 3. Snout as long as the diameter of the eye, which is 3 times in length of head and equals 13 interorbital width; maxillary extending to below anterior border of eye; cheeks naked; opercles and occiput scaled. Gill-rakers short, 12 on lower part of anterior arch. Dorsal XIX 8; spines increasing in length to the last, which measures 2 length of head and nearly equals longest soft rays. Pectoral 4 length of head. Ventral reaching vent. Anal X 6; spines increasing in length to the last, which slightly exceeds longest dorsal. Caudal rounded. Caudal peduncle as long as deep. Scales 46 2s lat. 1. — Yellowish, with 11 dark brown bars, the first across the vertex ; fins greyish, dorsal and anal edged with blackish. Total length 70 millim. A single specimen from Kinyamkolo. 3. LAMPROLOGUS COMPRESSICEPS, sp. n. (Plate I. fig. 3.) A few moderately large curved canine teeth in front of each jaw, followed by a narrow band of minute teeth; lateral teeth very small, curved. Depth of body 25 to 23 times in total length, length of head 23? to 23. Head very strongly compressed, with concave upper profile; snout a little longer than diameter of eye, which is 4 times in length of head and equals 14 interorbital width ; maxillary extending to below anterior border of eye; cheeks naked; opercles and occiput with small scales. Gill-rakers moderately long, 15 on lower part of anterior arch. Dorsal XX-XXI 6; 8 MR. G. A. BOULENGER ON A COLLECTION spines increasing in length to the sixth, which measures half length of head, the posterior a little shorter; longest soft rays a little longer than longest spines. Pectoral 4 to 3 length of head. Ventral produced into a filament. Anal X 5; spines increasing in length to the last, which equals the last dorsal ; longest soft rays 2 length of head. Caudal rounded. Caudal peduncle as long as deep. Scales 32-33 2; 12) lat. 1. a . Brown, with indistinct traces of five darker vertical bars; pectoral bright yellow, other fins blackish towards the border. Total length $3 millim. Two specimens from Kinyamkolo. 4, LAMPROLOGUS MOORI, sp. n. (Plate I. fig. 4.) 9 or 10 equal, moderately large conical teeth in front of each jaw, followed by a narrow band of minute teeth; lateral teeth very small. Depth of body 24 times in total length, length of head 3 to 33. Snout as long as diameter of eye, which is 3 to 33 times in length of head and equals interorbital width; maxillary extending to below anterior border of eye; cheeks with small, deciduous scales; opercles and occiput scaled. Gill-rakers short, 9 or 10 on lower part of anterior arch. Dorsal XIX-XX 8-9; spines slightly increasing in length to the last, which measures nearly half length of head; middle soft rays prolonged, at least # length of head. Pectoral 2 to $ length of head. Ventral prolonged into a long filament. Anal VII-VIII 6-7 ; spines increasing in length to the last, which is a little longer than the longest dorsal ; middle soft rays prolonged into filaments. Caudal rounded. Caudal peduncle ; lat. 1. 2°. Dark brown ; fins blackish. 2? 9-13" 99 9 as long as deep. Scales 33-35 Total length 95 millim. Several specimens from Mbity Rocks and Kinyamkolo. 5-7 11-1 5. LAMPRGLOGUS MODESTUS, sp.n. (Plate I. fig. 5.) A few large curved canine teeth, tipped with brown, in front of each jaw, followed by a band of minute teeth; lateral teeth very small. Depth of body 3% to 32 times in total length, length of head 3 to 33. Snout a little longer than diameter of eye, which is 384 to 4 times in length of head and equals interorbital width; maxillary extending to below anterior border of eye; cheeks naked; opercles and occiput scaled. Gill-rakers very short, 7 increasing in length to the last, which is not quite half length of head; middle soft rays prolonged, % to # length of head. Pectoral about $ length of head. Ventral prolonged into a short filament. Anal V 6-7; spines increasing in length to the last, which is as long as middle dorsals; middle soft rays prolonged, like the dorsals. on lower part of anterior arch. Dorsal XX 8-9; spines Caudal truncate. Caudal peduncle as long as deep. Scales 36-40 = lat. 1. 7. OF FISHES FROM LAKE TANGANYIKA. 9 Uniform brown; soft dorsal and caudal fins with round black spots between the rays. Total length 75 millim. A single specimen from Mbity Rocks, and another from Kinyamkolo. 6. LAMPROLOGUS ELONGATUS, sp. n. (Plate I. fig. 6.) 6 to 8 large canine teeth in front of each jaw, followed by a broad band of minute villiform teeth ; lateral teeth very small. Depth of body 4 times in total length, length of head 2% to 2$. Snout twice as long as diameter of eye, which is 5 times in length of head and equals interorbital width ; maxillary extending to below anterior border of eye; cheeks naked; opercles and occiput scaled. Gill-rakers moderately long, 12 on lower part of anterior arch. Dorsal XVIII 10-11; spines slightly increasing in length to the last, which measures 4 length of head; longest soft rays half length of head. Pectoral half length of head. Ventral reaching vent. Anal V 8; spines increasing in length to the last, which equals longest dorsal. Caudal truncate. Caudal peduncle 14 as long as deep. Scales 90-95 aoe) lat. 1. = Brown, with darker spots having a tendency to form cross-bars ; caudal fin with round dark spots between the rays. Total length 115 millim. One specimen from Mbity Rocks, and one from Kinyamkolo. 7. LAMPROLOGUS FURCIFER, sp. n. (Plate II. fig. 1.) A few large curved canine teeth in front of each jaw, followed by a moderately broad band of minute villiform teeth ; lateral teeth very small. Depth of body 4 to 44 times in total length, length of head 2$ to 3. Snout as long as or a little longer than diameter of eye, which is 54 to 3% times in length of head and exceeds interorbital width; maxillary extending to below anterior fourth of eye; cheeks and opercles with deciduous scales. Gill-rakers short, 16 on lower part of anterior arch. Dorsal XX-XXI 7-8; spines increasing in length to the last, which measures > length of head; middle soft rays produced, 3 to ? length of head. Pectoral # length of head. Ventral reaching vent or origin of anal. Anal VI-VII 6; spines increasing in length, the last nearly as long as last dorsal; middle soft rays produced. Caudal deeply notched, crescentic. Caudal peduncle 14 as long as deep. Scales 50-54 ae lat. 1. a lower lateral line often nearly complete. Dark brown, with very indistinct blackish bars on the body; dorsal and caudal with round black spots between the rays; tips of the caudal lobes whitish. Total length 125 millim. Three specimens from Kinyamkolo, and one from Mbity Rocks. VOL. XV.—PaART I. No. 2.—December, 1898. 10 MR. G. A. BOULENGER ON A COLLECTION The genus Lamprologus was known from a single species L. congoensis Schilthuis, discovered a few years ago in the Congo. The species now described raise the number to 7, distinguishable by means of the following key :— I. Caudal rounded or truncate. A. Anal with 10 spines. D. XIX 8; Sc. 46 a depth of body 4 times in total length. . . . . . 1. L. fasciatus. D. XX-XXI 6; Se. 32-33 & ; depth of body 23 to 2? times in total length. 2. L. compressiceps. B. Anal with 6 to 8 spines. D. XIX-XX 8-9; Sc. 33-35 © ; depth of body 23 times in total length . 3, L. moorit. 11-12’ D. XVITI-XIX 8-10; Se. 42-53 wats depth of body 3} to 4 times in total length. 4. L. congoensis. C. Anal with 5 spines. D. XX 8-9; Sc. 36-40 56. depth of body 3} to 32 times in total length. 5. L. modestus. 11-14” D. XVIII 10-11; Se. 90-95 sat depth of body 4 times in total length . 6. L. elongatus. II. Caudal deeply notched, crescentic. : D, XX-XKI 7-8; A. VI-VII6; Sc.50-54 77. 2 2 2 1 1 1. . 7. LL. furcifer. TELMATOCHROMIS, g. 1. Body more or less elongate; scales ctenoid. Jaws with a series of conical teeth, followed by a broad band of minute tricuspid teeth; lateral teeth small, conical. Maxillary exposed. Dorsal with 20 to 22 spines, anal with 6 or 7. Vertebre 33 (16+17). This genus is closely allied to Lamprologus, differing in the small teeth forming a band behind the outer row being tricuspid instead of conical. It therefore stands somewhat in the same relation to Lamprologus as Tilapia does to Paratilapia. 8. TELMATOCHROMIS VITTATUS, sp. n. (Plate II. fig. 2.) 12 to 16 enlarged conical teeth, tipped with brown, in the outer row in each jaw. Depth of body 43 to 4% times in total length, length of head 4. Snout descending in a strong curve, as long as or a little longer than the diameter of the eye, which is 32 to 4 times in length of head and equals interorbital width ; maxillary extending to below the nostril; head naked, opercle with a few deciduous scales. Gill-rakers very short and tew. Dorsal XXI-XXII 8; spines increasing in length to the last, which equals } length of head; soft rays a little longer. Pectoral ¢ length of head. Ventral produced into a short filament, reaching origin of anal. Anal VII 5-6; spines increasing in length to the last, which equals last dorsal. Caudal rounded. Caudal peduncle as long as deep. Scales 45-02 +3 Natale ae Yellowish, with a dark OF FISHES FROM LAKE TANGANYIKA, 11 brown lateral stripe from the upper lip, through the eye, to the base of the caudal, where it expands into a spot; another dark brown stripe from the vertex along the base of the dorsal; a few brown spots on the dorsal; anal edged with dark brown; a black bar at the base of the pectoral, which is white. Total length 78 millim. ‘Two specimens from Mbity Rocks. 9. TELMATOCHROMIS TEMPORALIS, sp. n. (Plate IT. fig. 3.) 8 to 12 enlarged conical teeth, tipped with brown, in the outer row in each jaw. Depth of body 3 to 33 times in total length ; length of head 3 to 3. Snout descending in a strong curve, 14 as long as the diameter of the eye, which is 44 times in length of head and a little less than interorbital width; maxillary extending to below anterior border of eye; head naked, or with a few deciduous scales on the opercles. Gill- rakers very short and few. Dorsal XX-XXI 6-7; spines increasing in length to the last, which equals 3 to } length of head; middle soft rays produced, % to $ length of head. Pectoral 3 length of head. Ventral produced into a filament, extending beyond origin of anal. Anal VI-VII 6-7; spines increasing in length to the last, which equals or slightly exceeds last dorsal; soft rays produced, like the dorsals. Caudal rounded. Caudal peduncle as long as deep. Scales 43-46 <3 lat. 1. om Brown, with small round darker spots between the dorsal, anal, and caudal rays; a dark horizontal streak behind the eye; a dark bar at base of pectoral. Total length 85 millim. Three specimens from Kinyamkolo, and one from Mbity Rocks. The deep anterior groove of the skull, in which the ascending processes of the pre- maxillaries are received, extends to the anterior third of the orbits. and the strong occipital crest is prolonged forward to it; parietal crests are entirely absent; the chain of suborbital bones is very slender. None of the ribs are sessile, being inserted on a step at the back of the transverse processes of the vertebrae at a short distance from the centre; all bear epipleurals ; only the last pracaudal vertebra has a hemal bridge. JULIDOCHROMIS, g. n. Body elongate ; scales ctenoid. Jaws very narrow, with a few curved canines in front, the outer of which are very large and tusk-like, followed on the sides and behind by minute conical teeth forming a narrow band. Maxillary exposed. Dorsal with 22 to 24 spines, anal with 8 or 9. Vertebre 34 (17-+-17). This genus represents an exaggerated type of Lamprologus. ‘The very specialized dentition and the general aspect remind one of the Julidine Labridw, whence the name proposed. G bo 12 MR. G. A. BOULENGER ON A COLLECTION - 10. JuLIpocHRoMIS orNATUS, sp. n. (Plate II. fig. 4.) 4 or 6 canines in each jaw, tipped with brown. Depth of body 4 to 43 times in total length, length of head 34 to 34. Snout 14 to twice as long as diameter of eye, which is 44 to 5 times in length of head and 1} in interorbital width ; maxillary extending to below nostril ; cheeks naked ; opercles scaled. Gill-rakers very short and few. Dorsal XXII-XXIV 5; spines equal from the eighth or tenth, 4 length of head ; longest soft rays 4 to 3 length of head. Pectoral about 3 length of head. Ventral produced in a filament, reaching origin of anal. Anal VIII-IX 4-6 ; spines increasing in length to the last, which equals 2 length of head. Caudal rounded. Caudal peduncle as long asdeep. Scales 45-50 are: lat. 1. aoe. Yellowish, with three dark brown stripes on each side, the lower from the end of the snout to the base of the caudal, the upper along the base of the dorsal; a large round dark-brown spot on the base of the caudal; a small black spot at the base of the pectoral; anal edged with brown; caudal brown. Total length 85 millim. Five specimens from Mbity Rocks. The premaxillary groove is deep and extends to the anterior third of the orbits ; the occipital crest is low and continued forward to the premaxillary groove; parietal crests are present, but very feeble; the chain of suborbital bones is replaced by a ligament ; the mandible is very massive, the lower surface flat, projecting as a keel on the sides. The ribs are inserted on a step of the transverse processes, at a short distance from the centre; all bear epipleurals; only the last preecaudal vertebra has a hemal bridge. 11. PARATILAPIA PFEFFERI, sp. n. (Plate ITI. fig. 1.) Teeth small, in 3 series in each jaw, forming a narrow band, outer largest. Depth of body equal to length of head, 2% times in total length. Snout with straight upper profile, as long as eye, the diameter of which is 33 times in length of head and equals Id interorbital width ; maxillary extending to below anterior border of eye ; 3 series of scales on the cheek; large scales on the opercle. Gill-rakers rather long, 11 on lower part of anterior arch, the larger spatulate. Dorsal XVI &; spines increasing in length to the sixth, which measures a little less than } length of head and nearly equals longest soft rays. Pectoral $ length of head. Ventral reaching origin of anal. Anal III 7; third spine longest, as long as longest dorsal. Caudal feebly emarginate, Caudal peduncle 1 as long as deep. Scales very finely denticulate on the edge, 332 ; lat. 1. — Pale olive above, silvery beneath, with seven darker vertical bars; fins greyish brown. Total length 76 millim. A single specimen from Kinyamkolo. OF FISHES FROM LAKE TANGANYIKA, 13 This species, which I have much pleasure in naming after the distinguished zoologist of Hamburg, Dr. G. Pfeffer, who has much contributed to our knowledge of East African ichthyology, is nearest allied to P. intermedia Gthr., from Nyassa. 12. PaRaTILAPIA MACROpS, sp. n. (Plate III. fig. 2.) Teeth small, in 3 series in each jaw, forming a narrow band, outer largest. Depth of body three times in total length; length of head 22 to 8. Snout with straight upper profile, a little shorter than the eye, the diameter of which is 2% times in length of head and exceeds interorbital width; maxillary extending to below anterior border of eye; 2 or 3 series of scales on the cheek; large scales on the opercle. Gill-rakers short, 11 on lower part of anterior arch. Dorsal XVI 10-12; spines increasing in length to the sixth, which measures a little less than 3 length of head and equals longest soft rays. Pectoral as long as head. Ventral extending a little beyond origin of anal. Anal III 6-7; third spine longest, a little shorter than longest dorsal, Caudal with crescentic emargination. Caudal peduncle as long as deep. Scales very finely denticulate on the edge, 33-34 oe latayele oa} upper lateral line complete, extending to base of caudal. Pale brownish above, silvery beneath; a series of five indistinct dark blotches on each side of the body; spinous dorsal edged with brown above. Total length 70 millim. Two specimens from Kinyamkolo, and one from Mbity Rocks. Closely allied to the preceding species; distinguished by the larger eye, the complete upper lateral line, the longer pectoral, and the more strongly emarginate caudal. Connects the preceding species with the following, which represents a more aberrant type. 13, PARATILAPIA VENTRALIS, sp. n. (Plate IIT. fig. 3.) Teeth very small, in 2 series in both jaws, the outer larger and tipped with brown. Depth of body 22 to 3 times in total length, length of head 3, Snout with curved upper profile, a little shorter than the eye, the diameter of which is 2 to 27 times in length of head and exceeds interorbital width; maxillary extending to below anterior fourth of eye; 2 or 3 series of scales on the cheek; large scales on the opercle. Gill-rakers rather long, lanceolate, 17 to 19 on lower part of anterior arch. Dorsal XII-XIII 13-14; spines increasing in length to the eighth or ninth, which measures nearly 4 length of head and } or $ longest soft rays. Pectoral a little longer than head. Ventral much produced, extending far beyond the origin of the anal, especially in the males, in which it may reach the end of the caudal. Anal III 9-10; third spine longest, about 2 length of head; middle soft rays produced, nearly as long as head. Caudal deeply emarginate, crescentic. Caudal peduncle 1} as long as deep. 14 MK. G. A. BOULENGER ON A COLLECTION on? lat. 1. — complete or nearly so, usually reaching base of caudal. Grey-brown above, with or without irregular darker spots, silvery below ; fins grey-brown, the ends of the ventral filaments white. Total length 92 millim. 8 specimens from Kinyamkolo, and one from Mbity Rocks. ‘This species is remarkable among its congeners from continental Africa for the low number of its dorsal spines combined with the large eye, the crescentic caudal, the extremely produced ventrals, and the upper lateral line extending on the caudal peduncle. Scales finely denticulate on the edge, 34-36 upper lateral line 14, PARATILAPIA FURCIFER, sp. n. (Plate IV. fig. 1.) Teeth very small, in 3 series in both jaws, the outer largest and tipped with brown. Depth of body equal to length of head, 3 times in total length. Snout with curved upper profile, a little shorter than the eye, the diameter of which is 22 to 23 times in length of head and slightly exceeds interorbital width ; maxillary extending to below anterior border of eye; 2 or 3 series of scales on the cheek; large scales on the opercle. Gill-rakers rather long, lanceolate, 15 or 16 on lower part of anterior arch. Dorsal XIII 13-14; spines increasing in length to the ninth, which measures a little less than } length of head; some of the soft rays produced, nearly as long as head. Pectoral a little longer than head. Ventral much produced, extending nearly to caudal. Anal III] 9; third spine longest, 3 length of head; middle soft rays produced, like the dorsals. Caudal deeply emarginate, crescentic, the rays at the angles produced. Caudal peduncle 14 as long as deep. Scales finely denticulate on the edge, 60-63 on lat. 1. — upper lateral line nearly complete, extending on the caudal peduncle. Bluish above, white beneath ; a few ill-defined yellow streaks along the body; some yellow marblings on the postocular part of the head; fins white, with some yellow streaks on the dorsal and anal and between the ventral and caudal rays (according to a sketch by Mr. Moore). Total length 110 millim. ‘lwo specimens from Kinyamkolo. Closely allied to the preceding; distinguished by much smaller scales. 15. PaRATILAPIA LEPTOSOMA, sp. n. (Plate III. fig. 4.) Teeth small, in 3 series in the upper jaw, in 4 in the lower, outer largest. Depth of body 4 to 43 times in total length; length of head 3. Snout with straight upper profile, as long as or a little longer than the eye, the diameter of which is 33 to 33 times in length of head, and equals interorbital width; maxillary extending to below anterior border of eye or not quite so far; 2 series of scales on the cheek; opercle OF FISHES FROM LAKE TANGANYIKA. 15 covered with scales. Gill-rakers long, slender, and close-set, 20 on lower part of anterior arch. Dorsal XII 14-15; spines increasing in length to the last, which measures § length of head, and is nearly as long as the soft rays. Pectoral # length of head. Ventral reaching origin of anal. Anal III 10-12; third spine longest, a little shorter than longest dorsal. Caudal feebly emarginate. Caudal peduncle twice ee lat. 1. a Brown, lighter beneath ; dorsal and anal with or without brown longitudinal streaks ; caudal spotted with brown or black at the base. Total length 87 millim. Two specimens from Kinyamkolo, and two from Mbity Rocks. Approaches P. longiceps Gthr. in the elongate body, but differs greatly in the dentition, the shape of the snout, and the ftin-formula. Only the Malagassy species of the genus Paratilapia were known to have fewer than 13 spines in the dorsal fin; but one of the species described above (P. ventralis) has only 12 or 13. as long as deep. Scales very finely denticulate on the edge, 59-40 BATHYBATES, g. n. Body elongate ; scales cycloid, small and irregular. Several rows of large fang-like teeth in the jaws. Maxillary exposed. Dorsal with 14 spines, anal with 3. Vertebre 36 (17+19). This genus is closely related to Paratilapia, with which it is to a certain extent connected by P. longiceps Gthr., from Nyassa. However, the more formidable dentition, coupled with the character of the scales on the body, appear sufficient to warrant the establishment of a new genus. ‘The unique species appears in most respects more specialized than Paratilapia longiceps, although more primitive than any known species of that genus in the longer anal fin and the complete upper lateral line. 16. BAaTHYBATES FEROX, sp. n. (Plate IV. fig. 2.) Teeth long and sharp, fang-like, wide apart, in 4 series in the upper jaw, in 3 in the lower. Depth of body 4 times in total length, length of head 3 times. Snout long and strongly compressed, with convex upper profile; eye large, its diameter 12 times in length of snout, 34 in length of head, and a little greater than interorbital width ; maxillary not quite reaching to below anterior border of eye; 5 series of small scales on the cheek; large scales on the opercle. Gill-rakers moderately long and slender, 13 on lower part of anterior arch. Dorsal XIV 16; spines rather feeble, slender, subequal from the fifth, which measures 7 length of head; longest soft rays 4 length of head. Pectoral ? length of head. Ventral not reaching vent. Anal IT] 16; spines short and feeble. Caudal deeply forked, middle rays not hi lf as long as outer. Caudal peduncle nearly twice as long as deep. Scales small and irregular, 16 MR. G. A. BOULENGER ON A COLLECTION especially below the lateral lines, 78 oe lat. 1. E ; the upper extending from the opercle to the caudal, the lower from below the last dorsal spines to the caudal. Pale bluish green, iridescent above, white below; dorsal and anal bluish grey; ventrals and caudal yellowish ; pectorals yellow ; two dark streaks on the dorsal. Total length 275 millim, A single specimen from Kinyamkolo, taken at a depth of 400 feet. The specimen is a female with ripe ova; these are of large size, measuring 53 millim. in diameter. The stomach contains a small, partially-digested fish of the genus Paratilapia, as first ascertained by a sciagraph kindly prepared by Messrs. Gardiner and Green, which has enabled me to compare the structure of the vertebral column with that of other members of the family Cichlidw. The insertion of the ribs is typical, viz., sessile, except on the last three preecaudal vertebra ; the transverse processes are short and in front of the ribs. ‘The number of vertebra agrees with that of the more elongate species of Paratilapia. ERETMODUS, g. n. Body moderately elongate; scales ctenoid. Jaws with rather large spatulate teeth with truncated crowns disposed in oblique transverse rows of two or three. Lips much developed; maxillary entirely concealed when the mouth is closed. Dorsal with 23 to 25 spines; soft rays reduced to 3 to 5; anal with 3 spines. Vertebre 30 (15+15). A strongly aberrant form, remarkable for its dentition and the structure of the dorsal, formed almost entirely of spines. 17. ERermopus cyanostictus, sp. n. (Plate III. fig. 5.) 8 or 10 transverse series of teeth in each jaw, the crowns reddish brown. Depth of body equal to length of head, 3 times in totallength. Profile of snout curved ; length of snout 14 to twice diameter of eye, which is 43 to 5 times in length of head, and a little less than interorbital width; mouth extending to below nostril; cheeks and opercles naked. Gill-rakers short, 9 or 10 on lower part of anterior arch. Dorsal XXIII-XXV 3-5; spines subequal from the sixth, $ length of head, a little shorter than soft rays. Pectoral 3 length of head. Ventral reaching vent, Anal III 6-7; third spine longest, a little longer than dorsals; soft rays about 3 length of head, 22-23 6-9 ° Caudal rounded. Caudal peduncle deeper than long. Scales 32-35 aos lat. 1. Blackish brown, with scattered pale blue dots; belly yellowish. Total length 75 millim. Five specimens from Mbity Rocks, and three from Kinyamkolo. The toothed portions of the premaxillary and mandible are much developed in depth, in a manner suggestive of the Sparidw and Scaride, and the teeth are implanted in sockets. ‘The deep triangular groove for the reception of the premaxillaries extends OF FISHES FROM LAKE TANGANYIKA. Uf to between the orbits, and the occipital crest is prolonged to it ; the parietal crests are produced on the frontals ; the preeorbital is large, and the chain of suborbitals very slender. ‘The pharyngeal teeth have long slender shafts and conical brown cusps. The ribs are attached to the back of short transverse processes; all but the last one support epipleurals ; the last four preecaudal vertebrae have a hemal bridge. 18. Tinapia LaBiaTa, sp. n. (Plate V. fig. 1.) Outer teeth rather large, feebly notched ; inner teeth very small, tricuspid, in 3 or 4 series. Depth of body equal to length of head, 22 to 22 times in total length. Snout with straight upper profile, 14 to 1} diameter of eye, which is 34 to 44 times in length of head and equals interorbital width; maxillary not extending to below anterior border of eye; 3 or 4series of scales on the cheek; large scales on the opercle ; lips very strongly developed, both produced into a large triangular lobe in front. Gill- rakers moderate, 15 on lower part of anterior arch. Dorsal XVIII 10; middle dorsal spines longest, about ? length of head, and a little shorter than longest soft rays. Pectoral 3 to length of head. Ventral reaching origin of anal. Anal III 6-7; third spine longest, as long as longest dorsals, slightly shorter than longest soft rays. Caudal truncate. Caudal peduncle slightly longer than deep. Scales finely denticulate on the border, 33-35 ee lat. 1. aa Pale olive, with 10 more or less distinct darker cross-bars; fins greyish-brown ; dorsal sometimes with oblique dark and light streaks ; caudal with numerous round dark spots between the rays, Total length 170 millim. Four specimens from Kinyamkolo. This species is easily recognizable by the extraordinary development of the lips, which bears a curious resemblance to that observed in the Central American Heros labiatus*. It appears to be nearest allied to Ctenochromis nuchisquamulatus Hilg., and C. sauvagii Pfeffer, from the Victoria Nyanza. TROPHEUS, g. 0. Body moderately elongate ; scales ctenoid. Jaws angularly bent at the sides, with bands of minute tricuspid teeth, an outer row of bicuspid teeth, and enlarged conical teeth at the sides of the premaxillary; mouth transversely linear when closed; maxillary concealed under the preorbital. Dorsal with 21 spines, anal with 6. Vertebre 33 (17-+16). This genus differs from Zi/apia in the angularly-broken premaxilla armed with enlarged conical teeth at the sides, and the higher number of anal spines. The name proposed (zpo¢etc, one who rears, brings up, educates) refers to the peculiar mode of nursing the eggs described hereafter. 1 Giinther, P. Z. 8. 1864, pl. iv. fig. 1. voL. XV.— Part I. No. 3.—December, 1898. D 18 MR. G. A. BOULENGER ON A COLLECTION 19. TRopHEUS MooRI, sp. n. (Plate V. fig. 2.) Teeth minute, those of the outer series tipped with brown and numbering about 50. Depth of body 23 to 22 times in total length, length of head 34 to 33%. Snout descending in a strong curve, as long as or a little longer than the diameter of the eye. which is 34 to 4 times in length of head, and equals $ to ? interorbital width ; mouth extending to below anterior border of eye; 4 series of scales on the cheek ; large scales on the opercle. Gill-rakers short, 11 or 12 on lower part of anterior arch. Dorsal XXI 5-6; spines increasing in length to the sixth, which measures not quite half length of head; longest soft rays 2 to # length of head. Pectoral as long as head. Ventral produced into a short filament, reaching beyond origin of anal. Anal VI 5-6: spines increasing in length to the last, which slightly exceeds longest dorsal. Caudal OOF) slightly notched. Caudal peduncle as long as deep. Scales 30-32 as lets Il, Fe Dark brown; a large bluish-white blotch on each side; belly reddish brown ; fins blackish. Total length 110 millim. Five specimens from Kinyamkolo. The occipital crest is very strong and the parietal crests are produced on the frontals. The insertion of the ribs is as described in the preceding genera Telmatochromis ana Eretmodus. The mouth and pharynx of one of the specimens contains four eggs of very large size, the vitelline sphere measuring 4 millimetres in diameter, with an embryo in an advanced stage of development. ‘The egg of the Fifteen-spined Stickleback ( Gasterosteus spinachia), hitherto regarded as the largest Teleostean egg in proportion to the size of the animal, measures only 3 millimetres in diameter’. Besides the Siluroids of the genera Arius and Galeichthys, which have very large eggs, at least two species of Tilapia were known to give shelter to their eggs in the manner noticed above, viz.. T. simonis Gthr. (Chromis paterfanilias Lortet), as observed by Prof. Lortet in Lake Tiberias, and 7’. nilotica Cuv., as noticed by me ina specimen collected by Canon Tristram in the same lake. But these eggs, produced by fishes of the size of our common Perch, are very numerous, and measure only about 2 millim. in diameter. It has besides been observed in these Tilapia, as well as in the Siluroids, that the function of protecting the eggs devolves on the male sex, while, to my surprise, the Tanganyika fish proved on autopsy to be a female. Whether this is constantly so, or whether either parent takes to the nursing duties, remains to be ascertained by examination of a larger number of specimens. Iam all the more disposed to think the latter possibility will be confirmed, from the fact thata specimen of Z%lapia nilotica 1 The absolutely largest known Teleostean egg is that of the Siluroid Arius commersonz, which measures 17 millim. in diameter. The size of the Arius-egg is much underrated by Giinther (‘Study of Fishes,’ p. 160) and by McIntosh and Masterman (‘ British Marine Food Fishes,’ p. 12), who ascribe to it a maximum of 10 millim OF FISHES FROM LAKE TANGANYIKA. 19 with the pharynx filled with embryos belongs to the female sex, while Dr. Lortet’s observations on 7. simonis had led to the belief that specimens carrying eges in that manner are invariably males. It is here necessary to recall the observation contained in Livingstone’s ‘ Last Journals,’ vol. ii. p. 17, that the “ Dagala” or “ Nsipé,” of Lake Tanganyika, a small fish 2 or 3 inches long, and very like Whitebait, is said to emit eggs by the mouth. The comparison of this fish to Whitebait excludes the possibility of the one here described being the “Dagala” or “Nsipé,” which will probably prove to be a Cyprinodont, if not actually the Haplochilus tanganicanus described below. SIMOCHROMIS, g. n. Body moderately elongate; scales ctenoid. Jaws with broad bands of minute tricuspid teeth, with an outer series of larger bicuspid teeth, and a single series of conical teeth at the sides of the pramaxillary; maxillary concealed under the pre- orbital. Dorsal with 17 or 18 spines, anal with 8. Vertebree 31-32 (15+16-17). The type species differs from the other members of the genus to which it was originally referred in the differentiation of the teeth at the sides of the pramaxillary, a character in which it agrees with the genus Zropheus. 20. SIMOCHROMIS DIAGRAMMA Gthr. Chromis diagramma Ginth. Proc. Zool. Soc. 1893, p. 632, pl. lviii. fig. 3. The bands of teeth in both jaws consisting of 8 to 10 series, gradually decreasing in size from the outer border. Depth of body 25 to 2% times in total length, length of head 34 to 85. Snout descending in a strong curve ; diameter of eye 3 to 34 times in length of head, 14 in interorbital width ; mouth extending to below anterior border of eye; 4 series of scales on the cheek ; large scales on the opercle. Gill-rakers short, 12 or 13 on lower part of anterior arch. Dorsal XVII—XVIII 9-10; spines increasing in length to the sixth, which measures } to ¢ length of head; longest soft rays about 3 length of head. Pectoral as long as head. Ventral reaching vent or origin of anal. Anal III 7-8; third spine longest, as long as longest dorsal. Caudal truncate. Caudal peduncle as long as deep. Scales 31-33 z; lat. 1. a Olive, whitish beneath ; young with 9 rather indistinct vertical bars ; gill-membrane sometimes with dark brown spots; fins greyish; a dark brown stripe may be present along the spinous dorsal, which has a black edge. Total length 105 millim. Four specimens from Kinyamkolo. 20 MR. G. A. BOULENGER ON A COLLECTION PETROCHROMIS, g. n. Body moderately elongate; scales ctenoid. Jaws with very broad bands of minute, bristle-like teeth, with incurved bi- or tricuspid crowns; maxillary concealed under the preorbital. Dorsal with 17 or 18 spines, anal with 3. Vertebree 32 (16416). Distinguished from Tilapia by the very broad bands of minute teeth with incurved crowns, from Simochromis by the absence of conical teeth in the preemaxillary. 21. PETROCHROMIS POLYODON, sp. n. (Plate VI. fig. 1.) Crowns of teeth brown. Depth of body 21 to 22 times in total length, length of head 93 to 3 times. Snout with convex upper profile, 14 to 1% diameter of eye, which is 4 to 41 times in length of head and 14 in interorbital width; mouth hardly extending to below anterior border of eye; 4 or 5 series of scales on the cheek ; large scales on the opercle. Gill-rakers very short, 12 or 13 on lower part of anterior arch. Dorsal XVII-XVIII 8-9 ; spines increasing in length to the sixth or seventh, which measures about 2 length of head; longest soft rays $ to 3 length of head. Pectoral nearly as long as head. Ventral reaching vent or origin of anal. Anal III 7-8 ; third spine longest, as long as longest dorsal. Caudal truncate. Caudal peduncle as long as deep. 3-4 29-94 : as ‘ Scales 32-34 s543; lat. 1. {34,. Olive-brown, whitish beneath ; fins grey or blackish. Total length 185 millim. Two specimens from Kinyamkolo, and two from Mbity Rocks. The premaxillary and mandibular bones are very massive, and the maxillary is much reduced in size; the ascending processes of the premaxillaries extend to between the anterior borders of the orbits and are received in a deep excavation, to which the strong occipital crest extends; the parietal crests are produced forwards as far as the frontals ; the preorbital is large, and the chain of suborbitals very slender. Only the first rib is absolutely sessile, the following being attached to the back of the transverse processes at a short distance from the centre ; the epipleurals extend to the twelfth rib; the last two preecaudal vertebree form a hemal bridge. PERISSODUS, g. N. Body elongate; scales cycloid. ‘Teeth rather large, unequal in size, few, with swollen bases and low slightly-notched crowns, compressed transversely to the axis of the jaws, disposed in a single series; maxillary exposed. Dorsal with 18 spines, anal with 3. Vertebre 35 (17+ 18). The extraordinary dentition which characterizes this new genus may be looked upon as an extreme specialization of that exhibited by Ztlapia, a specialization in an opposite direction from that attained by Petrochromis. OF FISHES FROM LAKE TANGANYIKA. a1 22. PERISSODUS MICROLEPIS, sp.n. (Plate VI. fig. 2.) 10 teeth on each side of the premaxillary, 9 on each side of the mandible. Depth of body 3% times in total length, length of head 31. Snout a little longer than diameter of eye, which is 4 times in length of head, and almost equals interorbital width ; lower jaw projecting; maxillary extending to below anterior border of eye; 3 series of scales on the cheek ; large scales on the opercle. Gill-rakers rather long, 14 on lower part of anterior arch. Dorsal XVIII 10; spines increasing in length to the last, which measures } length of head; longest soft rays 3 length of head. Pectoral # length of head. Ventral not reaching vent. Anal III 8; third spine longest, nearly as long as longest dorsal ; longest soft rays half length of head. Caudal truncate. Caudal peduncle 13 as long as deep. Scales 65 a8 lat. 1. a Uniform dark reddish brown ; a blackish opercular spot. Total length 100 millim. A single specimen from Mbity Rocks’. 1 T think it desirable to allude here briefly to two remarkable genera of Cichlidw from Lake Tanganyika, which I have founded on specimens obtained by Capt. Descamps and sent to me for study by the Administration of the Congo Free State. The specimens are in so bad a condition that it will never be possible to draw up from them complete definitions of the species, but their dentitions, which indicate remarkable new genera, are well preserved, and I am sure the annexed figures together with the short notes here given will be sufficient to ensure their future recognition :— Ecropus, g. n.—Teeth very small, conical, in two series in both jaws, the outer larger; outer mandibular teeth pointing outward, perpendicular to the others; maxillary concealed under the preorbital when the mouth is closed. Scales rather large, ctenoid. Dorsal with 14 spines, anal with 3. Ectodus descampsi, sp.n. (Fig. 1, a.)\—D. XIV 14; A. IIL 8. Depth of body 34 times in total length, length of head 3. Eye very large, 24 in length of head, 12 interorbital width. Two series of scales on the cheek. Sc. 84 4; lat. 1.2. Total length 95 millim. Ectodus melanogenys, sp.n. (Fig. 1,0.)—D. XIV 16; A. III 13, Depth of body 5 times in total length, Fig. 1. + (A ZF A ll Zz a Gis b, length of head 34. Eye 4 times in length of head, 14 in length of snout, 14 interorbital width, Three series of scales on the cheek. About 35 scales in a longitudinal series. A black spot on the chin, Total length 95 millim. bo bo MR. G. A. BOULENGER GN A COLLECTION MASTACEMBELIDA. 23. MASTACEMBELUS MOORII, sp.n. (Plate VII. fig. 1.) Depth of body 14 times in total length, length of head (without rostral appendage) 3 times; vent equally distant from end of snout and base of caudal; length of head 24 to 3 times in its distance from vent, and 4 in its distance from first dorsal spine. Snout 3 times as long as eye, ending in a trifid dermal appendage which is a little longer than eye; cleft of mouth extending to below centre of eye; no pre- opercular spines. Vertical fins united with the rounded caudal. Dorsal XX V-XXVII 70-80 ; spines very short. Anal II 70-80. Pectoral + length of head. Scales very small, 30-35 between origin of soft dorsal and lateral line. Brown, tail with a wide-meshed blackish network; dorsal and anal whitish, with a vertical series of blackish spots or vertical bars; anal and caudal edged with blackish. Total length 440 millim. Two specimens from Mbity Rocks. This species has been compared with J/. marmoratus, Perugia, from the Congo, which differs in the more slender body (its depth contained 20 times in the total length) with larger scales, the shorter tail, the longer rostral appendage (measuring twice the diameter of the eye), and the higher number (30) of dorsal spines. I think as many as 13 African species of Mastacembelus may be distinguished, as shown in the following synopsis :— PLEcoDUs, g. n.——Teeth large and few, in a single series, dilated at the base, truncated at the end, compressed, slightly grooved in front, curved and directed backwards; 14 teeth in the upper jaw, 12 in the lower; maxillary exposed. Scales moderate, cycloid. Dorsal with 19 spines, anal with 3. Plecodus paradowus, sp.n. (Fig, 2..—D. XIX 14; anal III 12. Depth of body 4 times in total length, length of head 41. Eye large, longer than snout, 4 length of head, 17 interorbital width ; maxillary extending 5 lat. 1. °. Total length 90 millim. to below anterior third of eye; 3 series of scales on the cheek. Sc. 65 7; OF FISHES FROM LAKE TANGANYIKA, 23 I. Vent nearly equally distant from the end of the snout and the base of the caudal fin. A. Two strong preopercular spines. D. XXIH-XXIV 75; A. Il 75-76; length of head (measured to extremity of opercle and without rostral appendage) hardly twice in its distance from vent AS 6 times as great as its distance from the first dorsal spe . . . oe a elVarmanchi s alive D. XXIV 100; A. II 100; length of head 34 reat in its distance ro vent, equal to its distance from the first dorsal spine. . . See - . . 2. M. eryptacanthus Gthr. D. XXVII85; A. III 92; length of Tiel On in its Fivcnes from vent, 3 times as great as its distance from first dorsal spine . . . . =... =... 3. M. congicus Blgr." D. XXVI-XXX 80; A.TI180... . Bo 6 0 0 0 0 Sy WE inooaaGdS Wiksee D. XXIX-XXX 117-130; A. II 117- 17: length of head nearly 4 times in its distance from vent, and nearly twice as great as its distance from the first dorsal spine. 5. M. loennbergii Blgr.’ B. Preopercular spines minute (young) or absent. D. XXVII-XXVIII 70-80 ; A. IT 80-90; length of head 8 to 4 times in its distance from vent, equal to or a little less than its distance from the first dorsal spine. 6. M. flavomarginatus Blgy.’ D. XXVII 80-82; A. II 60-65 ; length of head equal to its distance from the first dorsal spine. 7. M. niger Sauv. D. XXVII-XXIX 65-70; A. II 68-75 ; length of head 24 to 3 times in its distance from the vent, and 4 in its distance from the first dorsal spine. . . 8. MZ. shiranus Gthr. D, XXIX 95; A. II 85; length of head 3 times inits distance from the vent, aud 4 in its distance ' from the first dorsal spme. . . . . . 9. M. nigromarginatus Blgr.’ D. XX V-XXVII 70-80; A. II 70-80 ; see of heed twice in its distance from the vent, and nearly 4 in its distance from the fir st dorsal spme. . . . 10. ML moorii Bler. II. Vent much nearer the caudal fin than the end of the snout; one or two preopercular spines. ID), 2OOS C3 AG INGO oo 6 68 5 4 685 6 og 6 JU Whoa Eee D. XXXITI-XXXV 50-60; A.IT50-60. . . . .. . . 12. MU. tanganice Gthr. III. Vent much nearer the end of the snout than the caudal fin. 1D, SOOM ROOOIL NOs AG INNG 5 6 2 6 6 6 eo 5 5 co Meh WE Caen Cisne ‘ A species from the Upper Congo founded on a single specimen with mutilated tail and regenerated caudal fin. I have now a perfect specimen. * M. marchii Steind. nec Sauy.—Liberia. * M. cryptacanthus Loennberg nec Gthr.—Cameroon. + A new species founded on specimens from the Gaboon previously referred to WM. cryptacanthus. 5 A new species founded on a single specimen from Ashantee. ° Type examined.—Vent equally distant from the head and from the caudal; head once and a halt as toug as its distance from the first dorsal spine; 17 scales between origin of soft dorsal and lateral line, 24 MR. G. A. BOULENGER ON A COLLECTION SILURID. 24. CLARIAS ANGUILLARIS L. This species occurs in the Nile as well as in the Niger and the Ogowe. It is one of those Siluroids which are known to live a considerable time out of water and to make excursions on dry land after the manner of Hels. 25. CLARIAS LIOCEPHALUS, sp.n. (Plate VII. fig. 2.) Vomerine teeth in a narrow band, without posterior process. Depth of body 52 times in total length, length of head 5 times. Head smooth, covered with sott skin, slightly longer than broad; occipital process very short, angular; diameter of eye 3 times in length of snout, 6 times in interorbital width; maxillary barbel as long as the head, nasal barbel a little shorter; inner mandibular barbel 2 length of head. Dorsal 70. Anal 50. Caudal free. Pectoral $ length of head, not extending to the vertical of origin of dorsal fin. Uniform blackish brown. Total length 80 millim. A single specimen from Kinyamkolo. 96. ANOPLOPTERUS PLATYCHTR Gthr. Two specimens from marshes near Mbity. I have recently recorded the occurrence of this species in Lake Nyassa, and it has also been described by Vaillant as Chimarrhoglanis leroy, from Mrogoro torrent, Urugaru Mountains, East Africa. 97. AUCHENASPIS BIScUTATA Geoffr. Known from the Nile and the Senegal. 28. SYNODONTIS MULTIPUNCTATUS, sp.n. (Plate VIII.) Premaxillary teeth in 5 or 6 irregular transverse series; mandibular teeth in a single series of 16, feebly curved, simple, measuring hardly 3 diameter of eye. Depth of body 34 times in total length, length of head 33. Head scarcely longer than broad, slightly convex on the crown; snout rounded, less than half length of head, twice as long as eye; eye supero-lateral, well visible from above, its diameter 44 times in length of head, twice in interorbital width. Gill-cleft very narrow, not extending below base of pectoral. Maxillary barbel simple, reaching a little beyond anterior third of pectoral spine; mandibular barbels strongly fringed, outer a little more than twice as long as inner, and half as long as maxillaries. Dorsal II 7; spine strong, a little shorter than the head, strongly serrated behind in its distal half. Adipose fin low, a little shorter than the head, twice as long as its distance from the dorsal. Humeral process simply granulate, sharply pointed, not OF FISHES FROM LAKE TANGANYIKA. bo n extending quite so far as the occipito-nuchal shield, which is 14 as long as broad and reaches the first soft ray of the dorsal. Anal III 7. Caudal deeply bifurcate. Skin smooth. Pale reddish brown above, with very numerous blackish-brown round spots, which are smallest on the head and nape; lower parts and barbels white, unspetted ; ventrals yellow ; dorsal and caudal broadly edged with yellow. Total length 240 millim. A single specimen from Sumbu. 29, MALAPTERURUS ELECTRICUS Gin. The Electric Cat-fish is widely distributed, being on record from the Nile, the Niger, the Ogowe, and the Congo. CYPRINIDS. 30. LaABEo, sp. inc. An unmistakable figure of a Zabeo is among Mr. Moore’s sketches, and although the specimen is not represented in the collection entrusted to me for description, I have included the fish as being the only representative of the Cyprinoids yet found in Tanganyika. The specimen from which the drawing was made was of large size, and therefore could not be preserved under the difficult circumstances to which allusion has been made in the introduction to this memoir. CHARACINIDA, 31. ALESTES MACROLEPIDOTUS C. & V. Occurs in the Nile, the Senegal, the Niger, and the Congo. 32. ALESTES MACROPHTHALMUS Gthr. This fish was known only from the Ogowe. 33. Hyprocyon FORSKALI Cuy. Occurs in the Nile, the Senegal, and the Niger. CYPRINODONTID4. 34. HAPLOCHILUS TANGANICANUS, sp.n. (Plate VI. fig. 3.) Body compressed, its depth 4 times in total length; length of head 4% times. Snout depressed ; lower jaw projecting beyond the upper; eye a little longer than snout, a little shorter than interorbital width, 3 times in length of head. Dorsal 13, originating at equal distance from the head and the caudal fin; the first ray corresponds to the 18th scale of the lateral line; posterior rays longest, § length of head. Anal 28, ariginating below extremity of pectoral. Pectoral $ length of head, extending far beyond root of ventral, Caudal feebly emarginate. Caudal peduncle twice as long as VoL. Xv.— PART I. No. 4.—December, 1898. E 26 MR. G. A. BOULENGER ON A COLLECTION deep. Scales 42 in a longitudinal line, 11 in a transverse line. Silvery, with a darker lateral stripe; dorsal and anal with greyish horizontal streaks. Total length 80 millim. A single specimen from Mbity Rocks. PoLYPTERIDA. 35. PoLyprerus BICHIR Geoffr. (2). Mr. Moore informs me that a Polypterus occurs in moderate abundance in the lake, and although he did not bring home specimens, he feels tolerably confident, from his recollection of them, that the fish belongs to the typical P. dichir, which occurs in the Nile, the Niger, and the Congo, and not to any of the closely-allied species that are often confounded with it. APPENDIX. By J. E. S. Moors. In the foregoing description of the fishes which I succeeded in bringing through from Tanganyika to the coast, Mr. Boulenger has already alluded to the difficulties that were experienced in transporting them in spirit, through several hundred miles of often trackless, always scorching, forest, and of the inevitable losses which this entailed. ‘The difficulties of transport, however, were by no means all. It must be nearly impossible for anyone who has not visited the African lakes to realize their huge size and oceanic character. One must be as heavily equipped for dredging in these waters as would be required for effective operations in the open sea. It will easily be understood, therefore, how incomplete our knowledge of the deep-water fauna of these lakes must be considered, when it is remembered that on Tanganyika I was of necessity forced to work with native dug-out boats, and with nothing better than the natives themselves as motive power. In the case of the fishes, moreover, there are no sources of collateral evidence from which we may obtain any insight into the nature of the deep-water forms, for dead fishes, unlike molluscs, leave no shells behind them, to be thrown upon the beaches of the lake, whereby, in the case of the molluscs, we gain some knowledge of the existence of forms which have not been seen alive; nor can we make use of the knowledge of the inhabitants in this matter, for the best of the Tanganyika natives are but wretched fishermen, merely using either surface traps, or light and inefficient drag-nets, which are thrown out a short distance from the shore and then hauled in to the land. Such nets are necessarily used only on smooth sandy beaches, and consequently the fishes caught in them are only of those s; ecies which inhabit places OF FISHES FROM LAKE TANGANYIKA. 27 of this description, and although they are numerically abundant, they belong to half a dozen species at the most. Nearly all the new forms which I obtained were killed by dynamite from the craggy ledges of the west coast of the lake, where the water was deep enough, about 20-35 feet, but not too deep for my men to dive and procure the greater number of the fishes, which, after every shot, were invariably found on the bottom. The number of fishes in such situations is really surprising, and on several occasions, after firing a single cartridge, I obtained more than two tall negroes could well carry, when slung in a bag between them on a pole. Only one fish was obtained from anything approaching to the deeper water of Tanganyika, about 400 feet, which was accidentally entangled in a basket dredge, and this was the specimen of the new species Bathybates feroa described on p. 15. Of the fishes which exist on the great mud-plains, which are to be found about 1000 feet below the surface in many portions of the lake, or of those which may occur in the vastly deeper portions, absolutely nothing is known. The fishes which have been collected, moreover, were obtained almost exclusively from the south-western extremity of Tanganyika, and to suppose that the fishes which occur in this locality are in any way representative of the fishes, even in similar situations, throughout the whole lake, would be as absurd as if we were to suppose that the shoal-water fish of the whole English Channel could be obtained from a few miles of rocky beach about the Needles and the Isle of Wight. ‘That the fishes differ in different portions of the lake I have myself observed, and it has also been shown in a most striking manner by the fact that of the six species previously described from ‘Tanganyika, which were collected by Captain Hore near Ujiji, I have only re- discovered one. The fishes which have been already obtained in ‘Tanganyika, or in Nyassa for that matter, must therefore be regarded as merely the surface skimmings of these lakes, and as in no sense representing the whole fish-fanna they may contain. Really deep water, 300 fathoms and upwards, is to be fonnd over large areas in Nyassa, and to what depth the water in Tanganyika may reach no one knows; but it is obvious that where water of such depth exists, if the lakes in which it is con- tained have been permanent for any great length of time, there ought to be modified forms of the freshwater fishes to suit such depths, and, with the exception of Bathybates ferox, these have not yet been obtained. In the case of Tanganyika, the interest which attaches to the future exploration of its deeper water is particularly great, since I have elsewhere shown that the majority of the Halolimnic animals, most of the molluscs, sponges, and so forth, are exclusively deep-water forms. In fact, it was only when my dredging capacities, so to speak, were giving out, that the more interesting of these forms, the genera T'yphobia, Bathanalia, and the like, were beginning to come in. 28 MR. G. A. BOULENGER ON A COLLECTION That there are no Halolimnic representatives among the fishes which have hitherto been obtained is no evidence that other fishes of a widely different and possibly of a Halolimnic type may not in future be secured. On the other hand, the fact that the Teleostean fishes now existing in Lake Tanganyika should not correspond with the Molluscan section of the Halolimnic group is really what one would expect; for, as I have recently shown ', the facies of the Molluscan section of the Halolimnic group is almost, if not quite, indistinguishable from that of the Jurassic seas. Except the Herring-like Leptolepide, few, if any, Teleostean fishes are represented in Jurassic beds, and we should therefore expect the piscine accompaniments of the Halolimnic molluscs to be entirely composed of Ganoids and the like. J found a species of Polypterus, which I took to be P. dichir, abundant on the southern shores of Lake Tanganyika, and it is quite possible that some of the active carnivorous fishes which inhabit the open water may be Ganoids too. What the fish is that so much surprised Glaive, when he crossed the lake, by attacking the paddles of his boat, is quite unknown, but I myself saw these same fishes attack the paddles of my own boat, not 20 miles from the spot where Glaive described them, on the west coast of the lake. Judging, therefore, from the incomplete character of our knowledge of the fish- fauna of Lake ‘Tanganyika, and from the vast antiquity of the lake, as evidenced by the Jurassic facies of its molluscan shells, it is only natural to expect that future exploration may reveal, among these fishes, forms that are of the highest scientific interest from a morphological point of view. I mean that Tanganyika and its neighbourhood present one of those few localities where it is legitimate to expect that we may discover many forms that in most places have become extinct. Our very slight acquaintance with the surface-forms existing in Nyassa and ‘Tan- ganyika, and the complete absence of all knowledge of the contents of the deep waters of these lakes, unfortunately by no means exhausts the sum of our present ignorance of these matters. We have no real knowledge of the extent of the Halolimnic fauna, beyond the lake in which it was originally discovered. I have shown only that this fauna is not present in Nyassa nor in Shirwa, nor yet even in the little lake Kela, not more than 20 miles from the Tanganyika shore, and that it is extremely probable that it does not extend to Mwero and Lake Bangweolo to the west. But for anything that can actually be shown to the contrary, it may be present in Lake Rukwa to the east, and in the Albert and Albert Edward Nyanzas to the north, and it is perhaps almost to be expected that some representatives of this stock should be found in Lake Kivu, which lies in the same great valley not more than 50 miles to the north of Tanganyika. But what is the relation of this lake, the effluent of which flows south into Tanganyika, to the Albert and Albert Edward Nyanzas, with effluents flowing into the Nile, towards the north? All that ‘ “ On the Hypothesis that Lake Tanganyika represents an old Jurassic Sea,’ Quart. Journ, Microse, Sci. xli. no. 162, June 1898. OF FISHES FROM LAKE TANGANYIKA, 29 we know at present is that the watershed between these lakes is situated in their common valley, and that the whole series of lakes, from Tanganyika in the south to the Albert Edward in the north, appears to lie in one continuous valley-system, which owes its origin to a single geological movement, that has affected, although it may be at very different times, an immense area of ground, and reaches at the present time actually to the shores of the Red Sea. The lakes which lie in these valleys immediately to the north of Tanganyika are therefore of especial zoological interest, for it may be that in these isolated sheets of water there exist animals analogous to those which constitute the Jurassic Halolimnic fauna of Tanganyika itself. The similarity of the Halolimnic gastropods to those of the Jurassic seas is, as I have insisted elsewhere, so complete and so remarkable that no effort should be spared which may throw light on the real nature and origin of these forms. The only geographical fact which is at present available is this, that there exists a certain similarity in the fish-fauna of Tanganyika to that of the lower Nile, and it may be that this is an indication that at some time the Nile watershed was further south, and drained out of Tanganyika to the north, in which case Halolimnic animals will without doubt be found among the lakes towards the north. All this is, however, little more than mere conjecture, but it will help my present purpose if it renders more apparent the extreme desirability of becoming better acquainted with the zoology of the lakes in the Rift valleys immediately to the north of Tanganyika. The presence of a Halolimnic fauna in Lake Tanganyika, as something superadded to the normal freshwater stock of the lake, is at present the central fact, and if the expectations which the morphological characters of those Halolimnic animals already known have raised should be fulfilled, we have something here which may completely change our views respecting the past history of the continent, and provide morphology with some most unique and valuable comparative material. EXPLANATION OF THE PLATES. PLATE 1. Fig. 1. Lates microlepis, p. 6. Fig. 2. Lamprologus fasciatus, p. 7. Fig. 3. compressiceps, p. 7. Fig. 4. 5 moori, p. 8. Fig. 5. Be modestus, p. 8. Fig. 6. _ elongatus, p. 9. PLATE IL. Fig. 1. Lamprologus furcifer, p. 9. Fig. 2. Telmatochromis vittatus, p. 10. VoL. XV.—PaRtT 1. Nu. 5.—December, 1898. P ON A COLLECTION OF FISHES FROM LAKE TANGANYIKA. Fig. 8. Telmatochromis temporalis, p. 11. Fig. 34. Ne es dentition. Fig. 3 0. Fe MS skeleton. Fig. 4. Julidochromis ornatus, p. 12. Fig. 44. . oF dentition. Fig. 44. 2" a skeleton. PAE lle Fig. 1. Paratilapia pfeffert, p. 12. Fig. 2. a macrops, p. 13. Fig. 3. Ba ventralis, p. 15. Fig. 4. im leptosoma, p. 14. Fig. 5. Hretmodus cyanostictus, p. 16. Fig. da. of ~ dentition. Fig. 5 2. es a skeleton. PLATE IV. Fig. 1. Paratilapia furcifer, p. 14. Fig. 2. Bathybates feroa, p. 15. Fig. 2a. o ee centition. PLATE V. Fig. 1. Tilapia labiata, p. 17. Fig. 2. Tropheus moorit, p. 18. IB Wis > » dentition. 1B Boop ,, Skeleton. PLATE VI. Fig. 1. Petrochromis polyodon, p. 20. Fig. La. die 45 dentition. ies 170: es = skeleton. Fig. 2. Perissodus microlepis, p. 21. Joie, BOS 5 es dentition. Fig. 3. Hapiochilus tanganicanus, p. 25. PLATE VII. Fig. 1. Mastacembelus moorit, p. 22. Fig to) Fig. 2a. +9 39 2. Clarias liocephalus, p. 24. upper view of head and nape. PLATE VIII. Magnified (x3). Synodontis multipunctatus, p. 24, with lower view of head. Sram, 00l. Soe Vel NV GCL. J.Green del. et hth. Mintern Bros.imp LATE S MICROLEPIS- 2.LAMPROLOGUS FASCIATUS. 38.LAMPROLOGUS COMPRESSICEPS. 4 LAMPROLOGUS MOORII . 5.LAMPROLOGUS MODESTUS. 6. LAMPROLOGUS ELONGATUS. Trans. Loo. Soe NOAV ICH | 3. 3b. WN Sal [; AR RA A BN SAR oe ip / Dee Aa. ba. NS, Lh. Mintern Bros im 1.LAMPROLOGUS FURCIFER. 2.TELMATOCHROMIS VITTATUS. 3.TELMATOCHROMIS TEMPORALIS. 4. JULIDOCHROMIS ORNATUS. J. Green del. et ith. Trans D0. Se Ul XVIGH. Peder MN > AON 5a 5b. J.Green del.et bth Mintern Bros.imp Tae ARCA Tu AE AS VB rE yBuniRa 2. PARATILAPIA MACROPS. 3.PARATILAPIA VENTRALIS 4,.PARATILAPIA LEPTOSOMA 5. ERETMODUS CYANOSTICTUS. Cs a ham ae Rn elo h MAC CC tal a { *XOUdA SHIVEAAHLVE S “URMIONNA VIdVILLVaVd T ‘OurcarD ‘so1gwrequgy “UATE © [ep usezg ¢ Ml OAK T0PC CPDL POL, af dur’ soag UtaquTAy “VLVIEWT VIGVILL 1 YATT 32 Sp wae. p ane yyy NGM ROA TO (A LAY, 1 Sace it Pe ee Tae ‘ J fe omelet eee CON AO Ge : i . . An Toy rh J ey i tee L 2 C \ . ¥ > . 5, ‘ . NN ; f ; a ; + Ps ‘ ‘ ‘ t rae ? + } . tx : S 4 \ ‘ x \ "| ue ‘ 7 4 2 ‘ i > ‘ ; i . * 4 ’ oy ¥ ‘ 1 F rm . ay , ni = sy ue Re ; . yi ty 2 4 ‘ i : Nis : ‘ 2 . aus (rae P f \ x + q Aes ad } x ’ ‘ ‘ H 4 a A ‘ ny f cs % : Gi < 5 : i 4 f Sat ‘ f FM a ory (A te * i \s F t : Oy . 1 ' , Srans Lol, Foe Nol. AV IE VI ae al y J Green del et ith. Montern Bros. imo ) $s 1.PETROCHROMIS POLYODON. 2.PERISSODUS MICROLEPIS. 3.HAPLOCHILUS TANGANICANUS. ‘SNIVHdHOOIT SVISV TO Z “ILHOOW SOTHENAIVISVIN i dua Souq wrapup, WNT 728 [ep use8ay p WA AY Plo °F PIE PY pra (sateen ah pce pmiet S ay talent baie Ah aman An eae Ta RO gi Ac dt Lary | ‘ OE be " A a ; ” ee ath Ri yas Ming 5 0 se ; cout “SNLVLONOdILTAN SILNOGONAS “OULOTY SOLE Uta qu “YAIL99 [ep useig ¢~ MA RENN IUPERYS FS PPL a} i a * a i : x : x TRANSACTIONS OF THE ZOOLOGICAL SOCIETY OF LONDON (continued). To Fellows. To the Public. eescuma cls ie eels VOLUME XIII. (1891-1895, contaming 62 Plates). . Price6 8 3 811 0 Part]. (1891, contaming 6 Plates) . ... . Pred Oil a) eek) Wigley 2 (WoO contaminoy. Gy elates)a ssues= ace wbi Be AO lay 2) sol Leek) >» o- (1891, contammg 6 Plates) . . . . . re OASYS) LAGS) » 4. (1892, contammg 1 Plate). aid OSs eh 0 6 O 5) on (L893, contaimime 9iPlates) 19). 3: pace Oe Tera Obes @ » 9. (1893, containing 4 Plates) ore) a) 012 0 5 7. (1898, containing 6 Plates) ees pp yes AB 012 0 » 8 (1894, containmg 6 Plates) . ... . eg cat) ie Dest), 012 0 » 9. (1894, contaming 6 Plates) . ... . pe On leas 015 O » 10. (1895, contamming 5 Plates) . . . . . Fire ete a) O12 0 5, Ll. (1895, containing 7 Platesand Titleand Index),, 0 15 9 lesa betG) VOLUME XIV. (1896-1898, containing 47 Plates) . Price5 5 0 7 6) Part 1. (1896, contammg 2 Plates) . 9. 2. . ave OcOue 10 Om. Sx20 » 2 (1896, containmg 6 Plates)... . . =O. OE sO) 012 0 », 9. (1897, containing 9 Plates) wo Oalomeg Er Alee 20 » 4. (1897, containing 10 Plates) . . . .. SS oe Omne0, 220-0 > 0» (1898; contaming 2 Plates)... . . pe LO SO) 0 8 0 5, Os (1898; contaming 7 Plates) 4) 7... ie Ol sig 0 1 4 E07. (1e08,fcontaimine 96 Plates) 12 ere tO LS 015 0 » 8. (1898, containing 5 Platesand TitleandIndex),, 0 9 O 012-0 VOLUME XV. Pani 1; e(s98scontammes:S)Plates )s eves. ou Price: Oh TOS kOe. ie ate ak gk a CONTENTS. I. Report on the Collection of Fishes made by Mr. J. E. 8. Moore in Lake Tanganycha during his Expedition, 1895-96. By G. A. Bounenesr, F.RS., F.ZS. With an Appendix by J. B.S. Moore, A.R.C.S. (Plates I-VI.) . . . page 1 THE PUBLICATIONS OF THE ZOOLOGICAL SOCIETY OF LONDON. Tun scientific publications of the Zoological Society of London are of two kinds—“ Proceedings,” published in an octavo form, and “ Transactions,” in quarto. According to the present arrangements, the “ Proceedings” contain not only notices of all busi- ness transacted at the scientific meetings, but also all the papers read at such meetings and recom- mended to be published in the “ Proceedings” by the Committee of Publication. A large number of coloured plates and engravings are attached to each annual volume of the “ Proceedings,” to illustrate the new or otherwise remarkable species of animals described in them. Among such illustrations, figures of the new or rare species acquired in a living state for the Society’s Gardens are often given. The “ Proceedings ”’ for each year are issued in four parts, on the first of the months of June, August, October, and April, the part published in April completing the volume for the preceding year. The “Transactions” contain such of the more important communications made to the scientific meetings of the Society as, on account of the nature of the plates required to illustrate them, are better adapted for publication in the quarto form. They are issued at irregular intervals. Wellows and Corresponding Members, upon payment of a Subscription of £1 1s. before the day of the Anniversary Meeting in each year, are entitled to receive all the Society’s Publications for the year. They are likewise entitled to purchase the Publications of the Society at 25 per cent. less than the price charged for them to the Public. A further reduction of 25 per cent. is made upon purchases of Publications issued prior to 1871, if they exceed the value of five pounds. Fellows also have the privilege of subscribing to the Annual Volume of the ‘ Zoological Record ” for a sum of £1 (which includes delivery in the United Kingdom only), but this privilege only holds good if the subscription is paid before the First of December in each year. Such of these publications as are in stock may be obtained at the Society’s Office (8 Hanover Square, W.), at Messrs. Longmans’, the Society’s publishers (Paternoster Row, H.C.), or through any bookseller. P. L. SCLATER, December 1898, Secretary. PR a ast ae TRANSACTIONS OF THE ZOOLOGICAL SOCIETY OF LONDON. Vor. XV.—Parr 2. LONDON: PRINTED FOR THE SOCIETY, SOLD AT THEIR HOUSE IN HANOVER-SQUARE ; AND BY MESSRS. LONGMANS, GREEN, AND CO., PATERNOSTER-ROW. August 1899. Price 12s. Taylor and Francis, Priaters,] {Red Lion Court, Fleet Street TRANSACTIONS OF THE ZOOLOGICAL SOCTETY OF LONDON. To Fellows. To the Public. arsed BOK eh VOLUME I. (1883-1835, containing 59 Plates). . Price 313 6 . . . 418 O* VOLUME | II. (1835-1841, containing 71 Plates). . , 4 0 0 iO Oe VOLUME III. (1842-1849, containing 63 Plates). . ,, 3 8 3... 411 O* VOLUME IV. (1851-1862, containing 77 Plates). . ,, 6 20... 8 2 6+ VOLUME _ V. (1862-1866, containing 67 Plates). . , 5 4 3 . . . 619 OF VOLUME VI. (1866-1869, containing 92 Plates). . , ll 5 0 peal Ob Ont VOLUME VII. (1869-1872, containing 78 Plates). . ,, 10 4 0 . . .18 12 OF VOLUME VIII (1872-1874, containing 82 Plates). . ,, QO 8 se ahi te eedia: Ll Ost VOLUME IX. (1875-1877, containing 99 Plates). . ,, 12 1 6 .. .16 2 OF VOLUME X. (877-1879, containins 95 Plate) | 10 0 Be. 13 7 OF GENERAL INDEX, Vols. I. to X. (1833-1879) . ,, On 10s ae OOO VOLUME XI. (1880-1885, containing 97 Plates) $ D2 sO ek eee al 2 Oso VOLUME XII. (1886-1890, containing 65 Plates). . ,, DOr Orie rs eta come O) * No copies of these volumes in stock. + Only complete copies of these volumes left in stock. Continued on page 3 of Wrapper. Il. On the Marine Copepoda of New Zealand. By G. Stewarpson Brapy, W_D., D.Sc., LL.D., F.R.S., CM.ZS. Received February 9, 1899, read March 21, 1899. [Puates IX.—XIII.] IN this paper I complete, so far as possible, an account of two collections of New Zealand Entomostraca submitted to me by Mr. G. M. Thomson of Dunedin, and Drs. Meinert and H. J. Hansen of Copenhagen. A former contribution (already published in the ‘ Transactions ’ of this Society, vol. xiv. p. 429) deals with the Ostracoda; the present paper refers to the Copepoda. It is rather remarkable that no Cladocera were noticed in any of the numerous gatherings represented in these collections, and it is also unfortunate that of some of the most interesting species only one or two imperfect examples were found,—in some cases so extremely imperfect that I have made no attempt to describe or figure them. In some other cases of specimens preserved in picric acid, the opacity and extreme brittleness of the subjects made dissection very difficult, and recognition of anything beyond the coarser external features quite impossible. The collections were made for the most part by means of the tow-net at the surface and down to a depth of 6—8 fathoms; a few were from littoral alge. Not much systematic investigation of the marine microzoa of New Zealand appears hitherto to have been made; for the work already done we are indebted chiefly to Mr. G. M. Thomson, who has published papers on the Ostracoda and Copepoda in the ‘Transactions’ of the New Zealand Institute. There can be no doubt that a rich harvest awaits any observer who may be able to devote time and labour to the task. The memoirs quoted in this paper are as follows. They are indicated in the synonymy by the numbers assigned to them in the list :— Borcg, Axet. (1).— Oversigt over de ved Norges Kyster iagttagne Copepoder henhérende til Calanidernes, Cyclopidernes og Harpactidernes Familier. (Vidensk.-Selskabs Forhandlinger, 1864.) Brapy, G. S. (1).—A Monograph of the free and semi-parasitic Copepoda of the British Islands. (Ray Society, 1878-80.) (2).—Report on the Copepoda collected by H.M.S. ‘Challenger’ during the years 1873-76. (1883.) (3).—Description of a new Species of Cyclops. (Sixth Annual Report of the Fishery Board for Scotland.) 1888. (4).—A Revision of the British Species of Freshwater Cyclopide and Calanide. (Natural History Transactions of Northumberland, Durham, and Newcastle-upon-Tyne, vol. xi.) 1891. VoL. xv.— Part ul. No. 1.—August, 1899. G 52 DR. G. 8S. BRADY ON THE Craus, C. (1).—Die frei lebenden Copepoden, mit besonderer Berticksichtigung der Fauna Deutschlands, der Nordsee und des Mittelmeeres. Leipzig, 1863. Dana, J. D. (1)—American Journal of Science, vol. viii. (1849). (2).—Crustacea of the United States Exploring Expedition (1852). Girssrecut, W. (1).—Systematik und Faunistik der pelagischen Copepoden des Golfes von Neapel und der angrenzenden Meeres-abschnitte. Berlin, 1892. Kramer, A. (1).—On the most frequent Pelagic Copepoda and Cladocera of the Hauraki Gulf. (Transactions and Proceedings of the New Zealand Institute, 1894, vol. xxvii.) Lussock, J. (1).—On two new Subgenera of Calanide. (Annals & Magazine of Natural History, March, 1853.) Scorr, THomas (1t).—Report on Fntomostraca from the Gulf of Guinea, collected by John Rattray, B.Sc. (Transactions of the Linnean Society of London, 2nd ser. Zool. vol. vi. January, 1894.) Tuomson, G. M. (1).—On the New Zealand Copepoda. (Transactions and Proceedings of the New Zealand Institute, vol. xv.) 1882. Section I. GNATHOSTOMA ‘Yhorell. Family CALANID. Genus Cauanus Leach. 1. CaLaNus FINMARCHICUS (Giinner). Taken in the surface-net, Otago Harbour. Genus Paracauanus Boeck. 1. ParacaLanus parvus (Claus). (Plate IX. figs. 1, 2.) In the surface-net, Otago Harbour and Port Chalmers; Hauraki Gulf (Dr. Kramer). Genus CLAUSOCALANUS Giesbrecht. 1. CLAUSOCALANUS ARCUICORNIS (Dana). 1849. Calanus arcuicornis Dana (1). 1852. Calanus arcuicornis Dana (2), p. 1056, pl. Ixxii. figs. 7 a, 5. 1892. Clausocalanus arcuicornis Giesbrecht (1), pp. 185, 1938, pl. i. fig. 14, pl. x. figs. 1-19, pl. xxxvi. figs. 29-31. 1894. Clausocalanus arcuicornis Scott (1), p. 78, pl. vill. figs. 88-47. Taken in the surface-net, Port Chalmers. Recorded also by Dr. Kramer from the Hauraki Gulf. This species has been sufficiently figured by Giesbrecht and Scott (oe. cit.). 2, CLAUSOCALANUS FuRCATUS (G. S. Brady). (Plate IX. figs. 3-7.) 1883. Drepanopus furcatus G. S. Brady (2), p. 77, pl. iv. figs. 1, 2, pl. xxiv. figs. 12-15. 1892. Clausocalanus furcatus Giesbrecht (1), p. 194, pl. xxxvi. figs. 32, 33, 35. MARINE COPEPODA OF NEW ZEALAND. 33 The genus Drepanopus which, as proposed by the present writer, included two species, D. furcatus and D. pectinatus, has been divided by Dr. Giesbrecht into two genera—Drepanopus and Clausocalanus, to the latter of which D. furcatus is transferred. The original description dealt with females only, no males having been observed in the ‘Challenger’ collections. I therefore give here figures of some of the more important diagnostic structures. In some of the New Zealand gatherings both of these species occur together, and when associated with C. areuicornis and Paracalanus parvus in various stages of growth it becomes by no means an easy matter to separate them. The most conspicuous characters are to be found in the fusions of the antennal joints, the relative lengths of the abdominal segments, and in the structure of the fifth pair of feet in both sexes. The anterior antenne of the female in C. furcatus reach somewhat beyond the extremity of the cephalothorax and are 24-jointed, having all the joints perfectly distinct; in the male (tig. 5) the first and second joints are coalescent, so also are the eighth, ninth, and tenth ; all the rest are distinct. ‘The fifth pair of feet in the female (fig. 4) are very short, two-jointed and ‘alike on both sides ; in the male (fig. 7) the foot of the right side is five-jointed, and about half as long as the abdomen (considerably shorter than in C. arcuicornis), that of the left side is two-jointed and only as long as the first joint of the right foot. The abdomen is four-jointed in both sexes (figs. 3, 7), the first segment being very short, the remaining segments nearly equal in the female, while in the male the third is rather reduced and the fourth rather increased in length. Hab. Taken plentifully in the surface-net in Otago Harbour and at Port Chalmers, and in the net at 7 fathoms off Gisborne. Genus AcaRrTIA Dana. 1. ACARTIA ENSIFERA, sp. nov. (Plate IX. figs. 8-15.) Female.—Seen from above (fig. §) the cephalothorax is elongated, slightly narrowed towards the front, which is rounded but obscurely angulated in the middle, posterior angles rounded off and entirely without spines ; rostrum obtuse, not furcate ; abdomen about one-third as long as the cephalothorax (fig. 9), composed of three segments, first segment very wide in the adult and equal to the united lengths of the second and third segments; caudal stylets slender, more than twice as long as broad, longer than either of the two preceding segments. ‘The antennules reach slightly beyond the posterior extremity of the cephalothorax. Fifth pair of feet (fig. 10) two-jointed, basal joint broad, quadrate, its outer margin prominent and mucronate in the middie, whence arises a single very long seta; second joint bulbous at the base and tapering to a long setiform apex. ‘Terminal spines of the swimming-feet (fig. 11) very slender and sword-shaped, with finely pectinate margins. Length 1-2 millim. G2 34 DR. G. 8. BRADY ON THE Male.—Abdomen composed of four segments (fig. 14), the first very short, second twice as long, third and fourth rather shorter; caudal stylets rather longer than the preceding segment. Antennules more nodose than in the female owing to the distal dilatation of many of the joints (figs. 12, 13), some of which are more or less distinctly coalescent. ‘The fifth foot of the right side is longer than that of the left (fig. 15) ; the inner margins of the first and second joints have large lobose protuberances, and the last joint is bent upon itself at a right angle, ending in a sharp point; on the left side the first joint is broad and quadrangular and has a long seta attached to its angulated outer margin; second joint rather longer and narrower, third dilated at the base and bearing on the inner margin of the narrowed distal extremity three small setiform lamine. Hab. Plentiful in surface gatherings from Otago Harbour, Port Chalmers, Bay of Islands, and Napier; also in the net at a depth of 7 fathoms off Gisborne. ‘These gatherings consisted almost entirely of females, males being extremely rare; almost all had attached spermatophores. A form described in the ‘ Challenger’ Report under the name of Acartia denticornis is very nearly allied to this species, but the antennules of A. ensifera possess no well-marked teeth, though these are constantly present in both sexes of A. denticornis. The antennules in A. denticornis are fully as long as the animal itself, while in A. ensifera they reach scarcely further than the extremity of the cephalothorax. The peculiar enlargement of the first segment of the abdomen in the female A. ensifera is likewise characteristic. Genus Temora Baird. 1. TEMORA TENUICAUDA, sp. nov. (Plate IX. figs. 16-23.) Female.—Cephalothorax robust, width great in proportion to the length (fig. 16), broadly rounded in front, abruptly narrowed behind ; angles of the posterior extremity rounded off, not at all produced or spinous. Abdomen short, four-jointed, first segment about twice as long as either of the following three, and in the adult having a pouch-like process on the ventral surface; caudal stylets cylindrical, extremely long and slender, from eight to ten times as long as broad, bearing one small seta on the outer margin beyond the middle, and three at the apex, one of which is spathulate at its base and nearly as wide as the stylet itself (fig. 18). Antennules sparingly clothed with extremely short sete, 24-jointed, slender, about as long as the cephalothorax (figs. 16,17). Feet of the fifth pair (fig. 19) simple, rudimentary, three-jointed, the last joint nearly twice as long as either of the preceding two and ending in three small claw-like sete. Terminal spines of the outer branch of the swimming-feet (fig. 20) well-developed and stout, the principal one strongly serrated with about fourteen teeth on its inner margin. Length 1°75 millim. MARINE COPEPODA OF NEW ZEALAND. 35 Male.—Cephalothorax and abdomen as in the female except that the abdomen is five-jointed ; none of the caudal sete have dilated bases (fig. 23). Eighteenth joint of the right antennule (fig. 21) very faintly serrulated. Fifth foot (fig. 22) of the right side strongly prehensile, basal joint large, subquadrate, and produced into a Jong curved claw which opposes the three-jointed outer ramus; first joint of the outer ramus about twice as long as broad, second and third not much longer than broad and bearing respectively one and two short marginal spines ; fifth foot of the left side short, simple, three-jointed, last joint forming a short curved unguis. Hab. Taken plentifully in the surface-net in the Bay of Islands and Otago Harbour, also at a depth of 6 fathoms in Auckland Harbour. Of all the specimens examined not one was found in perfect condition, and all were very brittle and opaque owing to the action of picric acid preservative. The species is very closely allied to the well-known 7’. longicornis of European seas, but there is no serration of the right male antennule, except on the eighteenth joint ; the widely- dilated tail-seta of the female is characteristic, and the serratures of the spines of the swimming-feet are much fewer and larger than in 7. longicornis. Genus CENTROPAGES Kroyer. 1. CENTROPAGES DISCAUDATUS, sp. nov. (Plate X. figs. 1-7.) ? Centropages typicus Kroyer, var. aucklandicus Krimer (1), pl. xv. figs. 1-5, p. 217. Female.—Antennules very slender, 24-jointed, reaching considerably beyond the posterior end of the cephalothorax. Infero-posteal angle of the cephalothorax produced on each side into a long spine. Abdomen (fig. 6) three-jointed; first joint dilated proximally and bearing on its dorsal margin a slender plumose seta; second joint longer, much dilated proximally ; third very short, constricted proximally ; furcal joints short, not much longer than broad, almost square as seen from above (fig. 7), distally truncated and bearing five rather short terminal sete, three of which are bulbously dilated at the base. Outer branch of the fifth pair of feet (fig. 5) having its second joint produced inwardly into a very stout and long spine. Male.—Infero-posteal angles of the cephalothorax spinous, but not so strongly as in the female; abdomen four-jointed, all the joints subcylindrical (fig. 1), furcal joints cylindrical, about twice as long as broad, terminal sete not swollen at the base; the first abdominal segment short, the following three nearly equal, not very much longer ‘han broad and altogether devoid of spines. Basal joints of the antennules (fig. 1) almost devoid of marginal processes or teeth; joints from the 15th to the 18th much swollen, the 15th and 16th each bearing a strong marginal spine; the three following joints densely pectinated with fine seta (figs. 2, 3), which, at the base of the 17th joint, are replaced by a series of small squared tubercles. Outer branch of the fifth pair of feet of the right side (fig. 4) terminating in a chelate joint, the blades of which are 36 DR. G. S. BRADY ON THE long, slender, only moderately curved, and not at all crenulated, the outer blade having a markedly constricted median neck ; outer branch of the left foot devoid of marginal Length 1°6 millim. Hab. Taken in the surface-net at Port Chalmers and in Otago Harbour. sete, simple, and terminating in two very small apical sete. This may perhaps be the form referred to by Dr. Kramer as Centropages typicus, var. aucklandicus, though in some respects it does not agree accurately with his figures and description. It cannot, I think, be properly referred to any variety of C. typicus, though the pectinated armature of the male right antennule agrees exactly with that species. ‘To make my meaning clearer I give some of the more important characters of the two species in a tabular form. It is, however, very closely allied to C. brachiatus (Dana), and the females of the two species may not be easily distinguishable, but the armature of the antennule, together with the structure of the fifth pair of feet in the male, presents very sufficient distinctive characters. Basal joints of the antennules... 15th and 16th joints of the right male antennule. First abdominal somite in female. Caudal segments in female...... Prehensile branch of right fifth foot in male. 2. CENTROPAGES PECTINATUS, Sp. NOV. C. typicus. First, second, and fifth joints in both sexes have strong mar- ginal spines. Sixteenth joint has a strong spine.’ Two slender spine-like set on ventral and a shorter one on dorsal aspect. Twice as long as broad; setz not swollen at base. Inner claw slightly pectinated ; outer not distinctly con- stricted in middle. C. discaudatus. No well-developed spines on the basal joints in either sex. A strong spine on both joints. One dorsal seta. Scarecly longer than broad ; sete swollen at base. Inner claw not pectinated ; outer distinctly constricted. (Plate IX. figs. 24-27.) Female.—Abdomen slender, four-jointed (figs. 26, 27), the first joint wider than the rest, bearing on its dorsal surface a brush of several fine hairs and at the posterior angle a single slender spine; caudal segments about thrice as long as broad, two or three of the principal terminal sete dilated basally (fig. 27 0). Lateral spines of the outer branches of the swimming-feet lancet-shaped, with very finely pectinated margins ; median joint of the outer branch of the fifth pair (fig. 25) produced internally into a ‘ Giesbrecht figures a rather feeble spine also on the fifteenth joint, but I have not been able to find it in any of my specimens. MARINE COPEPODA OF NEW ZEALAND. 37 long falcate spine, the concave margin of which is distinctly pectinated; the basal joint of the fourth (?) pair of feet bears also a very long falcate spine (fig. 24). Male unknown. This description is extremely imperfect, being based upon only two battered specimens which were taken—one off Napier in the net at a depth of 8 fathoms, the other at the surface in Otago Harbour. The characters seem clearly to distinguish this from any described species. Subfamily PoNTELLIN «A, Genus Lasibocera Lubbock. 1. Lasrbocura cervi Kramer. (Plate X. figs. 8-14.) Labidocera cervi Kramer (1), p. 218, pl. xvi. figs. 6, 7. Female.—Cephalothorax subtrunecate posteriorly, the ventral angles produced and mucronate (fig. 9). Abdomen four-jointed, the first, second, and third segments about equal, fourth about half as long as the preceding; caudal lamine rather longer than broad, obliquely quadrate (figs. 9, 15); terminal sete scarcely as long as the abdomen. Antennules reaching to the posterior extremity of the cephalothorax. Fifth pair of feet (fig. 10) alike on both sides, two-branched, the internal branch composed of one small joint, the outer of one much larger joint which bears three moderately large lateral teeth and two smaller and much more slender ones at the apex. In the immature condition, however (fig. 11), the outer branch consists first of three and afterwards of two joints. Length 2:5 millim. Male.—The infero-posteal angles of the cephalothorax (fig. 15) are not quite so strongly spined as in the female; the abdomen five-jointed (fig. 13). The armature of the right antennule consists of a fine pectination of the seventeenth and eighteenth and of the coalescent nineteenth, twentieth, and twenty-first joints: the twenty-second joint is prolonged internally, forming a sharp process which extends beyond the apex of the penultimate joint. The right foot of the fifth pair (fig. 14) is strongly prehensile, the basal joint simple, second joint attached near the middle by a hinge- articulation, its outer portion forming a slender flexuous process, its inner portion forming a broad lamina and giving attachment at its apex to a long and slender falcate joint which opposes the outer process of the second joint; the foot of the left side is simple, slender, three-jointed, nearly as long as that of the right side, its terminal joint having a few short apical teeth and near the distal extremity a diffuse pubescent patch. Length 2°8 millim. Hab. In the surface-net, from Otago and Akaroa. Though my observations differ in some respects from those of Dr. Kramer, I hav« 38 DR. G. 8S. BRADY ON THE no doubt that this is the species named by him Labidocera cervi. Dr. Kramer says that the abdomen of the female is two-jointed, and in some specimens I have found the abdominal segments so interfused that it is difficult to make out the number: there may, indeed, be no visible separation, but in some there are distinctly four segments. In the allocation of joints to the right male antennule, I have followed Dr. Giesbrecht (Labidocera nerii, &c.), not having myself been able to make them out distinctly in the New Zealand specimens. Family CYCLOPID. Genus Cyctoprs (O. F. Miiller). 1. Cyctors Ewart G.S. Brady. (Plate X. figs. 15-17.) 1888. Cyclops ewarti G. S. Brady (3), pl. vii. figs. 1-6. 1891. 55 » idem (4) pl. vu. figs. 4-7. One specimen of a Cyclops which I cannot in any way distinguish from C. ewarti was found in a surface-net gathering from Otago Harbour. This capture is peculiarly interesting, inasmuch as the species was originally described from specimens taken in the Firth of Forth, and it seemed doubtful whether they might not have made their way thither from some neighbouring freshwater habitat. No other instance is on record, so far as I know, of a true Cyclops having been found living in the sea, and it is very remarkable that this New Zealand example, of the purely pelagic character of which there can be no doubt, should belong to the same species. I have thought it well to figure here some of the parts of the Otago specimen. The only difference between it and the Scottish examples is the greater width of the abdomen, which may, however, be accounted for by pressure. An interesting peculiarity of the species, in which both northern and southern forms agree, is the bipectinate character of the terminal spine of the inner branch of the fourth pair of feet (Pl. II. fig. 17a). In all other species these pectinations are represented by very fine sete. Genus O1tHona Baird. 1, OITHONA SPINIFRONS Boeck. Oithona spinifrons Boeck (1), p. 25. 9 » G. 8. Brady (2), p. 90, pl. xiv. figs. 1-9, pl. xxiv A. figs. 1, 2. Frequent in surface gatherings from Otago Harbour, and in the tow-net at 7 fathoms off Gisborne. I cannot distinguish these specimens from those which I have already (doc. cit.) described and figured under the above name. Dr. Giesbrecht disagrees with my MARINE COPEPODA OF NEW ZEALAND. 39 reference and thinks that they belong to 0. spinifrons Claus, but in the length of the antennules and some other characters they seem to me to agree more closely with Boeck’s species. Family HARPACTICID. Subfamily LoNGIPEDIINA. Genus Ecrinesoma Boeck. 1. EcTINosoMA AUSTRALE, sp. nov. (Plate X. figs. 18-25.) Antennules seven-jointed (?); first three joints stout and short, the succeeding joints much more slender (fig. 18); antenne (fig. 19) nearly as long as the antennules, three- jointed, secondary branch three-jointed, the second joint very short. Mandibles (fig. 20) with one large terminal tooth and four short spinules; palp long, two-jointed, and bearing on the first joint a minute two-jointed branchlet. The mouth-organs (figs. 21, 22) and swimming-feet (fig. 23) present no special features. Feet of the fifth pair (fig. 24) much subdivided ; inner segment digitiform and bearing two apical spine-like sete, one of which is long, the other very short; outer segment laciniated, its inner lobe bearing, like the neighbouring segment, two unequal sete, external to which are three short processes bearing single apical sete, the first and third short, the second long and slender. Candal lamine (fig. 25) very short and wide; sete slender, non-plumose, the central one about twice as long as the two principal laterals. Length ‘66 millim. Hab. Otago Harbour, between tide-marks. The distinctions between many so-called species of Hvtinosoma are extremely slight, and in this case the fifth foot is the only member which presents very definite characters. In the present state of our knowledge it seems very doubtful whether this, as well as some others, should not rank rather as varieties than as distinct species. Subfamily TacHIDIIN &. Genus Evurrerre Claus. 1. EvuvTerre craciuis Claus. 18638. Euterpe gracilis Claus (1), p. 110, pl. xiv. figs. 1-13. 1880. ea » G.S8. Brady (1), p. 22, pl. xl. figs. 1-16. 1892. », acutifrons Giesbrecht (1), p. 555, pl. xliv. 1852. ? Harpacticus acutifrons Dana (2), p. 1192, pl. Ixxxiil. figs. 11 a, 6. This species was found plentifully in most of the gatherings. In the surface-net from Otago Harbour and Port Chalmers; in the net at 6 fathoms, Auckland Harbour ; 7 fathoms, off Gisborne; 8 fathoms, off Napier. VoL. XV.—PaRT 11. No, 2.— August, 1899. H 40 DR. G. §. BRADY ON THE Subfamily AMYMONIN &. Genus AmMyMonz Claus. 1. AMYMONE cLAUsII ‘Thomson. 1882. Amymone clausii Thomson (1), p. 98, pl. v. figs. 1-8. Taken in the surface-net, Bay of Islands. Subfamily CANTHOCAMPTIN &. Genus LaopHontE Philippi. 1. LAOPHONTE SERRATA (Claus). 1863. Cleta serrata Claus (1), p. 123, pl. xv. figs. 18-20. 1880. Laophonte serrata G. S. Brady (1), vol. ii. p. 71, pl. Ixxiii. figs. 1-14, Hab. On alge in Lyttelton Harbour. 2. LAOPHONTE MEINERTI, sp. nov. (Plate XI. figs. 1-10.) Antennules of the female short, seven-jointed (fig. 2); the comparative lengths of go Bio eo Oo G aid CoS 23 6 slender ; hand narrow, its external margin obscurely angulated near the middle ; terminal unguis slender, slightly curved, rather longer than the hand. First pair of feet (fig. 5) very slender; outer branch three-jointed, scarcely more than one-third as long as the inner branch, exclusive of the terminal unguis; inner branch much elongated, destitute of setee, with a long and slender apical unguis. Swimming-feet (fig. 6) small and slender. Feet of the fifth pair (fig. 8) small, foliaceous, each joint bearing several marginal setee, one of which is much longer than the rest. Length ‘75 millim. the joints as in the following formula: Posterior foot-jaws (fig. 4) Second joint of the male antennule produced posteriorly into a short conical spine ; fourth joint very much enlarged and forming a quadrate vesiculiform swelling, from the anterior border of which arises a stout curved sensory filament (fig. 3). Hab. On alge in Lyttelton Harbour. Two specimens. Two further specimens, which may perhaps belong to this species, were found in a gathering from between tide-marks in Otago Harbour. Figures of the posterior foot- jaws and of the first and fifth feet of these specimens are given in Plate XI. figs. 9, 10. These do not altogether correspond with similar parts of LZ. meinerti, but the differences may possibly be varietal only. MARINE COPEPODA OF NEW ZEALAND, 4] Subfamily HARPACTICINA. Genus Dactyopus Claus. 1. Dacrytopus tisBorpgEs Claus. 1863. Dactylopus tishoides Claus (1), p. 127, pl. xvi. figs. 24-28. 1880. 0 #0 G. S. Brady (1), p. 106, pl. liv. figs. 1-16. Hab. Between tide-marks, Otago Harbour. 2. DACTYLOPUS HANSENI, sp. nov. (Plate XI. figs. 11-18.) Very similar to D. tisboides in general appearance (fig. 11), but rather larger. Antennules of the female (fig. 12) eight-jointed, rather densely setiferous; fourth joint bearing a stout sensory filament. Posterior foot-jaw (fig. 14 a) slender; hand bearing on each margin a single seta and a fringe of minute cilia; terminal claw slender, about as long as the hand itself. Terminal claws of the first pair of feet (fig. 15) long and slender, as are also the spines of the outer branch; penultimate joint of the inner branch bearing near the distal extremity a single small plumose seta. Fifth pair of feet of the female (fig. 18) large, the inner lamina considerably shorter than the outer one ; apex finely ciliated, subtruncated, and bearing two long and two short sete; distal extremity of the outer lamina bearing six sete of various lengths and widely separated, the lateral margins ciliated; no cilia between the large sete. Length -85 millim. Male.—The antennule (fig. 13) is geniculated, but none of the joints are very tumid ; terminal joint of the inner branch of the second pair of feet (fig. 16) forming two curved finger-like processes, from near the base of which spring two long plumose sete : fifth pair small (fig. 17); inner lamina shorter than the outer, with two short spine- like apical sete; outer bearing six sete, the two apical ones much longer than the rest. Hab. Lyttelton Harbour: on alge, 1 specimen; tow-net at 6 fathoms, 1 specimen. The nearest allies of this species seem to be D. cinctus (Claus) and D. stromii Baird, but both of them differ distinctly in minor points. Genus FLAVIA, gen. nov. Like Dactylopus, except that the inner branch of the second pair of feet consists of one greatly elongated joint, and is entirely destitute of spines or claws (ELE X. fig. 30). Anterior and posterior foot-jaws simple, prehensile, with long terminal ungues. H 2 42 DR. G. S. BRADY ON THE 1. FLAVIA CRASSICORNIS, sp. nov. (Plate X. figs. 26-31.) Antennules very short and stout, two-jointed, moderately setiferous (fig. 26). Feet of the first pair as in Dactylopus, the terminal ungues thick and hook-like (fig. 29): inner branch of the second pair (fig. 30) longer than the outer branch, very slender, with truncate unarmed apex; near the base there is on each margin an angular prominence, from which springs a minute seta, and near the distal extremity on the outer margin a rounded non-setiferous tubercle, nearly opposite to which, from the inner margin, arise two long and delicate plumose sete. Anterior foot-jaw simple, elongated, with a slightly curved terminal unguis (fig. 27); posterior somewhat similar, the hand armed with a long slender marginal spine and a strong curved terminal unguis (fig. 28). A single specimen of this species occurred in a gathering from among alge in Lyttelton Harbour. The posterior antenne, mandibles, and maxilla were not seen distinctly enough to permit of their being figured, but in general conformation are similar to those of most Harpacticide. Genus PHROS0, gen. nov. Body slender; in general appearance like Canthocamptus. First four pairs of feet with both branches three-jointed ; feet of fifth pair foliaceous, unbranched. Internal branch of antenne two-jointed. Mandible strongly toothed, palp two-branched. Anterior and posterior foot-jaws simple, with large terminal ungues and no palp. 1. PHROSO GRACILIS, sp. nov. (Plate XI. figs. 19-27.) Female.—Antennules eight-jointed, rather densely setiferous ; first four joints stout, last three much more slender (fig. 20): lengths of the joints as in the following 15 Oe BO. hoe. NPGS eyhaney | fy) 4 AEN Sy Antenne (fig. 21) nearly as large as the antennules, three-jointed ; to the basal joint is attached a two-jointed internal branch. Mandibles well developed (fig. 22), with sharp slender teeth and a palp composed of a large basal joint and two single-jointed branches. Maxille (fig. 23) small, lobose (imperfectly seen). Anterior foot-jaw (fig. 24) slender, simple, geniculated, with a stout terminal unguis and three plumose sete attached to the second joint. Posterior foot-jaw (fig. 25) larger than the anterior, composed of a single stout joint which bears a long falcate, terminal unguis. Basal joint of the fifth pair of feet (fig. 27) short and broad, and bearing on each lateral angle a long seta; terminal joint subovate, twice as long as broad; apex narrow and truncate; external formula : from the fifth joint arises a stout sensory filament. margin ciliated above, and below the middle having two long widely separated sete. Length 1:3 milliim. Male unknown. Hab. One specimen taken in the surface-net, Port Chalmers. MARINE COPEPODA OF NEW ZEALAND. 45 Genus THALESTRIS Claus. 1. ‘THALESTRIS FORFICULA Claus. 1863. Thalestris forficula Claus (1), p. 181, pl. xvii. figs. 7-11. 3 o Thomson (1), p. 104, pl. x. figs. 16-21. Otago Harbour, between tide-marks; “ dredged abundantly in Dunedin Harbour ” (Thomson). 2. THALESTRIS CILIATA, sp. nov. (Plate XI. figs. 28-33.) Robust ; lateral margins of the abdominal segments clothed with short sete (fig. 33). Antennules nine-jointed (fig. 28); joints subequal, the length represented by the ; | LO Bes BeB07 58.0 TA eae ; following formula: 3 S66 46-33 q Posterior foot-jaw forming a subovate, elongated hand, the anterior margin of which bears a continuous fringe of short, closely-set sete ; ‘terminal unguis stout, falcate, more than half as long as the hand (fig. 29). Outer branch of the first pair of feet (fig. 30) with densely pectinated margin: one long and stout spine at the apex of the first and one on the middle of the second joint; the last joint has two fine marginal sete and one very long, slender seta at the apex, between which are two stout faleate claws with hook-like projections at their bases: the outer branch has a single seta near the distal extremity of the first joint, and at its apex are two ungues, one large and one small. The marginal spines of the swimming-feet (fig. 51) are long and slender, and the outer edges of the joints of both branches are ciliated. The two lamine of the fifth pair of feet (fig. 52) are of equal length; the outer lamina ovate, with finely ciliated margins, four long sete at the apex and two on the outer edge; the inner lamina has five principal sete, two marginal and three apical, the intervals between which are ciliated. Male unknown. Hab. Akaroa Harbour. One specimen only. 3. THALESTRIS AUSTRALIS, sp. nov. (Plate XII. figs. 1-5.) Autennules nine-jointed, slender, the comparative lengths of the joints as in the following formula: : a! : : : : ; : : : i : : : : (fig. 1). (fig. 2) subovate, bearing near the middle of the anterior margin a single spine-like seta, terminal unguis moderately stout and reaching a little beyond the middle of the hand. Outer branch of the first pair of feet (fig. 5) shorter than the inner, stout, its inner margin bearing rather in front of the middle a stont plumose seta, outer margin a large median spine and four apical ungues which successively increase in length from first to last; the inner branch has a long. slender, apical unguis and a much shorter seta; anterior margins of the larger ungues of both branches finely pectinated. Marginal spines of the swimming-feet (fig. 4) long and slender, with Hand of the posterior foot-jaw 44. DR. G. S. BRADY ON THE blunt apices ; external margins of all the joints ciliated. LLamine of the fifth pair of feet (fig. 5) broad, subovate, nearly equal in length; margins of the external lamina finely ciliated, and bearing six subequal apical sete; inner lamina with five sete on its subtruncate apical margin. Abdominal segments slightly overlapping each other, but altogether devoid of sete or serrulations Male unknown. Hab. Otago Harbour, between tide-marks. One specimen only. Genus Harpacticus Milne-Edwards. 1. HARPACTICUS GLABER, sp. nov. (Plate XII. figs. 6-11.) Female.—In general appearance like H. chelifer. Antennules (fig. 6) slender, moderately setiferous, eight-jointed ; lengths of the joints as in the following formula : Wee Be Bo Gav S IO i Gat Ba a small two-jointed branch, extremity of the last joint abruptly truncated and bearing two long, rigid, finely pectinated spines and four slender geniculated sete ; posterior foot-jaw (fig. 8) almost exactly similar to that of H. chelifer. Outer branch of the first pair of feet elongated, two-jointed, terminating in three subequal, blunt, slightly curved and very finely pectinated claws; outer margins very sparingly ciliated ; each joint bearing within a short distance of its distal end a single short marginal seta: inner branch terminating in a strong slightly curved claw; the long first joint slightly ciliated and having near its apex a slender seta. Feet of the fifth pair two-jointed (fig. 10), basal joint wide, its inner portion distally produced and giving attachment to three long sete; terminal joint much narrower, margins ciliated, with three long sete at the apex and one on the outer edge. ‘The spines of the swimming-feet, like those of the first pair and of the antenne, are minutely pectinated. ‘The margins of the abdominal segments (fig. 11) are smooth except for a few minute hairs at the lateral angles and on the caudal lamin; principal tail-setee very long. Hab, Otago Harbour. One specimen taken in the surface-uet. From H. chelifer this species may be distinguished by the number and comparative lengths of the joints of the antennules, by the armature of the antenne, the non- spinous margins of the first pair of feet, the different build of the fifth pair, and the almost smooth margins of the abdominal somites. There seem to be points of difference also in the mandibles and maxille, but these I have not clearly made out. Antenne (fig. 7) stout, nearly as long as the antennules, bearing Genus Pevripium Philippi. 1. P<IDIUM NOVAI-ZEALANDIM, sp. nov. (Plate IX. figs. 28, 29; Plate XII. figs. 12-15.) This is in general build and appearance exactly like the common European species P. interruptum, Goodsir, the only important differences being found in the posterior MARINE COPEPODA OF NEW ZEALAND. 45 foot-jaw and fifth pair of feet. The manus of the second pair of foot-jaws (Pl. IX. fig. 29) is ovate, elongated, the anterior margin fringed with a continuous series of short sete ; terminal unguis long and slender. Feet of the fifth pair (PI. XII. fig. 14) stout, simple, with truncate apex, which bears three short, stout, and very blunt spines, the outermost of which is attached at a little distance from the apex: the inner margin bears beyond the middle a single seta; the outer margin has also a single seta much nearer the base, and is slightly pubescent between that point and the apex. The shell is extremely hard and thick, and closely beset with small circular puncta. Colour dark brown. Length ‘85 millim. Hab. Taken in the surface-net, Port Chalmers and Otago. Subfamily Ipyin &. Genus Ipya Philippi. 1. Ipya Furcata (Baird). Tisbe furcata Claus, Lilljeborg. Canthocamptus furcatus Baird. Idya furcata Boeck, Brady, Thomson. On alge in Lyttelton Harbour. ‘Common in shore-kelp and rock-pools near Dunedin ; also on kelp in Paterson’s Inlet” (Zhomson). Genus ScUTELLIDIUM Claus. 1. ScUTELLIDIUM PLUMOSUM, sp. nov. (Plate XII. figs. 16-21, 25-25.) In general form and appearance indistinguishable from the British S. tishoides. The setose investiture of the first foot (fig. 20) is, however, more dense, and the spinous armature of the last joint of the outer branch of the swimming-feet (fig. 21) is also different, there being on the outer margin towards the apex two plumose lancet-shaped spines with intermediate small hairs, while in S. tésbotdes (fig. 22) there are three stout, non-plumose, dagger-shaped spines with no intermediate hairs. Fifth foot in the female (fig. 25) elongated, slender, subcrescentic, pointed at the extremity, with setose margins, but not pubescent on the surface; in the male (fig. 23) small, subquadrate, with obliquely truncated apex, which bears three long and _ several shorter setee. ‘The antennules in the male are geniculated (fig. 16): from the second joint springs a small biarticulate peduncle, which bears three stout sensory filaments, and to the antepenultimate joint is attached a much longer and stouter sense-organ. Length :88 millim. Hab. Vhis species occurred sparingly in surface gatherings from Otago and Port Chalmers, and in gatherings from seaweeds in Akaroa and Lyttelton Harbours. 46 DR. G. S. BRADY ON THE The males of §. plwmosum seem, like those of S. tisboides, to be scarce. I have seen only one specimen, some parts of which are here figured. Of S. tisboides I have hitherto seen only one imperfect male specimen; and as Dr. Claus has only very partially described or figured it, one may suppose that his experience is similar. Subfamily PoRCELLIDIIN A. Genus PorcELLIDIUM Claus. 1. PorceLLipium FULVUM Thomson. 1882. Porcellidum fulvum Thomson (1), p. 107, pl. vi. figs. 10, 11, pl. vu. figs. 8-13. On alge in Lyttelton Harbour. Section I]. PH&CTLOSTOMA Thorell. Family CORYCAIDAi Dana. Genus Corycaus Dana. 1. Corycamus Ropusta Dana (?). Specimens which I refer doubtfully to this species were taken in the surface-net off Port Chalmers and in Otago Harbour. Genus PatRocopg, gen. nov.! Antennules five-jointed ; antenne three-jointed, simple; mandibles simply toothed ; posterior foot-jaw forming a prehensile hand with two slender terminal ungues. ‘I'wo pairs of swimming-feet, each foot consisting of two one-jointed branches. 1. PavRocopE RoBusta, sp. nov. (Plate XIII. figs. 1-9.) Antennules (fig. 3) short, five-jointed, sparingly setiferous, joints nearly equal in length. Antenne nearly as long as the antennules, simple, three-jointed, last joint shorter than the preceding, subquadrate, with six sete arising from the subtruncate apex, and on the proximal half of the posterior margin a series of about six small, equal, pellucid, lancet-shaped sete; on the distal half two short and two rather longer, stout, curved, plumose sete (fig. 4). Trophi very small; mandibles without a palp (?), ‘ quvpos, few; Kw, an oar. MARINE COPEPODA OF NEW ZEALAND. 47 apex broad and bearing an even series of small teeth (fig. 5). Maxillee composed of four digitiform lobes (fig. 6), which are armed with numerous spine-like divaricate seta. Anterior foot-jaws laminar, bearing numerous lateral sete, but not divided into distinct lobes (fig. 7): posterior foot-jaws two-jointed, basal joint short and stout, with three rigid spiniferous sete at the anterior distal angle; second joint forming a strong subtriangular hand, to the anterior angulated margin of which are attached two long stout sete, to the narrow distal extremity are hinged two long, biarticulate, slender ungues, which are armed with spine-like marginal sete (fig. 8). Two pairs of swimming-feet (fig. 9), two-branched, each branch consisting of a single broad lamina; distal and outer margins of the inner branch setiferous; outer branch with four large lancet-shaped spines on the outer margin, at the apex a larger spine, and on the inner margin three long sete. The cephalothorax is equal in length to the abdomen, and consists of three segments only, the first two being coalescent (figs. 1,2); the abdomen is two- jointed and suddenly narrower than the cephalothorax: to the posterior angles of the last thoracic segment are attached two small sete ; angles of the first abdominal segments produced backwards, second segment twice as long as the preceding, con- stricted in front; caudal segments very short. Length ‘88 millim. Hab. Of this very interesting species two specimens were found in a surface-net gathering from Otago Harbour, and one in the net at 7 fathoms from off Gisborne. The specimen described by Mr. T. Scott under the generic name Saphirella seems to be, except as to the mouth-organs, very similar to the present, and I regret that, owing to the Saphirella having passed out of Mr. Scott’s hands, I have not been able to compare the actual specimens. Mr. Scott's descriptions may be found in his paper on Entomostraca from the Gulf of Guinea (Transactions of the Linnean Society, 2nd ser. Zool. vol. vi. pt. 1, 1894). [Closely allied to Pauwrocope is a form represented in these gatherings by only two imperfect specimens, which, like Pawrocope, were taken at the surface in Otago Harbour. I am unable, for want of suitable material, to give a complete account of this species, and it seems not unlikely that it may really represent the male of Paurocope, the general resemblance being very close, more especially in the swimming- feet and caudal lamine. ‘The antennules and antenne are, however, very different, but this difference, to a lesser extent, is likewise seen in the closely related genera Sapphirina and Coryceus. A very remarkable character is found in the presence of two simple lenses (ocelli) at the bases of the second pair of foot-jaws. Figures of the animal and some of its appendages are given in Plate XIII. figs. 10-17. Provisionally I propose the name Centromma thomsoni for this species. ] VOL. Xv.— Part u. No. 38.—August, 1899. I 48 DR. G. S. BRADY ON THE Section III. STPHONOSTOMA Thorell. Family ENTOMOLEPID A, fam. nov. Genus ENToMOLEPIS ', gen. nov. Dorsal surface covered by a thin, oval, scale-like shield, beyond which the extremities of the limbs project only slightly. Antennules slender, simple; antenne prehensile ; swimming-feet two pairs (?). Abdomen short, slender, and, like the cephalothorax, completely covered by the scale-like dorsal investment. 1. ENTOMOLEPIS OVALIS, sp. nov. (Plate XIII. figs. 18-21.) Outline, as seen from above, oval, widest in the middle, width equal to more than half the length (fig. 18). Antennules ten-jointed, slender, the penultimate joint bearing a long and stout sensory filament (fig. 19); the comparative lengths of the sie 5 9 boo Bot oO oOo To Oo Wo NO joints as in the following formula: 5545-74-35 7 7 10 8° (fig. 20), the basal joint bearing at its apex a minute secondary branch ; second joint with a row of fine cilia on its distal half; third joint small and having a stout terminal unguis. Maxilla simple, two-jointed, the second joint pubescent and bearing three long terminal sete. Two pairs of foot-jaws (figs. 18 d, ¢), stout, prehensile, like those of Dyspontius. Siphon slender and of moderate length; swimming-feet two pairs (?) two-branched, each branch composed of three joints (fig. 18 /). Length 1:3 millim. Antenne three-jointed One specimen only was found among mud from a depth of 1-0 fathoms in Lyttelton Harbour. The foregoing description is incomplete, owing to some of the parts having been imperfectly seen, and it is to be hoped that some future observer may be enabied by the help of further specimens to describe more fully what seems to be a very remarkable form. Family ARTOTROGID A. Genus Artotrocus Boeck, 1. ArtTotTroaus ovatus Thomson. Artotrogus ovatus Thomson (1), p. 118, pl. x1. figs. 11-14. In the surface-net off Port Chalmers. One specimen. 2. ARTOTROGUS BREVICAUDATUS, sp. nov. (Plate XII. figs. 27-29; Plate XIII. figs. 22-26.) Cephalothorax much narrowed in front, widest behind the middle; abdomen very short, only about one third as long as the cephalothorax (Plate XII. fig. 22); siphon * &vropor, an insect; dems, a scale. MARINE COPEPODA OF NEW ZEALAND. 49 long and slender (fig. 24). Antennules very short, fourteen-jointed (2) (fig. 23), the basal joints very indistinctly marked; comparative lengths of the joints as in the following formula: 3 ; - ae ee = : - : = : = : = : = : = First segment of the abdomen much longer than the following segments, inclusive of the furca (fig. 26). The mouth-organs and swimming-feet present no distinctive characters. Length 1-1 millim. Hab, Akaroa Harbour, 6 fathoms. Genus Cariaus. 1. CaLiGus LONGICAUDATUS, sp.nov. (Plate XII. fig. 26.) A single specimen of a Caligus, which seems distinct from any described species, occurred in a surface-net gathering from Port Chalmers. The much elongated abdomen, the abruptly angular carapace, together with the slenderness and length of the last pair of feet, are the chief distinctive characters. Length 6 millim, EXPLANATION OF THE PLATES. jd BF. Ml Oe BG Paracalanus parvus, p. 32. Fig. 1. Antennule of male, x 140. Fig. 2. Foot of fifth pair of male, x 210. Clausocalanus furcatus, p. 32. . Female, seen from right side, « 84. Fifth pair of feet of female, x 250. . Antennule of male, x 84. . Fifth pair of feet of male, x 210. . Abdomen and foot of fifth pair, ¢, seen laterally, x 165. eo) S og ID Or oo Acartia ensifera, p. 33. Fig. 8. Female, seen from above, x 84. Fig. 9. Abdomen of female, seen from right side, x 110. Fig. 10. Fifth pair of feet of female, x 210. . 11. Terminal spine cf one of the swimming-feet, x 210. Fig. 12. Right antennule of male, x 140. Fig. 15. Left antennule of male, x 140. Fig. 14. Abdomen of male, x 110. Fig. 15. Fifth pair of feet of male, x 250. DR. G. S. BRADY ON THE Temora tenuicauda, p. 34. . Female, seen from left side, x 50. . Antennule of female, x 84. . Abdomen of female, x 84. . Foot of fifth pair, 29, X 210. . Terminal spine of swimming-feet, x 300. . Right antennule of male, x 84. . Fifth pair of feet of male, x 110. . Abdomen of male, seen from front, x 210. Centropages pectinatus, 2, p. 36. . Foot of fourth pair (?), x 84. . Foot of fifth pair, x 210. . Abdomen, seen from front, x 84. . Do oh AIL SS eee a. Sete of first segment, more highly magnified. 6. Sete of tail, more highly magnified. Peltidium novee-zealandie, p. 44. . Antennule of female, x 140. , 2A), Posterior foot-jaw, X 210. PLATE X. Centropages discaudatus, p. 35. . Male, seen from right side, x 65. . Armature of 17th segment of right antennule, ¢, x 210. = 17th, 18th, and 19th segments of same, x 210. Fifth pair of feet of male, x 84. female, x 100. 99 99 . Abdomen of female, seen from side, x 100. . Furca of female, x 100. Labidocera cervi, p. 37. . Antennule of female, x 50. . Abdomen of female, seen from below, x 50. > Hoot of fifth pain.) 95) >< 100! . The same, immature, x 84. . Right antennule of male, X 50. . Abdomen of male, x 50. . Fifth pair of feet of male, X 50. Fig. 15. Fig. 16. . Fourth foot of same, x 140. 23) (eye) Oo OMS OP OO De MARINE COPEPODA OF NEW ZEALAND. 51 Cyclops ewarti, p. 38. Antennule of female, x 140. Abdomen of same, x 84. a. Terminal spine of inner branch of the same, x 300. Ectinosoma australe, ¢ , p. 39. . Antennule, x 250. . Antenna, x 210. . Mandible, x 250. . Anterior foot-jaw, x 250. . Posterior foot-jaw, x 250. . One of the swimming-feet, x 210. . Foot of fifth pair, x 210. d. Caudal segments and sete, x 210. Flavia crassicornis, 2, p. 42. . Antennule, x 250. . Anterior foot-jaw, x 250. . Posterior foot-jaw, x 250. . Foot of first pair, x 210. second pair, x 210. 99 . Abdomen, x 100. PLATE XI. Laophonte meinerti, p. 40. Female, seen from left side, x 84. . Antennule of female, x 300. ms male, x 300. . Posterior foot-jaw, x 300. . Foot of first pair, x 210. . One of the swimming-feet, x 210. . Inner branch of foot of second pair, ¢, 300. . Foot of fifth pair, 9, x 210. first pair, X 226. fifth pair, variety, x 210. 39 39 52 DR. G. 8S. BRADY ON THE Dactylopus hanseni, p. 41. Fig. 11. Female, seen from left side, x 100. Fig. 12. Antennule of female, x 210. Fig. 13. na male, x 210. Fig. 14. Inner branch of antenna, x 210. Fig. 14 a. Posterior foot-jaw, x 280. Fig. 15. Foot of first pair, xX 210 Fig. 16. Inner branch of foot of second pair, ¢, x 300. Fig. 17. Foot of fifth pair, ¢, Xx 210. ate 1 ». 85% MO, Phroso gracilis, 2, p. 42. Fig. 19. Female, seen from right side, x 68. Fig. 20. Antennule, x 210. Fig. 21. Antenna, x 210. Fig. 22. Mandible, x 210. Fig. 23. Maxilla, x 210. Fig. 24. Anterior foot-jaw, X 250. Fig. 25, Posterior foot-jaw, x 250. Fig. 26. Foot of fourth pair, x 120. Fig. 27. a) wuithypair, x 210: Thalestris ciliata, 2, p. 43. Fig. 28. Antennule, x 210. ig. 29. Posterior foot-jaw, xX 260. . Foot of first pair, x 210. Se thirdipairs livia: sy sed od joys SIMON), . Abdomen, seen from front, x 84. i=) BS er 0a wo v2 © wD — oo PLATE XII. Thalestris australis, 9, p. 43. Fig. 1. Antennule, x 210. Fig. 2. Posterior foot-jaw, x 219. Fig. 3. Foot of first pair, X 175. Fig. 4. One of the swimming-feet, x 175. Fig. 5. Foot of fifth pair, x 130. bw wb ro Ww -e co tj KH COO © = Or eB of bo ee (Se) i) MARINE COPEPODA OF NEW ZEALAND. 55 Harpacticus glaber, 2, p. 44. . Antennule, x 200. . Antenna, x 200. . Posterior foot-jaw, x 210. . Foot of first pair, x 140. , fifth pair, x 210. . Abdomen and caudal sete, x 84. Peltidium nove-zealandie, ¢, p. 44. . Foot of first pair, x 210. peeetourth: pair, << 120: tthe pair ><) 20, . Abdomen, x 130. Scutellidium plumosum, p. 46. . Antennule of male, x 300. . Antenna of male, x 300. . Anterior foot-jaw of male, x 210. . Posterior foot-jaw of male, x 210. . Foot of first pair of female, x 210. . Outer branch of one of the swimming-feet, x 300. i a a (D. tisboides: Trish specimen), x 300. . Foot of fifth pair of male, x 250. . First abdominal somite of male, «x 210. . Abdomen and tail of female with fifth pair of feet, x 140. Caligus longicaudatus, p. 49. . Female, seen from below, x 16. Artotrogus brevicaudatus, p. 48. . Mandible (?), x 120. . Maxilla, x 210. . Foot-jaw, x 120. PLATE XIII. Paurocope robusta, 2, p. 46. . Animal, seen from left side, BA ; s) ss above, : . Antennule, x 140. oon o oe 10. aie e. 12. as: 14, g. 15. 16. Sy, LS: ON THE MARINE COPEPODA OF NEW ZEALAND. . Antenna, x 210. . Mandible, x 500. . Maxilla, x 500. . Anterior foot-jaw, xX 500. . Posterior foot-jaw, < 210. . Foot of first pair, x 210. Centromma thomsoni, p. 47. Animal, seen from below, x 140. Antennule, o- (1891, containing 6 Plates) . . . . . Ay 0) ites aD Le 400 ae (892. coutaming sal Plate)iee @ ug.y a: eo ee sr Owain 0 6 O Bon logecontarmimenn9) lates) ivncas au.ss 0 sian: aye OAL 8) 015 O » 6. (1893, containing 4 Plates) . . .-. . ee Oe) 012 0 » 7. (1893, containing 6 Plates) . . .. . ORONO 012 0 » 8 (1894, containmg 6 Plates) . . .. . ee One PO: O) 012 0 » 9. (1894, contains 6 Plates) . . . . . 5 Onis s Onl be » 10. (1895, contaming 5 Plates) .... . Ore GeO 012 0 » ll. (1895, containing 7 Plates and Titleand Index),, 015 9 i evo VOLUME XIV. (1896-1898, containing 47 Plates) . Price5 5 O 7 0 Part]. (1896, contaming 2 Plates)... . . i Ore Oa.0 0 0 5 es S96, contamimes ‘G6, Plates)i.) ay an Ono O 012 0 oe (Leo contamina s OPP lates)) os soe <1 om Bo pe Lose O » 4. (1897, containing 10 Plates) . . . . . el LOO. 2 0 0 fo oa (S98. contamimes 2) Plates)iise oneness PnOes OrwO 0 0 067898; contaming (Plates) ase alin eles 0) 1 0 72 (1898, contaming 46, -Plates) ios 2. | sles he O00 Ni bi ics} 0 15.0 » 8. (1898, containing 5 Plates and TitleandIndex),, 0 9 O 012 0 VOLUME XV. Parr 1. (1898, containing 8 Plates) .-. .. . PrceO 15 9 . . . I 1 a.) 22 (1899, contaming a Plates) wher se iran 0) ake HO teen cep al Oa P70) CONTENTS. Il. On the Marine Copepoda of New Zealand. By G. Srmwarpson Brapy, M.D., D.Sc., Lb. Di TLRS. OMS. (Plates: UX) ae ee pagelok THE PUBLICATIONS OF THE ZOOLOGICAL SOCIETY OF LONDON. Tue scientific publications of the Zoological Society of London are of two kinds—“ Proceedings,” published in an octayo form, and “Transactions,” in quarto. According to the present arrangements, the “ Proceedings”’ contain not only notices of all busi- ness transacted at the scientific meetings, but also all the papers read at such meetings and recom- mended to be published in the “ Proceedings” by the Committee of Publication. A large number of coloured plates and engravings are attached to each annual volume of the “ Proceedings,” to illustrate the new or otherwise remarkable species of animals described in them. Among such illustrations, figures of the new or rare species acquired in a living state for the Society’s Gardens are often given. The “ Proceedings ”’ for each year are issued in four parts, on the first of the months of June, August, October, and April, the part published in April completing the volume for the preceding year. The “ Transactions” contain such of the more important communications made to the scientific meetings of the Society as, on account of the nature of the plates required to illustrate them, are better adapted for publication in the quarto form. They are issued at irregular intervals. Fellows and Corresponding Members, upon payment of a Subscription of £1 1s. before the day of the Anniversary Meeting in each year, are entitled to receive all the Society’s Publications for the year. They are likewise entitled to purchase the Publications of the Society at 25 per cent. less than the price charged for them to the Public. A further reduction of 25 per cent. is made upon purchases of Publications issued prior to 1871, if they exceed the value of five pounds. Fellows also have the privilege of subscribing to the Annual Volume of the ‘ Zoological Record? for a sum of £1 (which includes delivery in the United Kingdom only), but this privilege only holds good if the subscription is paid before the First of December in each year. Such of these publications as are in stock may be obtained at the Society’s Office (3 Hanover Square, W.), at Messrs. Longmans’, the Society’s publishers (Paternoster Row, E.C.), or through any bookseller. P. LL. SCLATER, at 1899. : Aacu 3} Seerciary. TRANSACTIONS OF THE ZOOLOGICAL SOCIETY OK. LON DON. VoL. XV.—Parr 3. LONDON: PRINTED FOR THE SOCIETY, SOLD AT THEIR HOUSE IN HANOVER-SQUARBH; AND BY MESSRS. LONGMANS, GREEN, AND CO., PATERNOSTER-ROW. October 1899. Price 12s. Taylor and Francis, Printeys,] [Red Lion Court, Fleet Street. TRANSACTIONS OF THE ZOOLOGICAL SOCIETY OF LONDON. To Fellows. To the Public. oS. d, BORE Mh VOLUME I. (1833-1835, containing 59 Plates). . Price 313 6 . . . 418 O* VOLUME II. (1835-1841, containing 71 Plates). . 5, 400... 5 6 6% VOLUME III. (1842-1849, containing 63 Plates). . 5 8 8 3... 411 O* VOLUME IV. (1851-1862, contaming 77 Plates). . 4 6 2 0... 8 2 6f VOLUME _ V. (1862-1866, containing 67 Plates). . , 5 4 38 . . . 619 OF VOLUME VI. (1866-1869, containing 92 Plates). . ,, 11 5 0 .. .15 O OF VOLUME VII. (1869-1872, containing 73 Plates). . ,, 10 4 0 . . .1812 OF VOLUME VIII. (1872-1874, containing 82 Plates). . ,, QS Orne a ahe EO VOLUME IX. (1875-1877, containing 99 Plates). . ,, 12 16. ..16 2 OF VOLUME X. (1877-1879, containing 95 Plates). . ,, 10 0 3 .. .18 7 OF GENERAL INDEX, Vols. I. to X. (1833-1879) . ,, OCG pane OOO VOLUME XI. (1880-1885, containing 97 Plates). A Oe Oe es 30 VOLUME XII. (1886-1890, containing 65 Plates). . ,, Byeitoh Omang g eerfan yO) * No copies of these volumes in stock. T Only complete copies of these volumes left in stock. Continued on page 3 of Wrapper. I. On the Eatinet Birds of Patagonia.—l. The Skull and Skeleton of Phororhacos inflatus Ameghino. By C. W. Annrews, B.Sc., F.Z.8., Assistant in the British Museum (Natural History). Received March 21, read April 18, 1899. [Puares XIV.—XVII.] INa paper read at the Meeting of the British Association at Ipswich, in 1895, and subsequently published in the ‘ Ibis’ !, the present writer gave a brief review of a very interesting memoir by Dr. Florentino Ameghino on some remarkable fossil birds from the Tertiary deposits of Patagonia. Since that time, the specimens described by Ameghino have come to the Geological Department of the Natural History Museum, so that an opportunity of examining the bones themselves has arisen ; and in the case of the most remarkable of the specimens, the skull and skeleton of Phororhacos inflatus, the careful removal of the adherent matrix has revealed for the first time many points of great importance, e.g. the structure of the palate. It seems, therefore, desirable to give as complete an account as possible of the osteological characters of this bird, in order, if possible, to arrive at some conclusion as to its precise relationship to recent forms; and although the account given in Ameghino’s excellent paper is an accurate one as far as it goes, still it seems best, even at the risk of some repetition, to describe in detail all the parts of the skeleton available, in order to facilitate the comparison of them with recent types. It may be at once stated that if the remaius described by Ameghino really formed parts of one individual skeleton (and there seems to be no reason to doubt this), then this bird presents a most extraordinary combination of characters, to which no close parallel can be found among recent forms. A brief summary of the various opinions that have been expressed relative to the systematic position of the Stereornithes will be found in the above-mentioned review. The present paper will be confined to a description and comparison of the skeleton of Phororhacos inflatus Ameg., that species being represented by much the best series of remains. SKULL AND SKELETON OF PHORORHACOS INFLATUS AMEGHINO. Of this species there is a fine series of bones, including the skull and mandible, scapula, coracoid, bones of wing, pelvis, and bones of hind limb, as well as some vertebra: these are said to belong to a single individual. ‘There is also a number of 1 « Remarks on the Stereornithes, a Group of Extinct Birds from Patagonia,” Ibis, 1896, p. 1, VOL. XV.—PART UI. No. 1.-—October, 1899. K 56 MR. C. W. ANDREWS ON THE SKULL AND odd bones referred to this species, but the following description is founded as far as possible on the associated set, which is the type-specimen. The Skull and Mandible. (Plates XIV. & XV.) The skull and mandible are in a wonderfully perfect state of preservation, and, now the matrix has been removed, all the important details of their structure can be made out. The skull has been slightly crushed, so that there is some dislocation of the bones of the roof, and also a slight compression of the beak, but otherwise it retains its original form. Occipital surface of skull of Phororhacos inflatus Ameghino, # natural size. Lv., lambdoidal ridge ; m.¢., mammillary tuberosities; oc.c., occipital condyle ; oc.f., occipital foramen ; p-p., paroccipital processes ; sf.r., supraforaminal ridge; q., quadrate. The Occipital Region (text-figure, above).—The occipital surface is remarkable for its flatness, its great width from side to side, and the distinctness with which it is marked off from the other regions. This last peculiarity is due to the great development of the lambdoidal ridge (/.r.), particularly towards the sides. In its middle portion this ridge merely forms the angle (of about 100°) between the occipital and dorsal regions, but laterally it is produced outward, forming prominent lateral crests which are continuous ventrally with the broad paroccipital processes (p.p.). These latter are convex posteriorly, both from above downward and from side to side; their inferior angles extend somewhat below the occipital condyle, and are formed by the ventral prolongations of the very well-marked supraforaminal ridges. These ridges run downward and somewhat outward from the upper border of the foramen magnum (oc.f.), and form the inner borders of the paroccipital processes, separating them from a SKELETON OF PHORORHACOS INFLATUS. 57 depressed area lying between them and the foramen magnum. ‘This valley-like depression is closed above by the meeting of the supraforaminal ridges (sf.r.), but opens ventrally on either side of the occipital condyle. The inferior angles of the paroccipital processes are united to the mammillary tuberosities (m.t.) of the basi- temporal platform by prominent buttresses of bone, behind and internal to which the vagus foramina open. The foramen magnum (oc.f.) is relatively very small; it is oval in outline, the long axis being vertical. The plane of the opening is inclined backward, making an angle of rather more than 45° with the long axis of the skull. The transverse diameter of the opening is 9 mm., the vertical 16 mm. The slightly pedunculate occipital condyle (oc.c.) is oval, or rather reniform, ils upper border being nearly straight, with a vertical groove running down from its middle point nearly to the centre of the condyle. There is a small but deep pre- condylar fossa (pe.f., Plate XV. fig. 1). The base of the Skull (Plate XV. fig. 1)—The basitemporal platform (0.f.) is a triangular area which is very slightly elevated, and below the level of which the occipital condyle projects. It is concave from side to side, and the postero-lateral angles are produced downward into the very prominent mammillary tuberosities (m.t., fig. 1), which, as mentioned above, are united to the inferior angles of the paroccipitals by buttresses of bone. ‘The posterior margin of the platform is defined by a shallow groove separating it from the occipital region. The lateral borders are thickened ridges, forming anterior prolongations from the mammillary tuberosities (m.t.), and converging so that they appear to have met in front in a rounded median angle. ‘This, no doubt, lies immediately below the common opening of the eustachian tubes, which are immediately dorsal to the thickened ridges just described, and open posteriorly by wide apertures into the tympanic cavity. In front of the basitemporal platform the narrow rostrum (7.) is seen for a short distance, but it is here badly preserved, and anteriorly is completely invested by the bones of the palate. On the sides of the basitemporal platform, rather high up and behind its anterior angle, are a pair of basipterygoid facets, with which the pterygoids articulate by means of stout processes borne on their inner surfaces rather behind their middle point (5.p7.). Below and behind the articular facets for the pterygoids the basitemporal forms a thin plate of bone (pretemporal wing), which looks outward and downward; the lower edge of this forms the floor of the eustachian canal, while its upper portion is the outer wall of the large pretympanic recess, which opens widely into the tympanic cavity immediately above the posterior aperture of the eustachian tube. The side of the Skull (Plate XI1V.).—The tympanic cavity (ty.) is relatively small ; posteriorly it is bounded by the concave anterior face of the paroccipital process; the inferior portion of its inner wall is formed by the buttress of bone which joins the paroccipital to the mammillary tuberosity (m.t.). Its outer edge is formed by a downwardly-projecting flange of the squamosal, the anterior edge of which runs down K 2 58 MR. C. W. ANDREWS ON THE SKULL AND between the anterior and posterior quadrate facets. Of these, the antero-external (0.f.) is very large, oval in outline, and concave in all directions, but more deeply so from before backward; its outer edge forms a prominent lip on the lower edge of the squamosal prominence. The other facet (¢.f.) is placed internal and posterior to the Jast; in form it is a very elongate oval, the long axis being antero-posterior, in which direction it is deeply concave: it locks outward and downward. The opening of a large pneumatic fossa separates it from the downwardly projecting process of the squamosal and also from the anterior facet. Mesiad of the posterior facet, but separated from it by a narrow flat ledge of bone, is a very deep pocket-like depression, into which open numerous foramina. ‘The most posterior of these seems to be the fenestra ovalis; the others are probably pneumatic. Ventrally the tympanic cavity communicates by wide openings with the anterior tympanic recess and the eustachian canal. The squamosal region (Plate XIV., sq.) is deeply concave from before backward, and slightly convex from above downward. Posteriorly it forms the anterivr face of the prominent lambdoidal crest; ventrally, as already mentioned, it is produced downward into a pointed process, which helps to form the outer boundary of the tympanic cavity. Anteriorly its ventral edge is formed by the prominent outer lip of the anterior quadrate facet (0.f.). About 7 mm. above this there is a prominent elongated tuberosity, the long axis of which is directed downward and inward; this I take to be the zygomatic process (zy.). From its upper end a slight ridge runs upward and backward to the lambdoidal crest (d.7.), and forms the lower boundary of the temporal fossa. Internal to the zygomatic process, and connected with it by a short ridge, is another angular projection, the pretympanic process; from this also a slight ridge runs upward and backward, but it soon dies away without reaching the main temporal ridge. As might be expected from the great massiveness of the mandible, the temporal fosse are very large and deep. Their lower opening is bounded posteriorly by the zygomatic, anteriorly by the postorbital process (p.ord.p.), which approach to within 17 mm. of each other. The fosse are much deeper than broad, and greatly constrict the side of the cranium. Superiorly they are bounded by the temporal ridges, which, as above mentioned, run upward and backward from the zygomatic process till they meet the lambdoidal crest (/.7.), with which they are confluent for 28 mm. Leaving the lambdoidal ridge, they sweep forward, first towards the middle line, then outward, terminating anteriorly on the upper surface of the postorbital process. On the roof of the skull the fossee are separated one from another by an interval of about 12 mm. only. On the side of the cranium, immediately below the constriction caused by the temporal fosse, and about on the same level as the quadrate facet, there is a deep infundibuliform depression, at the bottom of which the trigeminal foramen opens. This depression is marked off from the orbit by a ridge which runs upward and SKELETON OF PHORORHACOS INFLATUS. 59 outward, terminating above in a prominent cristiform ridge on the ventral surface of the postorbital process. The orbits are very large; they appear to be divided one from another by a complete bony septum (7.0.s.), or there was at most a small vacuity in front of the optic foramen. These latter appear to be confluent, and lie on exactly the same level as the trigeminal foramina. In the upper posterior region of the orbit there is a deep depression, which no doubt lodged a gland; from this a shallow groove runs forward along the upper edge of the interorbital region. This groove marks the course of the olfactory nerve, and immediately above it the great penthouse-like roof of the orbit arises. The whole of the inner and posterior part of this is formed by the orbital plates of the frontals, while externally it is completed by the enormous supraorbital plate of the lachrymal, which in this specimen has been somewhat displaced downward on both sides, so that the skull, seen from above, appears as if it possessed large supraorbital depressions, such as occur in the Petrels, &c.; though, as a matter of fact, no such depressions existed (see Plate XIV. and Plate XV. fig. 2). The lachrymal (/ac.), as just mentioned, bears a very large supraorbital plate, which extends nearly to the postorbital process, terminating posteriorly in a blunt rounded angle. Its inner border was closely applied to, if not united with, the outer border of the frontal, and its outer border forms the upper margin of the orbit. Anteriorly its ventral surface is deflected, and the body of the bone turns inward and downward, and is produced ventrally into a nearly vertical bar of bone, which in its ventral portion is flattened laterally to a thin plate, and at its lower extremity articulates with the upper border of the jugal. A little above its middle point this bar has a small backwardly projecting process, but there does not seem to be any trace of a separation into two elements at this point, such as was formerly suggested!. The form and relations of this structure are almost precisely the same as in Serpentarius, or even more Accipitrine, as in the Eagles, e.g. the Harpy Eagle; and I can see no justification for Ameghino’s description of the orbit as open anteriorly in Phororhacos any more than it is in the birds just mentioned, in which also this process reaches the jugal. In front of the preorbital process of the lachrymal is the antorbital fossa, which is triangular in outline, its antertor and upper angles being rounded off. Its anterior border seems to be formed by a downgrowth of the nasal meeting an upward process from the maxilla; there are, however, a number of cracks in this region, which render its interpretation difficult, but by comparison of the two sides some degree of certainty can be attained. The nasals consist of a posterior body, which unites behind with the frontals and supraorbitals ; but it cannot be determined whether the two meet in the median | < This,’ 1896, p. 6. 60 MR. C. W. ANDREWS ON THE SKULL AND line or whether they are separated by the facial process of the premavxille, the junction with which is quite indistinguishable. It appears very probable that the premaxille ran in between the nasals and formed at least the upper portion of the high arch of the culmen between the nostrils, while the upper borders of those openings are formed by processes of the nasals; these bones also send processes downward and forward, forming the posterior border of the nares, and meeting the maxilla below as already mentioned. The angle between the anterior and inferior processes is rounded and has a thickened edge. The form and relations of the nasal are best seen on the left side of the specimen. Immediately within the anterior border of the antorbital fossa, and united with the inner face of the downward process of the nasal, is the base of a broad band of bone (eth.), which runs vertically down the middle of the fossa parallel to the downward process of the lachrymal. The upper end of this band of bone widens out into a fan-shaped expansion, the upper edge of which is slightly overlapped by the anterior margin of the lachrymal. ‘The lower end, also somewhat expanded, is cancellous, and unites with the upturned inner edges of the maxillo-palatine plates, with which the anterior ends of the palatines are indistinguishably fused. The narrowest point of this bar of bone is about its middle, below which it bears a backwardly-directed triangular process. A similar and similarly-situated bar in the ethmoidal region occurs in Cariama and also in some Accipitrine birds: e. g., in a modified form, in Serpentarius. The jugal arch is oval in section posteriorly, but in front it is compressed so as to form a broad, thin, vertical bar of bone. On the inner side of its posterior end there is a rounded knob which fits into a corresponding pit on the outer surface of the quadrate. Its anterior end seems to overlap the posterior prolongation of the maxilla. The downward process of the lachrymal articulates with the upper edge of the jugal close to its anterior end. The form and relations of the maxilla to the surrounding bones are very difficult to make out. As just mentioned, its posterior extremity seems to be overlapped by the jugal, and in front of this it no doubt formed the edge of the beak for some distance. On its inner side it widens out suddenly, the expanded portion being directed downward and inward, and forming the large maxillo-palatine plate, with which the anterior end of the palatine unites completely. The maxillo-palatine plates have a thickened hinder border and are highly convex from side to side, and their upturned inner margins unite with the downwardly-directed bar of the ethmoid. ‘The posterior border of each maxillo-palatine plate forms a conunuous curve with the lower edge of the jugal arch. On either side of the middle line of the palate they form a long ridge which projects considerably below the tomium ; the fused anterior ends of the palatines must also take part in the formation of this ridge. In the middle line they unite for a space of about 3°5 cm., in front of which they are separated by a narrow cleft 2-5 em. long; how far they extend in front of this cannot be determined, nor is it SKELETON OF PHORORHACOS INFLATUS. 6] possible to make out the extent and relations of the facial parts of the maxille, and it is therefore uncertain what share, if any, they may take in forming the boundary of the nostrils. The premazille no doubt form the greater part of the enormous beak, but their exact boundaries are not distinct. _ Superiorly they send out fused nasal processes, which form at least the upper part of the high bar of bone between the nostrils and perhaps extended to the frontals. Anteriorly the bodies of the premaxille form the high laterally- compressed anterior section of the beak, the tip of which forms a downwardly-directed hook, about 3:5 cm. long and small compared to the bulk of the whole beak. From the sides of the base of this hook a pair of parallel ridges run back on to the palate ; their posterior portion is separated from the tomium by a deep groove, into which the edge of the mandible fits when closed ; on their inner side also there is a deep channel, divided by a median ridge, which just behind the decurved hook rises into a prominent tuberosity, at first described as a tooth. The cutting-edge of the beak forms a double curve; it commence a little behind the point and first forms a strongly convex, prominent edge, sharp in front, but more rounded behind ; posteriorly it is very slightly concave, and is a continuation of the line of the lower edge of the jugal arch. The anterior convex portion is no doubt formed by the premaxille ; the posterior is maxillary. In the premaxillary region the narrow palate is closed, except for a few small median foramina, and is deeply concave from side to side. The beak, as a whole, is very deep from above downward, strongly compressed laterally, and its whole surface covered with angular and vascular impressions, which indicate that in life it was covered with a very thick horny sheath ; the whole must have formed an extremely formidable weapon, whether for attack or defence. The upper portion of the cranial region of the skull is much flattened (Plate XV. fig. 2). Posteriorly it is greatly constricted by the great temporal fossa, which are separated in the middle line by an interval of about 12 mm. only. In front of these it widens out and reaches its greatest width at the level of the postorbital processes. In front of these it again narrows, and anteriorly the frontals join the nasals, but their limits are not very clear. The orbital borders of the frontals have thick roughened edges, and probably the great supraorbital plates of the lachrymals, which now lie slightly below the level of the frontals, may, in their normal condition, have united with them throughout their whole extent; but, since their position is symmetrical on the two sides, it is possible they may be in their natural position. As already mentioned, this depression of the suprasrbital plates gives the skull, when looked at from above, the appearance of having possessed deep supraorbital fossi. The Palate (Plate XV. fig. 1).—The most important result of the careful removal of the matrix from the skull is that the structure of the palate is completely revealed, the bones being perfect and very little displaced. The taxonomic importauce of this 62 MR C. W. ANDREWS ON THE SKULL AND region is very great, and examination of it in this bird on the whole tends to support the views previously put forward. The pterygoids (pt.) are stout, laterally compressed bones, of considerable depth. At their posterior end they bear a cup-like depression, which receives a corresponding tuberosity on the inner border of the quadrate, and on their upper border, one-third of their length from the posterior end, there is a stout process, oval in section and projecting backward, upward, and inward; it terminates in a facet, which articulates with a corresponding surface on the base of the skull. These upper processes of the pterygoids occupy the position, and perform the function, of the basipterygoid processes of some birds. Anteriorly the pterygoids converge and just meet below the rostrum: their anterior ends are abruptly truncated at right angles to their long axis, and form facets for articulation with the palatines; these latter are very large bones; posteriorly they bear short, broad pterygoid processes, which meet in the middle line and articulate with the pterygoid facets just mentioned. The median union of the palatines extends forward about 3 cm., and is marked by a slight ridge: in front of this their median borders diverge and are notched by the hinder borders of the internal narial slits. In the interval between the palatines no trace of the vomer is preserved, except that close to the anterior end, where they join the maxillo-palatine plates, there are traces of a median plate, which is probably a portion of that bone. The body of the palatine is very large, and consists of two lamella separated by a deep valley and projecting almost vertically downward. Of these the outer is much the larger, and forms a great ventrally- directed flange. Its inferior border is somewhat thickened, and meets. the posterior border, which also is thickened, nearly at right angles. The angle thus formed is slightly everted, and a short distance above it the posterior border slopes gently backward to the pterygoid process. The inner lamella is very small; its ventral border is thickened and is continued backward to meet its fellow on the middle line, forming the median ridge above noticed. ‘The anterior portion (maxillary process of the palatine is formed by the widening out of the external lamella; this fuses with the maxillo-palatine plates, and together with them forms the prominent longitudinal ridge projecting below the level of the tomium, as before described. ‘The bird in which the palate most nearly resembles the fossil in this respect is Yiomedea, in which the disposition and relative proportions of these ridges and the form of the cutting-edges of the beak are very similar. Dorsally the palatines send up a thin triangular plate of bone on either side of the rostrum and embracing it. The quadrate (Plates XIV. and XV. fig. 1, also text-figure, p. 63) articulates with the skull by two distinct condyles sharply separated one from another by a deep pit, at the bottom of which is a large pneumatic fossa, hollowed out in the neck of the bone. ‘The antero-external capitulum (O.H.) is much the larger; it is criangular in outline, convex from above downward, and slightly concave from side to side ‘The SKELETON OF PHORORHACOS INFLATUS. 63 inner capitulum (I.H.) is borne on a thin, backwardly-directed buttress of bone; it has two facets, one nearly flat, looking upward and inward, the other convex and directed downward. ‘The neck of the bone is deeply excavated by pneumatic fossz, both on its anterior and posterior face, and also on the inner side at the base of the orbital process (O.P.). This latter is large; its upper and lower borders are nearly parallel, and it terminates in a roughened surface; its length is about 17 mm. Fig, 2. Left quadrate of Phororhacos inflatus Ameghino. 1. From outer side. 2. Articular surface for mandible. 3. From front. LH., inner head; I., inner articular surface for mandible ; J.C., cup for end of jugal arch; O., outer articular surface for mandible; O.H., outer head; O.P., orbital process; P.T., tuberosity for articulation with the pterygoid. The lower end is massive and bears two articular surfaces (I. and O.) for the mandible, which are separated by a shallow groove running outward and forward from the postero- internal border. ‘The form of the facets will be best understood from the figure (fig. 2, above). On the inner border of the bone, just beneath the base of the orbital process, there is an elongated tubercle (P.T.) for union with the hinder end of the pterygoid ; and on the outer side, immediately above the surface for the mandible, is the deep cirenlar cup (J.C.) with clearly defined border, which receives the posterior end of the quadrato-jugal. The mandible (Plate XIV.) is specially remarkable for its very massive structure. The relatively narrow symphysis is about 9 cm. long; its ventral border is convex, its dorsal nearly straight, and but slightly upturned at the top. The upper surface VoL. XV.—ParT 11. No. 2,—October, 1899. I 64 MR. ©. W. ANDREWS ON THE SKULL AND is deeply concave from side to side, giving the symphysial region a spout-like form. Behind the convexity of the symphysial region the ventral edge of each ramus is anteriorly slightly concave and posteriorly rather strongly convex. The upper borders of the rami are at first convex and sharp-edged, behind this, as far as the small coronoid process, they are straight and blunt. Behind the coronoid process the edge slopes down to the articular cup for the quadrate. The facets on this are two in number, one external and elongated, one inner and rounded in outline ; both are concave from before backward, and they are separated by a deep fossa, but no pneumatic foramina are present. There is a large and very prominent internal angular process; the inferior angle is broken away, but it can only have formed merely a slight projection. Just beneath and in front of the coronoid process the mandible is perforated by a very large vacuity. Comparison of the Skull of Phororhacos with that of other Birds. Comparison of the skull of Phororhacos with that of the Struthious birds at once excludes the possibility of any affinity between it and them. The form of the palate, in which, as already described, the palatines and pterygoids both articulate with the rostrum, the reduction or absence of the vomer, and the double head by which the quadrate articulates with the skull, are all points not occurring in any Ratite bird. It then remains to determine what are the closest allies of Phororhacos among the Carinate, a question of which any satisfactory solution is very difficult. Taking the palate first for comparison, we find that it presents in some respects a very remarkable resemblance to that of certain ‘Tubinares, especially the Albatross (Diomedea). In this bird, as in Phororhacos, the maxillo-palatine plates together with the anterior region of the palatines form on either side a very prominent ridge, projecting considerably below the level of the tomium, the form of which also is very similar in the two birds, The palatines and pterygoids are also similar to those of the fossil in their relations one to another and to the rostrum ; but the palate differs much in form, the internal lamina being much the larger, while in Phororhacos the reverse is the case. There are a number of other important differences: thus, in Diomedea the palate is schizognathous and the anterior end of the large vomer appears upon it; there are large supraorbital fossee. In Phororhacos the palate is desmognathous, the vomer is inconspicuous, and there seem to have been no supraorbital fossee whatever, although slight displacement of the large supraorbital plates of the lachrymal at first sight gives the skull the appearance of having possessed them. Moreover, the form of the antorbital fossa and that of the quadrate are different in many respects. In spite of a certain similarity between the skulls, it does not seem that any close affinity exists between Phororhacos and the Tubinares, and in fact, as will be seen below, the structure of the limb-bones and pelvis lends no support whatever to any such relationship. Comparing the fossil with the skulls of various desmognathous types, one is at once SKELETON OF PHORORHACOS INFLATUS. 65 struck with its Accipitrine aspect. This, of course, depends chiefly on the form of the beak, but is heightened by similarity in several points of detail, e. g. the form of the lachrymal and of the antorbital fossee. Closer examination shows, however, that the structure of the palate is very different, particularly in the relations of the anterior end of the palatines to the maxillo-palatine plates; the form of the articulation of the quadrate with the mandible is also widely different. Of the Accipitrine birds, Serpentarius approaches the fossil most closely both in structure of the anterior region of the palate and in the existence of an articulation between the pterygoids and the basis cranii. In the Storks the palate is in several respects like that of the fossil, particularly in the relations of the palatines to the maxillo-palatine plates; the distal articulation of the quadrate is also similarly constructed, but is relatively much wider from before backward. On the other hand, the lachrymal bears neither a long descending antorbital process nor a large supraorbital plate, and the antorbital fossa is different. Although the Gruiformes are typically both schizognathous and schizorbinal, it is remarkable that Cariama and Chunga, which in external appearance and habits so much resemble Serpentarius, should have become both desmognathous and holorhinal, like Accipitrine birds. The existence of these characters, therefore, in the skull of Phororhacos cannot be regarded as a bar to the relationship between it and the Gruiformes, suggested in a former paper and supported by many structural peculiarities of the pelvis and hind limb. In fact, since the general line of specialization of Phoro- rhacos is similar to that which produced the Cariamide, the occurrence of these characters might have been expected. In some points, indeed, ¢.g. in the form of the quadrate, Phororhacos differs less from the typical Gruiformes than Cariama does. Further discussion of the affinities of this remarkable form is best deferred till after the remainder of the skeleton, as far as we know, has been described. The dimensions of the skull are :— cm. Length from anterior end of premaxillze to paroccipital process. . . . 337 5 555 to centre of occipital condyle . 30:0 \RVicliln eae MOC OMA OROIGESITIS 6 co 56 o 6 09 G9 6 6 o 0 o Go Oo Ane 5 Ci Gli a Womyporell OSES 6 o o 5 0 6 op oo oe oO 5 Gs WOStOnomnll ROGENSES 9 co np » 0 0 6 5 0.8 oc 0 0 oo EK 5, (least) between temporal fosse@ - 5 2 2 s © « es ow es (OO » Opposite posterior edge of nostril . . . . . «. - » «. - « 388 Depth of upper jaw at middle of nostril » . . . . «. ~~ « - ~« ILS * i ay is Tmale OWNES 6 6 6 50 0 60 0 6 6 6 oF Ailey ILE Or Quays Orns ¢ @ s 96 o 6 ¢ 60005 60 0 Zl WadthvoL upper aw,atuiid dleypornitien sien) utc ci onnte! tli EER OsO SIGN OF /ORHMED UG oo 9 0 09 0 06 9 Po OK CNY Witch or clistall Gael or @uachenti® 5 5 56 o 6 56 6 06 6 6 5 6 Go Bh L2 66 MR. C. W. ANDREWS ON THE SKULL AND Fig. 3. My Ip Zp ORL Paya SB, Uf D Ny ah: Poe Be Diagram of skeleton of Phororhacos inflatus Ameghino, to show the relative proportions of the bones. The portions of the skeleton preserved are shaded. (About 3 natural size.) SKELETON OF PHORORHACOS INFLATUS. 67 The dimensions of the mandible are :— cm. WEG 6 6 1G eR 6 Boe 2 de 6 oo oe UY IDEN OMNIS 5 6 5 6 6 16 6 6 56 0 6 oD oo a) 6D Depth of ramus at coracoid process . . . . . ..... 49 Wichin@renmonle emt og ¢ 6 o 6 6 6 6 6 6 oO 8 The hyoid (text-figure, below) is very well preserved, a most exceptional, if not unique, circumstance in a fossil bird. The anterior element, the entoglossal (ENT.GL.), is very large and clearly has originated from the union of two rod-like elements, the ceratohyals, which are united by a thin plate of bone. On the ventral surface the division between the rods is marked by a deep furrow, deepening posteriorly, but the dorsal surface seems to have been regularly convex from side to side. ‘The anterior end is somewhat broken; the posterior is deeply divided by a double Fig. 4. Hyoid of Phororhacos inflatus Ameghino, slightly reduced. 1. From below. 2. Basihyal from side. 2. Section of entoglossal at point A. ENT.GL., entoglossal ; B.Hy., basihyal (first basibranchial) ; c.pr., ceratobranchial. articulation for the basihyal (basibranchial 1) (B.uy.), and laterad to this the outer angles are produced into short backwardly-directed processes. The first basibranchial (B.Hy,) (basihyal) is a short bar of bone, the thinner lateral borders of which seem to have been broken away. Its dorsal surface is concave from before backward, and on its anterior half from side to side also; ventrally it is convex from before backward. ‘The anterior end bears a double surface for articulation with the entoglossal, separated by a slight ridge, which at its ventral extremity is produced into an anteriorly-projecting prominence. Posteriorly there are two irregular surfaces for the ceratobranchials, meeting at an angle of about 120°. The first basibranchial is not prolonged backward into a urohyal, nor does it, as far as can be seen, bear any surface for the articulation of a distinct median element: if such were present it must 68 MR. C. W. ANDREWS ON THE SKULL AND have been small and have articulated with the bases of the ceratobranchials (c.BR.). These are thin, slightly curved rods of bone, the anterior ends of which are enlarged and bear the surface for the articulation with the first basibranchial. The posterior ends are also slightly expanded and terminate in convex facets, apparently for union with another element. In the Albatross the entoglossal is relatively very much smaller, the first basibranchial (basihyal) broad and short, and prolonged backward into a long urohyal, so that the articular surfaces for the ceratobranchials are widely separated. In Grus the entoglossal is very long, slender, and unossified ; the first basibranchial (basihyal) is also long and slender, but, as in the fossil, the ceratobranchials articulate with two contiguous surfaces on its hinder end, and the uroglossal is distinct and articulates between the bases of the ceratohyals. In Cariama the two elements forming the entoglossal are partly separated by a median foramen and are prolonged into postero-lateral processes behind the articulation with the first basibranchial. his is a short bar ef bone terminating in a pair of surfaces for the ceratobranchials. The urohyal is very small and is not in contact with the first basibranchial. In the Birds of Prey the ossified portion of the entoglossal is, as a rule, deeply bifid anteriorly, the two elements forming it being united for a short distance only. The first basibranchial is somewhat elongate and is prolonged backward into a uroglossal process. In the Storks the entoglossal is imperfectly ossified, and the first basibranchial is similar to that of the Raptores. On the whole, the hyoid of Cartama approacnes the fossil most nearly. The Coracoid. (Plate XVII. fig. 1.) The coracoid is particularly remarkable for its great length and slenderness, and for the reduction of the acrocoracoid (a¢.) process. At the sternal end the bone is almost 32 mm. wide; its anterior face is convex from side to side, and the posterior flat or rather concave. ‘The articular facet for the coracoid groove of the sternum is large and well defined; it extends from side to side of the distal end of the bone both on the anterior and posterior face. At about its middle it is overhung by a very prominent tuberosity, from which a shelf-like projection runs to the inner border of the bone; the union with the sternum must have been very strong. There is no distinct processus lateralis (p.l.), its place being occupied by the convex outer border (p./.) of the expanded distal end. On the anterior surface a linea aspera runs from the inner angle upward and inward, joining the ridge marking the outer border at about + of the length of the bone from its distal end. ‘The shaft towards its upper end is slightly compressed laterally. ‘The total length of the bone is 157 mm. SKELETON OF PHORORHACOS INFLATUS. 69 The proximal end of the bone is peculiar. The acrocoracoid is almost completely absent, the upper end of the bone rising scarcely at all above the scapular surface. From the acrocoracoid region a ridge runs down the antero-internal border of the bone and forms the inner wall of a deep channel, which is bordered externally by the prominence formed by the surface for the scapula and the small procoracoid process. On this descending ridge there is, at some distance from its upper end, an elongate flat facet with which probably the clavicle articulated. The surface for the scapula is a large, deeply concave area, nearly circular in outline ; below its internal border arisesa small process or hook of bone (broken in the specimen figured), which overhangs the groove above described and is the procoracoidal process. The glenoid surface is a smooth, slightly convex area, the lower border of which forms a prominent overhanging lip on the postero-external side. The coracoid just described is one of the most remarkable boues of the skeleton. Its extreme elongation and the peculiar structure of its upper end, particularly the almost complete suppression of the acrocoracoid, are paralleled, so far as I am aware, only in the coracoid of the flightless Aptornis. The coracoid of this bird differs, however, in several important particulars, ¢.g.in the large size of its procoracoidal process, the presence of a supracoracoid foramen, and in the complete fusion of the lower extremity with the sternum. Fiirbringer has regarded the absence of the acrocoracoid as diagnostic of the Ratite, and has suggested that the term “ Platycoracoidee” should be used for them in contradistinction to the Acrocoracoidee, or Carinate. The oceurrence of this character in two such distinctly Carinate types as Apfornis and Phororhacos, however, invalidates this distinction, and tends to show that the presence or absence of the acrocoracoid, like that of the keel of the sternum, is in some way dependent on the loss of flight, and probably also on the increase in the bulk of the body. The articulation with the sternum is particularly well developed, and in no bird could the prominent tubercle and shelf above described be found developed to the same extent. No trace of the sternum is preserved, but, judging from the coracoids, it was probably of considerable size. Scapula. (Plate XVII. fig. 2.) Of the scapula no perfect specimen is preserved, but the most complete, here figured, wants the distal portion of the blade only. The surface for union with the coracoid is highly convex and roughly circular in outline; it is much roughened, and the union of the two bones must have been a very close one. The glenoid surface, which is slightly convex, is raised on a prominence which projects considerably below the inferior border of the blade. There is a very large acromium process. The portion of the blade preserved is somewhat rounded externally and flat internally. The width of the proximal end of this bone is 30 mm. 70 MR. C. W. ANDREWS ON THE SKULL AND The Bones of the Wing. The exact length of the wing cannot be determined, owing partly to the incomplete state of the humerus and partly to the absence of the phalanges ; but it is certain that, in proportion to the bulk of the bird, it had undergone extreme reduction and must have been entirely useless for purposes of flight. At the same time the stoutness of the bones and the distinctness of the muscle-impressions indicate that it was in some way functional, and the presence of a series of tubercles along the posterior border of the ulna shows that it was provided with a series of large quill-feathers, the remiges. It may have been employed to assist the bird in running, or possibly as a shield, somewhat in the same way as the Secretary-bird is said to use its wings, although for this latter purpose it would appear to have been too small. Of the humerus (Plate XVII. fig. 3) only the distal end is preserved in the type- skeleton. Init the radial (r.) and ulnar condyles (w.) are of the ordinary avian form, but are less distinctly separated than usual, the intercondylar groove being almost obsolete. The surfaces of the condyles are rough, and probably a considerable pad ot cartilage intervened between them and the bones of the forearm. The distal border of the bone is very oblique, owing to the presence on the ulnar side of a pointed process projecting some distance beyond the ulnar condyle (w.). ‘The anterior face of the bone above the condyles is deeply concave, and the portion of the concavity close to the ulnar border of the bone contains the well-marked elongated impression of the brachialis anticus (b.a.). Posteriorly there is a shallow olecranon fossa. The ulna (Plate XVII. fig. 4) is a short, very stout bone ; its shaft is slightly curved and triangular in section, the sharpest angle forming its posterior border. At its proximal end the impression of the brachialis anticus is not very clearly marked, but the surface for the humero-cubital ligament is very distinct. There is a well-developed olecranon process (o0/.), the end of which bears a rugose surface for the insertion of the triceps tendon, and from this a slight ridge runs to another rugosity (for the insertion of the long head of the triceps) lying immediately below the edge of the radial glenoidal surface. On the postero-internal border, immediately below the articular surfaces, is a tuberosity for the insertion of the biceps tendon. The sharp posterior border of the bone bears a series of well-marked tubercles, indicating the existence of large quill-feathers (secondary remiges) in the wing: there are about seven of these tubercles. Towards its distal end the shaft loses its triangular form and becomes circular in section. The distal articular head is large, and on its outer surface there are impressions marking the origin of various flexor muscles of the digits, and a groove for the passage of tendons. ‘The length of the ulna is approximately 110 mm. The radius (Plate XVII. fig. 5) is represented by the proximal end only. In the metacarpus (Plate XVII. fig. 6) the proximal articular surface is broad and SKELETON OF PHORORHACOS INFLATUS. 71 only slightly grooved at its posterior end; there is no fossa into which the cuneiform fits when the manus is flexed upon the forearm; the so-called pisiform process is almost absent. The prominence formed by the fused first metacarpal is stout and short, and terminates in a roughened surface which may possibly have formed the base of a short horny spur or knob. Distally the first metacarpal terminates in a tubercle for articulation with the phalangeal. The second metacarpal (me. 2) is short and thick, compressed laterally and slightly curved. On its upper surface there is a nearly longitudinal, faintly marked groove, and on the anterior surface of the distal end is another short groove for tendons. The third metacarpal (me. 3) is much more slender than the second; it is strongly curved and so compressed that it forms a mere band of bone. On its ventral surface at the proximal end is a prominent tubercle (¢.), which will be referred to below ; distally it is fused with metacarpal 2 in the usual mauner. The distal extremities of the fused metacarpals bear each an articular surface for their phalangeals, as in most birds. The length of the metacarpus is 76 mm. The reduction that has been undergone by the bones of the wing makes their comparison with those of other birds difficult, The distal portion of the humerus differs widely from the Albatross, in which the impression of the drachialis anticus is in a different position and there is a large ectepicondylar process. In the Birds of Prey the depression on the anterior face above the articular surfaces is always much less than in Phororhacos, and the impression of the drachialis anticus is more median ; in some the distal border is somewhat oblique and there is a trace of the distal process. In Ciconia the distal border is not oblique and the position of the drachialis impres- sion is quite different. In Grus the differences are the same as in Ciconia. In both Cariama and Psophia also the distal border is only slightly oblique (in Cariama there is a trace of a blunt distal process on the ulnar side). The humerus to which I find most resemblance in the form of the distal end is that of Diaphorapteryx, the large extinct Rail of the Chatham Islands. In this the distal border is oblique, there is a fairly prominent distal process, and the form and position of the drachialis anticus impression is similar. The humerus of Apftornis is also some- what like the fossil, but in it reduction has gone further, the division between the radial and ulnar condyles being still more indistinct and the distal process small. The short, stout ulna is not very much like that of any of the birds referred to in the comparison of the humerus, but approaches that of Psophia most nearly. The proximal end of the right radius is preserved. Its articular surface is oval in outline, and the bicipital tuberosity is strongly developed. It is almost identical with the same element in Cartama, but the characters of the radius are not sufficiently different in the various groups to be of much importance in determining affinities. VOL. XV.—PaRT 11. No. 3.—October, 1899. M 12 MR. C. W. ANDREWS ON THE SKULL AND The metacarpus differs widely from that of Diomedea, Grus, and Ciconia, and is most like that of Psophia and Cariama. In both these birds, as in Phororhacos, there is on the ventral edge of the third metacarpal, close to its base, a small process forming a projection (¢., fig. 6, Plate X VIL.) on the palmar aspect of the manus. I have not observed this in any but these birds, and of them it is most strongly developed in Cariama, in which, however, there is a prominent pisiform process. On the whole, the evidence of the wing-bones seems to point to relationship to the aberrant Gruiform birds Cariama and Psophia: but in the humerus there is resemblance to such Rails as Ocydromus, Diaphorapteryx, and Aptornis, in which also the wings have undergone more or less reduction; in this case the similarity may merely be due to convergence in degeneration. Sacrum and Pelvic Girdle. (Plate XVI.) The vertebre uniting to form the so-called sacrum (fig. 3) (synsacrum, Parker) are about fifteen in number. The most anterior articulated with the last free dorsal by the broad saddle-shaped surface of the centrum and by very large anterior zygapophyses. The neural spine of this vertebra is high and its summit is overlapped by and fused with ossified fascize which form a forward extension of the supero-anterior angles of the ilia; looked at from the side a great part of the spine is exposed to view, the anterior borders of the ilia only just overlapping its hinder edge; between the spine and the ilia there is on either side a chink-like ilio-neural canal. Tubercular and capitular facets for a free rib are present, and both are roughly semicircular, the convexity being directed upward and forward. The capitular facet, borne on a slight parapophysial elevation, is close to the anterior end of the centrum and at the lower end of a ridge which runs upward and backward, bearing near its upper end the tubercular facet. From this a bar of bone, apparently equivalent to the zygapophysial bar of a free vertebra, runs back and unites with the antero-inferior margin of the ilium; from the tubercular facet a metapophysial ridge runs forward on to the zygapophysis. Just behind and above the capitular facet there is a pneumatic fossa of moderate size. The ventral surface of the centrum is pinched up so as to form a sharp median ridge. In the second vertebra the neural spine is completely concealed by the ilia, and running upward and backward from the anterior end by the centrum there is a thick ridge which, near its lower end, bearsa facet for the capitulum of a rib, and at its upper end abuts against the lower border of the ilium; here also there is a median ventral ridge. These vertebre are probably thoracic. The centra of the next few vertebra have been destroyed, but opposite the acetabulum they are again preserved. In this region the median ventral ridge is replaced by a median groove defined by slight ridges, and disappearing opposite on the second of the true sacrals As far as opposite the middle of the acetabulum the vertebre may be SKELETON OF PHORORHACOS INFLATUS, 73 regarded as lumbar; the last of these bears a broad parapophysial process which abuts against the inner border of the ilium. Behind this there seem to be two or three Jumbo-sacrals (Mivart) in which no parapophyses are present and, except perhaps on the hindermost, no diapophyses, unless these are directed dorsad. In this region the pelvis is very narrow, so that the lateral acetabular or anterior renal fosse are extremely small. Behind the lumbo-sacrals are the true sacrals, two in number. Both these possess ventral processes, which, no doubt, are formed by parapophyses + sacral ribs. In the first the process is stout, much expanded at its outer end, and directed outward, upward, and somewhat forward. In the second this process is very slender and quickly fuses with the middle of the upper process (diapophysis), which in this vertebra is stout and directed outward and forward. The diapophysis of the anterior sacral is slightly in front of, and about 1 cm. above, the parapophysis ; it is directed backward, outward, and upward. ‘The two diapophyses are separated by a nearly circular space, and their expanded outer ends, together with that of the ventral process of the anterior sacral fuse into a common mass of bone, which abuts against and fuses with the inner surface of the ilium immediately behind the acetabulum. On the vertebral centra, at the level of the second sacral rib, there is a median hemal ridge, which is continuous as far as the third of the urosacrals which bear transverse processes. Behind the last sacral there is an interval in which no transverse processes are present: then they reappear, the first being very slender and directed backward at a very acute angle with the vertebral column; at its distal end it fuses with the succeeding process. ‘This also is directed backward, but is much stouter than the last and greatly expanded at its distal end, where it abuts against the inflected portion of the ilium forming the floor of the pocket-like renal fossa. The next transverse process is very broad, and likewise unites externally with the ilium ; it is separated from the processes in front of and behind it by oval foramina. Behind this there are four pairs of similar processes borne by vertebree with long, narrow centra having a slight hemal ridge. The free end of the last fused centrum is flat or slightly convex. On the dorsal surface the neural spines of the postacetabular “ sacrals” form a prominent median ridge, separated from the ilia, anteriorly at least, by a narrow groove ; posteriorly there are several pairs of slit-like interosseous foramina. The pelvis, seen from above (Plate XVI. fig. 1), appears very long and narrow, the postacetabular region not being expanded to any great extent. Along the middle line runs a prominent ridge, formed in front by the fused upper edges of the ilia and behind by the united neural spines of the postacetabular vertebre. Opposite the hinder border of the acetabulum is a transverse ridge which, at its outer ends, rises into a pair of very prominent supra-trochanteric crests; these, with the anti-trochanters, form prominent lateral projections. On each side of the median ridge in the post- acetabular region there is a narrow groove separating it from the upper edges of the M2 74 Mk. C. W. ANDREWS ON THE SKULL AND ilia, and at the bottom of this there are traces of one or two pairs of interosseous foramina. The ilia terminate posteriorly in prominent ilio-caudal processes, which project some distance beyond the last fused caudal. From these processes there runs inward and forward on either side a ridge which, after continuing a short distance parallel to the vertebral axis, runs downward and outward to the upper angle of the ischiadic foramen; this ridge seems to be the hinder part of the ilio-lateral ridge, which is so strongly developed in many birds, é. g. Rails. Anterior to it the dorsal surface of the ilia are evenly convex from side to side. Viewed from the front, the most conspicuous points are (a) the relatively large size of the centra of the sacral vertebre; (8) the extreme lateral compression, the ilia meeting in the iliac crest at an angle of not more than 10°; (y) the very narrow slit- like opening of the ilio-neural canals; (8) the very prominent supra-trochanteric crests which, from this point of view, completely hide the whole of the dorsal surface of the postacetabular region. In a posterior view the supra-trochanteric crests and anti-trochanters completely conceal the whole pre-acetabular region. Behind them we have first the median ridge of the sacrum, on each side of this a shallow groove, and external to this again the ilia, the regular downward curve of which forms an are of a circle, and below these the ischia curving outward and diverging one from another. This portion of the pelvis has been crushed laterally, but not to any great extent. Looking at the pelvis from the side (Plate XVI. fig. 2) it will be seen that the acetabulum is considerably in front of the middle of the pelvis, and measured from the middle of the acetabulum the pre-acetabular portion of the ilium is 140 mm. in length, the postacetabular 265 mm., or roughly as 10 to 19. ‘The acetabulum itself is nearly circular in outline, its antero-posterior diameter being slightly the greater (acet.); through it the arches of the lumbo-sacral vertebre are visible. Its antero- superior border forms a projecting lip, and the anti-trochanter, the lower border of which is excavated by a narrow fossa, is large and very prominent, and, except for a narrow groove, is continuous above with the supra-trochanteric crest. Anterior to the acetabulum is the slightly concave and nearly vertical gluteal fossa of the ilium, below the edge of which the centra of the pre-acetabular sacral vertebre are visible. The ischiadic foramen (¢s.f.) is very large and roughly ovoid in outline. Just within its antero-superior angle, and bordered anteriorly by the anti-trochanter and superiorly by the edge of the ilium, there is a pocket-like fossa, which internally is separated by a slight ridge from the mass of bone formed by the fusion of the processes of the true sacrals. The anterior edge of thei schiadic foramen is continued downward as a ridge crossing the shaft of the ischium and terminating below in a prominent process (styliform process), against the inner side of which the pubis was in close contact, thus enclosing an obturator foramen (0.f.) which forms a narrow oval opening extending from beneath the posterior half of the acetabulum as far back as the front of the ischiadic foramen. SKELETON OF PHORORHACOS 1INFLATUS. 75 Posteriorly the pelvis ends in the sharp hinder angle of the ischium, which is separated by a semicircular bay from the ilio-caudal process of the ilium. Seen from below (Plate XVI. fig. 3) the narrowness of the pelvis is even more noticeable than in the dorsal view. ‘There is practically no internal iliac fossa, the centra of the anterior “sacral” vertebre being very large and projecting far below the edges of the ilia. The anterior renal fosse are extremely narrow, and are scarcely visible from this point of view; they are separated by the processes of the two sacrals from the enormous posterior renal fossee which, as in Rails and some other birds, are prolonged backward into long pocket-like extensions, floored by an ingrowth of the ilium. The only bird in which this ventral ingrowth of the ilium is developed to anything like the length seen in Phororhacos is Fulica, in which the broad transverse processes of three or four urosacral vertebre unite with it, while in the fossil there are six or seven such vertebre. The Ilium. (Plate XVL., il.) In its pre-acetabular region the dorsal border of the ilium is convex, the ventral concave, and in front the two are united by a nearly straight anterior border which is slightly inclined forward. The ilio-pectineal process is broken, but seems to have been small. As already mentioned, the postacetabular region is considerably longer than the pre-acetabular, from which it is sharply separated by the supra-trochanteric crest. It is nearly equal in width throughout its length. Posteriorly it terminates in the prominent ilio-caudal process, the lower part of which, however, may be formed by the ischium: from the end of this process runs a ridge which seems to mark the junction of the two bones, and certainly is continuous in front with the suture between them. A second ridge, commencing at the hinder border of the bone close to its sacral border, runs forward and then downward to join that just described close to the ischiadic foramen. The Ischium. (Plate XVI., zs.) The share which this bone takes in the formation of the acetabulum cannot be determined, owing to the complete fusion of the pelvic elements in that region. Beneath the ischiadic foramen it forms a bar of bone 15 mm. wide, which, near its proximal end, bears on its ventral edge a short, stout process which touches the pubis, thus enclosing an obturator foramen. Beneath the ischiadic foramen the ischium expands into a broad plate of bone, the outer surface of which is concave from above downward, and the inner traversed by a prominent rounded ridge marking the prolongation of the axis of the bone. The upper edge unites closely with the ilium, the lower curves downward and outward and terminates posteriorly in an angular process which projects slightly further back than the ilio-caudal process, 76 MR. C. W. ANDREWS ON THE SKULL AND The Pubis. (Plate XVI. fig. 2, pw.) The pubis is probably imperfectly preserved ; all that now remains is a slender rod of bone arising beneath the middle point of the acetabulum, and running back and terminating against the inner side of the styliform process of the ischium, thus closing an obturator foramen as above mentioned. Whether when complete it extended further back in the usual way cannot be determined. Comparison with the Pelvis of other Birds. Comparison of this pelvis with that of the Ratitee shows at once that Phororhacos certainly does not belong to that group, though it may have been, and probably was, ‘‘Ratite” in the strict sense of the term. It is true that in its length, narrowness, and the large development of the supra-trochanteric processes there is some similarity to the pelves of Dromeus and Struthio, a similarity which, no doubt, is merely the consequence of adaptive modification due to a like mode of progression. On the other hand, in essential points of structure, such as the form and relations of the ischia and pubes to each other and to the ilia, the structure of the ‘‘ synsacrum,” particularly in the distinctive form of the two true sacrals and in the form of the renal fossz, the pelvis in the fossil is very different from that of any Struthious bird. In Apteryx alone the sacrum shows slight points of similarity. Among the Carinate birds the pelves which show most resemblance in general outline to the fossil are those of the Grebes and Hesperornis; but in both of these the pre- acetabular portions of the ilia do not unite with the spines of the “sacrals ” to form an ilio-neural crest, but remain separated from them by a considerable interval, while, on the other hand, the postacetabular portions approach one another very closely, and may even unite in the middle dorsal line; the exact reverse is the case in Phororhacos and most other birds. To the pelves of the Cranes and Rails the similarity is in many ways remarkable, and probably indicates a real relationship with those birds. The form and character of the renal fosse, particularly the peculiar pocket-like prolongations of the posterior fossee, and the general structure of the “ synsacrum,” especially the distinctness of the true sacrals, are almost identical with those described in the fossil. One difference, however, must be pointed out, viz. that while in Phororhacos the postacetabular region is the longer, in nearly all the Gruiformes the reverse is the case. In Fulica, however, the post- and pre-acetabular regions are of nearly equal length, and in Cariama the postacetabular portion is the longer, as in Phororhacos. - The pelvis of Cariama is also similar in the form of its posterior border, the presence of prominent supra-trochanteric processes, and, so far as can be ascertained, in the relations of the pubes. It has, however, a well-developed pelvic escutcheon, the whole postacetabular region being relatively wider than in Phororhacos. On the whole, so far as the pelvis is concerned, I see no reason for changing the opinion expressed in a SKELETON OF PHORORHACOS INFLATUS. ul former paper ', that Cariama is probably a not very remote modern representative of the extinct type. From the Birds of Prey the chief points of difference are, that in them the post- acetabular region is very short and is deflected, and there is no ilio-caudal process to the ilium. In Serpentarius, however, the first two of these characters are much less marked, and in this respect, therefore, it approaches the fossil, these characters being probably merely correlated with the fact that it is a much better walker than the other Accipitrine birds. In the Storks the pre-acetabular portions of the ilia meet at a very obtuse angle, the postacetabular region is very wide, and there are no recesses to the post-renal fosse. In the Herons these fossee, though present in some, are small, the ilio-neural canals are widely open behind, and the pelvic escutcheon is wide; certain Galliformes, e. g. Phasianus and Francolinus, in which renal pockets are present, differ from the fossil in much the same respect. In the account of the skull some points of similarity with that of the Albatross were referred to, but in the pelvis no similarity whatever can be detected. Among the more notable differences may be mentioned the complete absence of pocket-like extensions of the renal fossee, almost complete absence of supra-trochanteric processes, the backward prolongation of the ischia and their mode of union with the ilia, and the indistinctness of the true sacral vertebra. Ameghino, in his manuscript catalogue, distinctly states that this pelvis and other bones were found associated with the skull above described, and formed part of one individual; and if this is the case, as there is no reason to doubt, the structure of the pelvis seems completely to} outweigh whatever evidence of relationship with the Albatrosses may be found in the skull. The dimensions of the pelvis are :— cm. Hxtremeylen ge chimenne nc Wew col bls eeTh MCT MUliaayleg 200 (sm faa (om ata cee aol (MLO Length ofilium .. . : Sy AS iota RM GLiccm ee athei cae ween ele) Longitudinal diameter of menedie fone tieht A aE MEN er itt) Bae Ryser on NOs) Wertical diameter ofaschiadie foramen.) 4) 4 = Length of obturator foramen. . . . . Fae ey incites 2: eet ESO Greatest height from hemal ridge of sacrum to ton Oealbeyocxess 6 6 o hs) WAGED Oo EMOTES 5 5 6 6 6 6 6 9 6 0 0 6 9 o oo oo op lb SURO NMSe OAEIS Go 5 6 6 0 0 0 5 6 0 4 Oo OD 10°6 » between upper edges of acetabula . . . ...... =. . «495 5) (least)) behind! supra-trochantericicrests) 50 3. 9) «| =.) 1) 638 r (UGASD) jorenestonlemmrio 6 6 5 5 5 6 5 8 bp oo oo BY pe icentrum oftrstitused sacral vertebray. |e ll) tell leeks OLS anetin @e emvemin 5b go 6 op ole eo Opole. Oo BSD 1 «The Ibis,’ 1896, p. 1; see also ‘Science Progress,’ vol. vy. (1896) p. 398. 78 _ MR. Cc. W. ANDREWS ON THE SKULL AND Vertebre. Of the vertebra, other than those forming the synsacrum, only three are preserved in this specimen. Of these one is the last free dorsal and the others anterior free caudals. The last dorsal has a short centrum, terminated by the usual saddle-shaped articular surfaces, of about equal height and breadth. The ventral portion of the anterior two- thirds of the centrum is pinched up to form a prominent heemal ridge, the summit of which has been broken away. Laterally the anterior end of the centrum is widened out by stout parapophysial prominences, bearing each an oval cup for the head of the rib. The diapophysis is broken away: immediately beneath its base is a large pneumatic fossa, and another occurs on the side of the centrum. The zygapophyses are of the usual form; the neural spine is very massive, and has its anterior and posterior surfaces roughened for ligamentous union with the spine in front and behind it: its upper portion is broken away. The caudal vertebree (Plate XVII. fig. 7) have centra about as long as their anterior face is wide, slightly constricted in the middle, and bearing towards their hinder end a pair of strong, backwardly-directed transverse processes, only the bases of which remain. The anterior articular face is wider than high, the posterior about equal in the two directions; the anterior surface is slightly concave, the posterior convex, but with a median pit, which probably marks the primitive position of the notochord. The neural arch does not extend quite to the hinder end of the centrum, and the neural spine is greatly thickened at its upper end, which forms a flat bilobate surface, which was evidently connected with the vertebr before and behind by tendons, probably more or less ossified. — The Femur. (Plate XVII. fig. 8.) The head of the femur rises to a marked degree above the level of the trochanter, the upper portion of the elevation being formed by a large blunt-pointed process, which partly divides the deep pit for the ligamentum teres into two parts. Ventrally also the articular surface is divided by a shallow groove into an anterior and a posterior lobe, so that when looked at from the inner side the head appears to be imperfectly trilobate. The neck is much hollowed out on its ventral surface, and the head is clearly peduncu- late. The articular surface of the trochanter is continuous with that of the head, at least posteriorly. On the anterior, and particularly on the outer, face of the trochanter are a number of extremely distinct impressions of muscle-insertions. The shaft is as nearly as possible straight, there being only a slight forward convexity in its lower portion. In the middle it is cylindrical in section, but just above the condyles it is somewhat flattened from before backward. The postero-internal border is defined by an extremely conspicuous linea aspera, which, as Ameghino remarks, is a veritable projecting lamina of bone. It runs from end to end of the shaft, terminating below in the upper angle of the inner condyle. The anterior face of the shaft is likewise SKELETON OF PHORORHACOS INFLATUS. 79 marked by an intermuscular ridge running obliquely downward and inward from the outer edge of the trochanter: in its lower third it forks, one branch running to the upper anterior angle of each of the condyles; in its upper portion this line no doubt separated the surfaces for the crurwus and the vastus eaternus. There is a deep popliteal fossa, which is separated by a prominent bar of bone from the intercondylar fossa, which is marked by two deep pits for ligaments. The outer condyle projects considerably below the inner; its fibular ridge is very prominent, and the surface for the fibula slightly concave from side to side. On the lower end of the outer condyle is a very distinct facet for the tendon of the outer head of the tibialis anticus. Anteriorly the condyles project considerably, and the rotular channel is comparatively deep. Comparing the femur of Déomedea with the fossil, we find that it differs, among other points, in the shallowness and form of its popliteal fossa and in the great antero-lateral compression of the lower end of the shaft. ‘The head, moreover, is much less distinctly pedunculate. The femur of Grus differs in the somewhat more curved shaft, the elevation of the trochanter above the head, the somewhat shallower popliteal fossa, and the smaller degree of obliquity of the distal articulation. In Psophia the trochanter is high and the popliteal fossa shallower than in the fossil. In Cariama the shaft is rather curved and the trochanter is somewhat raised above the head, which is very similar to that of the fossil: the distal end differs only in the somewhat greater shallowness of the popliteal fossa. In the Rails the shaft of the femur is always much curved, and the trochanter rises considerably above the head. In the Birds of Prey (including Serpentarius) the femur is pneumatic, the trochanter rises above the head, and the neck, as a rule, is very short. The dimensions of the femur are :— mm. er ot tues Myr semerpeme saree rea Seats Cole Voile tap, cms Py US Kel Sey cus. [etaiecicicietee fh Windih at ronal’ Gl 6 9 G6 5 6 8 oe bb ee Go 8 HY Pearavadistalven Gemma: wea mere MLN ue McK Mc aagl jl Ow oy | Geramaloval? ee gg a) oo la a he lela 6 Bian GD Cmenmtaanca orrmnilla Organ 56.5 5 6 68s 8 o 0 8 5» 5 © ell The Tibio-tarsus. (Plate XVII. figs. 9 & 10.) The tibio-tarsus is a comparatively long and slender bone. The shaft is sonewhat flattened on its anterior and posterior surfaces, and tapers slightly towards its distal end. It is straight, but the median border near its distal end curves slightly iuward, so that the inner condyle projects considerably inward ; the outer condyle is ina line with the outer border of the shaft. The anterior surface is bordered internally by a strongly- marked Jinea aspera, which is continuous above with the procnemial crest, and towards you. xv.— Parr m1. No. 4.—October, 1899. N 80 Mk. C. W. ANDREWS ON THE SKULL AND the lower end forms the inner border of the very deep groove for the extensor tendons ; it terminates at the inner end of the extensor bridge. This latter lies obliquely ; near the upper border of its inner end there is a deep pit for the attachment of the oblique ligamentous sling through which the tendon of the tibialis anticus passes: the other end of the ligament is inserted lower down on a well-marked tubercle just above the outer condyle; the lower border of the bridge passes externally into a prominent tuberosity, such as occurs in the Storks, &c., and serves for the insertion of the ligament bridging the extensor tendons. The condyles are about equal in size, and are separated by a deep intercondylar gorge which opens above into a median concavity lying in the middle line immediately below the last-mentioned tuberosity, and, when the leg is flexed, receiving the large inter- condylar tuberosity of the metatarsus. Posteriorly the intercondylar groove is short and shallow, but is still sharply defined, being bounded by the prominent posterior projections of the condyles. Looked at from the side the outer condyle is nearly circular in outline, while the inner is much elongated from before backward ; both are concave on their outer surfaces, The upper end of the bone is relatively small, but the pro- and ectocnemial crests are fairly well developed and rise considerably above the proximal articular surface, which slopes upward and forward. The inner glenoid surface is flat or only slightly convex; posteriorly it forms a considerable projection and is separated by a notch from the outer or fibular surface, which is small and very convex in all directions. There are no pneumatic foramina. ‘The centre of the shaft is occupied by a large smooth-walled cavity, and the thickness of the bone in the middle of the shaft does not exceed 4 mm. The tibio-tarsus of Diomedea differs from that of the fossil in the much greater breadth of the intercondylar fossa, the nearly transverse direction of the extensor bridge and the absence of a tuberosity at its outer end, and in possessing a much Jarger cnemial crest. In Grus the tibia differs widely in the relatively greater width of the distal articulation, in which the condyles are small and the intercondylar groove broad; on the other hand, there is a median tuberosity. In Psophia the extensor bridge is much more transverse, and the outer condyle. considerably the larger, as in Rails. In Cariama the condyles are relatively smaller and the intercondylar groove shallower and broader. In Ciconia the tibia presents considerable resemblance to the fossil, but the articular surfaces for the femur are less distinctly separated, the extensor bridge is less oblique, and the groove for the tendons occupies the whole width of the shaft. In the Accipitrine birds, including Serpentarius, the cnemial crests are very small, SKELETON OF PHORORHACOS INFELATUS. 81 the groove for the extensor tendons nearly median, there is no distinct tuberosity at the outer side of the extensor bridge, and the condyles are comparatively small and separated by a broad intercondylar gorge; while, in correlation with the small size of the intercondylar tuberosity of the metacarpus, there is no depression at the upper end of the intercondylar groove, into which it fits when the bones are flexed one upon another. The dimensions of the tibio-tarsus are :— em, Length, including cnemial crest 895 s without cnemial crest . 37-5 Width at distal end . Sige Wee Gn oly ces Meh yeu Came a Meant 3 Width of shaft at narrowest point, 7. e. about 3 cm. above extensor bridge. 2:7 Circumference at same point . 8:2 Width from front to back at same point 2-0 The Metatarsus. (Plate XVII. figs. 11-17.) The glenoidal facets for articulation with the tibia are deeply concave; the outer (0.) is slightly below the inner, and has a bluntly pointed process on its outer border. There is a prominent rounded intercondylar process (¢.c.), which on the outer side, at its base, bears a shallow pit for the insertion of the outer semicircular ligament. On the outer side of the bone, immediately below the anterior angle of the outer glenoid facet, is a rough surface for the insertion of the lateral tibio-tarsal ligament. Posteriorly the hypotarsus (hy.) forms a broad projecting mass, consisting of an outer broad ridge and a narrow inner one, separated by a shallow groove; the outer and inner surfaces of the hypotarsus are also slightly grooved, but there are no closed or nearly closed channels for tendons. The whole hypotarsus is short, and the two ridges constituting it converge below into a simple median ridge, which soon dies away on the hinder surface of the shaft. In the slight depressions on either side of the hypotarsus are the posterior openings of the interosseous foramina, the inner being slightly higher up than the outer. The anterior face of the shaft is deeply concave in its upper part, so that the interosseous foramina open anteriorly at the bottom of a deep fossa, the outer a little above the inner. Immediately below these foramina is a pair of oval rugosities, the outer slightly above the inner, for the insertion of the tibialis anticus. Internal to these is a deep groove, over which the inner border of the bone forms an overhanging lip; this groove is for the tendon of the ealensor communis digitorum. The concavity of the upper portion of the anterior face is continued as a broad channel extending down about the upper half of the shaft; but there is no specimen in which this region is complete, so that the exact length of the bone is unknown. The posterior surface is occupied at its upper end by a ridge forming a prolongation Na 82 MR. C. W. ANDREWS ON THE SKULL AND of the hypotarsus. Lower down the shaft it is narrowed and flattened, its outer border being formed by a prominent ridge, which is continued upward and outward, and terminates above in the process which forms the outer border of the external glenoid fossa ; its inner border is formed by a less prominent ridge, which is not, as in Cariama, continuous above with the inner ridge of the hypotarsus. The outer face of the shaft is bordered by the ridges which form the outer edges of the anterior and posterior faces. This surface widens out towards the middle of the bone, and is flat or even slightly concave. The outer surface is slightly convex, and is not very sharply marked off from the posterior surface. As above mentioned, the shaft is incomplete in both the metatarsi of the type- skeleton, but the distal portion (three or four inches) of the right metatarsus is very well preserved. The upper portion of the shaft preserved in this specimen is very slender and roughly triangular in section, one angle of the triangle being a rugose ridge on the posterior surface. A short distance above the inner trochlea, and at the lower end of the ridge just mentioned, is the surface for the attachment of the hallux (ha/.). Below this the posterior surface is slightly concave from side to side. The anterior surface has near its outer side a shallow groove deepening distally and having at its lower end the foramen for the tendon of the adductor digitt extern. ‘This foramen leads into two channels, one opening between the outer and middle trochlea, the other on the posterior surface of the bone immediately above the trochlee. The distal trochlee are arranged in a slightly curved line (fig. 17). The median trochlea (tr. 3) is very large and its articular surface is sharply delimited, both anteriorly and posteriorly; the median groove is much deeper behind than in front. The inner trochlea (¢. 2) is relatively small, and is considerably shorter than the middle one. Its articular surface is evenly convex from side to side, except on its posterior side, where it is slightly grooved. It is situated a little posterior to the middle trochlea, behind which it projects considerably. The outer trochlea (¢r. 4) is rather larger than the inner and is slightly longer; it also projects considerably behind the median, and its articular surface is grooved posteriorly only. The lateral surfaces of all the trochleew are deeply excavated by pits for the insertion of the ligaments of the toes. In comparing the metatarsus of Phororhacos with that of other birds, it must be remembered that it has probably been profoundly modified in correlation with the loss of the power of flight and the consequently exclusively pedestrian progression of the bird. Among the peculiarities which may probably be thus accounted for are the simple hypotarsus, the relatively great length of the bone, and the disposition of the distal trochlee. The metatarsus of Diomedea differs in several respects, the chief of which are complete absence of all trace of a hind toe, relatively large size of the articular ends compared with the shaft, breadth and complication of the hypotarsus, and absence of a pit for ligament at the outer side of the base of the intercondylar tuberosity. SKELETON OF PHORORHACOS INFLATUS. 83 In Grus the wetatarsus differs in the complication of the hypotarsus and the position of the inner (2nd) trochlea, which is high up and much deflected backward. In Psophia there is a close canal in the hypotarsus, otherwise the bone is much like the fossil. In Cariama the metatarsus in most respects closely resembles that of Phororhacos, but it is proportionately more slender, and at its distal end the lateral trochlez (particularly the inner 2nd) are less reduced. The metatarsus of Ciconia is similar in many points, but at the proximal end the hypotarsus is longer, its crests much more prominent, and it is situated further below the glenoid surfaces than is the case in Phororhacos. At the distal end the chief difference is that the lateral trochlez are larger in proportion to the median one. In most Accipitrine birds the metatarsus is widely different, both in the form of the hypotarsus and the deep excavation of the hinder surface of the shaft for the various muscles of the toes and the arrangement of the trochlee. Serpentarius, however, has undergone such modification in correlation with its terrestrial habits that it approaches Phororhacos very much more closely than do the rest of the group: this is chiefly the result of the elongation of the shaft and the alterations in the form of the talon and in the arrangement of the trochlee. Nevertheless, there are several points, such as the almost complete absence of an intercondylar tuberosity, in which it differs widely from the fossil. The dimensions of the metatarsus are :— carl \ninin om jain gind) eg 5 th 6 oo G8 eG oo gt 8 Se Pee atistaliendayey ace ciate nowy 0 lap isy RU ty fe oe emcee vane ak » Obmickile Gromit 5 oo 56 0 0 6 6 6 0 6 69 6 6 9 o ily Only one of the phalanges of the hind foot belonging to the skeleton above described is preserved, but there is a nearly complete left pes of another individual of the same species. In thisit is seen that the hallux was well developed and terminated ina strong hooked claw: the metatarsal articulates with the tarso-metatarsus by a simple knob, and terminates distally in a trochlear surface with a deep pit for ligament on its outer side. The total length of this toe is about 53 mm. In the second digit there are, as usual, three phalanges, the terminal (3rd) one being sharp-pointed and slightly hooked; the remaining two, though shorter than the phalanges of the middle digit, are nearly as stout. The lengths of the phalanges are: Ist, 31 mm. ; 2nd, 32 mm.; 3rd (ungual), 36 mm.; the total length of the toe is about 93 mm. The middle toe consists of four phalanges, the first of which is nearly as long as the two proximal phalanges of the second digit together ; both it and the two succeeding digits are somewhat compressed from above downward. The ungual phalangeal is a powerful hooked claw. The lengths of the phalanges are: Ist, 55 mm.; 2nd, 38 mm.; 5rd, 27 mm.; 4th (ungual), 44 mm. Of the inner toe only the ungual and one of the other phalanges is preserved; the length of the former is about 27 mm. §4 MR. C. W. ANDREWS ON THE SKULL AND ConcLupING REMARKS. In the preceding pages only a few of the types to which the fossils have been compared are mentioned, they being the only types to which any resemblances pointing to possible affinities could be made out. And even among these there are some to which the similarity is so slight that they also might perhaps have been omitted. For instance, in the case of Diomedea, it is only in the structure of the palate and one or two other points in the skull that any similarity with Phororhacos can be detected, the rest of the skeleton being strongly against any such relationship. In fact, it seems to the writer that the only groups that really come into question are the Falconiformes and aberrant Gruiformes, Cariama, Chunga, and to a less degree Psophia. It is toa relationship with the former group that the general appearance of the skull and feet of Phororhacos seems to point ; but, as above shown, the general structure of the skeleton, particularly of the pelvis and hind limb, is opposed to this, while, on the other hand, it is strongly in favour of affinity with the Gruiformes. In this, however, the skull at first sight seems to oppose a number of serious difficulties, differing widely from the skull of the typical Crane both in general appearance and in many points of structure, ¢.g. in its holorhinal nares and desmognathous palate ; but it is precisely in these points that the Cariamide also differ from the typical members of the group and therefore resemble the fossil. Still the differences between the skull of Phororhacos and that of Cariama are so great that some explanation of them is necessary if the close affinity of the two forms is to be maintained. This explanation seems to be found in the extraordinary size of the beak in the fossil, a specialization that has led to great modifications in several regions of the skull. ‘Thus the size and weight of the upper jaw have led to changes in the palate in the direction of greater rigidity, resulting in the extensive union in the middle line of the maxillo-palatine plates and their complete fusion with the anterior ends of the palatines (this occurs also in ‘Toucans and Hornbills). Moreover, the development of the peculiar prominent ridges formed by these combined elements, and, as already described, projecting below the tomium on either side the middle line, also tends greatly to increase the rigidity of the beak, and possibly the modes of union of the pterygoids with the basis cranii may still further add to it. Again, with the great massiveness of the mandible is correlated the large size of the temporal fosse, and the weight of the skull as a whole has led to an increase in the area of the occipital surface by the extension of the lambdoidal and paroccipital crests, to which the muscles supporting the head are attached. ‘The characters thus accounted for give the skull of Phororhacos its peculiar form, so different from that of Cariama, and would be evidence of a high degree of specialization even if the rest of the skeleton were unknown. It was no doubt this high degree of specialization that brought about the extinction of this giant bird as soon as the conditions to which it was adapted underwent some SKELETON OF PHORORHACOS INFLATUS. 85 change, while the smaller more generalized Cariama and Chunga, or rather their ancestors, survived. In fact, it appears that Phororhacos stands in somewhat the same relation to the Cariamide that such forms as Glyptodon and Panochthus stand to the modern Armadillos. All the specimens described are from the Santa Cruz Beds of Monte Observacion, Patagonia. The age of these deposits is doubtful, but they probably correspond to some part of the Miocene of the Northern Hemisphere. EXPLANATION OF THE PLATES. PLATE XIV. Skull and mandible of Phororhacos inflatus Amegh., p. 56. From side. # nat. size. PLATE XV. Skull of Phororhacos inflatus Amegh., p. 56. Fig. 1. From below. # nat. size. Fig. 2. From above. 3 nat. size. b.pt. Process of pterygoid articulating | o0.g. Orbital process of quadrate, p. 63. with base of the skull, p. 57. pal, Palatine, p. 62. 6.t. Basitemporai platform, p. 57. pe,f. Precondylar fossa, p. 57. eth. Ethmoid (?), p. 60. p.mez. Premaxilla, p. 61. if. Facet for inner head of quadrate, p.orb.p. Postorbital process, p. 58. p. 58. | pt. Pterygoid, p. 62. i.0.s. Interorbital septum, p. 59. g. Quadrate, p. 62. lac. Lacrymal, p. 58. gj» Quadrato-jugal, p. 60. lr, Lambdoidal ridge, p. 58. r. Rostrum, p. 57. m.t. Mammillary tuberosities, p. 57. sq. Squamosal, p. 58. max.pal. Maxillo-palatine plates, p. 62. | sup.o. Supra-orbital plate, p. 59. n. Nasals, p. 59. | ty. Tympanic cavity, p. 57. oc.c. Occipital condyle, p. 57. zy. Zygomatic process, p. 58. of. Facet for outer head of quadrate,p.58. | PLATE XVI. Pelvis of Phororhacos inflatus Amegh., p. 72. } nat. size. Fig. 1. From above, Fig. 2. From side. Fig. 3. From below. acet. Acetabulum, p. 74. o,f. Obturator foramen, p. 74. a.t. Ante-trochanter, p. 73. p. Pubis, p. 76. il. Ilium, p. 75. s. Sacral vertebree, p. 72. as. Ischium, p. 75. u.s. Urosacral vertebre, p. 73. is,f. Ischiadic foramen, p. 74. 86 ac, b.a. hyp. mMC.2, MC.3. oO. ON THE SKULL AND SKELETON OF PHORORHACOS INFLATUS. PLATE XVII. Limb-bones of Phororhacos inflatus Amegh. 3 nat. size. ics! — ge AAO on FP oh . Left femur, p. 78. co co . Right coracoid, anterior surface, p. 68. . Right scapula, inner surface, p. 69. . Distal end of left humerus, p. 70. . Right ulna, olecranon restored from left side, p. 70. . Right radius, proximal end, p. 70. Right metacarpus, palmar surface, p. 70. . Caudal vertebra, p. 78. . Right tibio-tarsus, p. 79. Tene, WD6 ep . distal end, p. 80. Fie. 11. Left tarso-metatarsus, postero-internal view of proximal end, p. 81. Fig. 12. + PA antero-external view of proximal end, p. 81. Fig. 13. 5 5 anterior view of proximal end, p. 81. Fig. 14. Right tarso-metatarsus, Fig. 15. oe a Fig. 16. # 3 Fig. 17. 99 99 Acrocoracoid, p. 69. Impression of brachialis anticus, p. 76. . Inner articular facet at proximal end of metatarsus, p. 81. . Intercondylar tuberosity, p. 81. Inner condyle of tibia, p. 80. . Surface for hallux, p. 82. Hypotarsus, p. 81. Second and third metacarpals, p. 71. Outer articular surface at proximal end of metatarsus, p. 81. posterior view of distal end, p. 82. anterior view of distal end, p. 82. upper articular surface, p. 81. lower articular surface, p. 82. o.c. Outer condyle of tibia, p. 80. ol. Olecranon process of ulna, p. 70. p.l. Processus lateralis of coracoid, p- 68. r, Radial condyle of humerus, p. 70. t. Tuberosity at base of the third metacarpal, p. 71. tr.2, tr.8, tr.4. Trochles of 2nd, 3rd, and 4th metatarsals, p. 82. u. Ulnar condyle of humerus, p. 70. SONLWTANI SOOVHYOUOH ‘dure soig we typ “Ua 32 TPP Usein ALE Ob AY 920 2A 9° 2E PIL inv apane ie “SNLVTANI SOOVHYOUOH * dure qwta yA “Tap u rp C sotg ita Fury quad Perea « | 1 anu d , ’ 1 i} Z i} i} 1 y good. IOMOUL / AY Ob AX 90 °F GPG CUaYG aa b ate aw y Trans. L006. Soo Vl. XV. G6 AV. J.Green. del.et hth Mintern Bros.imp PHOROREACOS INHEATUS: Trans. 200k. Soo Vo6. AV. FE. AVL. Mintern.Bros imp. J.Green del.et hth. PHORORHACOS INFLATUS. : sue f TRANSACTIONS OF THE ZOOLOGICAL SOCIETY OF LONDON (continued). To Fellows. To the Public. oy Ss) ds Ba GU VOLUME XIII. (1891-1895, containing 62 Plates). . Price6 8 38 811 0 Partl. (1891, containing 6 Plates) . . .. . » O15 9 Uae Cae) » 2. (1891, contaming 6 Plates) . Ge AU aa!) desea eel) » 9 (1891, contaiming 6 Plates) » O18 O 1 4 0 » 4 (1892, containing 1 Plate) . » O 4 6 0 6 0 » 9. (18938, containing 9 Plates) se Oras 015 0 » 6. (1893, containing 4 Plates) sO (0) 012 0 », 4%. (1893, containing 6 Plates) . . . 0) 0 012 0 », 8. (1894, contaiming 6 Plates) » 0 9 O 012 0 » 9. (1894, containing 6 Plates) . . . » Oll 38 015 0 » 10. (1895, containing 5 Plates) . . .. . » 9 9 O 012 0 », 11. (1895, containing 7 Platesand TitleandIndex),, 015 9 Tee a O VOLUME XIV. (1896-1898, containing 47 Plates) . Price5 5 O LAO eG) Part 1. (1896, containing 2 Plates) . . .. . » O 6 O O-) S220 » 2 (1896, containing 6 Plates) . - . . . sar OOO 012 0 » 9 (1897, containing 9 Plates) Fad OL ea Ieee) » 4. (1897, containing 10 Plates) sel lOO 2 0 0 >» 0 (1898, containing 2 Plates) 5 OO) 08 0 » 6. (1898, containing 7 Plates) Ee OSH O Le Aor » @. (1898, contaming 6 Plates) ..... er Oca 015 0 », 8. (1898, containing 5 Platesand TitleandIndex),, 0 9 O 012 0 VOLUME XV. Part 1. (1898, containing 8 Plates) . . . . . Price O 15 Tee 0 » 2 (1899, containing 5 Plates) ae Oke Olea Ol aes » 9 (1899, containing 4 Plates) . . .. ., poe) Oa ai Onah caeaerey ress Oye Tues CO CONTENTS. Ill. On the Extinct Birds of Patagonia.—I. The Skull and Skeleton of Phororhacos inflatus Ameghino. By C. W. Anprews, B.Sc., F.Z.S., Assistant in the British Museum, Natural History. (Plates XIV.—XVIIL). . . . page 95 THE PUBLICATIONS OF THE ZOOLOGICAL SOCIETY OF LONDON. Tue scientific publications of the Zoological Society of London are of two kinds—“ Proceedings,” published in an octavo form, and ‘ Transactions,” in quarto. According to the present arrangements, the “ Proceedings” contain not only notices of all busi- ness transacted at the scientific meetings, but o all the papers read at such meetings and recom- y the Committee of Publication. A largé number of colonred plates and engravings are attached to each annual volume of the “ Proceedings,” to illustrate the new or otherwise remarkable species of animals described in them. Among such illustrations, figures of the new or rare species acyuired in a living state for the Society’s Gardens are often given. The “ Proceedings’ for each year are issued in four parts, on the first of the months of June, Angust, October, and April, the part published in April completing the volume for the precedmg year. : mended to be published in the “ Proceedings” The “ Transactions’ contain such of the more important communications made to the scientific meetings of the Society as, on account of the nature of the ~2ates required to illustrate them, are better adapted for publication in the quarto form. They are issued at irregular intervals. Fellows and Corresponding Members, upon payment of a Subscription of £1 1s. before the day of the Anniversary Meeting in each year, are entitled to receive all the Society’s Publications for the year. hey are likewise entitled to purchase the Publications of the Society at 25 per cent. less than the price charged for them to the Public. A further reduction of 25 per cent. is made upon purchases of Publications issued prior to 1871, if they exceed the value of five pounds. Fellows also have the privilege of subscribing to the Annual Volume of the ‘ Zoological Record ” for a sum of £1 (which includes delivery in the United Kingdom only), but this privilege ouly holds good if the subscription is paid before the First of December in each year. Such of these publications as are in stock may be obtained at the Society’s Office (38 Hanover Square, W.), at Messrs. Longmans’, the Society’s publishers (Paternoster Row, E.C.), or through any bookseller. October 1899. P. L. SCLATER, Secretary, TRANSACTIONS OF THE ZOOLOGICAL SOCIETY OF LONDON. Vou. XV.—Part 4. LONDON: PRINTED FOR THE SOCIETY, SOLD AT THEIR HOUSE IN HANOVER-SQUARE ; AND BY MESSRS. LONGMANS, GREEN, AND CO., PATERNOSTER-ROW. December 1899. Price 12s. Taylor and Francis, Printers, | {Red Lion Court, Fleet Street. TRANSACTIONS OF THE ZOOLOGICAL SOCIETY OF LONDON. To Fellows. To the Publie. ey. th DiSey O VOLUME I. (1833-1835, containing 59 Plates). . Price 813 6 . . . 418 O* VOLUME II. (1835-1841, containing 71 Plates). . , 4 0 O DOr Ors VOLUME III. (1842-1849, contaming 63 Plates). . , 3 8 3 . . . 4 11 0% VOLUME IY. (1851-1862, containing 77 Plates). . ,, 6 2 0... 8 2 6+ VOLUME _ V. (1862-1866, containing 67 Plates). . ,, Diy Ape daokcime ee OmlOee Ost VOLUME VI. Gereuecs: contamines92 slates) 2 ll oO re elo tO) aOr: VOLUME VII. (1869-1872, containing 73 Plates). . ,, 10 4 0 . . .1812 OF VOLUME VIII. (1872-1874, containing 82 Plates). .,, Fate y os) ail eiarremme ober abla (0)5p VOLUME IX. (1875-1877, containing 99 Plates). . ,, 12 16. . .16 2 OF VOLUME X. (1877-1879, containing 95 Plates). . ,, 10 0 8 . . .18 7 OF GENERAL INDEX, Vols. I. to X. (1883-1879) . ,, O2ni, FO OO a0 VOLUME XI. (1880-1885, containing 97 Plates). a QAO 5 hy SAL ANG VOLUME XII. (1886-1890, containing 65 Plates). . ,, De 8. 20 lone iad O * No copies of these volumes remain in stock. rt Only complete copies of these volumes are left in stock. Continued on page 3 of Wrapper. IV. Second Contribution to the Ichthyology of Lake Tanganyika.—On the Fishes obtained by the Congo Free State Expedition under Lieut. Lemaire in 1898. By G. A. Bouuencer, F.R.S., F.Z.8. Received May 13, read May 16, 1899, [Puates XVIII.-XX.] THe extraordinary richness of Lake Tanganyika in Perciform fishes of the family Cichlide, first revealed by Mr. Moore’s collection, reported upon in these ‘Transactions, vol. xv. 1898, pp. 1-30, pls. i.—viii., is further shown by the collection made in July— August 1898 by the Lemaire Expedition at Moliro, at the southern extremity of the Lake, already explored by Mr. Moore. Although consisting of no more than 34 specimens, the collection entrusted to me for description by the Government of the Congo Free State contains the types of ten new species, three of which warrant the establishment of new genera. These new generaare of further interest in emphasizing a feature of the Tanganyika Cichlidw to which I have drawn special attention in my previous contribution, p. 2, viz. the extent of the lateral lines in many of the forms discovered by Mr. Moore. ‘The knowledge of forms with three lateral lines, a condition previously unrepresented in the Cichlidw, adds force to the remarks I have made with regard to the morphological significance of the so-called ‘interrupted lateral line.” The Lemaire collection is further valuable for the care with which coloured sketches of most of the fishes have been taken on the spot by M. Dardenne, the excellent artist attached to the expedition. These sketches have enabled me to represent some of the more strikingly coloured forms in chromolithography. In addition to the definitions of the new species, I am able to draw up descriptions of the two species of Hctoedus of which very incomplete definitions were given in the previous contribution, owing to the bad condition of the type-specimens. I have also added notes on the known species of which specimens are contained in the Lemaire collection, or of which coloured drawings, accompanied by an indication of the native names, have been supplied by M. Dardenne. VOL. XV.—PArT Iv. No. 1.— December, 1899. ) 88 MR. G. A. BOULENGER ON A COLLECTION SERRANID &. 1. Lates microueris Blgr. Originally described from young specimens, 155 millim. long. The adult, measuring 800 millim. and weighing 14 Ibs., has, as could be expected, very different proportions and is of a uniform dark silvery colour. The eye is contained 9 times in the length of the head, which is 4 of the total; longest dorsal spine 4 length of head. Caudal peduncle and caudal fin as in the young. Native name: “ Sangala.” CICHLID4. 2. LAMPROLOGUS LEMAIRI, sp. n. (Plate XVIII. fig. 1.) A few moderately large curved canine teeth in front of each jaw, followed by a narrow band of minute teeth; lateral teeth very small. Depth of body 33 times in total length, length of head 22. Snout slightly longer than the diameter of the eye, which is 34 times in length of head and nearly double interorbital width; maxillary extending to slightly beyond vertical of anterior border of eye; cheeks and occiput naked; a few small deciduous scales on the opercles. Gill-rakers short, 9 on lower part of anterior arch. Dorsal XIX 7; spines equal from the fifth, which measures a little more than 4 length of head and ? longest soft rays. Pectoral $ length of head. Ventral reaching origin of anal. Anal VIII 5; spines increasing in length to the last, which slightly exceeds longest dorsals. Caudal rounded. Caudal peduncle a little “pt edged; a blackish oblique bar from below the anterior third of the eye to the maxillary ; a large blackish opercular spot ; three oblique blackish bands, descending forwards, on each side of the back, extending on the base of the dorsal; dorsal and anal tipped with blackish. Total length 107 millim. A single specimen. This species takes its place in the series between Z. moorti and L. congoensis. longer than deep. Scales 48 lat. 1. > Pale brown, most of the scales dark- 3. TELMATOCHROMIS TEMPORALIS Bler. Native name: ‘“ Muchétila.” 4, PARATILAPIA DEWINDTI, sp.n. (Plate XVIII. fig. 2.) 9 Teeth very small, in 5 or 4 series in both jaws, the outer scarcely larger and not tipped with brown. Depth of body equal to length of head, nearly 5 times in total OF FISHES FROM LAKE TANGANYIKA. 89 length. Snout with curved upper profile, shorter than the eye, the diameter of which is 23 to 22 in length of head and exceeds interorbital width; maxillary extending to below anterior fourth of eye ; 2 or 5 series of scales on the cheek; large scales on the opercle. Gill-rakers rather long, lanceolate, 17 or 18 on lower part of anterior arch. Dorsal XII-XIII 12-13; spines increasing in length to the eighth or ninth, which measures 2 or 4 length of head and ? longest soft rays. Pectoral a little shorter than head, extending as far as origin of anal. Ventral produced into a long filament, extending beyond origin of anal. Anal III 9; third spine longest, 7 length of head; middle soft rays produced, as long as or a little shorter than head. Caudal deeply emarginate, crescentic. Caudal peduncle 14 as long as deep. Scales 0 finely denticulate on the edge, 37-58 ng lat. 1. ae upper lateral line not reaching base of caudal. Grey above, white beneath; four yellowish stripes along each side ; pectorals yellowish ; other fins dark grey or blackish. Total length 100 millim. Three specimens.—Native name: “ Likuko.” This species is named in memory of the distinguished young geologist, Dr. De Windt, attached to Lieut. Lemaire’s expedition, who was accidentally drowned in Lake Tanganyika. It is very closely allied to P. ventralis Blgr., from which it differs in the dentition and in the shorter pectoral fin. 5. BAatHYBATES FEROX Bley. Native name: ‘‘ Musupa.” ‘TREMATOCARA, g, Nn. Closely allied to Paratilapia, but with the teeth very minute, in a narrow band in both jaws, the maxillary entirely concealed under the preorbital when the mouth is closed, the muciferous cavities of the top and sides of the skull and of the mandible extremely large, as in Acerina, and with a single, incomplete, lateral line. 6. TREMATOCARA MARGINATUM, sp. n. (Plate XIX. fig. 1.) 9 Depth of body 34 times in total length, length of head 23 to 8. Snout with curved upper profile, shorter than the eye, which is 24 in length of head and exceeds inter- ocular width ; mouth extending to below anterior border of eye; cheek naked ; a few deciduous scales on the opercle; nasal, frontal, pre- and suborbital, preopercular, and mandibular bones with very large and deep cavities separated by narrow septa and 02 90 MR. G. A. BOULENGER ON A COLLECTION covered with a thin skin. Gill-rakers short, 10 on lower part of anterior arch. Dorsal X 11; spines? length of head, a little shorter than longest soft rays. Pectoral acutely pointed, as long as head. Ventral reaching origin of anal. Anal III 10; third spine nearly as long as dorsals. Caudal with deep crescentic notch. Caudal peduncle 14 as long as deep. Scales cycloid, 302; lateral line reduced to a few (6 or 7) short tubes in the upper series. Pale brownish above, white beneath ; a bluish lateral stripe; fins yellowish, dorsal and anal edged with blackish. Total length 65 millim. Two specimens.—Native name: “ Lilowé.” GRAMMATOTRIA, g. 0. Body elongate; scales ctenoid; three incomplete lateral lines. Jaws narrow, with a band of minute conical teeth and an outer series of rather longer ones ; maxillary concealed under the preorbital when the mouth is closed. A large papillose pad on each side of the pharynx, between the gills. Dorsal with 15 spines, anal with 3. 7. GRAMMATOTRIA LEMAIRI, sp. n. (Plate XVIII. fig. 3. Depth of body 4 times in total length, length of head 3. Snout with slightly convex upper profile, 14 diameter of eye, which is 52 in length of head and equals interocular width ; maxillary extending to between nostril and eye; 3 series of scales on the cheek ; opercle covered with scales. Gill-rakers short, 12 on lower part of anterior arch. Dorsal XV 14; spines slender, equal from the fifth, which measures nearly 3 length of head; soft rays not longer than the spines. Pectoral acutely pointed, nearly as long as head, extending as far as origin of anal. Ventral reaching vent. Anal III 10; third spine a little stronger and shorter than longest dorsals. Jaudal with deep crescentic notch. Caudal peduncle 24 as long as deep. Scales 48-52 6-7 i ; 55 yg; lat. 1. 26. Pale brown above, yellowish beneath; a small dark brown 13-15 opercular spot ; a round brown spot on caudal peduncle at root of caudal fin; dorsal fin greyish, the soft portion with round white spots; other fins yellowish. Total length 175 millim. A single specimen.—Native name: ‘“ Murungi.” 8. Ecropus! pescampsi Bler. (Plate XIX. fig. 2.) Depth of body 54 times in total length, length of head 3, Snout short, with curved * To the generic characters as given in the original description must be added the presence of a large dermal pad in front of the upper part of the gill-arches, same as in Pelmatochromis and Ohromidotilapia. OF FISHES FROM LAKE TANGANYIKA. 91 upper profile, slightly shorter than the eye, the diameter of which is 24 in length of head and equals 14 interorbital width; maxillary extending to between nostril and eye; two series of scales on the cheek; opercle naked. Gill-rakers short, 11 on lower part of anterior arch, Dorsal XIII 13 ; spines slender, increasing in length to the last, which is about 3 length of head and little shorter than the soft rays. Anal III 8; third spine slightly shorter than longest dorsals. Pectoral obtusely pointed, 2 length of head. Ventral reaching vent. Caudal emarginate. Caudal peduncle 14 a slong as deep. Scales 35 fe ati I = Pale brown above, yellowish beneath ; fins yellow ; a round blackish spot on the hinder part of the spinous dorsal. Total length 60 millim. A single specimen. 9. Ecropus MELANoGENYS Bler. (Plate XIX. fig. 3.) Depth of body 5 to 54 times in total length, length of head 34. Snout long, with nearly straight upper profile, 13 diameter of eye, which is 4 times in length of head and equals or slightly exceeds interorbital width; maxillary extending to between nostril and eye; 3 series of scales on the cheek; deciduous scales on the opercle. Gill-rakers short, 12-13 on lower part of anterior arch. Dorsal XIII-XIV 17; spines slender, increasing in length to the last, which is about 2 length of head; last soft rays produced, at least half length of head. Anal III 13; third spine 3 length of head. Pectoral pointed, a little shorter than head. Ventral reaching origin of anal. 34, Caudal deeply emarginate. Caudal peduncle twice as long as deep. Scales 43-44 75 3 lat. 1. —o Grey above, white below; dorsal scales with a pale blue central spot; a blackish opercular spot ; chin and branchiostegal membrane blackish ; dorsal grey, with whitish streaks and spots, and a large oval blackish spot in the middle of the spinous portion; anal grey, streaked with whitish ; pectoral and caudal yellowish, the latter with crescentic dark bands; ventral blackish at the end. Total length 110 millim. Two specimens.—Native name; ‘ Losorella.” 10, Ecropus LONGIANALIS, sp. n. (Plate XIX. fig. 4.) Depth of body 5 times in total length, length of head 3. Snout long, with slightly convex upper profile, 14 diameter of eye, which is 3} times in length of head and nearly 14 interorbital width ; maxillary extending to between nostril and eye; 3 series of scales on the cheek; deciduous scales on the opercle. Gill-rakers short, 12 on lower part of anteriorarch. Dorsal XY 15; spines slender, increasing in length to the 92 MR. G. A. BOULENGER ON A COLLECTION last, which is 4 length of head; soft rays slightly longer, the last not produced. Anal III 17; third spine 4 length of head. Pectoral pointed, $ length of head. Ventral nearly reaching origin of anal. ae deeply emarginate. Caudal peduncle twice as long as deep. Scales 442; lat. 1. aaa Brownish above, whitish beneath; a blackish opercular spot; a lateral series of rather indistinct dark spots; dorsal greyish; other fins yellow. Total length 97 millim. A single specimen, with the mouth and pharynx full of advanced embryos. The characters of the 5 species of Hcetodus may be contrasted as follows :— EMdescampsi. WO, Xe anya 14.) VAC ANS eaSqegs4= 35h mle Depth of body 33-54 in total length. Snout slightly shorter than ae EL. melanogenys. D. XIII-X1IV 16-17. A.TII 13. Sq. 43- ME Nie, jl, eeno 7 3 14-16 Depth of body 5—54 in total length. Snout much longer than eye. Fi. longianalis, D.XV 15. A.TIL17. Sq. 443; lat. 1. roe Depth of body 6 in total length. Snout much longer than eye. XENOTILAPIA, g. 0 Body moderately elongate; scales ctenoid; three lateral lines. Mouth small, very protractile; upper jaw with series of minute conical teeth ; two mandibular teeth turned outwards, in a single series; maxillary concealed under the preorbital when the mouth is closed. A large papillose pad on each side of the PETS between the gills. Dorsal with 14 or 15 spines, anai with 3. 11. XENOTILAPIA SIMA, sp.n. (Plate XIX. fig. 5.) Depth of body 4 times in total length, length of head 33. Snout very short and deep, with very steep convex upper profile; eye very large, oval, its diameter ? length of head and much greater than interorbital width ; mouth nearly straight, Non soneall extending to below anterior border of eye; 3 or 4 series of scales on the cheek; deciduous scales on the opercle. Gill-rakers very short, broad, truncate, 9 on lower part of anterior arch. Dorsal XIV—XV 12; spines subequal from the fifth, } length of head; longest soft rays ? length of head. Pectoral acutely pointed, as long as head. Ventral with the inner ray produced and reaching a little beyond origin of anal. Anal III 11; third spine 4 length of head. Caudal with deep Creseeuve notch. 82-36 Caudal peduncle twice as longasdeep. Scales strongly ciliated, 40-41 55 oh 3; lat. 1. 21-22. OF FISHES FROM LAKE TANGANYIKA. 93 Pale brownish, with a few round darker spots; a blackish opercular spot; a shining golden spot on the subopercle ; dorsal greyish, other fins yellowish. Total length 105 millim. Two specimens.—Native name: * Lufuina.” 12. Trnapia LaBrata Bler. Native name: ‘ Kobo.” 13. TILAPIA DARDENNH, sp. n. (Plate XX. fig. 1. ] gs Teeth very small, in 4 or 5 series in both jaws, outer bicuspid, separated from the series of smaller tricuspid teeth by a rather wide interspace. Depth of body 5 to 53 times in total length, length of head 54 to 53. Snout with strongly curved upper profile, little longer than the diameter of the eye, which is 35 times in length of head and equal to or a little Jess than interorbital width ; mouth small, 2 width of head, extending to between nostril and eye; 5 or 6 series of scales on the cheek; large scales on the opercle. Gill-rakers short and thick, 13 on lower part of anterior arch. Dorsal XIX 10; sixth to ninth spines longest, not quite half length of head, a little shorter than longest soft rays. Pectoral acutely pointed, as long as or slightly shorter than the head, not extending to origin of anal. Ventral reaching vent or not quite so far. Anal III 8; third spine as long as and much stronger than longest dorsals. Caudal feebly emarginate. Caudal peduncle 1% or 1# as long as deep. Scales mostly 24-25 17-19" dorsal cross-bars, the first between the eyes; yellowish streaks along the series of ctenoid, 57 eg lattes Yellowish olive above, silvery beneath, with 10 or 11 dark scales ; fins yellowish, dorsal with some olive marblings, pectoral and anal red at the base. Total length 155 millim. Two specimens.—Native name: “ Sangani.” 14. TILAPIA RUBROPUNCTATA, sp. n. (Plate XX. fig. 2.) Teeth very small, in 4 or 5 series in both jaws, outer bicuspid, separated from the series of smaller tricuspid teeth by a rather wide interspace. Depth of body 3 times in total length, length of head 22. Snout with straight upper profile, twice as long as diameter of eye, which is 43 times in length of head and equals interorbital width ; mouth large, # width of head, extending to between nostril and eye; a few deciduous scales on the cheek ; large scales on the opercle. Gill-rakers short, rather slender, 12 or 13 on lower part of anterior arch. Dorsal XVI 9; spines equal in length from the 94 MR. G. A. BOULENGER ON A COLLECTION sixth or seventh, measuring + length of head and 2 longest soft rays. Pectoral obtusely pointed, % length of head, not extending to origin of anal. Ventral reaching vent. Anal III 7; third spine a little shorter than longest dorsals. Caudal truncate. Caudal peduncle as long as deep. Scales mostly cycloid, a few on the sides of the body below the lateral line finely denticulate, 32-33 a lat. 1. ae Olive-brown above, pale yellow beneath ; dark cross-bars on the back; each scale of the back and sides with a central vermilion spot; head spotted and marbled with dark purplish brown; lower jaw and lower part of opercular region bright yellow ; dorsal and caudal fins bright yellow, spotted with dark brown; pectoral yellow; branchiostegal membrane, pectoral region, outer edge of ventrals, and anal vermilion-red. Total length 120 millim. Two specimens of this handsomely-coloured fish, which bears the native name “Kasanga Malengi” on M. Dardenne’s coloured sketch. 15. TILAPIA MICROLEPIS, sp. n. (Plate XX. fig. 3.) Teeth very small, in 4 series close together in both jaws, outer larger, bicuspid, with a principal and a small lateral cusp. Depth of body 33 to 4 times in total length, length of head 3. Snout with straight or slightly convex upper profile, once and a half the diameter of the eye, which is nearly 4 times in length of head and equals inter- orbital width ; mouth moderate, its width 2 that of the head, extending to between nostril and eye; 7 or 8 rows of scales on the cheek; larger scales on the opercle. Gill-rakers short, rather thick, 13 or 14 on lower part of anterior arch. Dorsal XVI- XVII 14-15; spines subequal in length from the fifth or sixth, measuring ¢ length of head and alittle shorter than longest soft rays. Pectoral acutely pointed, ? length of head, not extending to origin of anal. Ventral widely separated from vent. Anal Ill 9; third spine as long as and a little stronger than longest dorsals. Caudal with deep crescentic notch. Caudal peduncle twice as long as deep. Scales cycloid, 80-90 om: lat. 1. —- Pale olive-brown above, white below; faint dark bars across the back and four round dark spots on each side, the last at the root of the caudal; fins yellowish. Total length 1165 millim. Two specimens.—Native name: ‘‘ Mocupi.” As in Tilapia desfontainesi, the dentition of this species may be regarded as connecting Tilapia with Paretilapia. 16. TILAPIA GRANDOCULIS, sp. n. (Plate XIX. fig. 6.) Teeth very small, in 4 or 5 series in both jaws, of outer series larger, bi- or tricuspid, OF FISHES FROM LAKE TANGANYIKA. 95 and very obtuse. Depth of body 32 times in total length, length of head 3. Snout short, with rounded upper profile; eye very large, a little longer than the snout, its diameter 24 in length of head, and slightly greater than the interorbital width ; mouth small, $ width of head, extending to between nostril and eye; a few deciduous scales on the cheek; larger scales on the opercle. Gill-rakers very short, rather thick, 17 on lower part of anterior arch. Dorsal XIII 14; spines slender, equal in length from the seventh, measuring ¢ length of head, and a little shorter than longest soft rays. Pectoral falciform, slightly longer than the head, extending as far as origin of anal. Ventral prolonged in a long filament, extending beyond origin of anal. Anal III 10; third spine shorter but stronger than longest dorsals. Caudal with deep crescentic notch. Caudal peduncle a little longer than deep. Scales mostly ctenoid, 63 =>; 25 lat. 1. — Brown above, with ill-defined darker spots, whitish beneath ; pectorals yellowish ; other fins blackish towards the edge. Total length 115 millim. A single specimen. SILURIDA. 17. AUCHENASPIS BISCUTATA Geoffr. Native name: “ Porocco.” 18. CurysicHTHYS CRANCHII Leach. Two young specimens are in the collection. A larger specimen is represented in the series of drawings made by Mr. Moore. 19. SynopontTIs MuLTIPUNCTATUS Bler. The young (150 millim.) have the spots on the head and body larger and fewer. Native name: “ Katétia.” CHARACINIDSA. 20. ALESTES MACROLEPIDOTUS C, & V. Native name: “ TIlala.” 21. ALESTES MACROPHTHALMUS Gthr. Native name: “ Laala.” VOL. XV.—PART Iv. No. 2.—December, 1899. P 96 ON A COLLECTION OF FISHES FROM LAKE TANGANYIKA. CYPRINODONTIDA. 22. HAPLOCHILUS TANGANICANUS Bley. Native name: “ Mohanga.” > ore © bo EXPLANATION OF THE PLATES. PLATE XVIII. . Lamprologus lemairii, p. 88. . Paratilapia dewindti, p. 88. . Grammatotria lemairii, p. 90. PLATE XIX. . Trematocara marginatum, p. 89. 1a. Side view of head; 1 6. Lower view of head, enlarged. . Ectodus descampsi, p. 90. » meélanogenys, p. 91. oe longianalis, p. 91. . Xenotilapia sima, p. 92. . Tilapia grandoculis, p. 94. PLATE XX. . Tilapia dardennit, p. 93. os rubropunctata, p. 93. » mécrolepis, p. 94. Trans.Loct, Soc Wt AV. F¢, XVM. De OOOO ree J.Green del.et kth. Mintern Bros.imp. I1LAMPROLOGUS LEMAIRII. 2.PARATILAPIA DEWINDTI. 3.GRAMMATOTRIA LEMAIRITI. ej he Suse eh i.) Troms. Loot. Soc. WA AV. G¢_ATL. oe a, ir) J Green del.et lth. Mintern Bros.imp. LTREMATOCARA MARGINATUM. 2ECTODUS DESCAMPSI. 3.ECTODUS MELANOGENYS. 4.ECTODUS LONGIANALIS. 5.XENOTILAPIA SIMA. 6.TILAPIA GRANDOCULIS. Trand. Look Soo Vt AVG, XX. dsGreen del.et lth. : Maintern Bros.Chromoe. 1.TILAPIA DARDENNII. 2.T. RUBROPUNCTATA. 3.T.MICROLEPIS. 9 THE FOREST-BED OF HAST ANGLIA. 103 III. Arias. (Nearly complete, but wanting the end of the left transverse process. The second series gives the measurements of the articulated specimen of O. giganteus (Irish) in the Woodwardian Museum at Cambridge.) C. be'grandi C.giganteus( Wood- (Pakefield). wardian Museum), inches, millim, | inches. millim. ContrumMvereatestalenethvor meee cie ieee ere 2°5 64 BRT 68 Neural arch, length in median dorsal line..............-...-- 25 64 3-0) 77 Greatest wih of vertebra (estimated at twice the width of el 7.3 185 8-7 99] PAM). cococcooegorosegoeHUnaNnDODE bUnDOODUON SE Width between outer edges of facets for condyles of skull ...... 4-0 102 4-5 115 Width between outer edges of facets for odontoid process ...... 4-0 102 45 115 Front edge of facet for condyle to middle of lateral part of facet for 40 102 45 115 odontordiprocessineraie mirc cir rier en iran iar: Tip of neural spine to a point vertically below it on the ot 3-0 76 3. Bl 93 HUI) Ot WING CMI Uhagcqscgageccg0ndc nade goavabodes Tip of BYyPaDOD ELS veRneehy to the level of the hinder edge of 1 30 76 29 74 themeuralearchynethesmiddlcilinenassscwrt-tsiteriacieacnormetene tata f Deepest point in the depression behind the ventral nerve-foramen \ 6 ral 18 4G to the dorsal edge of the facet for the odontoid process........ Width of centrum between the two depressions .............. 19-21 47-53 2:8 72 Wadthkotmeuralycanalsnmeironbariiersclesicterlcicicnciercneeiotorene tts ie7/ 44 as 3 ame OCHUNGL Matera cnc yatgersrnsie varie arses ors 17 44 Vertical height of neural canal, in front .................... 12 3 9 5 behind weerrrcriiccrra tenner Weg 44 IV. Axis. (Practically complete.) C. belgrandi C. giganteus ( Wood- (Pakefield). wardian Museum). inches, millim, | inches. millim. Centrum, greatest length (without odontoid process) .......... 4:4. 112 o1 129 ie 53 io (with a pea) men ei wore Byrd 137 Neural arch, length at the top of the neural canal.............. 4-0 102 4:7 120 «9 SONA, LEMS OF ooogoobvovcancso0g00000 cad0DDga0N 5:0 127 Bess) 140 Hacotmoratlasswid thiol ieieitincthrsi icin nrenoiaeiclon: 3:9 100 44 112 Odontordyprocess;widthvat) tips errseeraciycileerienierioiie crn 16 40 Top of neural canal to bottom of facet for atlas .............. 2-9 74 3:2 82 Length of facet for anterior zygupophysis of 8rd vertebra ...-.. 16 40 1:9 45 Greatest vertical height (at hind end) ................-+.00- 54 137 6:2 157 a swadthy (ean) aiid tend) Mareintrrsrrcrsstrtrieicraiekceercre si 4:0 102 6:8 173 Posterior zygapophyses, width across............+0seeeeeeees 3:0 76 4-0 102 Posterior end of centrum, vertical height of articulating surface... 1°9 48 Top of neural spine to top of neural arch ....-........+.0000- 2-4. 61 3:6 92 vou. XvV.—PART Iv. No. 3.—December, 1899. Q 104 DR. S. F. HARMER ON CERVUS BELGRANDI FROM Tue AFFINITIES OF CERVUS BELGRANDI. It is now generally accepted that the Irish Deer (C. giganteus)! and its allies are members of the Damine group. The particular position which C. belgrandi should occupy in this group has not hitherto received adequate consideration, in consequence of the imperfect condition of all the previously described examples. It appears to me that Cervus belgrandi combines in a most interesting manner the characters of C. dama aud its fossil allies (C. brown, &c.) and of the series of forms including the irish Deer. Many specimens of C. dama agree precisely with C. belgrandi in the number and position of the tines and in the general character of the palmation. The principal differences between the two forms are: (1) the more horizontal position of the beam in C. belgrandi; (2) the position and characteristic downward curve of the first tine in that species; and (3) the existence of a scalloped edge of the crown in QC. belgrandi, as distinguished from that of C. dama, which.is prolonged into more definite tines. Both species agree in the absence of points on the anterior edge of the crown. The palmated region includes the back-tine in C. dama, whereas in C. belgrandi there is a distinct constriction between the back-tine and the palmation. ‘The comparative suddenness with which the expansion begins in this species is probably the reason why its real character has not hitherto been suspected. Right antler of Cervus dama, believed to be of preglacial age, from Belzig, 10 miles 8.W. of Berlin ; seen from the inner side. (After Keilhack.) 1 «Manuel d’Hist. Nat., traduit de l’allemand de J. F. Blumenbach,’ ii. 1803, p. 407. THE FOREST-BED OF EAST ANGLIA. 105 A most interesting specimen, referred to Cervus dama, has been described by Keilhack } from Belzig, 10 miles south-west of Berlin. It would not be easy to find a specimen affording a more complete transition from CO. belgrandi to the modern C. dama than this one, and I have accordingly reproduced one of Keilhack’s figures of the right antler (woodcut, p. 104). The specimen far exceeds in size all the Fallow-deer measured by Keilhack, the base of the beam being no less than 180 mm. in circumference, as com- pared with 132 mm., the mean of the measurements of the recent individuals given by the same autbor. It further approaches C. belgrandi in the scalloped edge of the posterior part of the palmation and in the downward curvature of the large brow-tine, which, however, originates immediately above the burr. The brow-tine appears from the figures to spring from the beam somewhat on its posterior side, as in the Forest-Bed form. ‘The affinity of Keilhack’s specimen to C, dama is, however, shown by the angle made by the antlers with the skull, the direction being much less horizontal than in C. belgrandi, and by the inclusion of the back-tine in the palmation, from which it projects in much the same way as in the recent Fallow-deer. The beam terminates in a short but distinct snag at the distal and anterior edge of the crown, but the interval between this and the back-tine has a scalloped edge closely resembling that of the Pakefield specimen. Nehring’, in remarking on the Belzig specimen, ina paper which preceded Keilhack’s memoir, expresses the opinion that it must be regarded, if not as an actual C. dama, at least as a direct ancestor of that species. The Belzig specimen indicates in the clearest way the probable derivation of the modern C. dama from forms resembling C€. belgrandi, the principal changes which have taken place being apparently the diminution in size, the approximation of the crowns of the two antlers by the assumption of a less horizontal position, the inclusion of the back-tine in the palmation, and the replacement of the serrations of the crown by the snags which characterize the recent Fallow-deer. The discovery of the Belzig specimen and that of the form here described go far towards filling up the gap between C. giganteus and C. dama, to the existence of which attention is called by Riitimeyer *, in insisting on the affinity of these two forms. It may be noted that the immature antlers 4 of C. dama have the posterior edge of the crown denticulated in a way strikingly suggestive of C. belgrandi. The resemblance of C. belgrandi to C. giganteus is no less striking. Pohlig® has ' K. Keilhack, ‘‘ Ueber einen Damhirseh aus dem deutschen Diluvium,” Jahrb. k. Preuss. geolog. Landes- anstalt (1887), 1888, p. 283, pl. x1. 2 « Ueber das fossile Vorkommen von Cervus dama .... m Norddeutschland,” SB. Ges. naturf. Berlin, 1883, p. 69. . * “ Beitrige zu einer natiirlichen Geschichte der Hirsche,” Abhandl. Schweiz. pal. Gesellsch. x. 1883, p. 111. * Of. Cuvier, ‘‘ Recherches sur les Ossemens Fossiles,” 3rd ed. t. iv. 1826, pl. ui. figs. 28, 29. 5 « Tie Cerviden des thiiringischen Diluyial-Travertines, mit Beitrigen tiber andere diluviale und iiber recente Hirschformen,” Paleontographica, xxxix. 1892, p. 215. 106 DR. S. F. HARMER ON CERVUS BELGRANDI FROM recently defined several forms of the latter species, which are accepted with certain modifications by Lydekker!. ‘Taking the nomenclature of the latter, we have 1. Cervus giganteus carnutorum Laug., of which C. verticornis is regarded as a synonym. C. verticornis occurs typically in the Forest-Bed (early Pleistocene), and C. carnutorum in beds of equivalent age in France. Boyd Dawkins records the occurrence of one specimen of C. verticornis in the Pliocene (Red Crag). 2. C. giganteus belgrandi Lart., from the Mid-Pleistocene of France and Germany. 3. C. giganteus italie Pohl., from similar and later deposits of Italy, Hungary, &c. 4. C. giganteus ruffi Nehr. (=C. euryceros germanie Pohl.), the typical German race, found also in other parts of the Continent and in England, contemporaneously with the Cave-Bear, although first recorded from the Mid-Pleistocene. 5. C. giganteus typicus Lyd. (=C. euryceros hibernie Pohl.), the Irish Deer, the most recent form. The conclusion that C. verticornis is a synonym of C. carnutorum appears to me in the highest degree doubtful. The latter, found at Saint-Prest, near Chartres, was described by Laugel? from some upper molars, some portions of the antlers, a metatarsal, and a lumbar vertebra, which were all believed to belong to one species, comparable in point of size with the Irish Deer. Laugel does not figure the antlers, but he calls attention to the fact that their bases are remarkably approximated. The crown was probably palmated, and is believed to have borne about five curved snags, the longest of which measured 26 cm. ‘These characters are quite sufficient to show the difference between Laugel’s species and C. verticornis, even without the illustrated account of the Saint-Prest fossils which has been given by Gervais®. rom this memoir it appears that the vertex of the skull of C. carnutorum is much raised, so as to resemble that of the * Antilope Bubale” ; this being an important difference between it and the Forest-Bed Stag. Theantlers are closely approximated at their bases (cf. pl. xvi. fig. 4 of Gervais) and diverge from one another at an acute angle. An equally striking feature of this species is that, when looked at from in front, the brow-tine appears to originate from the outer outline of the beam, while that of C. belgrandi, viewed from the same aspect, appears to spring from the inner outline, Professor Boyd Dawkins and Mr. E. T. Newton* have, however, referred several fragments from the Norwich Crag, and possibly from the Forest-Bed, to C. carnutorwm. I am not prepared to express any opinion with regard to these specimens. Tam indebted to Professor Boyd Dawkins for a letter in which he informs me that he is inclined to regard his Cervus verticornis (1872) as identical with C. belgrandi (Lartet) Belgrand (1869), and for the further information that the type-specimens of the latter are no longer in existence. C. belgrandi was described in Belgrand’s work 1 T. cit. pp. 184-141. * Bull. Soc. Géol. France, (2) xix. 1862, p. 711. * “ Zool. et Paléont. Générales.—Nouy. Recherches sur les Vertébrés Vivants et Fossiles,” 1° sér, 1867-1869, p- 84, pl. xvi. figs. 3-8, * For references, see EH. T. Newton, “The Vertebrata of the Pliocene Deposits of Britain,” Mem. Geol. Survey, 1891, p. 26. THE FOREST-BED OF EAST ANGLIA. 107 ‘La Seine ’’!, parts of two frontals, with the bases of the antlers, being figured. ‘These fossils appear to me to agree in all essential respects with the Pakefield specimen. The inclination of the antlers, the distance from one another of the pedicles, and in particular the position of the brow-tine agree in the two forms; the measurements correspond with a sufficient degree of accuracy; while it further appears from Belgrand’s plate xix. that the second tine of Cervus belgrandi does not originate in the same line with the brow-tine, but at a point 90° further forward than the line containing the base of the brow-tine, in exactly the way that is characteristic of C. verticornis. Polilig? regards the rudimentary condition of the brow-tine as a leading characteristic of C. belgrandi. Had he, however, copied Belgrand’s representation of the right antler in pl. xviii., instead of that of the left antler, this condition would have been much less apparent; and in the specimen 3 from Yaubach which he himself figures (pl. xxiv. fig. 1) the brow-t’ae has essentially the characters of that of the Forest-Bed form. Cervus giganteus italie also shows some interesting resemblances to the Forest-Bed form, The specimens in the Museums of Milan and Arezzo figured by Pohlig + have a palmated crown which shows a distinct resemblance to the Pakefield specimen, from which they differ, however, in the larger development of the points of the crown and in having the brow-tine in the position characteristic of the typical C. giganteus. Part of the definition of this form given by Lydekker—the upward inclination of the palmation, the front border of which is curved inwardly so that much of the outer surface is seen in a front view, and the occurrence of the points of the crown near the summit of the palmation—will apply equally well to C. delgrandi. The general direction of the antlers, the beams of which spread out at first nearly horizontally, the palmated part then curving upwards, is again strikingly suggestive of that form. C. megaceros ruffii was named and figured by Nehring®, who calls attention to the absence of points on. the anterior edge of the palmated part as an indication of affinity to C. dama. Although agreeing in this respect with C. belgrandi, it has already acquired long tines springing from the distal border of the crown in a manner suggestive of the Irish Deer. Some of the specimens of this race, for instance that from the Worms Museum figured by Pohlig (p. 222, figs. 4¢ and 4d), show a much closer approach to the Forest-Bed species, in the way in which the crown is set on to the beam and in the slight development of the terminal snags. The general conclusion seems to be that C. belgrandi is the most central type of the Giant Deer, showing as it does a distinct affinity to the Fallow-deer as well as to the 1 « Hist. Gén. de Paris. La Seine.—I. Le Bassin Parisien,” par E. Belgrand. ‘ Planches de Paléontologie,’ 1869, p. 13, pls. xvili.—xx. 2 2. ct. p. 232: * I do not feel conyinced that this really belongs to O. belgrandi. * T. cit. p. 228. * “Uber eine besondere Riesenhirsch-Rasse aus der Gegend yon Kottbus,” Sitzungsber. Ges. naturf. Fr. Berlin, 1891, p. 151. See also papers by the same author, “ Neue Notizen iiber Cervus megaceros, var. ruffir, Nhrg.,” ibid. 1892, p. 3; and ‘ Deutsche Jiger-Zeitung,’ xxxii. 1899, p. 631; xxxiii. 1899, p. 413. VOL. XV.— PART Iv. No. 4.—December, 1899. R 108 ON CERVUS BELGRANDI FROM EAST ANGLIA. Cervus giganteus series. The large size of the Pakefield specimen and the condition of the sutures of the skull preclude the idea that it can be a young form, and there is every reason to suppose that the characters shown by it are those of the adult animal. The crown has become conspicuously palmated, but the long tines found in C. giganteus are at present indicated merely by a slight serration of its distal border. Another point which seems to me to deserve special notice is the direction of the vascular grooves on the beam of the antler. In the Forest-Bed specimens these run in an accurately longi- tudinal course up the posterior and ventral surface of the beam, while in the Irish Deer, as well as in the specimens of C. giganteus from the Barrington gravel, the grooves have an elongated spiral course. This may indicate a torsion of the whole antler. Itisa noteworthy fact that the palmated crown of the Irish Deer is much more horizontally placed than that of C. belgrandi, in which the direction is very oblique, the dorsal surface looking somewhat backwards. If the anterior edge of the crown in C. belgrandi were depressed by the torsion of the beam, so as to bring the palmation into a horizontal position, this new position would not only agree with that found in the Irish Deer, but it would result in a torsion of the vascular grooves, which would thereupon assume the spiral course characteristic of the latter. C. dama agrees with the Irish Deer in this respect. Although the brow-tine of the Irish Deer may have the same curvature at its base as that in C. be/grandi, it comes off immediately above the. burr and from the extreme front edge of the beam. ‘he second tine of the Irish Deer is more nearly opposite the back-tine than in C. belgrandi, and the palmation often begins soon enough to include it, instead of commencing above the back-tine, as in that species. The geological occurrence of C. belgrandi agrees well with the conclusion above indicated. Its occurrence in the Forest-Bed is itself an evidence of its relative antiquity, while the existence of the species in the Red Crag (Boyd Dawkins) carries this form further back than any of the other races of Giant Deer. EXPLANATION OF PLATE XXT. 1 Fig. 1. Cervus belgrandi.—Skull and antlers, from behind. X 7. ‘The specimen was obtained at Pakefield, near Lowesteft. a, burr; J, first tine; d,d’, back-tine. The tines c, which are not visible in this position, come off from the front edge of the beam, midway between 6 and d. . View of the same specimen from the right side. ‘The figure having been drawn from a photograph, the parts of the right antler appear much larger than they really are. 0, first tine ; ¢, second tine; d, d’, back-tine. . Atlas vertebra of the same specimen, from the ventral side. ‘The end of ‘the right transverse process is partially obscured by the matrix. Fig. 4, Axis vertebra of the same specimen, from the right side. ‘The end of the neural spine is partially obscured by the matrix. I ee Q b> = ge w9 eb piiguing Yl 72 [22 UOSfi 7 ‘TIQNVYOTHE SNAXHO EDOM ITIP B PL VO Ys V. On a Specimen of Cervus belgrandi Lart. (C. verticornis Dawk.) from the Forest- Bed of East Anglia. By Stoney F. Harwer, Sc.D., FR.S., Superintendent of the University Musewm of Zoology, Cambridge. Received April 24, read June 6, 1899. [Puate X XT. ] THE specimen which forms the subject of the present paper is from the Forest-Bed series of Pakefield, near Lowestoft, where it was found as a result of the great destruction of the cliff due to the high tides of November 1897. The pieces were obtained from the men who had dug it out by my assistant, Ernest Lane, to whom is due the credit of recognizing the importance of the specimen, which is now in the University Museum of Zoology at Cambridge. I have been unable to find a record of any specimen belonging to this species which will compare with the present one in completeness. The posterior part of the skull, including the whole of the brain-case, is in excellent condition. The anterior parts, from the orbital region, are wanting. The left antler is nearly complete, and the parts which are missing are to some extent represented in the right antler, a large part of which is also present. ‘The atlas and axis were found associated with the skull. The specimen throws considerable light on the real characters of the antlers of the form usually known as Cervus verticornis; and in particular it demonstrates the incorrectness of the diagnosis which is usually given of this species. C. verticornis was originally described by Professor W. Boyd Dawkins ?, and a fuller account of it was given by the same author in a later publication *®. The species is commonly represented by basal parts of the antlers only, in collections from the Forest-Bed, and is readily recognized by its massive cylindrical beam, with the brow-tine (usually situated about 2 inches above the burr) curving forwards and downwards. ‘The restorations given by Boyd Dawkins indicate an antler which slightly widens out immediately above the back-tine, and ends in a short bifurcated tip, which was not, however, really represented in any of the specimens. Professor Dawkins’s opinion, stated in his fuller account, that he cannot recognize any affinity between the antlers of C. verticornis and those of the 1 The abstract of this paper which appears on pp. 715, 716 of the ‘ Proceedings’ for 1899 was published without giving me an opportunity of correcting the proof. he locality of the fossil and my conclusion with regard to the nomenclature are incorrectly stated in it—S. F. H., Nov. 13, 1899. 2 « On the Cervidee of the Forest-Bed of Norfolk and Suffolk,” Quart. J. Geol. Soc. xxviii. (1872), p. 406. 8 «British Pleistocene Mammalia.—VI. British Pleistocene Cervidie,” Palazontographical Society, 1887. 1 98 DR. 8S. F. HARMER ON CERVUS BELGRANDI FROM Irish Deer (C. giganteus) was no doubt the result of this inadequate restoration, which is also responsible for such statements as those of Flower and Lydekker! that the antlers are very short and thick, or of Lydekker? that their palmation is generally slight. The Pakefield specimen shows, on the contrary, that C. verticornis belongs to a type comparable with the Irish Deer not only in its size but in the amount of the palmation of the antlers, though differing from it in the absence of prominent snags springing from the palmated extremity. It fully bears out the conclusion stated by Lydekker ® that this form indicates the ancestral stock of C. giganteus. The specimen measures 6 feet in a straight line between the extreme points of the antlers, but this measurement hardly gives a correct idea of its real size. ‘The crown of the right antler rises nearly vertically, while that of the left antler is much more horizontally expanded, besides being more perfect. The distance of the distal extremity of the left antler from the middle line of the skull is thus no less than 4 feet 5 inches. If both antlers had been equally well preserved and if the direction of the right antler had been as horizontal as that of the left antler, the total span, in a straight line, would have been 7 feet 3 inches. The more complete left antler is no less than 52°6 inches, measured along the curve ; its pedicle is 2:9 inches in diameter (dorso-ventral) and the burr is 3:6 inches. The brow-tine (b) is 2 inches above the burr and is perfectly typical, springing from the back of the beam in such a way that its posterior surface is level with the posterior surface of the beam. ‘The anterior surface of its base is hollowed out, and beneath it there is a slight tubercle corresponding with the accessory tines or with the “ offers” described by Boyd Dawkins‘ in the same position. This is hardly indicated in the right antler. The beam is nearly cylindrical and is strongly furrowed by grooves for blood-vessels, which run in a longitudinal direction without showing any tendency to become spiral. ‘Two inches above the brow-tine, the beam measures 2°6 inches in greatest diameter, its anterior part soon afterwards becoming flattened in a dorso-ventral direction to form the second tine (¢), which is only represented by its base, the proximal edge of which is about 7 inches from the distal edge of the base of the brow-tine. Beyond the second tine, the beam nearly regains its cylindrical shape, although having a low ridge on its anterior aspect. It has hitherto been nearly straight, though slightly concave below and pointing somewhat backwards; but beyond the second tine it curves rapidly upwards, expanding and becoming flattened. At the origin of the back-tine (d) it is convex on its under surface and slightly concave above. ‘The back-tine is fairly long (6 inches, though incomplete at its tip) and is directed backwards and inwards; the distance of the proximal part of its base from > “Yntroduction to the Study of Mammals,” 1891, p. 323. * «The Deer of all Lands,” 1898, p. 140. Se Lxcutspy Lak Pale SOG Compa onal envalea tig wleNO nr. THE FOREST-BED OF EAST ANGLIA. we) a point on the beam opposite the distal end of the base of the second tine is about 7 inches ; and from the posterior edge of the burr, in a straight line, about 19 inches. Opposite the back-tine, the beam is 2°1 inches in greatest dorso-ventral diameter; and it is still 1:9 inches thick, becoming flat above, but remaining convex below, at a point 9 inches above the middle of the base of the tine. The front edge of the antler is wanting from a point 6°5 inches above the tine ¢, and for 11 inches from its lower end the fractured edge varies from 1°4 to 1:7 inches in thickness. Posteriorly the edge has been getting thinner from the back-tine, about 10 inches above which it expands into a large palmated crown, the plane of which is different from that of the part which precedes it. ‘The beam, from the second tine to the base of the palmation, was extended upwards in a plane about 45° from the median vertical plane, its edge being seen in a view of the skull from the front. The posterior edge of the crown is, on the contrary, rotated outwards and downwards, so that much of its lower surface becomes visible from in front. The posterior edge of the base of the crown is very concave, and passes obliquely upwards and backwards, the palmation extending further backwards than the end of the back-tine. The crown is quite complete except in front, and is everywhere thin, scarcely exceeding an inch in thickness in any part; its distal border is gently scalloped, about 7 points or rounded lobes being distinguishable, none of them being more than mere denticulations. The thinness of the palmated crown at its antero- distal extremity indicates that no other tines were present in this region. The part between the main palmation and the back-tine is, however, thick (1:7 inches) on its fractured anterior border; and another anterior tine may have been present in this region, or the beam may have terminated in a distinct point, as in the Belzig specimen (see woodcut, p. 104), mentioned below, or as in C. dama. From a comparison with the latter, I do not, however, think it likely that an anterior tine was present. The right antler agrees closely with the left, except that the region above the secornt s tine turns dorsally much more sharply than that of the other side. It is indeed almost at right angles to the beam, whereas that of the left side forms a much more open curve with the beam. ‘The first and second tines agree almost exactly in position with those of the left side, but the second tine is much better represented, measuring 7:5 inches in length and being practically complete. At its base it is flattened and is a good deal broader than thick (breadth 2°8 inches, thickness 1‘Sinches). ‘The back-tine is lost. ‘The anterior edge of the antler is complete for about 13:5 inches (measured round the curve) above the base of the second tine, and the antler is prolonged for another 10 inches above the point where the anterior edge ceases to be complete. Its thinness makes it practically certain that no other tine could have existed. The part of the skull which is preserved measures 9 inches in greatest length, and extends as far forwards as the posterior portion of the orbit. The occipital region and the base of the skull are in good condition. ‘The interval between the 100 DR. 8S. F. HARMER ON CERVUS BELGRANDI FROM pedicles is wide (3 inches) and the pedicles are much more horizontally placed than in C. dama, to which the species appears to be related ; they pass with an even curve into the top of the frontal, whereas in another specimen of C. belgrandi in the Cam- bridge collection they make a distinct angle with the frontal and are somewhat more vertical. The pedicles are nearly cylindrical above, but have a sharp ridge descending from their outer borders to the posterior margin of the orbit; their bases are deeply hollowed out behind, as in other species of Cervus. Below the pedicles, the skull has a median ridge, with a marked depression on either side of it, thus differing from a part of a skull described by Boyd Dawkins’. The posterior part of the parietals (apparently just in front of the suture with the occipital) has a round foramen nearly in the middle line, and a smaller one slightly to the left side of it and rather further forwards (the other side being here injured). I have found a similar vacuity in the roof of the skull in another Forest-Bed specimen which I refer to C. savini Dawk. The transverse occipital crest is strongly marked, in correlation with the great size of the antlers, and the occipital surface is nearly vertical. Ventrally there is a very deep emargination between the condyles. There was apparently a large supra-orbital foramen, but the skull is deficient in this region. The specimens of the Irish Deer (C. giganteus) in the Cambridge collection differ from the above in having much shorter and less marked pedicles, in the comparatively slight median ridge on the frontals and their more depressed character in this region, and in the smaller size of the supra-orbital foramen. The vertex of the skull, between the two pedicles, is less raised in C. belgrandi, the roof of the skull of which is practi- cally flat from the front of the occipital crest to a point between the two pedicles. In the Irish Deer, on the contrary, the skull between the pedicles is much higher than the most prominent part of the occipital crest, from which it is separated by a distinct concavity. A striking difference, which appears to me of some importance, is that whereas in C. giganteus the vascular grooves on the beam of the antlers have a markedly spiral course, those of C. delgrandi run quite straight up the beams. The race of C. giganteus found in the Barrington River-gravel agrees with the Irish Deer in this respect. The occurrence of the atlas and axis vertebre associated with the skull is of special interest. Their size is not far short of that of the corresponding vertebre in an articulated specimen of the Irish Deer in the Woodwardian Museum at Cambridge, the spread of whose antlers is 8 feet 4 inches. This is brought out in the tables of measurements which follow. I have noted the following differences between the two forms, although some of them may be no more than individual peculiarities. In the Artas:—The neural arch is flatter above in C. giganteus, while it rises externally more abruptly on the dorsal surface in C. belgrandi. The tip of the 1 Pal. Soe., ¢. cit. p. 28. THE FOREST-BED OF EAST ANGLIA. 101 rudimentary neural spine is at the middle of the upper surface in the former, and in front of the middle in the latter. The dorsal nerve-foramina of C. giganteus are constricted into a figure-of-8 shape (and completely divided by a bridge in a specimen from the Barrington gravel in the Woodwardian collection), while they are long slits in C. belgrandi. ‘The neural arch of C. belgrandi has a wider sinus behind on its dorsal side ; its anterior dorsal emargination is more square-cut: the ventral depression behind the nerve-foramen is much deeper; the hypapophysis is most prominent behind, with a median ridge passing towards the anterior end of the vertebra (that of the Woodwardian specimen of C. giganteus having no ridge). In some of these latter points the Barrington specimen holds an intermediate position between the two others. In the Axis :—The dorsal edge of the neural spine is thinner in C. belgrandi; the transverse processes, which are incomplete, were probably less projecting ; the posterior zygapophyses are narrower (1:05 inches, as compared with 1:5-1:8 inches) and more raised; there is a distinct emargination between the pesterior zygapophysis and the neural spine; and the ventral median ridge is distinctly higher and bounded by deeper grooves at its sides than in C. giganteus. In the tables which follow, the measurements have been taken in such a way as to permit of easy comparison with those given by Boyd Dawkins’, Belgrand?, and others. I. SKULL. inches. | millim. Greatest length of the part of the skull preserved .................. > | BRO 228 Ghrsaicss Dmealiin OF ne wl, cose coocngecenc0p oes en oodneuocecaaon 81 206 Occipital crest to fronto-parietal suture ........-... see cece cere eee 4] 104 is 5 asl (Ghowasn Gis) Gocobbeaocaedonccodupos a 3°0 76 iGavel lxanyeem joeClales, cooogbhooudoacpeducunDasoc0K bp acho sco Rr 3:0 76 bp o supra-orpitaly foraminay) -e2 jee se eee eos 3 99 From burr to burr (shortest distance)...........--2- sees eee eee 54 137 i, 3 (distance between outer edges).......-.-.-----++- 10-4 265 Top of foramen magnum to top of occipital crest ..............---45. 3 | 76 Foramen magnum, height of .... 2.2... 2... eect e ee cet ee ilhealt 28 3 5 SU AKONHIaYtop2 | yin Blo, io orosBlatausiese core e Meceer eee tece ea necetuG ot canic 14 35 Frontal width, just below pedicles ....-.-..... 66.6, reels on baits 00d O 73 | 185 Deshdle, Mangia OPogosoodneodcnvccyegognobaoce dean eomuog c000 000% 18 46 re circumiferenceyOf sero i) SERN ESE TSUN arae runt apar Leientae te 8:5 | 215 ,. antero-posterior diameter of .... 2... 6s se cece eee eee eee ee 24 | 61 ,», dorso-ventral diameter of .........- 211s ee eect tee eee eine 2:9 74 Interval between bases of R. & L. brow-times. ....--.- ee eee eee eens 9:5 | 250 Brain-case, diameter (external) behind pedicles...........+..-+0e ses +6 > ality Distance between outer edges of condyles ........ ++ ese eee eee eres +8 i 1 Pal. Soc., t. cit. pp. 25-28. 2 « Ta Seine.—I. Le Bassin Parisien. Planches de Paléont.,” 1869, p. 13. 102 DR. S. F. HARMER ON CERVUS BELGRANDI FROM II. ANTLERS. Left. | Right. inches. millim. | inches. millim. Circumference of baselof beam! Banas: eeee caine 9:1. 231 8:9 226 Diameter ae RESIN Cha PRR MA Abe Cloths bra Boley 2°9 74. Diameter OF ULI. tis siaicvenese alee eheic tele eevee SCO ROR: 36 91 Circumferencesoigburrepmeee eee eee eee 11:0 280 Bure ito tame Wy (G)i issue swereneiereos iartolerstorn ceeraern rate rere 2:0 51 Burr to fork of tine 1 (measured to distal side of base) ........ 4:0 102 Circumference of beampaboyve tine sme cree arclereiiiccietoiae 8:2 208 Tine a ibasalicircumierencevo Lane ener 6:8 173 Tine 1 to tine 2 (c) (distal side of 1 to proximal side of 2)...... 7-0 178 Tine 2ucircumferencelorbaselolamee ene eee ae oe 7:0 178 Tine: 2) lon eth obs ce caraeicuesokce nae bsten evar womans Shoes rae Ie a Ss 75 190 Circumference of beam abovetine 20. .0: 0.15 j.es ese see nsec 9°5 241 10°5 261 iBurratoydistaltsidevotebaseiofatinesse meen eee ec eine 14-0 356 14:0 356 Tine 2 (¢) to tine 3 (d) (point opposite distal side of base of 2 to IDaSO/O8 13) tke a iietee ciStande erausuieropeu tava eden reer ese eae 7:0 178 Posterior edge of burr to tine 3 (in a straight line) ............ 19-0 482 Burritoydistalisidevorpasclontinelol een eioeiee 22:0 560 ines Saccircumberencerotmbaselo Leaner eee 6:5 165 Miners Men othioie weymeree yee Coe ee ea Tae. 6-0 152 Circumference of beam above tine 8 ..............20cceeeees 12:2 310 Motalilenstha(Gntagstraightiline) ence een cr 50°5 1283 30:0 762 Hotalvlensthy(a@loneicunye) emer 52:6 1335 38:0 965 : we ————.\-—__—-—_—~Y Total span (straight line) ............ VAG ah ace omaoad 6 feet. 1830 millim. The variability in the position of the tines in this species is illustrated by the following measurements of the interval between tine 1 and tine 2, taken from four other specimens in the Cambridge collection :— Catalogue No. inches. millim. WIGAN GN) aerate peo etctapeuerenepatercvencees tae 3:5 89 MSA. ecient taa teleneponertencteeeiee arrester 4:0 102 GANG crevite ie teeatareterertetelarerawiousts tere i eet 4:5 114 ISO (ob BaddovodCboe US 6G00000040 5:0 127 TRANSACTIONS OF THE ZOOLOGICAL SOCIETY OF LONDON (continued). To Fellows. To the Public. ET Gil Be EBL VOLUME XIII. (1891-1895, containing 62 Plates). . Price6 8 38 Silla 0 Parr. (1891, contaming 6 Plates) =... . . 5 ON © I a ae » 2. (1891, contaming 6 Plates) . ... =. i Oe) Is has » 0. (1891, containing 6 Plates) . . . . . oOo sO 1 4 0 » 4 (1892, contaming 1 Plate). . . .. . gy oO tle @ 0 6 0 » 9. (1898, containmg 9 Plates) . . ... eee Omllinere Onto 0 » 6. (1898, containing 4 Plates) . . . . . LUE OO 0 12,20 » @. (1898, containing 6 Plates) . ... . Bee) 0 012 0 » 8. (1894, contaiming 6 Plates) . . . . . pk) 0 012 0 , 9. (1894, contaming 6 Plates) . . . . . pp Oy 8 0.15, 0 » 10. (1895, containing 5 Plates) . . . . . at Ose OO) 012 0 » ll. (1895, containing 7 Plates and Title and Index),, 015 9 Aa ® VOLUME XIV. (1896-1898, containing 47 Plates) . Price5 5 O fe -ORAO Parr l. (1896, contaming 2 Plates) . ... . Oa ee (0) 0 8 0 » 2 (1896, containing 6 Plates) . . . . . see Ono. O 012 0 5 9. (1897, containing 9 Plates) Op 9 Seley. O 5, 4. (1897, containing 10 Plates) . . . . . Sed Oss-O 22026 3} py (1898; contaming 2 Plates). 9. 2. ay Ww 0 0 8 0 2/6, (1898) contaming 7 Plates) se. pe Omelie () Le AO 3 dé (L898; contamme) 6 Plates) 2 2°. 5). BeOS 015 0 » 8. (1898, containing 5 Platesand Titleand Index),, 0 9 O 012 0 VOLUME XV. Part 1. (1898, containing 8 Plates) . . . . . PriceO 15 9 al sine 2 (L899; (contaimine5) Plates) 6 0) sue. © POs Oke O 012 0 » 9 (1899, containing 4 Plates) . . .. . Oise) a0) 012 6 » 4. (1899, containmg 4 Plates) . . shes ee Olean 012 0 CONTENTS. IV. Second Contribution to the Ichthyology of Lake Tanganyika —On the Fishes obtained by the Congo Free State Lupedition under Lieut. Lemaire in 1898. By G. A. Bovtunerr, F.R.S., 7.Z.8. (Plates XVIIL-XX.). . . page 87 V. On a Specimen of Cervus belgrandi Lart. (C. verticornis Dawk.) from the Forest- Bed of Hast Anglia. By Sipyuy F. Harmer, Sc.D., Y.R.S., Superintendent of the University Musewm of Zoology, Cambridge. (Plate XXI.). . . page 97 THE PUBLICATIONS OF THE ZOOLOGICAL SOCIETY OF LONDON. Tue scientific publications of the Zoological Society of London are of two kinds—“ Proceedings,” published in an octavo form, and “Transactions,” in quarto. According to the present arrangements, the “ Proceedings”’ contain not only notices of all busi- ness transacted at the scientific meetings, but also all the papers read at such meetings and recom- mended to be published im the “ Proceedings’ by the Committee of Publication. A large number of coloured plates and engravings are attached to each annual volume of the “ Proceedings,” to illustrate the new or otherwise remarkable species of animals described in them. Among such illustrations, figures of the new or rare species acquired in a living state for the Society’s Gardens are often given. The “ Proceedings ” for cach year are issued in four parts, on the first of the months of June, August, October, and April, the part published in April completing the volume for the precedmg year. The ‘‘ Transactions” contain such of the more important communications made to the scientific meetings of the Society as, on account of the nature of the plates required to illustrate them, are better adapted for publication in the quarto form. They are issued at irregular intervals. Fellows and Corresponding Members, upon payment of a Subscription of £1 1s. before the day of the Anniversary Meeting in each year, are entitled to receive all the Society’s Publications for the year. They are likewise entitled to purchase the Publications of the Society at 25 per cent. less than the price charged for them to the Public. A further reduction of 25 per cent. is made upon purchases of Publications issued prior to 1871, if they exceed the value of five pounds. Fellows also have the privilege of subscribing to the Annual Volume of the ‘ Zoological Record ” for a sum of £1 (which includes delivery in the United Kingdom only), but this privilege only holds good if the subscription is paid hefore the First of December in each year. Such of these publications as are in stock may be obtained at the Society’s Office (3 Hanover Square, W.), at Messrs. Longmans’, the Society’s publishers (Paternoster Row, H.C.), or through any bookseller. December 1899, P. L. SCLATER, Secretary. TRANSACTIONS THE ZOOLOGICAL SOCIETY OF LONDON. VoL. XV.—Parr 5. LONDON: PRINTED FOR THE SOCIETY, SOLD AT THEIR HOUSE IN HANOVER-SQUARE ; AND BY MESSRS. LONGMANS, GREEN, AND CO., PATERNOSTER-ROW. December 1900. Price £3 10s. Od. Payl or and Francis, Printers, ] {Red Lion Court, Fleet Street. TRANSACTIONS OF THE ZOOLOGICAL SOCIETY OF LONDON. To Fellows. To the Public. £s. d. Bs aa VOLUME I. (1833-1835, contamimg 59 Plates). . Price 313 6 . . . 418 OF VOLUME | II. (1835-1841, containing 71 Plates). . . 400... 5 6 6* VOLUME III. (1842-1849, contaiming 63 Plates). . , 8 8 8 . ..4 211 O* VOLUME IV. (1851-1862, containing 77 Plates). . ,, 6 2 Operas sui Ouee a Os: VOLUME _ V. (1862-1866, containing 67 Plates). . ,, Die AS ete Oe Os VOLUME VI. (1866-1869, containing 92 Plates). . ,, ll 5 Oe ee 1b 0 OF VOLUME VII. (1869-1872, containing 73 Plates). . ,, 10 4 0 . . .1812 OF VOLUME VIII. (1872-1874, containing 82 Plates). . ,, Oe Begs 3 We Os VOLUME IX. (1875-1877, containing 99 Plates). . , 12 16... .16 2 OF VOLUME X. (1877-1879, containing 95 Plates). . ,, 10 0 38 . . .18 7 OF GENERAL INDEX, Vols. I. to X. (1833-1879) . ,, OLN OF Oe On a0 VOLUME XI. (1880-1885, containing 97 Plates). ._,, PI Ors ot cl 1G Ow VOLUME XII. (1886-1890, containing 65 Plates). . ,, 5 SO a ie 40 * No cupies of these volumes remain in stock, r Only complete copies of these yolumes are left in stock. Continued on page 3 of Wrapper. fp 200 VI. A Monograph of the Genus Casuarius. By the Hon. Wauter Roruscuitp, Ph.D., F.ZS. Witha Dissertation on the Morphology and Phylogeny of the PALMOGNATHA (Ratira and Crypturi) and Neognatua (Carinata). By W. P. Pycrart. Received and read June 20, 1899. [Puates XXII.—XLV. | Part I.—A Mownocraru or THE Genus Casvarius. By the Hon. WatteR Roruscainp, Ph.D., F.Z.S. Genus CASUARIUS Briss. 1790. Casuarius Brisson, Orn. v. p. 10. 1801. Rhea (non Latham, 1790) Lacépéde, Mém., de l’Inst. i. p. 519. 1842. Hippalectryo Gloger, Hand- u. Hilfsbuch, pp. xxxxili, 452. The genus Casuarius is the typical genus of the group Casuarii of the Palwognathe (Ratite of former authors), and inhabits the Papuan subregion, 7. ¢., New Guinea with the islands in Geelvink Bay, Salwatty, New Britain, the islands of the Aru group, Northern Queensland, and the island of Ceram in the Moluccas. The Cassowaries are evidently separable into a great many local forms, apart from the few very distinct good species ; but owing to their large size, to the uncertainty about the localities of the living specimens that are brought to Europe, and to the disappearance after death of the most characteristic coloration and structure of the bare skin on the head and neck, our knowledge of the species is doubtless still limited. The material which I have been able to study is very large compared with that available in most Museums. For several years I have tried to procure a great number of living specimens and of skins as well, and I have been able to place numerous fresh bodies and skeletons before Mr. Pycraft, whose appended paper will, I trust, be of great value for the study of the anatomy of the Palwognathe. Outwardly the genus Caswarius may easily be recognized by the following peculiarities :— The bill is short, generally shorter than the head, laterally compressed, strong, and with the culmen curved downwards near the tip. ‘The nostrils are placed in the middle of the maxilla, in a wide membranous groove, and open anteriorly. Head and neck bare and mostly brightly coloured, the skin much carunculated and wattled in various places. ‘The forehead is ornamented with a more or less developed helmet of various shapes. Wings quite rudimentary, with about five round black shafts, without any indication of webs in the place of remiges; no rectrices apparent. The metatarsus VOL. XV.—ParRT v. No. 1.—December, 1900. s 110 HON. W. ROTHSCHILD ON THE GENUS CASUARIUS. _ is shorter than in most other Ratite, very robust, and covered in front with hexagonal scutes, and transverse ones near the toes. ‘Toes three, the two outer ones with obtuse, curved, and short claws; the inner toe with a long, straight, powerful, pointed claw, which is a dangerous weapon. The body is covered with stiff, hair-like feathers, in which the after-shaft is as long as the principal shaft. The old birds are black, the young ones brown, the nestling, when hatched, is striped longitudinally above. The eggs are about six to eight in number, with a strong and very coarsely granulated surface. When fresh they are evidently all of a light green colour, but when exposed to the light they become first more bluish, then greyish, and at last almost cream- coloured. The male alone incubates. Herr Schalow (Journ. f. Orn. 1894) has attempted to make a key to distinguish the eggs of the various species; but his key is a total failure and misleading, being principally based on the colour, which is the same in all, and on the transparency of the shell, which is not specifically different. All Cassowaries are inhabitants of forests, while the rest of the large living Palwo- gnathe (or Ratitw) ave denizens of steppes and deserts. Their food seems to consist of all sorts of vegetable matter and fruits ; but they also pick up insects and any creeping thing that comes in their way. In captivity, at least, they kill and devour chicks and small birds when they come across them. ‘They also, like Ostriches, Rheas, and others, swallow quantities of stones and gravel to assist digestion. ‘They are entirely diurnal, sleeping from sunset till morning. The voice of the Cassowaries is a curious sort of snorting, grunting, and bellowing, usually not very loud, and differing according to the species. Their temper is generally sullen and treacherous, and they are extremely pugnacious, even the different sexes often fighting at other seasons than the breeding-season. Exceptions are rather rare, but there is an adult Cassowary, now in the Society’s Gardens, which is quite tame, and was always so, since I had him; while another young bird used to follow the keeper who fed him like a dog when I had him in Tring, but has since become somewhat shy. It seems that the Dutch navigators in 1596 were the first Europeans who became acquainted with the Cassowary, at least there does not appear to be any indication that the Portuguese, who visited the East-Indian Archipelago long before, came across the bird. For more than two and a half centuries there was no certainty as to the existence of more than one species, but in 1854 Mr. Thomas Wall procured a specimen of Caswarius casuarius australis near Cape York, which was described by his brother, for the first time, in the ‘ Hlustrated Sydney Herald’; but the description (made from memory) is as bad asit can be. In 1857 Gould described the New Britain Cassowary, in 186) Blyth and Sclater described two new species, and since then the described forms have constantly been increasing, till they now number 20. The best general accounts are those by Salvadori in his monograph of the genus in the HON. W. ROTHSCHILD ON THE GENUS CASUARIUS. ea ‘Memorie della Accademia di Torino,’ and in his famous ‘ Ornitologia della Papuasia,’ iil. pp. 473-503. The Cassowaries are easily divided into three groups, as follows:—1. The typical group, containing two species, Caswarius bicarunculatus and C. casuarius with its seven subspecies ; 2. ‘The one-wattled Cassowaries, containing two species, C. philipi and C. unappendiculatus, with four subspecies; 3. The Mooruks, which number four species—C. papuanus with two subspecies, C. picticollis with two subspecies, C. bennetti with two subspecies, and C. /orie. These three groups are readily separable externally and anatomically. Mr. Pycraft’s memoir will detail the anatomical features. The external characters may be roughly designated thus :— Casque compressed laterally, two wattles on fore-neck . . . . Typicat Group. Casque depressed posteriorly, sometimes compressed laterally, one wattle on fore-neck. . . ..'...’. . .. .'. . » ONE-waTTLED CAssowaRizs. Casque depressed posteriorly, no wattle on fore-neck . . . . . Tuer Mooruxs. In the typical group the species Casuarius casuarius and its subspecies vary enormously in the height and slope of the casque; and although in the “ Key” to the species I have stated that certain forms are distinguished by a high or low casque, the fact remains that we often have greater differences among individuals of one subspecies than between two subspecies. Also we often find that males of one given subspecies show, as a rule, much higher and more erect casques, while the females have the easque lower and sloping back over the occiput. ‘There is no doubt that certain differences exist in these two characters among the subspecies and species of the Casuarius casuarius section; but too much weight cannot and must not be placed on these differences: first, because there is considerable individual variation ; secondly, because there is sexual variation; and, thirdly, because the period of growth seems variable (at least in captivity). I have seen young birds: still in brown plumage with enormous casques, while fully adult birds with black plumage showed hardly any casque at all. I have to thank Drs, Finsch, Gestro, Giglioli, and Heck, Count Salvadori, Dr. Sclater, Mr. Blaauw, and others, for their kind help in lending me specimens, drawings, &c. Key to the Species and Subspecies of the Genus Casuarius. ( Plumage striped. Chicks. 1 { Plumage brown. Immature. { Plumage black. Adult. Seen ae 2. / Two wattles on fore-neck ; casque not depressed posteriorly, | lone Conajnressecl leery 5 6 56 56 6 60 o 6 0 6 Sk 24 One wattle on fore-neck ; casque depressed posteriorly or | compressed laterally \ No wattle on fore-neck, casque depressed posteriorly . . 11. 112 HON. W. ROTHSCHILD ON THE GENUS CASUARIUS. | Wattles wide apart on sides of neck . Wattles close together on front ofneck. . . . . . ( Size large; wattles very large, above 5 inches long . Size large ; wattles medium, 3 inches or less, or small . 4.4 Size small; casque low, sloping, and less erect; sides of neck anteriorly bluish purple, posteriorly red ! ( Size small; casque very high . sui 6 Sides of neck anteriorly bright blue, posteriorly soanlet Sides of neck anteriorly bluish purple, posteriorly scarlet . ( Casque low but very large, wattles joined at base ; sides of | neck anteriorly blue, posteriorly scarlet . . 6 6 { : ' Casque low, not very large, wattles separate at base ; sides L of neck purplish mauve Sides of neck entirely blue . : : Sides of neck anteriorly blue, nostaniorle fcaler Fore-neck blue, wattle crimson and blue; casque very high, compressed laterally, depressed posteriorly . . 84 Fore-neck blue in front, orange-scarlet at the sides; wattle blue | Fore-neck orange or yellow, wattle yellow or iid / Casque depressed posteriorly, not compressed laterally . 9 { Casque compressed laterally, fore- and hind-neck and sides ( of the neck uniform orange . 5 Casque low, the blue not extending below ai, faved ile | of neck yellow veined with red . 5 | Casque very high, the blue extending on to the Hone nalred t sides of neck crimson . u { Fore-neck blue . Hh "| Fore-neck red, throat pink . 10 Hind-neck scarlet . “™ | Hind-neck blue : Baa Petavek dire eet | Large black patch on side of rent below ear surrounded 13 by abroad pink band. . .. . No black and pink patch on sides of throat ( Head and neck blue, sides of neck black with pink Between the;wriukles 20) — 2) : 5 0 ¢ 50 Head and neck blue, round seitth on fore! neck nl sides 14 of neck rosy purple. Pi ete ae Head whitish blue, fore- and hide neck Whe sides of nN neck purplish red Fore-neck entirely blue . Fore-neck with red patch in centre C. bicarunculatus. 4 5. 6 C. casuarius. We C. casuarius sclateri. C. casuarius australis. C. casuarius beccarii. C. casuarius violicollis. C. casuarius intensus. C. casuarius salvadorit. C. philipi. C. unappendiculatus rufotinctus. 9. 10. C. unappendiculatus aurantiacus. C. unappendiculatus. C. unappendiculatus occipitalis. 12: C. lorie. 13. 14. C. papuanus edwardsi. C. papuanus. C. bennetti. C. bennetti maculatus. 15. C, picticollis hecki. C. picticollis. Py Ie 1603. 1605. 1635. 1666. 1666- 1770. 1676. 1726. 1738. ' eo NNN WNW ov or oF =): x —) 4 ov Ol oo HON. W. ROTHSCHILD ON THE GENUS CASUARIUS. 115 CasvaRIvs casuaRtus (L.). Common or Ceram Cassowary. (Plate XXII.) Avis Eme Aldrovand., Orn. iii. p. 541 (Banda, brought to Holland by Schellinger, who obtained one in Java: very bad figure). Emeu Clusius, Exotic. lib. v. p. 98 (good figure, taken from same specimen). Emeu Nieremberg, Hist. Nat. p. 218 (fig.). Casuar Olearius, Gottorffische Kunstkammer, p. 23, pl. xiii. (alive in Gottorff; figure copied from Clusius), 1699. Casoar Perr. & Duvern., Mém. Ac. Se. depuis 1666-1699, iii. 2, p. 155, pls. 56, 57 (teste Salvadori) ; Buffon, Hist. Nat. Ois. i. p. 464. Emeu Ray & Willughby, Orn. p. 105, pl. xxi. Kazuwaris (Ceram), Valentyn, Oud en nieuw Oost-Indien, iii. (Amboina), p. 298. Cassowary or Emeu Albin & Derham, Nat. Hist. B. ii. p. 56, pl. Ix. (alive at George Tavern, Charing Cross). . Emeu Barrére, Orn. spec. nov. class. 3, gen. xxxviii. 1. . Casearius Klein, Hist. Av. Prodr. p. 17. . Cela Moehring, Avium Genera, no. 56. . Emeu Worm, Mus. Worm. p. 292. . Avis Emeu Johnston, Theatr. Univ. de Avibus, p. 180, pl. 56. . Struthio casuarius Linneus, Syst. Nat. ed. x. i. p. 155 ; Id., Syst. Nat. ed. xii. 1. p. 265 ; Gmelin, Syst. Nat. 1. p. 726, no. 2. Casoar Brisson, Orn. v. p. 10, pl. i. fig. 2. Casuarius Frisch, Vorstell. Vog. ix. Hauptart. 1, pl. 105, 105 @ (good fig. bird and egg). Casuarius emeu Latham, Ind. Orn. ii. p. 664 ; Dum., Dict. Sc. Nat. vii. p. 199; Lesson, Voy. Coq., Zool. i. pt. 2, p. 711 (partim !) ; Id., Traité d’Orn. p. 7, pl. 11. fig. 1 (partim !) ; Thienemann, Fortpfl. ges. Vogel, p. 2, pl. i. fig. 1 (egg) ; Schlegel, Mus. P.-B., Struthiones, p. 9. Casuarius galeatus Bonn. Enc. Méth. 1. p. 4, pl. iv. fig. 2; Vieillot, Nouv. Dict. Hist. Nat. v. p. 345, pl. B. 11. fig. 1; Merrem, Abhandl. Berl. Akad. p. 179; Stephens in Shaw’s Gen. Zool. xiv. 2, p. 432, pl. 29 ; Vieillot, Gal. Ois. 11. p. 77, pl. 225 ; Merrem in Ersch & Gruber’s Encycl. xv. p. 348; A. v. Homeyer, Journ. f. Orn. p. 365 ; Sclater, Proc. Zool. Soc. Lond. pp. 210, 250 ; Des Murs, Tr. d’Oologie, p. 362 ; Rosenberg, Journ. f. Orn. p. 45; Crisp, Proc. Zool. Soc. Lond. p. 137 ; Sclater, Trans. Zool. Soc. Lond. iv. p. 358, fig. a, p. 8360 (note), pl. 71; 114 HON. W. ROTHSCHILD ON THE GENUS CASUARIUS. 1862. Schlegel, Jaarb. zool. Genootsch. Nat. Art. Mag. p. 196, 1863. Sclater, Proc. Zool. Soc. Lond. p. 234 ; 1864. Schlegel, Dierent., Vogels, p. 238; 1866. Sclater, Proc. Zoo]. Soc. Lond. p. 168; 1867. Id., op. cit. p. 179 (hatched in Society’s Gardens) ; 1869. Wallace, Malay Archip. n. p. 156 ; 1871. Flower, Proc. Zool. Soc. Lond. pp. 82-35 ; 1873. Garrod, Proc. Zool. Soc. Lond. p. 644; 1873. Rosenberg, Journ. f. Orn. p. 390 ; 1875. Sclater, Proc. Zool. Soe. Lond. pp. 86, 87 ; 1877. Harting, Ostriches & Ostr. Farm. p. 103, plate ; 1879. Rosenberg, Malay. Arch. p. 323 ; 1881. Sclater, Voy. ‘Challenger,’ Birds, p. 64; 1881. Salvadori, Mem. R. Ac. Se. Torino, (2) xxxiv. p. 188, pl. i. fig. 2; 1882. Id., Orn. Pap. e Moluce. iii. p. 479 ; 1891. Thébault, Bull. Soc. Phil. (8) iii. pp. 198-210; 1891. Salvadori, Mem. R. Ac. Sc. Torino, (2) xlii. p. 213 ; 1894. Schalow, Journ. f. Orn. p. 18 (egg) ; 1895. Salvadori, Cat. B. Brit. Mus. xxvii. p. 592 ; 1896. Oustalet, Nouv. Arch. Mus. Paris, (3) vill. p. 263 (one menagerie bird only in Museum). 1811. Casuarius casuarius Miger, Prodr. Mamm. et Av. p. 247. 1842. Hippalectryo indicus Gloger, Hand-u. Hilfsb. p. 452 ; 1844. Casuarius orientalis 8. MuJl., Land- en Volkenk. p. 109 (partim !). 1872. Hippalectryo casuarius Sundev., Tentamen, p. 152. 1875. Casuarius gavanensis (sic!) Gulliv., Proc. Zool. Soc. Lond. pp. 478, 488. 1879. Casuarius beccarit (non Sclater!) Pelzeln, Ibis, 1879, pp. 376, 377 ; 1881. Id., Ibis, p. 401 (egg). Native name: called Suwuri by the Malays; but it is not certain if this isa local name on Ceram, or, what is more likely, a word used for all Cassowaries in the trade and of Papuan origin. Adult. Bill long; casque large, not very high, and somewhat sloping backwards, compressed laterally, of a dark brownish horn-colour. Head and occiput Nile-blue, becoming darker on the upper part of the hind-neck. Lower two-thirds of hind-neck scarlet. Chin, throat, and fore-neck dark blue. On the lower fore-neck are two large lappet-shaped wattles, joined at their base, rugose, granulated, and of a deep pink colour. Naked lower sides of neck bluish purple anteriorly and bright scarlet posteriorly. Lores black ; iris dark brown ; legs brownish grey. Plumage black. Size medium, : Total length about 13 m., bill from gape 98 mm., tarsus 250 to 275 mm., claw of inner toe 70-78 mm. Young (two-fifths grown). Plumage of various shades of yellowish brown to rufous brown. Head and fore-neck dull leaden blue, hind-neck dull yellow, naked lower sides of neck a mixture of dull blue and livid purple. HON. W. ROTHSCHILD ON THE GENUS CASUARIUS. 115 Young (three-fifths grown). Plumage yellowish or rufous brown mixed with black. Fore-neck dull indigo-blue ; head and occiput pale dull blue; hind-neck dull orange- red, naked lower sides of neck blue, posteriorly mixed livid purple and dull red. Young (full-grown). Plumage black, mixed with a few rufous feathers. Fore-neck indigo-blue; head and occiput pale blue ; hind-neck scarlet, naked lower sides of neck blue anteriorly, passing through plum-purple to scarlet posteriorly. Chick. Head and neck rufous, paler below; three broad stripes on back, one irregular stripe on each side reaching from the anus to the wing, and two other lateral stripes extending to the thighs, the last two broken up into irregular blotches. Hab. Ceram. As shown in the synonymy, the Ceram Cassowary is the oldest known member of the genus. In the year 1595 a number of merchants in Amsterdam formed a company, and sent out four vessels to open communication with the Eastern Archipelago and to bring home spices and other valuable merchandise. In December 1596 the ships were anchored at Sydayo, in Java, and it was there that Jan Jacobsz Schellinger, the skipper of the ship ‘ Amsterdam,’ was presented with a Cassowary, which had been brought to Java from Banda Island. This was a day or two before Schellinger was murdered by the chief of Sydayo. Although the ship ‘Amsterdam’ was left and burnt, the wonderful “ large fowl” was brought on board of one of the other vessels and was landed alive at Amsterdam in 1597. It was at first exhibited to the public for some months, then came into the hands of Count George Everard Solms, of s’Gravenhage, who kept it for a long time at the Hague, and afterwards presented it to the Elector, Prince Ernestus of Cologne, who, again, subsequently gave it to the Emperor Rudolphus the Second. Count Solms, before parting with the bird, had an excellent coloured picture made of it, from which the very good figure in Clusius, representing the bird, one of its feathers, and an egg, was taken. ‘This figure has been copied into several other works. A wretched figure of the same bird was also published in the diary of the long and dangerous voyage during which it was first observed by Europeans, and this figure is reproduced by Aldrovandus. In 1666 Olearius, in his somewhat despicable work ‘ Die Gottorffische Kunstkammer, assures us that one had been kept alive at Gottorff in Schleswig; but he does not say how it was procured. Olearius’s description is very amusing, but short and not very accurate. Valentyn (1726) seems to have been the first to give the actual home of this bird, which is Ceram, He says that some Dutchmen found it sitting on three eggs on that island as long ago as 1660. Valentyn also says that the Cassowary occurring in Aru differs from that of Ceram. Other writers have stated Sumatra, Ceylon, and all sorts of places to be inhabited by Cassowaries, and even Linneus said: ‘‘ Habitat in Asia, Sumatra, Molucca, Banda.” Up to the present day erroneous statements respecting the home of Caswarius casuarius have often been made, but it is evident that of all the Moluccan islands only Ceram is inhabited by a Cassowary. 116 HON. W. ROTHSCHILD ON THE GENUS CASUARIUS. Linneus, in 1758, bestowed our present specific name on this bird, calling it Struthio casuarius, and diagnosing it as “ Struthio pedibus tridactylis, vertice paleari- busque nudis.’ He then quotes Olearius, Albin, Dodart, Ray, Clusius, Bontius, Willughby, Aldrovandus, and finally adds: “Callus elevatus quasi diadema verticis. Palearia duo sub collo; collum a tergo nudum. Ungue intermedia majore ferit ; ova punctis excavatis.” The species is admitted by Wallace and Rosenberg to be not at all rare in the interior of Ceram, but all travellers say that it 1s extremely shy and difficult to approach, and it seems that no European naturalist has ever seen it in its wild state. All we know about its life-history comes from native sources, and may be more or less trustworthy, but it seems certain that it feeds principally on fallen fruit. ‘The eggs are said to be three to five in number; but we have had, at Tring Park, six to eight laid by one female, so that we cannot believe that its clutch can regularly contain three eggs only. A young Casuarius casuarius was hatched in the Society's Gardens in 1867, but observations respecting the time of incubation and other interesting points were apparently not made. I have never succeeded in getting a male that would pair, although I have had a female which laid many eggs. The egg described by Schalow (Journ. f. Orn, 1894, p. 19) is faded. ‘There is, I am convinced, not a single form of Cassowary that lays ‘‘ pale greyish-green ” eggs with “ dark grey-green” glazed granulations. In any case Caswarius casuarius lays light apple-green eggs, with darker bright green corrugations. ‘They are elliptical, the thickest zone about the centre, though very often a thick and a thin end can be distinguished. The elevated glazed granulations are all more or less connected, forming a labyrinthic network. The measurements are:—150 : 94, 142: 93, 140: 90, 131: 90, 129: 87, 123: 84, 146: 85 mm. The eggs are excellent food. Owing to their size it is difficult to boil them well in the shell, and besides it requires a number of persons to finish one; but for scrambled eggs, omelettes, cakes, and other purposes they are, like those of the Rhea and Emu, most excellent. bo . CASUARIUS CASUARIUS BECCARII Scl. Beccari’s Cassowary. (Plate XXIII.) 1875. Casuarius beccarii Scl., Proc. Zool. Soc. Lond. p. 87, figs. 1, 2 on p. 86 (Vokan, Aru Islands !) ; 1877. Forbes, Proc. Zool. Soc. Lond. pp. 307, 316 (anatomy) ; 1878. A. B. Meyer, Journ. f. Orn. p. 300; Salvad. Ann. Mus. Civ. Genoa, xii. p. 421; 1881. Salvad., Mem. R. Ac. Se. Torino, (2) xxxiv. p. 197 (partim: Aru only) ; 1882. Id., Orn. Papuas. e Moluce. iii. p. 484 (partim: Aru) ; 1894. Schalow, Journ. f. Orn. p. 21 (egg, Aru) ; 1895. Salvad., Cat. B. Brit. Mus. xxvii. p. 596 (partim: Aru). 1875. (7) Casuarius bicarunculatus (non Scl.!) Beccari, Ann. Mus. Civ. Gen. vii. p. 717 (Aru). HON. W. ROTHSCHILD ON THE GENUS CASUARIUS. ALY 1881. (?) Caswarius galeatus (non Bonn.!) Scl., Zool. Voy. ‘Challenger,’ Birds, p. 64 (young, bought in Ternate—seem to be young of C. c. beccarii from the Aru Islands; ef. Cat. B. Brit. Mus. xxvii. p. 597). 1884. (?) Caswarius sp. A. B. Meyer, Zeitschr. ges. Orn. i. p. 296 (egg from Aru—as there are, however, several Cassowaries inhabiting the Aru Islands, there is no certainty whatever about an egg without history !). Adult. Plumage black; casque very high, broad and massive, dark brown; bill long and not pointed. Head and occiput pale greenish blue. Upper part of hind-neck deeper blue, followed by a space of mixed blue and red, producing a somewhat livid pale purple effect. Lower part of hind-neck scarlet. Chin, throat, and fore-neck dark blue. Wattles large, broad, and very conspicuously joined at base, of a deep fleshy-red colour, which colour runs up in a narrow line nearly to below the ear. Naked lower sides of neck mixed red and blue. A very large bird, but I have not been able to measure a fully adult dead example. Young (three-fourths grown). Fore-neck indigo-blue, head and occiput pale blue ; hind-neck orange-yellow, with a purple patch between blue and orange. Naked lower sides of neck, anterior half dull blue, posterior half orange-yellow. Plumage brown. Young (about full-grown). Same as adult, only colours less bright and plumage brown. Hab. Vokan, Aru Islands. Dr. Sclater (P. Z. 8. 1875, p. 87) first recognized the Aru Cassowary as a distinct species, and gave a woodcut, showing a side view of the head and the wattles trom in front, accompanied by the following diagnosis :— “ Casuarius similis C. australi et crista part modo elevata: sed caruncula cervicis una media, ad apicem divisa. “ Hab. Wokan, inss. Aroensium (Beccari). “ Mus. civit. Genoeusis.” In the same volume our author also mentions a Cassowary from the south coast of New Guinea, of which he believes that it ‘* probably is of the same species as that of the Aroo Islands,” namely C. beccarii. ‘This specimen was afterwards described as C. sclateri, but later on reunited with C. beccarii, and so it has remained also in the ‘Catalogue of Birds.’ However, I am fully convinced that the form from Southern New Guinea can be separated from that from the Aru Islands, where it inhabits, so far as we know, only the northern island of the Vokan group. I am not able to give descriptions of well authenticated eggs of Beccari’s Cassowary, but there is an enormous egg in the British Museum from the Aru Islands, which is named C. bicarunculatus, which, however, from its large size I am more inclined to believe is that of either C. ¢. beccarii or C.c. violicollis. It measures 140 : 98 mm. Its granulations are low. Schalow (/. c.) describes an egg brought from the Aru Islands by C. Ribbe, measuring VOL. XV.—PaRT Vv. No. 2.— December, 1900. T 118 HON. W. ROTHSCHILD ON THE GENUS CASUARIUS. 128: 89 mm. It is, however, not ascertained that it is an egg of C. ¢. beccarw, which, as we know, is not the only Cassowary living in the Aru group. That the eggs of any Cassowary are more transparent than those of any other species is evidently not the case. 3. CASUARIUS CASUARIUS SCLATERI Salvad. D?’Albertis’s Cassowary. 1844. Casuarius sp. S. Miill., Land- en Volkenkunde, p. 22 (Utanata, Princess Marianne Strait). 1875. Casuarius beccarii (non antea, p. 87, quod Aru!) Scl., Proc. Zool. Soc. Lond. p. 527, pl. lviii. (S. New Guinea) ; 1880. Casuarius beccarii (non Scl., P. Z. S. 1875, p. 87!) D’Albertis, Nuova Guinea, pp. 494, 588 ; 1881. Salvadori, Mem. R. Ac. Sc. Torino, (2) sxxiv. p. 197 (partim: New Guinea only), pl. i. fig. 5 ; 1882. Id., Orn. Papuas. e Moluce. iii. p. 484 (partim: New Guinea only !) ; 1895. Id., Cat. B. Brit. Mus. xxvii. p. 596 (partim: New Guinea only). 1877. Casuarius australis (non Wall!) D’Albertis, ‘Sydney Mail,’ p. 143; id., Ann. Mus. Civ. Genova, x. p. 19 (Fly River) ; id., Ibis, p. 372. 1878. Casuarius sclateri Salvad., Ann. Mus. Civ. Gen. xii. p. 422 (coast opposite to Touan or Cornwallis Island, S.E. New Guinea) ; A. B. Meyer, Journ. f. Orn. p. 300 ; 1879. Sharpe, Ibis, p. 116 (type now preserved in Brit. Mus. !). Adult. Casque dull brown, very high and much compressed laterally, very thin, and when fully adult curved over to one side. Bill long and stout, black. Wattles very large and long, sometimes over 7 inches, but in the old birds they get torn and damaged either in fighting or when creeping through the thick scrub, so that often they are very short or split into a number of ragged ribbon-like appendages. ‘This is well illustrated in the series of figures of D’Albertis’s fine specimens in Count Salva- dori’s monograph. This is the largest form of the C. caswarius group aud also the largest Cassowary, standing, when erect, as much as 6 feet high and the body being fully half as big again as that of an Emu (Dromq@us). Plumage black, bead and occiput pale whitish blue. Fore-neck and nape brilliant indigo-blue. Hind-neck of a most intense bright scarlet. Naked lower sides of the neck, posterior 3 brilliant scarlet, anterior 4 bright blue. Total length about 1600 mm., bill from gape 130 to 140 and even 145 mm., tarsus 280 to 300 mm., inner claw 80 to 95 mm. Juv. (two-thirds grown). Plumage dark brown, often nearly as black as adult, also occasionally pale yellowish brown. Fore-neck leaden blue; head and occiput pale blue. Hind-neck dull orange-yellow. Naked lower sides of neck, anterior 4 blue, posterior % yellow. Chick. Head and neck dark rufous, variegated with black bands and _ spots. Longitudinal dorsal bands very broad, about 13 inches wide, black, slightly variegated with rufous, and 5 in number—~z. é., a central one and two lateral pairs; one band on flank extending down the thigh on each side, and a second one indicated by a number of dark brown irregular spots. Hab, Southern New Guinea from Port Moresby to the Princess Marianne Strait, and perhaps even up to McCluer’s Inlet. I have had several examples of this bird alive, though none quite adult. They HON. W. ROTHSCHILD ON THE GENUS CASUARIUS. UY) appear to be delicate and never to live long, and none ever seem to reach the size in captivity which the species attains in a wild state. Adult wild specimens of this race are the largest Cassowaries known. The fine adult bird from which the description and accompanying photograph were taken was shot by Mr. Emil Weiske on the Brown River in 8.E. New Guinea. Fig. 1. Casuarius casuarius sclatert. Sclater’s Cassowary was at first thought to belong probably to Beccari’s Cassowary ; it was then separated, and afterwards united with the latter by Count Salvadori. ‘The original diagnosis of this form is as follows :— ‘“‘ Medius, niger, casside parum alta, crassa, superne rotundata, latere posteriore fere verticali, brevissimo, crasso; paleare unico, mediocri, ad apicem diviso ; ungue digiti interni breviuscula ; loris et mento violaceis ; capite et gula ceruleis, paleare rubro. “Long. tot. 1™ 500, rostri hiat. 147 mm., tarsi 300, ungu. dig. int. 80 mm.” TZ 120 HON. W. ROTHSCHILD ON THE GENUS CASUARIUS. A most magnificent series of this bird was procured on the Fly River by D’Albertis, and a number of woodcuts of the heads and necks of them are given in the Annals of the Genoa Museum and in Count Salvadori’s splendid monograph of the genus Casuarius, and a coloured figure, from a sketch of D’Albertis, on plate i. (fig. 4) of the same work. Some eggs from near Port Moresby, which I received together with a chick, are covered with glossy glazed granulations, which are not at all connected, but stand more or less separated from each other. They are bright green, and measure 150: 103, 146: 100, 146: 97, and 146: 95 mm. 4, CASUARIUS CASUARIUS SALVADORII Oust. Salvadori’s Cassowary. (Plate XXIV.) 1875 ?. Casuarius tricarunculatus Beccari, Ann. Mus. Civ. Genova, vii. p. 717 (Warbusi) ; 1878. Salvad., Ann. Mus. Civ. Genova, xii. pp. 419, 420 ; 1881. Salvad., Mem. R. Ac. Sc. Torino, (2) xxxiv. p. 184; 1882. Salvad., Ann. Mus. Civ. Genova, xviil. p. 413 ; 1882. Salvad., Orn. Pap. e Moluce. ii. p. 473 ; 1895. Salvad., Cat. B. Brit. Mus. xxvii. p. 591 ; 1896. Oustalet, Nouv. Arch. Mus. Paris, (3) vill. p. 263 (type in Mus.). 1878. (Feb. 28rd.) Casuarius salvadorii Oustalet, Bull. Assoc. Sc. de France, Bull. no. 539, p. 350 (“ Warbusi ”’”) ; 1878. Scl., Proc. Zool. Soe Lond. pp. 213, 214, fig. 213 (Wandammen; head and neck); A. B. Meyer, Journ. f. Orn. p. 202; Salvad. Ann. Mus. Civ. Genova, xii. p. 420; 1879. Salvad., Ibis, p. 105; 1881. Salvad., Mem. R. Ac. Se. Torino, (2) xxxiv. p. 204, pl. 1. fig. 5 ; 1882. Salvad., Orn. Papuas. e Moluce. ii. p. 488 (footnote) ; 1895. Salvad., Cat. B. Brit. Mus. xxvii. p. 595. 1878. Casuarius altijugus Scl., Nature, xvil. p. 875 (Wandammen) ; 1879. Salvad., Ibis, p. 105. Adult. Plumage black; casque high, brownish black; bill thick, long, and not pointed. Head and upper part of hind-neck pale greenish blue, rest of hind-neck scarlet. Naked lower sides of neck blue anteriorly, purplish crimson posteriorly. Chin, throat, and fore-neck deep blue. Wattles smal] and widely separated, base and inner half blue, rest dark fleshy pink. ‘Total length about 13 m., tarsus 300 to 316 mm., bill from gape 150 to 190 mm., inner claw about 80 mm. Chick. Head and neck rufous; rest of body creamy white or pale buff, with three longitudinal dorsal black bands variegated with rufous. The three lateral stripes along the flanks and thighs are broken up into a number of irregular brown spots and patches. Hab. North-west New Guinea. Casuarius casuarius salvadorii was originally described by Oustalet and Sclater as C, salvadorit and C. altijugus respectively, from skins obtained by Laglaize near Wandammen on the south-west coast of Geelvink Bay, in North-west New Guinea ; HON. W. ROTHSCHILD ON THE GENUS CASUARIUS. 121 and to this day we know nothing absolutely certain as to the colour of the naked parts of wild-shot Arfak birds; but there are a certain number of birds sent alive to Europe as coming from New Guinea, and which agree neither with Caswarius casuarius from Ceram, nor with C. ¢. beecarti from Aru, nor C. c. sclatert from Southern New Guinea in colour, and these birds I, for the present, take to be Casuarius casuarius salvadorii as they agree in other respects. The bird from which I made the above given description lived two years at Tring and one year in the Society’s Gardens. It is fully adult, and a male. I have had several younger ones, but none lived to maturity. The bird shortly mentioned by Beccari in his ornithological letter (Ann. Mus. Genova, vii.) as agreeing in the form of the helmet with C. caswarius, but having three wattles, namely, two lateral ones, and below these a smaller median third one, was a young bird bought at Warbusi, a little to the south of Dorei. It was left at Ternate to grow up, where Laglaize saw it still alive in July 1877, but it was eventually lost. It must, in my opinion, have been an aberration of C. ¢. salvadorii. Oustalet says that the type of C. salvadorii came also from Warbusi, at the bottom of Geelvink Bay. Warbusi, however, is near Dorei, while Wandammen is the place at the southern end of the Bay whence the other specimens from Laglaize were said to come. Therefore Salvadori (1881) thinks that Warbusi might be a mistake for Wandammen; but I see no reason why it should not be correct, since C. tricaruncu- latus 1s not a species according to my judgment, as I have a similar aberration of C.¢. sclateri. Laglaize probably bought his specimens from Bruijn or his hunters. When excited these birds extend the neck horizontally and emit five or six very loud grunting and grating sounds; frequently, when very angry, they spit several times, elevate their wing-rudiments and the feathers of the hind-quarters, and peck and kick most viciously. 5. CASUARIUS CASUARIUS INTENSUS Rothsch. Blue-necked Cassowary. (Plate XX VII.) 1898. Casuarius casuarius intensus Rothsch., Bull. B. O. Club, viii. (nos. lviii., Ixiv.), pp. xxi, lv. Adult. Casque very high and laterally compressed, dark brown; bill pointed, short, black. Plumage black. Size small. Occiput, nape, upper half of hind-neck, chin, throat, and fore-neck deep brilliant blue. Lower half of hind-neck brilliant scarlet ; between the blue and red of the hind-neck an irregular black patch. Naked lower sides of neck of an intense blue. Two wattles on fore-neck medium-sized, oval, close together, but separated their entire length, blue outer edge, and a few small spots pink. Bill from gape about 110 mm., tarsus 275 mm., greatest length of helmet 130 mm. Young (half-grown). Plumage brown; head and neck dull blue; lower hind-neck dull orange-yellow. Lower naked sides of neck blue. Young (three-fourths grown). Plumage brown, mixed with a few black feathers. Fore-neck dark blue; head and occiput indigo-blue, paler round ear. Lower hind-neck orange-red. Lower naked sides of neck bright blue. 122 HON. W. ROTHSCHILD ON THE GENUS CASUARIUS. Young (full-grown). Plumage black, mixed with brown feathers. Head and neck deep blue, paler around the ear ; lower hind-neck scarlet. Naked lower sides of neck very bright blue. Hab. unknown. In addition to the type I have had five others alive, and they all agreed in having the intense blue colour and the entirely blue naked lower sides of the neck; but this Cassowary remains still a very little-known form, and will do so till its habitat is discovered. The original description is as follows :— « ¢ ad. The casque differs from that of C. caswarius in being very high and much more erect ; the wattles almost entirely blue instead of dark red; the blue of the head and neck uniform and very dark; the orange of the hind-neck much restricted and separated at the upper end from the blue by a black crescent-shaped patch ; naked sides of lower neck entirely uniform blue instead of red, bordered anteriorly with blue.” 6. CASUARIUS CASUARIUS VIOLICOLLIS Rothsch. Violet-necked Cassowary. (Plate XX VI.) 1899. Casuarius casuarius violicollis Rothsch., Bull. B. O. C. viii. (mos. lix., Ixiv.) pp. xxvii, lv. Adult. Casque low, sloping back very much, laterally compressed, horny brown, green at base. Bill very long, straight and pointed. Plumage black. Face bluish green. Occiput and upper hind-neck pale greenish blue, throat and fore-neck brilliant deep blue. Hind-neck orange-scarlet. Naked lower sides of neck bright violaceous mauve, bordered anteriorly with a narrow line of blue, posteriorly with a narrow line of scarlet. Two wattles on fore-neck large, round, and close together in female; small, long, and wide apart in male, blue at base, rest fleshy pink. Juv. (half-grown). Plumage brown; neck and head dull leaden blue; hind-neck dull gamboge-yellow; lower naked sides of neck dirty flesh-pink. Juv. (three-fourths grown). Plumage brown, with a few black feathers; head and occiput pale dull blue; fore-neck dull indigo; hind-neck orange; lower naked sides of neck dark pink streaked with purplish red. Hab. ¢ Trangan Island, Aru Islands. The type of the species was imported along with two specimens of C. bicarunculatus from Aru, and is a female, for she has laid an egg; but I have since received seven others of this form, and they all agree in their characters. As the other islands of the eroup are inhabited by different forms of Cassowaries, I am inclined to think that Trangan Island must be the home of this interesting form. I described this form at the January meeting of the British Ornithologists’ Club in 1899 as follows :— “This apparently undescribed form of Cassowary is most nearly allied to Caswarvus casuartus salvadorii, but differs conspicuously in the colour of the naked parts and in the very large size, which fully equals that of C. c. australis. HON. W. ROTHSCHILD ON THE GENUS CASUARIUS. 123 ‘Bill much longer and straighter than in other Cassowaries. Casque horny brown, green at base. Face and a broad band running down the side of the bill bluish green. Base of lower mandible dark blue, with a yellow line running along one-third of the length of the mandible on each side. Wattles at base of fore-neck very large, round, and short, 3 x 2% inches, pale blue at base, otherwise pink all over, entirely separate for their whole length, but close together. Auricular orifice larger than in any other Cassowary. Throat and fore-neck bright ultramarine-blue. Occiput and upper hind- neck pale greenish or eau-de-Nil blue. Lower hind-neck brilliant orange-scarlet. Naked lower sides of neck magenta-purple, bordered anteriorly with ultramarine-blue, posteriorly with orange-scarlet; the magenta-purple space deeply carunculated and sharply cut off from the red and blue borders, which are plain and smooth.” The “ song” of this bird is an extraordinary performance, generally as follows: — It lowers its head and neck and remains in this position—with head and neck stretched out straight in front—for about fifteen seconds, with the bill open and gradually inflating its neck, without making a sound; then, bowing and jerking its head so that the bill and wattle clap together, it emits some barking grunts, apparently with great effort, the skin of the neck being then so much inflated that the wrinkles become quite obscure. The type specimen laid an egg last year, which is, needless to say, bright green; but its shell is not properly developed, and a description of the surface and size is therefore better omitted. 7. CASUARIUS CASUARIUS AUSTRALIS Wall. The Australian Cassowary. (Plate XXV.) 1854. Casuarius australis Wall, Illustr. Sydney Herald of June 3rd ; 1857. Gould, Proc. Zool. Soc. Lond. pp. 269, 270 ; 1858. G. R. Gray, Proc. Zool. Soc. Lond. p. 196; 1860. Scl., Proc. Zool. Soc. Lond. p. 210; Ibis, p. 310 ; 1862. Scl., Trans. Zool. Soc. Lond. iv. p. 360; Schleg. Jaarb. zool. Genootsch. Nat. Art. Mag. p. 200; 1865. Gould, Handb. B. Austr. ii. p. 206; 1866. Schleg., Zool. Gart. p. 180; Scl. Proc. Zool. Soc. Lond. p. 557 ; 1867. Scl., Proc. Zool. Soe. Lond. p. 242 (C. johnsoni=australis) ; Bennett, Proc. Zool. Soc. Lond. p. 473 ; 1868. Scl., Proc. Zool. Soc. Lond. p. 376; Ibis, p. 348 ; 1868. Ramsay, Proc. Zool. Soc. Lond. pp. 381, 388 ; 1869. Gould, B. Austr., Suppl. pls. 70, 71; Krefft, Ibis, p. 348 ; 1870. Editors, Ibis, pp. 119, 120 ; 1871. G. R. Gray, Hand-list B. iil. p. 2, no. 9851; Flower, Proc. Zool. Soc. Lond. p. 32 (skeleton) ; Scl., Proc. Zool. Soc. Lond. p. 547 (in confinement) ; 1873. Schleg., Mus. Pays-Bas, Struthiones, p. 9; 1874. Ramsay, Proc. Zool. Soc. Lond. p. 320 ; 1875. Scl., Proc. Zool. Soc. Lond. p. 85 ; 124 HON. W. ROTHSCHILD ON THE GENUS CASUARIUS. 1876. Ramsay, Proc. Zool. Soc. Lond. p. 119 (life-history) ; Rams., Proc, Linn. Soc. N. 8. Wales i. p. 186; 1877. Rams., Proc. Linn. Soc. N.S. Wales, ii. pp. 196 (distribution), 376 (woodcuts of head) ; Harting, Ostr. & Ostr. Farming, p. 95; Salv. & Sel., Ibis, p. 327 ; 1881. Salvad., Mem. R. Ac. Se. Torino, (2) xxxiv. p. 192, pl. 1. fig. 3 (head) ; 1882. Salvad., Orn. Pap. e Moluce. iii. p. 473 ; 1888. Rams., Tabul. List Austr. B. p. 19 (distribution) ; 1889. Lumholtz, Among Cannibals, pp. 98, 99; 1890. North, Nests & Eggs Austr. B. p. 294, pl. xv. (egg) ; 1891. Sharpe, Cat. Osteol. Spec. R. Coll. Surg. ii. p. 439; 1892. Schalow, Zeitschr. f. Ool. p. 11; 1894, Schalow, Journ. f. Orn. p. 24; 1895. Salvad., Cat. B. Brit. Mus. xxvii. p. 594; 1896. Meston, Proc. Roy. Soc. Queensland, x. pp. 59-64. 1866. Casuarius johnsonii F. Miller, Australasian of Dec. 15th (cf. Proc. Zool. Soc. Lond. 1867, pp. 241, 242) ; ; 1867. Carron & Bennett, Proc. Zool. Soc. Lond. pp. 473, 474; Krefft, Proc. Zool. Soc. Lond. p. 482; Diggles, Orn. Austr. pl. & text ; 1868. A. Newton, Ibis, p. 348 (critical note on Diggles) ; 1869. Krefft, Ibis, p. 348 (maintains that johnsoni is correct name) ; 1870. A. Newton, Ibis, p. 120 (contends with others that the name australis is to be used). 1873. ? Casuarius regalis (nomen nudum) Rosenberg, Journ. f. Orn. p. 390. Adult. Casque very high and massive, compressed laterally, deep brown; bill long and stout, black. Next to C. casuarius sclateri, the largest of the forms of Caswarius casuarius. Plumage black. Sides of head, occiput, and upper hind-neck pale greenish blue; lower hind-neck scarlet, running a little into the blue of upper portion. Chin, throat, and fore-neck deep indigo-blue. Lower naked sides of neck mixed red and blue, giving a purple hue. ‘Two wattles on fore-neck very large, deep pinkish red, mottled with paler pink, blue at base. Young (quarter-grown). Head and hind-neck chestnut rufous. Chin, throat, and fore-neck densely covered with short downy feathers of a brownish-buff colour. Rest of body clayey brownish yellow mixed with dark rufous. Young (two-thirds grown). Plumage brown. Head and occiput dull pale blue; lower hind-neck orange; fore-neck leaden blue; lower naked sides of neck blue, mixed with livid purple. Chick. Head and hind-neck pale rufous; fore-neck yellowish buff; rest of body yellowish buff. From base of neck to end of tail along the back run three broad longitudinal black bands, variegated with rufous, and each about seven-eighths of aninch wide. On the sides are three irregular wavy black bands extending from the shoulder-girdle down the sides to the beginning of the metatarsus. Hab. Northern Queensland. I have kept the adult female, described above, since 1890: she laid some 40 or 50 eggs at Tring, and has now lived nearly two years in the Society's Gardens. HON. W. ROTHSCHILD ON THE GENUS CASUARIUS. 125 The first discovery of the fact that a Cassowary existed in Australia is due to Mr. Thomas Wall, who, like so many other explorers, left his bones in the waterless deserts of the Australian continent. Wall procured a specimen, but this was lost, and only a poor description, evidently made from memory, was published in the ‘ Illustrated Sydney Herald’ of June 3rd, 1854, by Thomas Wall’s brother, William Sheridan Wall. The account ran as follows :— “The first specimen of this bird was procured by Mr. Thomas Wall, naturalist to the late expedition commanded by Mr. Kennedy. This was shot near Cape York, in one of those almost inaccessible gullies which abound in that part of the Australian continent. The Cassowary, when erect, stands about 5 feet high. The head is without feathers, but covered with a blue skin, and, like the Emu, is almost without wings, having mere rudiments. The body is thickly covered with dark brown wiry feathers. On the head is a large protuberance or helmet of a bright red colour, and to the neck are attached, like bells, six or eight round fleshy balls of bright blue and scarlet, which give the bird a very beautiful appearance. The first, and indeed the only, specimen of the Australian Cassowary was unfortunately left at Weymouth Bay, and has not been recovered. Mr. Wall being most anxious for its preservation had secured it in a canvas bag and carried it with him to the spot where, unfortunately for himself and for science, it was lost. In the ravine where the bird was killed, as well as in other deep valleys of that neighbourhood, they were seen running in companies of seven or eight. On that part of the North-eastern coast, therefore, they are probably plentiful, and will be met with in all the deep gullies at the base of the high hills. The flesh of this bird was eaten, and was found to be delicious; a single leg afforded more substantial food than ten or twelve hungry men could dispose of at one meal. The Cassowary possesses great strength in its legs, and makes use of this strength in the same manner as the Emu. ‘Their whole build is, however, more strong and heavy than that of the latter bird. They are very wary, but their presence may be easily detected by their utterance of a peculiarly loud note, which is taken up and echoed along the gullies; and it would be easy to kill them with a rifle.” Most incorrect as this description is, there can be no doubt that it actually refers to the Australian Cassowary, and we can therefore, in accordance with all other authorities, accept the name of australis for this bird, and need not substitute that of johnsonii for it. This form is most closely allied to the Ceram Cassowary and to Sclater’s Cassowary. The key, descriptions, and plates, however, will show the differences. This bird is known only from Northern Queensland, from Cape York south to Rockingham Bay, Of no other Cassowary are we so well acquainted with the life- history. The most complete description of the latter is that of Ramsay in the Proc. Zool. Soc. Lond. 1876, pp. 119-123. It runs as follows :— “ One of the chief objects of my visit to Rockingham Bay was to become acquainted VOL. XV.—PaRT y. No. 3.—December, 1900. U 126 HON. W. ROTHSCHILD ON THE GENUS CASUARIUS. a with the habits of this noble bird. In 1872 I had sent my collector, Edward Spalding, to this district for this purpose, but with very poor results. “While in Brisbane on my way up I purchased by telegram a fine young living specimen, the first that had been obtained and reared, and ultimately succeeded in bringing it alive to Sydney and shipping it to the Society, where I am glad to hear it arrived safely. I found also that several very young Cassowaries had been obtained, and, for the first time, a nest and eggs had been found. ‘This was great news; and I need not relate how I made all haste to the Herbert River Police Camp, where I was most hospitably entertained and welcomed by Inspector Johnstone, who was the first to rediscover and bring under the notice of others the existence of this remarkable species. I found Inspector Johnstone a true sportsman, as well as an ardent lover of nature, a zealous and energetic naturalist, and a careful observer. I am indebted to this gentleman for much valuable information respecting the manners and customs of the aborigines, and notes on the habits of many birds and animals new to me, and especially for information on the present species. “The Australian Cassowary is a denizen of the dense dark scrubs scattered over the district of Rockingham Bay, and extending as far north as the Endeavour River. It was tolerably plentiful only a few years ago even in the neighbourhood of Cardwell ; but since the advent of sugar-planters, etc. on the Herbert River and adjacent creeks, these fine birds have been most ruthlessly shot down and destroyed for the sake of their skins, several of which I saw used for hearth-rugs and door-mats. Formerly they were easily enough procured; but latterly so wary have they become, and their numbers so decreased, that it is only with the greatest amount of patience that even a stray shot can be obtained. I know of no bird so wary and timid; and although their fresh tracks may be plentiful enough, and easily found in the soft mud on the sides of the creeks or under their favourite feeding-trees, yet the birds themselves are seldom now seen. During the day they remain in the most dense parts of the scrubs, wandering about the sides of the watercourses and creeks, diving in through the bushes and vines at the slightest noise. Toward evening and early in the morning they usually visit their favourite feeding-trees, such as the native figs, Leichardt-trees, and various species of Acmena, Jambosa, Davidsonia, &c.; they appear to be particularly fond of the astringent fruit of the Leichardt-trees and of a species of Maranta, which produce bunches of large seed-pods filled with juicy pulp, resembling in appearance the inside of a ripe passion-fruit (Passiflora edulis). Fruits and berries of all kinds are eagerly sought after; the tame semiadult bird which I had the pleasure of forwarding to the Society (1875) became so fond of the fruit of the Cape-Mulberry that he would allow no one to come near the tree he had taken possession of. ‘This bird has frequently devoured at a time as much as three quarts of ‘loquats’ (fruit of Hriobotria japonica), and several fair-sized oranges whole, besides its usual amount of bread per diem (about 3 pounds). In nature, I found that in the afternoons they frequently came out and HON. W. ROTHSCHILD ON THE GENUS CASUARIUS. 12% walked along the scrubs, or along the sides of the river or creeks, and swallowed large quantities of pebbles and small rough-edged stones. In confinement, plantains and sweet potatoes (in large pieces, which they swallow whole) are a favourite food, while nothing seems to come amiss to them—grasshoppers, spiders, earthworms, cockroaches, caterpillars of all kinds, dough, and even raw meat. ‘They ascertain the flavour of their diet by first taking it up in the tip of their bill and giving it a slight pinch ; and if not suitable, they throw it away. “J found they invariably refused green loquats, but always picked them up in the bill first to try them. In confinement they become very tame, and may be allowed to walk about the place without restraint, coming when called, or more often running after and following any one accustomed to feed them. If disappointed or teased, they not unfrequently ‘show fight’ by bristling up their feathers, and kicking out sideways or in front with force sufficient to knock a strong man down—a feat I have witnessed on more than one occasion. ‘These birds are very powerful, and dangerous to approach when wounded. On more than one occasion a wounded bird has caused a naturalist to take toa tree; the sharp nail of the inner toe is a most dangerous weapon, quite equal to the claw of a large Kangaroo, and capable of doing quite as much execution. “T found the Cassowaries to be excellent swimmers, and frequently tracked them across a good-sized creek or river. On Hinchenbrook Island, situated about 14 mile from the mainland, they have been frequently met with; and I have myself heard them calling at night and early in the morning as I passed up the channel, at a distance of a least two miles from them. Mr. Johnstone informs me he met with one swimming across a river of considerable width during his explorations while on the ‘ North-east Coast Exploring Expedition.’ Their note, most usually emitted by the male, is a series of harsh guttural prolonged croakings quickly repeated, and continued for about three minutes; it is very loud, and may be detected across the water at a distance of at least three miles on a still night. I have listened to it resounding through the scrubs at a distance of a mile anda half on land, and then thought it close and one of the most unearthly noises I ever heard. They breed during the months of August and September. ‘The first nest procured was found by some of Inspector Johnstone’s black troopers, from whom Mr. Miller, a settler on the Herbert River, purchased some of the eggs. One which he kindly presented to me is of the light green variety mentioned hereafter. The nest consists of a depression among the fallen leaves and débris with which the ground in the scrubs is covered, with the addition of a few more dry leaves. The place selected is always in the most dense part, and well concealed by entangled masses of vegetation. The eggs were five in number in the only two instances recorded; and in both cases one of the eggs in each set differed from the others, being of a light green colour, and having a much smoother shell. The others all have a rough shell, covered rather sparingly with irregular raised patches of dark but bright green on a lighter green and smooth ground. In the pale (No. 1) variety u 2 128 HON. W. ROTHSCHILD ON THE GENUS CASUARIUS. these raisings on the shell are closer together, and not so well developed; in both varieties they are more thinly spread over the central portion than at the ends. On the whole they closely resemble the eggs of Caswarius bennettii, in which similar variations are noticeable; but they are larger and of a greater diameter, being greatest in the middle. I am indebted to Inspector Robert Johnstone for the fine series of the eges of this species which at present grace my collection. “The following are measurements of some of the specimens of the eggs of both species :— “ Casuarius australis. Length in inches. Width in inches. “No. 1. Light-green smooth shell . . . . 5:33 x 3°73 No. 2. Dark-green rough shell . . . . . 53 x 3°88 “ Casuarius bennettii. “No. 1. Light-green smooth shell . . . . 5°65 x 3°54 No. 2. Light-green rough shell . . . . . 5°82 x 3°31 No. 8. Light-green rough shell . . . . . 5:34 x 3°4 No. 4. Dark-green rough shell . . . . . 52 x 3°32 “The young of Caswarius australis are of a dull rusty brown, the feathers having frequently a blackish shaft-stripe, giving to the back a streaked appearance. After the first year the plumage takes a deeper, lighter brown hue, and black feathers begin to appear mixed with brown, some being parti-coloured. After the second season, at the age of from 18 to 24 months, the black feathers predominate, and the helmet, which has hitherto been undeveloped, more like the shield of a Coot (fulica), begins to show a keel or ridge in the centre, which rapidly increases in height. ‘The skin round the head, on which still remain a few brownish hair-like feathers, begins to become wrinkled and coloured, varying from bluish-green to orange on the lower part, and bright blue on the sides of the neck, the wattles becoming carmine. ‘The helmet still remains comparatively small and undeveloped long after the wattles and naked parts of the neck become coloured. I believe that the helmet does not attain its full size until the fourth or fifth year at least. “Tn traversing these scrubs the head is carried low to the ground, and the vines and branches of trees striking the helmet slide over it on to the back. Otherwise, in the dense vine-scrubs bordering the Herbert River and elsewhere, progress would be greatly impeded; but as it is, the Cassowaries traverse the scrubs with wonderful speed, jumping over fallen trees and logs when in the way. -shaped notch of the upper limb is greatly developed, forming a long, backwardly directed, spine-like supra-orbital process (Pl. XLII. fig. 2). The inferior limb is notched below the lachrymal foramen. In D. nove-hollandie the inner fork of this notch is short and confluent with the free edge of the ventral border of the antorbital plate. The outer fork is produced downwards far beyond this, and rests upon the quadrato-jugal bar. In D. irroratus and D. ater the outer and inner forks of the prong are subequal, and do not quite reach the level of the ventral border of the antorbital plate, which is continued downwards by means of a slender column on to the maxillo-palatal process, where it expands. In Struthio (Pl. XLII. fig. 3) the outer fork of the upper limb is comparatively short and blunt, and projects more laterally than in the above mentioned. Further- more, it differs in that it is joined by a series of large bony scales lying between its posterior extremity and the supra-orbital ledge of the frontal, recalling the supra- orbital chain of bones in the Tinamou. Late in life the scales fuse with the lachrymal in front and the frontal behind, thus enclosing a large supra-orbital foramen. The lower limb of the lachrymal takes the form of a slender backwardly curved rod, whose free end is curved sharply outwards to join the inner border of the quadrato-jugal bar. To the inner side of this outwardly directed free end is attached a larger ossiculum palatinum, which in the adult fuses with the lachrymal on the one hand, and the antorbital plate on the other. Below the ventral border of its inner end comes in contact with the palatine. AEpyornis and Dinornis have the lachrymal so completely fused with the frontal and antorbital as to be with difficulty distinguishable. ‘The upper limb in both has completely merged into the supra-orbital ledge. The lower in Dinornithide is much flattened antero-posteriorly. The lachrymal foramen is imperfect, its outer bar being wanting. Insome, e.g. Anomalopteryz, it is perfect. Its ventral extremity is squarely truncate and articulates with the maxillo-jugal bar. In the only Apyornis skulls I have been enabled to examine, this lower limb has been broken away. In both these groups the anterior border of the lachrymal rests against a well- marked, spine-like, maxillary nasal process. In Apteryx the lachrymal is described by Parker [71], from young specimens, as “a small irregular bone consisting of a shell-like central portion applied to the surface of the aliethmoid, and of an ascending portion which articulates with the descending process of the nasal. It is perforated obliquely by the lachrymal foramen.” In the adult it is completely fused with the aliethmoid. Its anterior border is often traceable as a thin suture between it and the maxillary process of the nasals. In the Crypturi the lachrymal is indistinguishably fused with the nasal, and forms therewith a prominent outstanding process bounding the orbit in front. There are no backwardly projecting spines. The ventral limb fuses with the antorbital plate. 180 MR. W. P. PYCRAFT ON THE MORPHOLOGY AND The Cranial Cavity.—The metencephalic fossa in all, save Dinornis and Afpyornis, takes the form of a moderately deep basin, the floor of which is tilted upwards toa very considerable extent. The anterior tilted portion forms the posterior wall of the pituitary fossa. Posteriorly it is continued backwards on to the occipital condyle. Its postero-lateral border, immediately below the pro-6tic is pierced by the vagus foramen, to the inner side of, and slightly posterior to, which lie one or two condyloid foramina. Mesiad of the internal auditory meatus, and anterior to the same, is the abducens foramen (Vv1.). In Zpyornis and Dinornis the floor of this fossa is almost flat. The ceredellar fossa is roofed by the parietal and supra-occipital bones. ‘The pro- dtics bound it laterally, ventrally it passes into the metencephalic fossa. It is bounded behind by the free edge of the occipital foramen, and in front by the median portion of the tentorial ridge. The late Prof. Jeffery Parker, in his masterly monograph, states that in Dinornithide the supra-occipital region of this fossa is marked by “ transverse _grooves corresponding with the gyri of the cerebellum.” Immediately above the internal auditory meatus lies the floccular fossa. The size of this, according to Parker, appears to vary individually in Dinornithide. It is of considerable depth in the other Palwognathe and in 'Tinamous. The mesencephalic fossa is a deep basin-like depression in all save Dinornithide and Apterys. It appears to be more sharply defined in Casuarius than in any other member of this group. It is bounded above by a very prominent shelf of bone, forming the lateral portion of the tentorial ridge, behind by the pro-dtic. It extends forwards considerably beyond the level of the pituitary fossa which bounds it in the middle line. ‘The trigeminal foramen (v.) leaves by a large aperture excavated out of the ventral border of this fossa, between it and the pro-6tic, and is of considerable size. Just within the mouth of this foramen, below the rim of its anterior border, lies the aperture of the orbito-nasal nerve (v.'), which has its exits in an almost obsolete lacerate fossa. In Dromeus the trigeminal and orbito-nasal foramina are distinct. The tentorial ridge, bounding the fossa superiorly, is almost as strongly developed as in Casuarius. In Struthio this fossa is relatively smaller and shallower, neither are its boundaries so sharply defined. The apertures of the trigeminal foramen and the orbito-nasal nerve lie close together. Rhea and Crypturi have the fossa somewhat more sharply defined than in Struthio, but in none of these does that portion of the tentorial ridge bounding the fossa superiorly form anything more than a low ridge. In Dinornithide this fossa is relatively ill-defined and shallow. The aperture for the trigeminal foramen steals away a large portion of its external wall. The orbito- nasal aperture opens as in Casuarius, is tunnel-like, and bounds the fossa anteriorly, separating it from the pituitary fossa. PHYLOGENY OF THE PALZOGNATHA AND NEOGNATH A. 181 In Apteryx it appears to be reduced to the vanishing point. It is extremely small, oblong in shape—its long axis horizontal,—and more or less sharply defined, when seen at its best. It lies entirely behind the level of the pituitary fossa. It is bounded above by a low tentorial ridge, in front by a ridge of bone dividing it from the pituitary fossa, below by the rim of the metencephalic fossa, and behind by a huge projecting shelf-like ledge from the tentorial ridge, The trigeminal foramen opens at its postero-dorsal border, the orbito-nasal in its anterior border. In A. haasti the outer wall of this fossa is almost entirely absorbed by a large aperture common to the trigeminal and orbito-nasal nerves. The pituitary fossa in Casuarius is a deep pocket-shaped depression. Its mouth is directed slightly forwards. The rim thereof forms the dorsum sell@ behind, and the pre-pituitary ridge in front. Within the mouth of this fossa (Dinornis) or on its rim (Dromeus) is the oculo-motor (11t.) foramen. The pathetic (1v.) foramen lies above and to the outer side of the oculo-motor, often it serves also for the passage of the internal ophthalmic artery. Anterior to the prepituitary ridge in the middle line is a narrow knife-like optic platform, on either side of which are the confluent optic foramina, The optic platform is continued upwards vertically and again expands to form the pre-optic ridge, which in Dromeus, and slightly less obviously so in Casuarius, joins the tentorial ridge. In Struthio the pituitary fossa is of great depth; the pre-pituitary ridge and optic platform are well marked. In Dinornithide the fossa is comparatively shallow, but of considerable circumference. The optic platform is very wide. In Apteryx the pre-pituitary ridge is but feebly developed. ‘The optic platform is not distinguishable from the pre-pituitary ridge, which divides the optic foramina one from another. ‘The pre-optic ridge is well developed and overhangs the fossa, the mouth of which, in consequence, has a semicircular outline. It should be mentioned here that the bottom of the fossa is produced forwards into a yet deeper, almost spherical, depression, overhung by the pre-pituitary ridge. In Tinamous this fossa is deep and tubular in form. The cerebral fosse are sharply defined in all the Palwognathe save Apteryx. In all, with the exception just mentioned, the cerebral lies entirely in front of the cerebellar fossa. In Apteryx, and to a lesser extent in the Crypturi, the cerebellar contour is scarcely distinguishable from the cerebral. ‘The tentorial ridge, which serves to cut off the cerebral from the remaining fosse, arises from the level of the pre-optic ridge and trends outwards and backwards to the level of the top of the pro-otic; from this point it turns sharply upwards and forwards to meet in the mid-dorsal line, over the region corresponding with a vertical line passing upwards from the dorsum selle, That portion of the ridge which forms the upper boundary of the mesencephalic fossa in Casuarius and Dromeus is of great depth, and projects into the cranial cavity like a voL. Xv.—Part Vv. No. 10.—December, 1900. 2D 182 MR. W. P. PYCRAFT ON THE MORPHOLOGY AND deep shelf. This is less marked in Rhea. In Dinornis and Crypturi it is but feebly developed. The cerebral fosse in Dinornithide do not taper forwards to finally pass into the infundibuliform olfactory fosse, but, on the contrary, are of great size and width in this region, reducing the olfactory fossz to the dimensions of shallow pits. The median bony falx, continued forwards from the tentorial ridge to the crista galli, is extremely well developed in Casuarius and Dromeus, less so in Rhea, and only feebly in Struthio and Dinornis and Tinamous. ‘There is no trace of it in Apteryx. In Casuarius and Dromeus the fossee bear distinct and deep impressions for the temporal and frontal lobes, a low ridge indicating an indistinct Sylvian fissure. The olfactory fosse in all, save Dinornis and Apteryx, are paired, more or less tubular, infundibuliform chambers, separated one from another by a crista galli. Anteriorly they are closed by a more or less perforated plate for the passage of the olfactory nerves. In Dinornis the olfactory fossee take the form of shallow pits lying at the anterior end of the large cerebral fossa. The cribriform plate is large. In Apterya these fossee are of great size, and pass insensibly backwards into the cerebral fosse. They are not, in the dried skull, shut off anteriorly from the turbinals by a cribriform plate as in other Radtitw, but continued forwards directly into the olfactory chamber. The Premaailla. The premaailia in all the Palewognathe is peculiar, in that the nasal process is median, and not paired as in all other birds (Pl. XLII. figs. 1-3). In all save Struthio there are well-developed palatine proceses. In Casuarius the beak is more or less laterally compressed. The body—that portion of the premaxilla from which the nasal, maxillary, and palatine processes are given off—is short, and more or less deeply grooved on either side, so that the distal end of the nasal process seems to be continued forwards into a strongly marked ridge. The maxillary processes extend backwards as a pair of slender splints to the level of the lachrymal on either side. ‘The palatine processes are a pair of short narrow lamine extending backwards to a point corresponding with a vertical line passing upwards behind the top of the free end of the median nasal process. They enclose anteriorly a small chink-like palatine notch. In C. australis there is a small pre- narial septum, corresponding to that described by Parker in Dinornis, as the ‘“ strong ascending keel, formed posteriorly of paired plates, but solid in front and gradually diminishing in height towards the top of the beak.” This keel is derived from the dorsal surface of the palatine processes. The nasal process is rod-like, its free end passes backwards into a tunnel-shape in the inflated, mesethmoid mass forming the characteristic casque of these birds. PHYLOGENY OF THE PLZOGNATHA AND NEOGNATHA. 183 In Dromeus the beak is flattened dorso-ventrally. The premaxillary body is larger than in Caswarius, and, as in this genus, the rod-like nasal process seems to be continued forwards to the tip of the beak by reason of paired lateral grooves which divide it from the maxillary processes. The nasal process is relatively longer than in Caswarius. Posteriorly it is received into a shallow groove excavated out of the median borders of the nasals. The tip of its free end rests upon the mesethmoid. The maxillary processes are splint-like, and extend backwards to the level of the anterior region of the orbit—beyond the body of the lachrymal. The palatine processes are short pointed lamelle, rather more widely separated one from another in the middle line than in Caswarius. The chink-like anteriorepalatine foramen is closed caudad, as in Caswarius, by the vomer. ‘This process extends backwards as far as the level of the anterior border of the posterior narial aperture. In Casuarius, it should be mentioned, it falls short of this. In Dinornis the median nasal process is markedly flattened and expanded dorso- ventrally. ‘The maxillary processes are very short, not extending backwards beyond the level of the hinder end of the antorbital fossa. The palatine processes take the form of flattened lamine, extending backwards, as in Dromeus, to the level of the anterior end of the posterior nares. The alveolar borders are deeply pitted. The anterior palatine foramen is small and cordiform. It is closed above by the parasphenoid rostrum, and posteriorly by the vomer. In Hmeus it is more completely closed than in any other Palwognathe by a pair of processes from the postero-internal angles of the palatine processes, which meet in the middle line, serving at the same time as articular surfaces for the anterior ends of the probably paired vomers. The prenarial septum is most clearly marked in Dinornis. Its postero-ventral border is bilaminate, but the lamine rapidly fuse to form the single median vertical septum. In Apyornis the nasal process rises sharply from the body of the bone, giving the beak a peculiar conical form. ‘The alveolar border is deeply pitted. The backward extensions of the palatine and maxillary processes cannot be made out in either the British Museum skull or that belonging to the Hon. Walter Rothschild. In the British Museum skull there is a prenarial septum, but the lamine composing this have not fused to form a median plate as in Dinornis. Instead, as indicated, they remain widely separated, and enclose between them a long tunnel-like cavity extending forwards to the tip of the beak. In Rhea the premaxillary body is of greater length than in any other Ratite, and much depressed dorso-ventrally. It is grooved dorsally on either side of the distal end of the nasal process, as in the other forms. The nasal process is very broad and flattened, but tapers more or less caudad 9 ‘*) 4 ~ 134 MR. W. P. PYCRAFT ON THE MORPHOLOGY AND The maxillary, which is confluent for about half of its length with the palatine process, is styliform, and terminates in the anterior region of the antorbital fossa, The palatine process is of great size, extending beyond the maxillary process, and terminating at, or falling but little short of, the anterior border of the posterior nares. I this it agrees with Dromeus. Struthio differs from all the other Ratite in the absence of a palatine process, and the consequent enormous size of the palatine foramen. The whole premaxilla has a unique triradiate form ; the radii are of great length, so that the body of the bone is reduced to the smallest possible dimensions. The nasal process is broad and flattened, tapering somewhat abruptly posteriorly. The maxillary processes are long and styliform, terminating at the level of the anterior region of the orbit. In Crypturi the form of the premaxilla is precisely similar to that of Rhea. The free end of the palatine process is brought to the level of the anterior border of the posterior nares, and interlocks with the distal end of the palatine. In this last point it differs from Rhea. The palatine foramen, as in Rhea, is long and narrow. Apteryx is the most abnormal of all the Palwognathe. The whole beak is produced forwards into an elongated probe, curved slightly downwards, in the tip of which the external nostrils he. In the adult skull the fusion of the separate elements, not only of this region of the beak alone, but of the skull generally, is so complete, that it is impossible to make out their boundaries. The body of the premaxilla is reduced to its smallest possible limits, being com- posed only by the confluent end of the rod-like nasal process, and of the combined maxillary and palatine processes, which for this reason are but delicate rods. The nasal process takes the form of a long slender rod, grooved ventrally and running backwards to the mesethmoid. Its posterior region rests upon the elongated nasal bones. The maxillary and palatine processes are fused throughout the greater part of their length, and form elongated flattened rods, divided in the middle line by a narrow groove representing the palatine foramen. Posteriorly there is given off a short splint-like palatine spur, representing the palatine process. The extremest limit of this falls considerably short of the anterior end of the antorbital fossa. The maxillary process can be traced backwards splint-wise as far as the level of the foot of the lachrymal. Thus the maxillary processes are of about equal length, ‘The nasal, and nasal process of the premaxilla above, and the maxillary process of the premaxilla below, serve to form a long narrow groove leading forwards into the anterior nasal fossa, this last having shifted forwards to the extreme end of the beak, This nasal furrow bears a strong resemblance to that of the Ibises, the difference being that the nasal fossa hes behind instead of in front of the groove, as in Apteryx. It seems to suggest that the shifting forward of the nasal aperture has followed the course PHYLOGENY OF THE PALZOGNATHA AND NEOGNATHA., 185 of the orbito-nasal nerve, accompanying it to its tip, so as to bring the sense of smell to aid the sense of touch found in the Ibises. It is interesting to note that in the very early embryo, as shown in Parker’s figures, the position of the external nares is near the middle of the beak. Its forward shifting seems to be involved in the growth of the elongating beak. The Maxillo-jugal Arch. The mazilla in Casuarius takes the form of an elongated flattened splint. Its interior.half has the outer and inner borders turned upwards to meet in the mid-dorsal line so as to form a long thin-walled tunnel tapering forwards to a point. This is the antrum. The rim of the mouth of this is produced backwards into two short spikes ; the outer forms a strong dorso-lateral and the inner a lateral spike, conspicuous when the skull is viewed ventrally. The floor of this tunnel, viewed ventrally, is twice as wide as the posterior region of the maxilla. This wider region—the floor of the antrum,— which in the ventral view presents a deeply hollowed free posterior border forming the anterior boundary of the posterior narial aperture, is the mavillo-palatine process. The whole inner border of this process is grooved for the reception of the vomer. The maxilla extends forwards above the maxillary and palatine processes of the pre- maxilla. The latter process is received into a shallow depression in the maxillo- palatine, a raised ridge of which bone can be seen lying between the palatine and maxillary process. The backward extension of the maxilla is considerable. It runs below the quadrato-jugal bar on the outer side and the palatine on its inner side to a point slightly beyond the level of the free border of the hinder end of the posterior nares, or, to take another standard, to within a short distance of the base of the vomer. The quadrato-jugal fossa is the name which I have applied to indicate the large space enclosed by the quadrato-jugal bar on the outside, the palatine and pterygoid in the inside, and the hinder end of the maxilla in front. In Casuarius it does not extend forwards beyond the hinder end of the posterior narial aperture. In Dromeus the maxilla is relatively much wider than in Caswarius, and does not extend so far backwards. The maxillo-palatine processes are wider but very much shorter antero-posteriorly than in Caswarius. The antrum, in consequence, is much reduced in length and capacity. Its dorsal wall, moreover, is fenestrated. The mesial borders of the maxillo-palatine processes rest upon the vomer. The quadrato-jugal fossa extends forwards to within a short distance of the middle of the posterior narial aperture. In Dinornithide the maxillo-palatine region of the maxilla is short and broad. Its anterior end is wedged in between the maxillary and palatine processes of the pre- maxilla. Ventrally it appears as an oblong flattened plate. Its postero-internal border bounds the anterior end of the posterior nares. Its posterior border is either notched or hollowed and forms the anterior boundary of the quadrato-jugal fossa. 186 MR. W. P. PYCRAFT ON THE MORPHOLOGY AND From its dorsal surface in Dinornis there arises a large hollow shell of bone, the inner wall of which embraces the triangular plate of the rostrum. ‘The cavity of this bony shell—the antrwm—opens posteriorly by a large foramen into the quadrato-jugal fossa. It articulates with the palatine by a short recurrent process from its postero-internal angle. In Emeus there is no antrum. What corresponds to its inner wall in Dinornis exists as a flat concavo-convex vertical plate applied to the anterior border of the mesethmoid triangular process and the rostrum. In Rhea the maxilla is almost entirely represented by the maxillo-palatine process. Viewed dorsally it is seen to take the form of a flat, triangular, perforate plate of bone, with its sharply pointed end running forwards to within a short distance of the anterior end of the external narial aperture. ‘This more pointed region rests upon the palatine process, and is bounded externally by the maxillary process of the premaxilla. Its postero-mesial border is bounded by the vomer. Its hinder end presents free postero-internal and external borders. The former lies more or less markedly forward than the latter, and bounds the posterior nares anteriorly. The latter lies more or less considerably caudad, and bounds the anterior end of the quadrato-jugal fossa. Its postero-internal angle is produced backwards into a long spine, which is closely applied to the outer border of the palatine, extending backwards for a considerable distance. Its postero-external border is also produced backwards spine-wise, and underlies the quadrato-jugal bar. From the dorsal surface of the maxillo-palatine plate there arises a long slender style which, directed upwards and backwards, ultimately articulates with the anterior angle of the lachrymal, and thus serves to divide the lachrymo-nasal (antorbital) from the nasal fossa (Pl. XLIII. fig. 9). In all other birds this is done by the descending maxillary process of the nasal. The antrum is represented only by a minute aperture at the base of the vertical rod—the lachrymo- nasal process of the maxillo-palatine—just described. The guadrato-jugal fossa extends forwards nearly as far as the level of the anterior end of the posterior nares. In Struthio the maxilla is splint-shaped, produced posteriorly into two spine-like processes, the inner of which runs dorsad over the distal end of the palatine, the outer running ventrally below the quadrato-jugal bar. Anteriorly it extends forwards nearly as far as the body of the premaxilla. The maxillo-palatine processes are small and hatchet-shaped ; their mesial borders clamp the vomer, which is grooved laterally for their reception. There is a small shallow antrum, the roof of which is fenestrated. The anterior border of the maxillo-palatine process bounds the enormous palatine foramen posteriorly. The ossified style, described in Rhea as dividing the lachrymo-nasal from the nasal fossa, is here represented by a ligament springing from the external dorsal angle of the mouth of the antrum and inserted into the short descending process of the nasal. PHYLOGENY OF THE PALEZOGNATHAH AND NEOGNATHA. 187 The quadrato-jugal fossa extends forwards to within a short distance of the anterior border of the posterior nares. In Crypturi the maxilla resembles that of Rhea, but its maxillo-palatine process is much narrower and is unfenestrated. There is no antrum. ‘The lachrymo-nasal fossa is bounded anteriorly by a descending process of the nasal and not by a spur from the maxillo-palatine as in Rhea. In Apterya the maxilla is of great size, its maxillo-palatine processes, seen ventrally, extending forwards to the middle of the elongated beak, and backwards along the outer border of the palatine to within a short distance of the level of the free end of the orbital process of the quadrate. They are separated one from another in the middle line, forwards, only by the narrow vomer. Posteriorly there is an external lateral spur for articulation with the quadrato-jugal bar. The proximal mesial border is closely fused with the external border of the short palatine. ‘There is no antrum. Immediately in front of the distal end of the palatine the maxillo-palatine process occasionally sends inwards a very short process to articulate with the vomer, which apparently corresponds with the much larger vomerine process found in Struthio. The quadrato-jugal is so reduced in size in Casuarius as to be represented by little more than a nodule of bone, scarcely projecting beyond the limits of the glenoid cup of the quadrate. The jugal in Casuwarius is connected by a close vertical suture with the quadrate portion, and projects above it at this part in the form of a more or less prominent spine. It extends forwards as far as the middle of the lachrymo-nasal fossa. In Dromeus, seen from the outside, the quadrato-jugal scarcely exceeds that of Casuarius (Pl. XLIII. fig. 4a). On the inner side, however, it extends forwards along the jugal for about one-third of its length. In Casuarius it does not extend beyond the rim of the quadrate cup on the inside. The forward extent of jugal is about the same as in Casuarius. Rhea and Struthio resemble Dromeus in the size and relations of these two bones. In Apterya the quadrato-jugal is a long slender style, reaching as far forwards as the middle of the zygomatic arch. The jugal is long and slender, extending forwards along the dorsal aspect of the quadrato-jugal as far as the level of the anterior extremity of the palatine, which corresponds with a line drawn transversely across the vomer about its middle. In Crypturi the quadrato-jugal is as much reduced as in Caswarius. The mesial border of the proximal end of the jugal articulates, as in this form, with the quadrate. The Vomer, Pterygoid, and Palatine (see also p. 206). (Pl. XLII. figs. 5-8.) In Casuarius the vomer is of great length and slenderness. In C. ¢. salvadorii, C. ec. sclateri, and C. c. beccarti it is cleft from behind forwards for more than half its 188 MR. W. P. PYCRAFT ON THE MORPHOLOGY AND length into two slender rods grooved dorsally; in C. bennetti only about one-fourth is so divided: the cleft not extending more than half the distance of the posterior nares, in the other forms it is continued forwards beyond this point. C. c. australis differs from both types, the cleft of the vomer being rather less than half the total length and terminating in the region of the anterior end of the posterior narial aperture. Anteriorly the vomer becomes laminate and wedged in between the maxillo-palatine processes, and passing forwards rests upon the palatine processes of the premaxilla. The palatine foramen by the great forward extension of the vomer is almost obliterated. In a ventral view—before complete anchylosis has taken place—the feet of the vomer appear to be wedged in between the pterygoid and palatine. Removal of this latter, or a dorsal view, however, reveals the true relations of these parts. It will then be seen that the right and left limbs are closely applied to the external ventral border of the twisted pterygoid, extending backwards to a point very near its middle. The pterygoid can be traced forwards along the mesial border of the vomer as a slender spike for some considerable distance. On a dorsal view, the right and left grooved limbs of the vomer are well displayed, and show that each has been twisted outwards so that its inner free border comes to look dorsally, and the latter ventrally. The flattened surface of this twisted region is closely-applied to the ventral surface of the anterior end of the pterygoid, so that vomer and pterygoid form one common rod articulating with the quadrate. The dorsal border of the twisted end of the vomer is slightly thickened and, traced backwards, trends downwards beneath the outer border of the pterygoid. From the vomer backwards, the sutural surface is continued as a shallow groove along the outer ventral border of the pterygoid, terminating at the level of the basipterygoid processes. It is this ventral union of palatine and pterygoid that gives the impression that the vomer is wedged in between these two bones, when the skull is viewed from this surface. In Dromeus (Pl. XLIT. fig. 7) the “feet” of the vomer extend forwards for about one-third of its total length. Anteriorly it expands into a broad oval plate tapering again forwards to terminate, as in Caswarius, near the body of the premaxilla. The maxillo-palatine process overlaps the distal end dorsally; the palatine processes underlie it ventrally. Its relations with the pterygoids and palatines posteriorly are precisely similar to those which obtain in Caswarius. As a matter of fact, it should be remarked, the palate of Dromeus is even simpler in type than in Casuarius, and consequently represents, as we shall see, the simplest arrangement of these bones of all the Palwognathe. After these, the simplest palate amongst the remaining Paleognathine forms is that of Struthio. The vomer of Struthio (Pl. XLII. fig. 8) is a degenerate structure. Anteriorly it does not extend beyond the middle of the enormous palatine foramen ; posteriorly it is cleft, but has entirely lost all connection with the pterygoid, falling short thereot PHYLOGENY OF THE PALZOGNATHE AND NEOGNATHA. 189 by some very considerable space. It has assumed a trough-like form, the hollow of the trough closely embracing the parasphenoidal rostrum. On either side the vomer is closely embraced by the maxillo-palatine processes, which fit into specially roughened facets. A similar, but less perfect, arrangement of these parts obtains in the Cassowary. If the vomer were produced backwards on the pterygoid we should have a type of palate more nearly resembling that of Casuwarius and Dromeus than any other Ratite. In Rhea the vomer is cleft posteriorly for rather less than half its length, and is pro- duced anteriorly into two large blade-like lamine, widely separated in the median line, and each terminating in a pointed extremity. These anterior lamine are embraced between the palatine processes of the premaxilla and the maxillo-palatines. The parasphenoidal rostrum appears in the median line, between these processes, terminating at the level of their free anterior ends. In Casuarius, Dromeus, and Struthio the anterior end of the rostrum is entirely concealed by the vomer. In the two former the free end of the vomer projects far beyond the rostrum, in the latter the free end of this and of the vomer terminate at the same point. The body of the vomer takes the form of a pair of vertical lamin, fused anteriorly, and moulded on to the rostrum, upon which it slides; the forward anterior horns lie in a horizontal plane ; the feet are obliquely truncated, their external lateral borders suturating with the pterygoid (Pl. XLII. fig. 5). The true relations of the vomer to the pterygoid cannot be made out except from a study of the dorsal view; they will then be seen to be precisely what obtains in Casuarius or Dromeus, the foot of the vomer underlying the anterior end of the pterygoid. This last is in Ahea rod-shaped in form, but the sutural surface is dorso-ventrally compressed and closely applied to the vomer; at least in young birds, in adults it seems to become less compressed and yet more pointed. Ina ventral view of the skull the vomer would appear to suturate with the palatine only and not the pterygoid, the former being apparently thrust in between them. Examined from the dorsal surface, however, it will be seen that the palatine—the exact form of which will be discussed presently—underlies the foot of the vomer on either side, and is continued backwards along the pterygoid in the form of a long pointed spur, concealing all but the proximal end of the pterygoid, so that, seen ventrally, as already remarked, the vomer and pterygoid would seem to be shut off one from the other. This is a fact of considerable significance, inasmuch as this isolation of the vomer is a point wherein all the Neognathew appear to agree, and differ from the Palwo- ynathe. But, as I have elsewhere shown, when discussing the palatine of Rhea (p. 206), this is only an apparent difference. In Crypturi the vomer is precisely similar in form and proportions to that of Rhea. VOL. xvV.—ParT V. No. 11.—December, 1900. 22 190 MR. W. P. PYCRAFT ON THE MORPHOLOGY AND Furthermore, it agrees with the same in the nature of its relations with the pterygoid. The palatine, in the part played by it with regard to the pterygoid and vomer, differs but little from that of Rhea. Its chief difference consists in the fact that it is much narrower from side to side, and that it terminates in a pair of spines, an outer and an inner, instead of an outer pterygcidal only as in Rhea. The form of the palatine, as a whole, will be described latter. In Dinornithide, e.g. Hmeus (Pl. XLII. fig. 3), the vomer takes the form of a pair of perfectly distinct vertical lamine embracing the parasphenoidal rostrum. They are continued forwards only for a very short distance beyond the level of the anterior end of the posterior narial aperture. It differs from that of Rhea (1) in the greater vertical extent of the paired lamin, (2) in that these are often free throughout, and (3) in that there are no anterior horns running forward between the palatine processes of the premaxilla and the maxillo-palatine. In Himeus, the palatine processes send inwards a pair of flattened spurs to embrace the sides of the anterior end of the vomerine plate. Posteriorly, the feet of the vomer suturate with the pterygoid exactly as in Rhea. The relations between these bones and the palatine is also the same. Indeed, the inturned under-flooring region of the palatine in Hmeus is more strongly developed than in Lhea, projecting backwards to the level of its pterygoidal spur. In Apteryx the vomer is relatively short, and cleft from behind forwards for about half its length. Its form and relation are described on p. 205. The pterygoid in Casuarius is, seen ventrally, au elongated, crescentic rod of bone dorso-ventrally compressed, with convex surface directed towards, and partly resting on, the parasphenoidal rostrum, and its concavity, for the most part, articulating with the palatine. Anteriorly, it runs forwards along the inner side of the hinder limb of the vomer—which is at this point wedged in between the pterygoid and palatine—and terminating thereon in a sharp spine. Its proximal end articulates with the quadrate, below the orbital process; mesiad and anterior to this articulation is a second for articulation with the basipterygoid process. Seen dorsally, the pterygoid is fusiform, not crescentic, a large median tube overlapping that part of the palatine which ona ventral view appears within a crescentic hollow. Thus the true shape of the pterygoid cannot be seen without the removal of the palatine, which discloses the dorsal overlapping lip. In Dromeus (Pl. XLII. fig. 7) the form and relations of the pterygoid are simpler. Its relative length is about the same, and it terminates anteriorly ina point. The anterior region of its external lateral border is directed upwards in the form of a thin vertical plate affording a sutural surface for the external vertical plate of the vomer. The posterior region of this border is straight and affords attachment to the palatine. Its internal, mesial border is convex, the convexity facing the angle between the PHYLOGENY OF THE PALZOGNATHA AND NEOGNATH A. 191 parasphenoidal rostrum and the basipterygoid processes. The palatine is but loosely attached to this bone. The free end of the vomer underfloors its anterior end. In Struthio the form of the pterygoid (Pl. XLII. fig. 8) closely resembles that of Dromeus: the points wherein it differs are points wherein it is also degenerate. It lacks the upstanding vertical plate from its antero-external lateral border, and it has fused with the palatine. In Khea the pterygoid (Pl. XLII. fig. 8) has undergone a very considerable reduction in size, and now takes the form of a \— -shaped rod. The short upper limb bears facets for articulation with the quadrate and basipterygoid processes. The longer is laterally twisted around its own axis, and continued forwards till it reaches the parasphenoidal rostrum. Ventrally it is grooved. This grooved surface is applied, anteriorly, to the superior border of the vomer, and posteriorly to the upper surface of the postero-external angle of the posterior extremity of the palatine. Thus the posterior end of the palatine comes to underlie the posterior extremity of the vomer. In the Dinornithide, e.g. Hmeus, the form of the pterygoid very closely resembles that of Rhea. The points wherein it differs are such as must be regarded as repre- senting a more primitive condition: the chief of these lies in the greater width dorso- ventrally; the pterygoid of Hmeus possessing a broad ledge along its mesial border. Its relations with the vomer and palatine are practically the same as in Rhea. In Crypturi the pterygoid is rod-like as in Rhea, relatively longer, much straighter, and more slender. In Apteryx the pterygoid is elongate, laminate, and pointed in front where it rests upon and fuses with the vomer. It anchyloses very early and completely with the palatine, so that the precise form of the pterygoid as a whole cannot be made out in the adults (see p. 204). Proximaily, it is tightly wedged in between the basipterygoid process on the one side and the orbital process of the quadrate on the other, the pterygoid having the appearance of being “‘morticed” in between these two points of contact. The palatine in Casuarius takes the form of a roughly triangular plate. Its base articulates with the pterygoid and vomer, and the anterior half of its outer side with the maxilla. Its connection with the pterygoid is by means of an overlapping suture, the pterygoid underlying a broad shelf of bone given off from the palatine. The vomerine suture is that of a simple apposition of contiguous edges, as also is the junction with the maxilla, In the adult the palatine and pterygoid become indistin- guishably fused proximally. In Dromeus the palatine resembles in form that of Casuarius. It is, however, fenestrated. Its suturations differ only in that there is no overlapping shelf above the pterygoid. ‘here is no fusion between pterygoid and palatine as in Caswarius. In Rhea the palatines (Pl. XLII. tig. 6) are broad thin lamine, roughly quadrangular 2E2 192 MR. W. P. PYCRAFT ON THE MORPHOLOGY AND in form, with the external border produced backward into a pointed spur which runs along the pterygoid, and forwards for a considerable distance along the maxilla. Its mesial border is cut away anteriorly so as to embrace the external border of the vomer instead of underlying it, as is done by the remaining posterior portion. Its form is seen best from the ventral aspect of the skull. In the Dinornithide, e. g. Emeus, the form of the palatine resembles that of Rhea, but by a little tension the body of the bone has come to be vertical instead of hori- zontal in position, thus giving a rod-shaped superficial appearance. In the Crypturi the main body of the palatine is, as in Rhea, horizontal, but the anteriorly produced external lateral border is of much greater length, and takes the form of an elongated and slender rod suturating anteriorly with the maxilla. In Struthio the palatine (Pl. XLII. fig. 8) resembles that of many Carinate in that the main body of the bone is twisted into the form of an almost vertical scroll. It is peculiar, however, in that it is fused at its base with the overlapping pterygoid. The external lateral border is produced forwards, as in Tinamous, into an elongated slender rod, but which fuses anteriorly with the base of the hastate maxillo-palatine process. In Apteryx the palatine (Pl. XLII. fig. 5) takes the form of a more or less flattened rod, differing in its relations from all the other associated forms. Anteriorly, it is fused with the maxillo-palatine process. The greater part of its external border is anchylosed with the maxilla. The remaining posterior end of this bone anchyloses with the external border of the pterygoid. What corresponds with the posterior region of the mesial border in Dromeus is here free, and runs in a backwardly transversely oblique direction across the skull. For the precise relations of the bone however, see p. 204. The Quadrate. (Pl. XLITI. figs. 10-12.) The quadrate in Caswarius has its various processes very sharply defined. The ofie process is very slender, tapering gradually downwards into the body of the bone. At its base, where it merges into the main body, isa large pneumatic foramen, it opens on the posterior surface. The head for articulation with the pro-6tic and squamosal is single, oblong in form, with its long axis directed backwards and inwards. The inner and outer mandibular condyles join mesially, the two projecting backwards to form a conspicuous spur. The difference in the level of the two condyles is but slight. The orbital process is well defined, moderately large and free, with a sinuous superior border. At its extreme base, and on its inner side, is a small demi-facet for the pterygoid. ‘The outer surface of the quadrate is very closely applied to the long stout zygomatic process of the squamosal, which terminates only a short distance above the level of the cup-shaped quadrato-jugal fossa. PHYLOGENY OF THE PALHOGNATHA AND NEOGNATHA, 195 The quadrate of Dromeus differs but slightly from that of Caswarius. The outer and inner mandibular condyles are on distinctly different levels, the inner being ventrad of the outer, shorter from without inwards and wider from before backwards. The backwardly projecting condylar spur is less marked and formed entirely by the outer condyle. The orbital process is less hollowed above. In the Dinornithide, e. g. Emeus, the quadrate resembles that of Dromeus in the difference of level between the outer and inner mandibular condyles and in the form of the orbital process. This process is, however, peculiar in that its whole inner surface is requistioned for the articulation of the pterygoid. The inner mandibular condyle is peculiar in that the middle region of its articular surface is produced downwards, whilst the outer condyle looks more directly downwards than in Dromeus. The pneumatic foramen is mesiad, and not posterior, A posterior foramen is, however, sometimes present in the Dinornithide. In Struthio the inner and outer mandibular condyles are widely separated ; otherwise they closely resemble those of Dromeus. The orbital process differs from that of the other forms so far described, in that its distal end turns sharply upwards, and in that there projects from its inner ventral border a sharply defined articular facet for the pterygoid. ‘The ofc process is relatively longer than in the other forms. Both the mesial and posterior foramina are present, though the latter may apparently be sometimes absent. In Aipyornis the mandibular condyles resemble in form and disposition those of Struthio, and in some respects Dinornis. The form of the orbital process seems to have differed from all other genera in its great breadth. Its free end is missing, but from the sharp edge of the proximal end it is probable that it articulated with the pterygoid as in Struthio. 'The pneumatic foramen is mesial. The specimen from which this description is taken was kindly lent me by Dr. Forsyth Major. In Rhea the quadrate resembles that of Dromeus rather closely. The external mandibular condyle is broken up into two distinct articular facets. The inner is similarly subdivided but not so distinctly ; of these two last the outer is vertical and looks forwards and outwards, the inner ventral. The otic process is very short. Its articular head has its long axis at right angles to the long axis of the skull. The posterior pneumatic foramen lies immediately below the articular condyle. The orbital process is very short, blunt, and very wide, it does not project beyond the level of its articulation with the quadrate. In Crypturi the form of the quadrate is quite peculiar in the development of a distinct cingulum projecting backwards above the external mandibular condyle. The otic process is bent backwards at a sharp angle, and forms with the orbital process a Y-shaped bone. The external form of the orbital process resembles that of Struth/o but is more slender; the method of articulation with the pterygoid closely resembles 194 MR. W. P. PYCRAFT ON THE MORPHOLOGY AND that of Dinornis. The form of the outer and inner mandibular condyles is not unlike that of Dinornis. The inner condyle is, however, more nearly spherical than in any of the other forms herein described. Apteryx possesses a readily distinguishable quadrate. ‘The height of its vertical axis is less than that of a line passing backwards through the orbital process, which is very long. The articular condyles for the mandible present the same ground type as in Dromeus and Casuarius, but the external condyle presents a tendency to split into two facets, the region nearest the quadrato-jugal bar being scooped out. The zygomatic process of the squamosal is closely applied to the outer side of the quadrate. In Struthio and Rhea, it should be mentioned, there is a considerable space between these two points. In all the Palwognathe the quadrate is peculiar in that the pterygoidal articulation extends upwards on to the orbital process for more than half of its length. In Apteryx the articular surface for the pterygoid is rather more sharply defined than in the other Palwognathw. Arising at the base of the internal mandibular condyle, it extends outwards on to the inferior border of the base of the orbital process, taking the form of a deep hollow into which the pterygoid closely fits. The otic process has two fairly distinct articular facets. Superiorly transversely elongated, its mesial extremity is expanded into an almost circular head and articulates with the pro-6tic. Its external extremity runs obliquely down to the shaft towards the middle line, and affords an articular surface for the alisphenoid and squamosal. The pneumatic posterior aperture is sometimes absent. The separation into pro-6tic and squamosal heads in the other Palwognathe is only very feebly indicated. The exoccipital does not contribute towards the articular surface for the quadrate in Apteryx. The Mandible. (Pl. XLIIL. fig. 9.) The mandible in Casuarius is long, slender, and sigmoidally curved. Posteriorly, it is expanded to afford an articular surface for the quadrate. This is moderately deep, and with sloping sides meeting at a sharp angle at the bottom of the fossa. Internally, it is produced into a blunt ¢nternal angular process. The surface for the external condyle forms a long narrow facet running parallel with the external border of the cup ; that for the inner forms a band-shaped facet, hollow from within outwards. Immediately without the inner end of this internal facet is a pneumatic foramen. ‘The posterior angular process is only feebly developed. In Dromeus the internal and external angular processes are larger than in Casuarius. The rami are almost straight, and the symphysis is very broad. In other respects the jaw agrees with that of Casuarius. In Dinornithide, e. g. Dinornis, the articular PHYLOGENY OF THE PALZOGNATHA AND NEOGNATHA., 195 surface for the external condyle of the quadrate is cup-shaped. The internal and external angular processes are feebly developed. In Rhea the form and disposition of the articular surfaces agree with those of Dromeus. The internal angular process is large. The symphysis is very broad, The ramus bears a small, perforate posterior lateral vacuity. Struthio in the form of its mandible rather closely resembles Dromeus. It differs, however, in its greater relative slenderness, in that it is sigmoidally curved, in the greater relative size of the expanded posterior portion, and in the elongated form of the articular facet for the external condyle of the quadrate. Aipyornis also resembles Dromeus. The rami are, however, more laterally com- pressed, have an elongated anterior lateral vacuity, a broad dentary border, and the articular facet for the internal condyle of the quadrate situated at the bottom of a deep cup. The internal angular process is well developed and bears a large pneumatic foramen. In Crypturi the mandible resembles that of Rhea. The posterior angular process is, however, longer in Tinamous, and the posterior lateral vacuity appears to be wanting. Aptery« differs from all the other forms in the great length of the symphysis, which is about half the length of the ramus. The internal angular process is well developed. The coracoid in all the Paleognathine forms is small and rod-shaped. Sometimes, as in Struthio, it is reduced to a mere vestige. The dentary suture is always distinct. The Hyoid. Except in Apteryx, only the cerato-branchials appear to ossify in the Paleognathine forms. Epibranchials and the true hyoid elements appear to remain permanently cartilaginous. In Apteryx the basihyal appears to ossify at least occasionally. Tue SKULL or THE NESTLING. The Cartilage-bones. The youngest of the Palzognathine skulls in my possession is that of an embryo Dromeus. The cartilage-bones are in all the stages herein described completely ossified. The basi-occipital, in the young Dromeus mentioned above, is perfectly distinguishable from the neighbouring bones (Pl. XLIII. fig. 5). It is bounded antero-laterally, and in front, by the basi-temporal plate of the para- sphenoid, and postero-laterally by the exoccipital. It is produced caudad to form the median portion of the occipital condyle. It is interesting to note, however, that it forms only the ventral part of this. It is entirely shut off from the occipital foramen 196 MR. W. P. PYCRAFT ON THE MORPHOLOGY AND by the exoccipital portions of this condyle, which meet in the median line above it. Seen dorsally, the basi-occipital is cut off from the pro-étic by a broad patch of cartilage. Anteriorly it only just touches the basisphenoid. In the youngest Cassowary and Rhea skulls in the Museum collection, the sutures of the basi-occipital have almost closed. It takes, however, a slightly larger share in the formation of the occipital condyle, inasmuch as it takes a small share in the formation of its free border. In a half-grown Apterya australis mantelli it agreed very closely with that of Dromeus, and, as in this form, the exoccipitals meet in the middle line above the basi- occipital. The exoccipital, externally, is relatively large. The upper portion of its posterior (inner) border bounds the infero-lateral border of the epiotic (fig. 4). The lower region of this border is hollowed to form the lateral boundary of the foramen magnum. The share which it takes in the formation of the occipital condyle is considerable: in Dromeus the right and left plates meeting in the middle line behind the basi-occipital. Its dorso-lateral border is bounded by the parietal and squamosal in Dromeus, Casu- arius, and Apteryx, and by the squamosal only in Rhea-—this last, in Rhea, extending backwards to join the supra-occipital. The lower portion of the antero-lateral border is free, and in the shape of a large paroccipital process bounds the tympanic recess posteriorly. Its internal dorsal angle lies within the tympanic cavity, and helps to form the inferior pro-6tic articulation for the quadrate. Internally, the exoccipital is almost entirely concealed by the opis- and pro-dtics. The supra-oceipital, in Dromewus and Rhea, runs upward and forward in the form of a blunt cone; in Apterysx the superior border is gently hollowed. In Dromwus, Casuarius, and Apteryx it is cut off from the squamosal by the exoccipital ; in Rhea it runs up- wards to join the squamosal. That portion of its outer (inferior lateral) border which comes into contact with the exoccipital really belongs to the epiotic. In the Penguin this last is easily distinguishable, but in the forms now under description it can only be made out on a reference to the interior of the skull. The epiotic is not indicated externally in the peculiarly distinct manner seen in the Penguins ; even in the skull of the youngest of the embryo Apteryxes in our collection its only indication is the venous canal which points the boundary line between this and the supra-occipital. Internally, it is fairly distinct in all save Apteryx. In all the skulls of this collection save those of the embryo Apterya and Dromeus, it has fused with the pro-dtic. The pro-otic, even in the youngest Apteryx (Pl. XLIII. fig. 3) and Dromeus (Pl. XLITT. fig. 5) skulls, is not visible externally above the tympanic cavity, being concealed by the squamosal. In this it differs markedly from the young Penguin, in which a large portion of this bone remains uncovered until the bird is quite half-grown (82). PHYLOGENY OF THE PALZOGNATHA AND NEOGNATH#. 197 The pro-6tic completely shuts out the squamosal from the interior of the skull in all but Apteryxr. In this type a large, roughly triangular plate appears immediately above the pro-étic and wedged in between the parietal and alisphenoid. The floccular fossa is deep, and oblong in all but Apterya, the long axis running vertical to that of the skull. In Apterya it is represented by a small but deep and perfectly circular pit. Externally, the pro-6tic appears only within the tympanic cavity and affords two more or less distinctly separated articular surfaces for the quadrate: the one lying slightly above and in front of the other. Distad of the superior glenoid surface lies a similar articular surface furnished by the alisphenoid; caudad of the inferior lies a second facet furnished by the exoccipital. These appear to be most distinctly traceable in Casuarius. Externally, articular surface for the quadrate is completed by the squamosal Thus there are two pro-6tic, one exoccipital, one alisphenoid, and one squamosal articular facet for articulation of the quadrate. The opisthotic is seen as a small oblong nodule of bone wedged in between the pro- otic and the exoccipital, and bounded above by the inferior end of the epiotic. It is not visible externally. The foramen for the vagus lies at its ventral extremity and between it and the exoccipital. The dasisphenoid has in every case fused completely with the underlying basi-temporal plate, and is therefore traceable only from the superior aspect of the skull. It is moderately thick in section, the body of the bone being thickened by a mass of diploé. It is bounded behind by the basi-occipital, laterally by the pro-dtic and alisphenoid. Anteriorly, in Rhea, it is produced forward into a thin vertical plate of bone, to join the cartilaginous presphenoid; in the other forms this vertical plate is much shorter. It serves to form the inner half of the circumference of the foramen for the orbito-nasal nerve (V.), the outer portion being contributed by the alisphenoid. Out of its anterior region is scooped the pituitary fossa. The abducent foramen pierces it on either side near its postero-lateral angle. It forms the floor of the anterior region of the metencephalic fossa. In Apteryx it is almost quadrangular. In Casuarius and Dromeus, and in Rhea, it is expanded laterally to form a pair of wings, and is narrower behind than in front. The alisphenoid is bounded by the parietal and squamosal behind, the orbital plate of the frontal above, and the combined basisphenoid and alisphenoid as rings of the parasphenoid. In Dromeus, Casuarius, and Rhea it bears a share in the formation of the post-orbital process. The alisphenoid lodges the greater part of the mesencephalic fossa, and contributes a share towards the formation of the anterior wall of the cerebral fossa. The orbitosphenoid is represented only by cartilage and is much reduced in size. The presphenoid is represented by a thin vertical plate of cartilage resting upon the VoL. xv.—Part v. No. 12.— December, 1900. 2 & 198 MR. W. P. PYCRAFT ON THE MORPHOLOGY AND parasphenoidal rostrum and wedged in between mesethmoid in front and the anterior, laterally compressed vertical plate of the basisphenoid just described. It extends upward and backward to the orbital plate of the frontal. From this region it sends outwards and backwards a pair of cartilaginous wings, which represent the orbito- sphenoids. The mesethmoid, in all but Apteryx, in common with the presphenoid region, forms a conspicuous interorbital septum. In the nestling Dromwus, Caswarius, Rhea, and VOstrich, and in the adult Tinamou, this is pierced by a large interorbital fenestra. This fenestra appears to be excavated partly out of the large, backwardly directed plate which unites with the orbital plate of the frontal, and partly (caudad) out of the anterior border of the presphenoid. In the nestling, the antero-ventral border is more or less linguiform, and terminates at the level of the anterior border of the lachrymal. From this, there runs forward a narrow, triangular, cartilaginous septum nasi. In the adult Casuarius, Dromeus, Dinornithide, and 'Tinamous, ossification extends forwards almost to the end of the rostrum, and terminates in a vertical truncate border, beyond which there is a small cartilaginous septum nasi. In Rhea and Séruthio the ossification extends quite to the free end of the rostrum, 7. ¢. relatively further forwards, and terminates in a sharp point instead of a vertical truncated border. In these, the cartilaginous septum nasi is wanting or greatly reduced. The dorsal border of the mesethmoid, in Dromeus, Caswarius, and Dinornithide, is of considerable thickness and very slightly convex ; it forms a complete floor to the olfactory fossa, extending outwards on either side to the orbital plate of the frontal (Pl. XLIV. fig. 4). In this particular, as will be seen later, it differs conspicuously from Apteryx, wherein this region forms a thin median partition between two extensive cavities. At the anterior end of the olfactory fossa, both in Dromeus and Casuarius, it sends upwards, and then backwards, a strong median pillar deeply grooved on either side for the passage of the olfactory nerves. This corresponds to the tegmen cranii described in Apteryx by 'T. J. Parker, and referred by Kitchen Parker to the tegmen cranii of the lower vertebrates. Thus the evista galli,in these two forms, is furnished entirely by this tegmen cranii, the “ tegminal process ” of Jeffery Parker. At the apex of this tegminal process, as has been well described by Parker, the mesethmoid reaches the outer surface of the skull; here it throws out lateral horizontal ecto-ethmoidal plates—the aliethmoids, as Jeffery Parker proposed to call them. In Dromeus the upper surface of these plates forms a lozenge-shaped area, bounded by the frontals behind, the nasals laterally, and the nasal process of the premaxilla in front. The free edges of these right and left aliethmoidal plates in the nestlings turn downwards and inwards (Pl. XLIV. fig. 4), and in the adult become further extended and intricately folded to form the lower region of the aliethmoid and the antorbital plates. In the nestling Dromeus it should have been remarked that the median vertical portion of the mesethmoid is pierced by a large vacuity, the cranio-facial fissure. PHYLOGENY OF THE PALZOGNATHZ AND NEOGNATHA. 18) Casuarius differs conspicuously from Dromeus and the other forms associated here- with in that, that portion of the mesethmoid which in Dromeus, for instance, reaches the outer surface of the skull to form the lozenge-shaped plate already described, is surmounted by an independently ossifying mass forming the centre of the characteristic casque of the adult—the lateral portions of the casque being furnished by the inflation of the nasals and that portion of the frontal bone which comes into juxtaposition with the nasal and median masses (Pl. XLIV. figs. 8 & 4.a—b). The median portion often extends so as to cover the parietal. In the adult this casque is seen to be built up of a homogeneous mass of cancellated tissue of great delicacy, and protected in the living bird by a horny sheath. The form and development of this casque varies considerably, and is one of the factors employed in the determination of species. The exact homology of this casque has yet to be made out. That the lateral portions are simply inflations of the nasal and frontal bones there can be no doubt, but the origin of the median portion is not so certain. It appears to be an independent structure superimposed upon the mesethmoid, rather than an outgrowth of that element. Stages in the growth and composition of this casque are shown in Pl. XLIV. figs. 3 and 4 a—0. In Rhea and the Tinamous, as in Dromeus and Casuarius, the dorsal border of the mesethmoid comes into close relation with the orbital plate of the frontal, but the relationship is even closer, this border being narrower, so that the crista galli is formed only by the tegminal process. But the passage for the olfactory nerves is restricted to an exceedingly narrow space lying on either side of this process. In Dromeus and Casuarius this passage, though formed in precisely the same way, is much larger. In Apterya the dorsal border of the mesethmoid is exceedingly narrow and knife- like, and slightly concave. It is, moreover, relatively much shorter antero-posteriorly than in Caswarius or even Dromeus. ‘The tegminal process is less well-developed. In the embryo, it appears on the surface of the skull as in Dromeus, and also, as in this form, sends out horizontal plates which form a lozenge-shaped shield between the frontal aud nasal bones. Later in life all trace of this shield is obliterated. The quadrate in the embryo of Dromeus differs in shape from that of the adult in the lack of an ossified orbital process, and in that the external mandibular condyle is <-shaped instead of transversely oblique. In Apteryx that of the embryo appears to differ in nowise from that of the adult. The other skulls in the Museum collection are too advanced to afford any intermediate characters. The articular is a pyramidal nodule of bone, bounded cephalad by the coronoid the base of which conceals its anterior face; externally (in Rhea) by the supra- 2F2 200 MR. W. P. PYCRAFT ON THE MORPHOLOGY AND angular. It furnishes the articular surface for the quadrate. In Dromeus and Apterys, both angular and supra-angular take part in the protection of its external face. The stapes is represented in the dried skull only by the bony base. It does not apparently offer any characters of value from a systematic point of view. The extra- columella has been already admirably described by Gadow, Parker, and others. The Membrane-bones. The parietal, in Casuarius, Dromeus, and ? Struthio, is a transversely oblong plate of bone, the anterior and outer borders of which are produced forward into a small but sharp point which is wedged in between that portion of the frontal which over- hangs the post-orbital process and the process itself. ‘This parietal spur thus helps to form the base of the post-orbital process, the main body of which is furnished by the alisphenoid. In Rhea and ?'Tinamous this antero-external parietal angle is not produced forwards, and fails entirely to reach the postorbital process (Pl. XLITI. fig. 9). Its postero- external angle is overlapped by an upstanding process of the squamosal. Internally the parietal carries on the tentorial ridge from the alisphenoid upwards to its junction with the falx. Behind this ridge it is gently scooped out to complete the roof of the cerebellar fossa, in front it is also hollowed out to form the posterior wall of the cerebral fossa. In Apteryx, as in the other forms, the parietal is irregularly four-sided. Its external border is bounded entirely by the squamosal. In all the other forms, it will be noted, it extends forward beyond this. Internally, it is found to lack the tentorial ridge. Its hinder and external borders pass insensibly the one into the other. ‘The former skirts the supra-occipital and a portion of the pro-6tic beyond, the latter rises gently from the hinder region of the superior border of the pro-6tic—where the hinder border may be said to cease—upwards to skirt the upper’ border of the small triangular area of the squamosal which enters into the formation of the inner wall of the skull. It is entirely cut off from the alisphenoid. In the other forms, e. g. Rhea, Dromeus, the superior borders of the pro-6tic and the alisphenoid form the boundaries ventrad of this bone. The frontal extends backwards so as to form the greater part of the roof of the cerebral fossa. Its hinder border is more or less sinuously curved, and traced from within outwards will be found, in Caswarius and Dromeus, to run transversely from the middle line to the base of the post-orbital process, from which, however, it is actually cut off by a very slender spur from the parietal. In Rhea, on reaching the antero- external angle of the parietal, it turns abruptly upwards and outwards, so as to rest upon the base of the post-orbital process itself. Externally, the frontal is bent in upon itself to form a deep orbital plate which articulates caudad with the alisphenoid and cephalad with the mesethmoid, eventually terminating in a long spike, in 2hea, resting upon the mesethmoid, and overlapped by a greatly elongated posterior extension of the PHYLOGENY OF THE PALZOGNATHA AND NEOGNATH A. 201 nasals and the free end of the inner spur of the lachrymal. In Dromeus the relations are similar, but the nasal spur is shorter. In Casuarius the forward extension of the frontal is curtailed, since it terminates with the mesethmoidal region of the orbital plate, at the level of the origin of the antorbital plate. The greater part of its mesial border is much thickened and inflated ; this region being involved in the formation of the casque, and representing its hindmost postero-lateral boundary. In some species this inflation of the bone extends backwards as far as the parietals. Furthermore, it is significant to note the fact that the frontals fail to meet in the middle line anteriorly, where they under-floor the median ossification of the casque (Pl. XLIV. fig. 4). Thus a portion of the ventral border of the casque comes to take part in the formation of the roof of the cerebral fossa. It would seem, from this, as though this secondary ossification was gradually engrafting itself more and more completely upon the skull, by absorption of the underlying bone in much the same way as bone has replaced cartilage in other cases in the cranial skeleton. The extreme anterior end of the mesial border of the frontal is gently hollowed out so as to form the external lateral boundary of the superficial lozenge-shaped plate of the mesethmoid. In Apteryx the frontal presents one or two points of difference from that of the remaining forms associated herewith. Its mesial border joins its fellow of the opposite side, and its hinder border is bounded by the parietal, as usual. Its postero-ventral or lateral border joins the aliethmoid. Immediately in front of this, the frontal sends downwards and inwards a narrow plate of bone which, eventually, meets its fellow of the opposite side in the middle line, immediately behind the crista galli. From the outside, this orbital plate of the frontal gives off a short slender plate to the aliethmoid. Cephalad of this orbital plate is a deep notch, the swpra-orbital fontanelle. This fontanelle is bounded anteriorly by a narrow bar of bone, called the descending process of the frontal, which at its ventral extremity joins the aliethmoid. Mesiad of this bar—between its upper and the outer wall of the aliethmoid—the orbito-nasal nerve passes from the orbit to the nasal cavity. The sgwamosal, in all the forms herein dealt with, entirely conceals the pro- and opisthotics. In the embryo Dromeus, which forms the subject of this paper, it is almost quadrangular, its antero-ventral angle is produced into a blunt process which extends downwards to the level of the quadrato-jugal articulation. ‘This downward process of the squamosal is thickened on its inner surface and so closely applied to the quadrate as to render it immovable. In the embryo skulls of Dromeus, the squamosal and the parietal are connected by means of a close suture, the juxtaposition of the two elements forming a perfectly level surface. Posteriorly the squamosal rests upon the outer border of the exoccipital, and later fuses therewith. Similarly the squamosal, anteriorly, rests upon the ali- 202 MR. W. P. PYCRAFT ON THE MORPHOLOGY AND sphenoid, anchylosis therewith taking place later. ‘Thus it is that the pro-6tic comes to be entirely concealed. Tn all this, Dromeus and Casuarius seem to be less primitive than the Penguin. Ina skull of this group recently described and figured by me [82] the squamosal was attached to the dried skull simply by its articulation with the parietal. Its hinder border was deeply emarginate and free, exposing the pro-6tic. Its anterior border was also free, not extending to the alisphenoid, but exposing between itself and this last a small portion also of the pro-dtic. There is one great point of difference, however, between the squamosal of Dromeus and that of the Penguin. In the former, its antero-ventral angle is produced downwards into a large “‘ zygomatic process,” to embrace the quadrate; in the latter, this spine is replaced by two small spines, as in all the Neognathe. Except in Apteryx the squamosal takes no part in the formation of the inner wall of the skull. In Apteryx the squamosal has entered into a slightly closer relation with the alisphenoid and exoccipital than in the forms described above, resembling in this respect the Neognathe. ‘Thus the anterior border is concave and fits the corresponding convex border of the alisphenoid; whilst the posterior border, though separated by a narrow line of cartilage from the exoccipital, yet gives signs of effecting a close union therewith immediately. The squamosal appears on the inner surface of the skull as a small triangular plate above the pro-6tic. The nasal in Dromeus is holorhinal. The maxillary process is greatly reduced, especially so in the adult, where it exists only as a vestige. It is produced caudad into a long slender process resting on the frontal, and projecting beyond the lozenge-shaped plate of the mesethmoid, to which it binds externally. Its premaxillary or internal process is of considerable length, extending forwards almost the whole length of the nasal process of the premaxilla, which protects them from above. Dinornis is holorhinal. A skull of Megalaptery«, in the collection of the Hon. Walter Rothschild, retains sufficient of the traces of sutures in this region to indicate the form of the nasal in the Dinornithide. In this skull, it is seen to be relatively much shorter than in Dromeus. Its backward extension does not pass the level of the horizontal mesethmoid plate and is not so pointed as in Dromeus. Its maxillary process is wanting. ‘The ossified aliethmoidal plate projects beyond its lateral border. In Casuarius the nasal is schizorhinal. ‘The maxillary process is vestigial, and the premaxillary process long, as in Dromeus. It is peculiar in that the body of the bone is greatly inflated and takes part in the formation of the casque as already described PHYLOGENY OF THE PALZOGNATHA AND NEOGNATH 2. 203 In Rhea the nasal must be regarded as holorhinal, though the maxillary process is absent. Inasmuch as this process, when present, serves as the boundary, caudad, of the anterior nares, its place must be regarded as being filled by a vertical spike of bone arising from the maxillo-palatine process to the anterior border of the lachrymal. In Struthio this vertical spike—in Rhea separating the anterior nares from the lachrymo- nasal fossa—is represented by a ligament which is attached dorsad to the free end of the maxillary process of the nasal. In Struthio the nasal is holorhinal (Pl. XLII. fig. 3). Its maxillary process is well developed, and continued downwards on the maxillo-palatine by ligament. The Zinamide may be regarded as holorhinal, though the nasal cleft has extended backwards so far as to lie level with the free end of the nasal process of the premaxilla. The maxillary process is not wanting asin Rhea, though reduced to a mere vestige. The nasal cleft is bounded externally by a vertical, bony rod, representing the maxillary spike found in Rhea. It differs therefrom in being rod-like, very slender, and is perfectly free. Above, it articulates, by means of a short inturned process, with the anterior border of the lachrymal and the vestigial maxillary nasal process, whilst ventrally it articulates with the extreme postero-external border of the maxiilo-palatine process where it joins with the rod-shaped portion of the maxilla. This vertical rod was regarded by Kitchen Parker as the maxillary process of the nasal. It is, however, in all the skulls I have examined, quite distinct from this. The lachrymal, in Dromeus, sends backwards a very long, curved, rod-like process to overhang, and thereby greatly increase the size of the orbit (Pl. XLII. fig. 2). It sends downwards a large antero-posteriorly flattened limb, the outer border of which is produced still further downwards into a short rod-like spine which articulates with the quadrato-jugal bar. The body of this ventral process of the lachrymal is perforated by a large foramen for the passage of the lachrymal duct. The inner border of this ventral process articulates with the antorbital plate. In Casuarius the lachrymal differs from that of Dromeus in the greater relative shortness of its external orbital process, which is scarcely longer than the internal. A further point of difference lies in the fact that the external orbital process passes insensibly forwards and downwards into a laterally compressed plate, which, after forming the boundary caudad of the posterior narial aperture, turns sharply backwards, and becomes the antero-posteriorly compressed ventral process bounding the lachrymo- nasai fossa as described in Dromeus. From this it differs in that the laminated portion is continued downwards to the level of the quadrato-jugal bar, instead of stopping short of this, and sending downwards thereto a spine from its outer border. The foramen for the lachrymal duct is smaller than in Dromeus. Internally it articulates with the antorbital plate. In Rhea the lachrymal resembles that of Dromeus. ‘The supra-orbital process is large. The lachrymo-nasal process is also antero-posteriorly compressed, ‘The foramen 204 MR. W. P. PYCRAFT ON THE MORPHOLOGY AND for the lachrymal duct is very large. This lachrymo-nasal process stops short of the quadrato-jugal bar and is connected therewith by a large subquadrate ossiculum lachrymo-palatinum. Internally the last eventually fuses with the antorbital plate. In Struthio the lachrymal resembles that of Rhea. Its supra-orbital process is, however, relatively shorter and thicker. In two nearly adult skulls (one of which is from the collection of the Hon. Walter Rothschild) this orbital process supports one end of a chain of irregularly shaped ossicles which is continued backwards along the frontal as far as the post-orbital process. Normally this chain of ossicles is so completely fused as to be indistinguishable from the frontals on the one hand, and the lachrymal on the other. There is a large supra-orbital fenestra bounded in front by the lachrymal and behind by the frontal ; externally by a portion of this chain of supra-orbitals. The lachrymo-nasal process is continued downwards as a sigmoidally curved rod to terminate on the inner side of the quadrato-jugal bar. On the inner side of this extremity of the lachrymal there lies a subcylindrical ossiculum lachrymo-palatinum articulating mesially, in very old specimens, with the antorbital plate. The lachrymal foramen is represented only by a deep notch. In Apteryx the lachrymal is vestigial, and takes the form of a small roughly triangular plate of bone closely applied to the outer surface of the aliethmoid and the maxillary process of the nasal. It is perforated by the lachrymal foramen. The premaailla, maxilla, jugal, quadratojugal, vomer, palatine, and pterygoid differ in no important respects from that of the adults already described (p. 187). The vomer, pterygoid, and palatine of Apterya demand, however, a further examina- tion, as these in the adult are too completely fused with the neighbouring bones to render it possible to make out their boundaries. The pterygoid is bifid cephalad, being divided into two moderately long, pointed and widely separated prongs. The inner prong, which constitutes the main body of the bone, is deeply grooved along its mesial border, producing two shelf-like projections one above the other: these converge near the tip of the prong to form a slender pointed style resting in a groove on the outer border of the posterior extremity of the vomer as in other Palwognathe, and as in the hemipterygoid of Neognathe, e. g. Impennes, Colymbi. The outer limb is closely applied to the ventral and external aspect of the extreme hinder end of the maxillo-palatine process, extending as far forwards as the point from which the maxillary portion of the quadrato-jugal bar is given off. In the cleft between the outer and inner limbs the palatine is interposed, as will be seen presently. The palatine is short and broad, with a slightly bifid free extremity. It can only be seen in its entirety from above. It is closely applied, throughout its entire length, to a backwardly projecting plate from the mavxillo-palatine process. This it entirely conceals from above. Its proximal extremity sends off, upward and inward, a flange- like plate to articulate with, and underlie, the extreme hinder end of the vomer. Thus this last has its right and left limbs sandwiched, on either side, between the PHYLOGENY OF THE PALZOGNATH® AND NEOGNATHA, 205 pterygoid above and the palatine below. A portion of this flange-like plate of the palatine helps to fill up the cleft between the pterygoid fork already described. As mentioned above, the palatine rests upon a backwardly directed plate from the maxillo- palatine process, the whole of which it conceals from the dorsal aspect. Ventrally, this process is found to run along, and below, the external palatine border, and to terminate near its extreme posterior end. ‘The forward extension of the palatine on the ventral surface corresponds with the level of the tip of the vomer. This extension is derived from the ventral surface-layers of the main body, the palatine terminating abruptly on the dorsal aspect in a V-shaped notch at the level of the middle of the vomer. ‘This ventral process rests in a groove along the mesial border of the maxillo- palatine process. The relations of the vomer have already been more or less indicated. Suffice it to say, that the free end of each posterior limb is wedged in between the spine-like extremity of the pterygoid above and the flange-like inward projection from the dorsal surface of the palatine below. The palate of Apteryx will be found, on a close comparison with that of the remaining Paleognathe, to have departed trom the typical Struthious form in the direction of that taken by Rhea, the relations between palatine and vomer being similar in both. The torm of the pterygoid differs from that of all other birds, in that it is bifid, or, in that it sends off a separate limb whose mesial border passes downwards to underlie the external border of the palatine. ‘This external, palatine limb is elsewhere unrepre- sented. ‘The internal, vomerine limb may be regarded as a flattened plate folded upon itself, the edge being inwards. ‘The superior edge, or upper lip, must be regarded as representing the mesial border of an originally flattened pterygoid such as that of Dromeus. The dentary is of great length. In Dromeus, Casuarius, Rhea, Apteryx, and ‘Tinamous it extends backwards to within one-fourth of the whole length of the ramus. It is relatively less in Struthio. Caudad each ramus divides into dorsal and ventral limbs, of which the ventral is the larger. In Rhea the dorsal limb is paired. The splenial is of great length in all, extending forwards to the symphysis of the jaw, and backwards to within a short distance of the level of the glenoid cavity for the quadrate. It is rod-like in Struthio, laterally compressed in the remaining forms. The angular, in all save Apterya, articulates, proximally, with the supra-angular, being embedded caudad in its external border; distally, it twists downwards so as to form the ventral border of the ramus. In Casuarius and Dromeus its extreme posterior and inner border comes into contact with the articular. In Rhea and Struthio, it falls considerably short of this. In Apteryx, the angular articulates proximally with the ventral border of the articular, which is hollowed out to receive it. The supra-angular rests upon its dorsal border. vou. xv.—Part vy. No. 13.—December, 1900. 26 206 MR. W. P. PYCRAFT ON THE MORPHOLOGY AND The supra-angular, in all, articulates proximally with the articular. Its dorsal border is received between the right and left moieties of the superior limb of the dentary. Its ventral border descends caudad to the level of the inferior border of the angular; distad it rises gradually so as eventually to cross the dorsal border of the same. The coronary is in all a short, slender, and rod-like bone. In many Neognathe it is often of considerable size and laterally compressed. In all the Palwognathe it appears to extend backwards so as to embrace the inner border of the articular and to form the internal boundary of the articular surface for the quadrate. In Rhea and Apterys it is very feebly developed distad, and may be described as vestigial. In Casuarius and Dromeus it is much larger. Struthio is intermediate. In no case does it extend forwards beyond the posterior third of the ramus. Such characters of the skull as are of use for systematic purposes will be found in the ‘ Keys” at the end of this paper. More material is needed before these can be made complete, since a “ Key” to the species has not been possible in the case of Caswarius. Whether the characters considered to be specific in the case of the other genera will stand the test of time remains to be seen. In several instances I have been obliged to rely upon the characters of a single skull, as in the cases of the skulls of Khea macrorhyncha, for instance. I feel justified in reviving Dr. Sclater’s Struthio meridionalis [go| on account of the many marked peculiarities which the skull and the trunk of this skeleton presents. The work of distinguishing the skulls of the four species of Apteryx has been a very difficult one. Indeed, as Mr. Beddard has already remarked [8], “the very slightest differences, apart from those of size, distinguish the Apteryges.” Some Pornts oF SIGNIFICANCE IN THE StRUTHIOUS PALATE, The following points appear to be worthy of special consideration and may fitly be discussed here. In comparison with the palatines of the Neognathe, it will be remarked that in Rhea, as in the other Palewognathe, the palatines are widely separated one from another by the laterally expanded feet of the vomer. In the Neognathe the palatines meet in the middle line, thrusting the vomer forwards, so that it now appears to articulate not with the pterygoid but with the palatines. In hea and Crypturi we have probably a clue to the way in which this has come about. ‘Turning once again to the skull of Dromeus, we notice that the palatines are only loosely attached to the outer border of the vomerine feet, whilst in Rhea (Pl. XLII. fig. 5) the palatines have moved inwards so as to overlap these feet and approach one another mesially. Their PHYLOGENY OF THE PALZOGNATHA AND NEOGNATH A. 207 relation to the pterygoid has also changed, for they no longer are connected with its outer border but by this same inward movement have come to underlie it. In the Neognathe, the inward movement has attained its maximum, the palatines meeting one another mesially, as we have just remarked, thrusting the vomer forwards in so doing. As a matter of fact, however, as we have already shown in earlier papers [82], this is not altogether the case. In the young bird the pterygoid extends forwards in a spike-like form, much as in Rhea, so as to articulate with the vomer, though but by the slightest contact. ‘The palatines have succeeded in moving inwards beneath these anterior pterygoid ends so as to all but entirely sever the original relations between them and the vomer. Later in life (soon after hatching) the severance is complete. At this stage, the anterior end of the pterygoid fractures at a point corresponding with the free end of the palatine. The fracture later becomes a true joint, and the anterior end of the pterygoid resting upon the palatine gradually merges with this bone so as to obliterate all traces of its original existence. Thus the free pterygoid of the Neognathw is a secondary feature, the palato-pterygoid connection in the late embryo not differing materially from that of the Palwognathew. Further, the apparent isolation of the vomer from the pterygoid in the Neognathe is seen to be a coenogenetic character, so that the palate of this group is brought into close relation with that of the Paleognathe. A further point of interest in this comparison between the Palzo- and Neognathine skull is the change which the vomer in the latter has undergone in relation to the parasphenoidal rostrum, a change which indicates a shortening both of vomer and rostrum. The vomer in the Weognathe rarely extends backwards beyond the base of the antorbital plate, in the Palwognathe it may reach nearly as far as the basipterygoid processes. This is an undoubted proof of the shortening of the vomer. That the rostrum has also undergone a considerable shortening is shown by the fact that in the Palwognathw it extends forwards for a very considerable distance beyond the level of the lachrymo-nasal fossa, in the Meognathew it commonly ends in the region of the antorbital plate. : Yet other evidences of shifting and modification of the dromeognathous palate reveal themselves in the Meognathw when we come to closely compare them, and whilst these show how closely the two groups are related they show still more the lower grade of type persistent in the Palwognathe. In the Palwognathe the free ends of the basipterygoid processes articulate with the extreme proximal end of the pterygoid quite close to the articulation with the quadrate. In Neognathe these processes, when present, have shifted forwards on to the rostrum, so as to articulate with the middle of the pterygoid. The forward shifting of the Neognathine vomer, which we have already noticed—a 262 208 MR. W. P. PYCRAFT ON THE MORPHOLOGY AND shifting which brings its proximal end more or less immediately under the antorbital plate—appears to have been accompanied by a similar shifting of the palatines, since these no longer are connected with the main body of the pterygoid bone but with its distal end. ‘This, as we have already shown, ultimately fractures and fuses with the palatine, a joint forming at the line of fracture. This connection of the palatine and pterygoid by means of a joint is a point of great difference between Paleo- and Neognathine skulls. In the Tinamidw we have an intermediate stage between the Palwo- and the Neo- gnathe, The vomer is undoubtedly relatively shorter posteriorly than in Rhea, its free end lying midway between the level of the antorbital plate and the basipterygoid processes. The pterygoid has increased in length, so that the vomer and palatine articulate with its distal extremity only. The quadrato-jugal fossa has also increased in length relatively, extending forwards now beyond the level of the antorbital plate as far as the vestigial maxillo-nasal process. A careful study of these points will greatly facilitate the conviction that the egithognathous and schizognathous skulls are but modifications of the dromeognathous type. The desmognathous is a further modification of the schizognathous palate. The single-headed otic process of the quadrate in the Palwognathe, upon which so much stress has hitherto been laid, appears to have less importance than the points to which attention has just been drawn, for in Apéeryx the otic process is two-headed, as in Neognathe. THe VERTEBRAL CoLuMN. The memoirs of Owen, Mivart, and T. J. Parker on the vertebral column of the struthious skeleton render it quite unnecessary to do more than briefly comment thereon here. a. The Presynsacral Vertebre. All the presynsacral vertebre of the Palwognathew are heterocelous, and all of this region are free except in the Yinamide, in which certain of the thoracic vertebree fuse. The vertebra of Dromcus are the least specialized in type; those of Caswarius are very similar. Both in Dromeus and Casuarius the cervical vertebre are conspicuously shortened antero-posteriorly. In Cuswarius the neural spine, traced from the head backwards, undergoes considerable change of form. Anteriorly it is little more then a median tubercle rising from the centre of a flat neural plate. At about the 6th vertebra it sends backwards a pair of low ridges which terminate in a pair of hyperapophyses. At about the 10th vertebra, the spine has exchanged its <-shape for a transversely, crescentic form. ‘This, for the next 4—5 vertebra, becomes broken across in the middle PHYLOGENY OF THE PALAIOGNATHA AND NEOGNATH A. 209 line to form two high curved ridges sloping downwards and backwards to terminate in metapophyses. At about the 15th vertebra the two ridges again coalesce and form a very high vertical transverse neural spine. From the 18th backwards this gradually becomes more and more laterally compressed, so as to pass insensibly into the typical neural spines of the thoracic vertebra. Every neural spine, from that of the axis backwards, bears a fossa at its base, both anteriorly and posteriorly. The latter is the deeper. The vertebre from the middle of the neck backwards have these fossz of very considerable size. They lodge a ligament. From 6-10 in C. casuarius and C. ¢. australis the diapophysis sends back a bar of bone to the hyperapophysis. The vertebrie are all pneumatic. In the hinder cervicals there is a large pneumatic aperture dorsad of the interzygapophysial ridge. In the thoracic there are several very large ones—one below the transverse process, oue between the transverse process and the postzygapophysis, and one dorsad, lying between the anterior zygapophysis and the base of the neural spine. This last is represented by a deep fossa in Dromeus. The aperture ventrad of the anterior zygapophysis is feebly developed or wanting in Dromeus. The cervical ribs (pleurosteites) and hypapophyses resemble those of Dromeus. In Dromeus the vertebre are less specialized than in Casuarius. The high transversely expanded neural spines are wanting, though the ligamental neural fosse, especially that caudad of the neural spine, are very deep. ‘The pneumatic fossa in, or above, the interzygapophysial ridge is very deep, as also is that lying at the base and in front of the neural spine. The sides of the fosse, moreover, are smooth, not obscured by cancellated tissue as in Casuarius. In Struthio the centra of the vertebree are relatively much longer than in Dromeus or Casuarius. The neural spines of the anterior cervicals are long, low, and rise to form a sharp median ridge. The ligamental fossee are narrow grooves channelled out of this ridge. The posterior cervicals have the neural spines wider and shorter, antero- posteriorly, and they are deeply hollowed for the ligament. The cervical ribs, as in Casuarius and Dromeus, are long, but more slender than in these. As in Dromeus they fuse with a plate of bone depending from the diapophysis and a lateral parapophysial outgrowth from the anterior end of the centrum below the prezygapophysis. The presence of this rib serves to enclose a canal for the vertebral artery. One great point of difference between this region of the vertebra and that in Drom@us and Casuarius lies in the fact that in Struthio the lamella depending from the diapophysis, and with which the rib articulates, is continued backwards along the centrum for a considerable distance, forming an extensive and tunnel-like passage for the artery. The pneumatic apertures are not so conspicuous as in Dromeus. There is no aperture in the interzygapophysial ridge of the anterior ceryicals as in Dromeus. In, 210 MR. W. P. PYCRAFT ON THE MORPHOLOGY AND the hinder cervical and thoracic vertebre there is a large aperture beneath the diapophysis, opening in the roof of the tunnel for the vertebral artery, or in the thoracic vertebra below the prezygapophysis and between the dia- and parapophysial articula- tions of the tuberculum and capitulum. The deep fossa at the base of the anterior border of the neural spine of the hinder thoracic vertebre is here yery shallow. In Rhea the vertebree have departed still more from the normal type. The cervicals are very long and cylindrical. Neural spines are present only on the 2nd to 5th vertebre. ‘The free hinder borders of the postzygapophyses of the 6th vertebra are received into a pair of shallow pits on the 7th, lying mesio-caudad of the prezyga- pophyses. Cervical ribs are only feebly developed. From the 2nd to 11th vertebre they are represented only by short styliform processes; from this vertebra backwards they are represented only by the extreme anterior articular end of the rib which serves to form the band-shaped external wall of the vertebrarterial canal. As in Struthio, the vertebrarterial canal is roofed by a backward extension of the lamelliform plate depending from the diapophysis and serving as an articulation for the cervical rib. This plate is described and figured as the “rudimentary rib” by Mivart [63]: doubtless by a slip. In Apterya the form of the atlas and axis vertebre bears some resemblance to Dromeus. The rest of its vertebra differ conspicuously from those of the forms herewith associated. The neural spines never bifurcate. On the anterior vertebre they form compressed blades from the 8th to the 12th; they are columnar and vertical in form and central in position with regard to the neural lamine. The diapophyses of the 6th and 7th vertebra project outwards and backwards beyond the level of the prezygapophysis; from the 7th to the 13th backwards and down- wards ; from this vertebra caudad they project directly outwards, and merge gradually into the transverse processes of the thoracic vertebra. The neural spines of the thoracic vertebre are equal in size, of great breadth antero-posteriorly, and interlock along the dorsal ridge by anterior and posterior bifurcations as in many Veognathe. There are no pneumatic apertures in either cervical or thoracic vertebra. Cervical ribs are vestigial. In Crypturi the thoracic vertebrae are anchylosed and bear hypapophyses. In the Paleognathe these occur only on the posterior cervicals (cervico-thoracic). The penultimate is free, the ultimate fused with the synsacrum. The interzygapophysial ridge is perforated by pneumatic apertures. In the cervical vertebre pneumatic apertures open into the roof of the vertebrarterial canal. The diapophyses of the cervicals project forwards beyond the prezygapophyses. The neural arch is constricted in the middle, immediately behind the prezygapophyses. The vertebree of Dinornis resemble those of Caswarius. The fosse lying beneath the transverse process of the thoracic vertebra, absent in Dromeus, were much larger relatively than in Caswarius. The cervicals resembled those of some species of PHYLOGENY OF THE PALZOGNATHA AND NEOGNATH A. 211 Casuarius in having a bony bar running from the diapophysis backwards to the hyperapophysis. The thoracic vertebre call for no special remark here. They have already been minutely and carefully described by Mivart. The vertebra known as the cervico-thoracic should probably be regarded as true thoracic vertebre, the ribs of which have lost their sternal segments through the backward shifting of the sternum, This view has already been promulgated by the late T. J. Parker. Writing of Apteryx, he says: “The occasional absence of the cervico-thoracic uncinates, taken in connection with the vestigial condition of the ribs in one case and with the fact that in the same skeleton the first thoracic rib of the right side terminates ventrad in a blunt free end and has no sternal portion, seems to point to an inclusion of anterior thoracic vertebre in the cervical region by atrophy of their ribs.” b. The Synsacral Vertebre. (Pl. XLIV. fig. 5.) In Casuarius there are 8 (eight) pre- and 8 (eight) post-sacral vertebrae, with the two true sacrals numbering eighteen synsacrals. Of these, two are thoracic. ‘The 1st lumbar has large parapophysial processes; the 2nd to 3rd of this series are much smaller; the 4th and 5th become larger and abut, the 4th against the pectineal process and the 5th against the ischium. Behind these follow two lumbo-sacral. The true sacrals are closely crowded and have slender parapophysial processes, which abut against the ilium and ischium immediately behind the acetabulum. Behind these in C. dicarunculatus lie two pseudo-sacrals with large parapophysial processes abutting against the ilium. But for the sacral plexus they might readily be mistaken for true sacrals. They are really the first two caudals, behind which follow six fused caudals with diminishing and backwardly directed transverse processes. In other Cassowaries there is but one pseudo-sacral. ‘The sacral vertebrae do not bear diapophyses. In Dromeus the number of synsacral vertebrae is the same as in Casuwarius. The 6th lumbar, however, does not send out a parapophysis to abut against the pectineal process. The neural spines of the synsacral vertebre are all highly pneumatic, and made up of loose cancellated tissue. ‘They remain separate for some time after hatching, but in the adult are fused into one mass, all traces of separate vertebra being lost. In the young bird the pre-sacral vertebre are just distinguishable between the closely approximated pre-acetabular regions of the ilium, The post-sacral have the neural spines laterally expanded so as to keep the post-acetabular ilium moderately widely separated, so that between each vertebra is a deep loculus (Pl. XLIV. fig. 5). In the adult the whole becomes roofed over by a thin bony plate extending across from the post-acetabular ilium of one side to that of the other. 212 MR. W. P. PYCRAFL ON THE MORPHOLOGY AND Seen dorsally in the half or even three-quarter grown birds the neural spines from above the acetabulum backwards for some considerable distance are found to take the form of vertical transverse plates, forming a series of loculi between the post-ilia. ‘Though further back in the series the transverse plates become columnar, the loculi remain. In the adult these loculi are completely covered in by a thin sheet of bone, formed by the fusion of a series of thin plates capping the neural spines. In the embryo Dromeus (Pl. XLIV. fig. 6) it is worth noting that the lofty transversely expanded neural spines of the sacral and anterior post-sacral vertebrae of the adult are represented only by low cartilaginous ridges, and bounded on either side by the free edge of the dorsal border of the ilium. In the dried skeleton the carti- laginous portion shrinks up, revealing (1) a pair of ossified neuroids, which fail to meet one another in the middle line, and (2) a low osseous ridge marking the foundation of the similar cartilaginous ridges already described. The increase in height of the neural spines is followed by a corresponding increase in depth of the ilium (p. 227). The centra of these synsacral vertebree in this early stage are, in the dried skeleton, represented only by a flat bony plate, so that a single vertebra, with its neural arches, is stirrup-shaped. ‘The thinness of the centrum makes the neural space of a relatively much greater size than in the adult. ‘The nerve-apertures are single and not paired as in the adult. This 1s well shown in Pl. XLIV. figs. 5-6, In Struthio there are ten pre-sacral and eight post-sacral vertebre. The series commences with the last two thoracic vertebre. Then follow seven lumbar vertebre. The parapophysial elements of these are all short and thick. The diapophyses of the lumbar vertebra from the 4th backwards extend upwards on to the inner side of the ilium as forwardly directed spikes. Short and triangular on the 4th, they gradually increase in length up to the 8th, where they form long rod-shaped bars of bone. On the 10th, this process is short and rod-shaped. The lumbo-sacral vertebra are possibly represented by a single vertebra which bears vestigial parapophyses. In one skeleton in the Museum collection the lumbo-sacrals were certainly unrepresented. ‘The last lumbar bore distinct parapophysial elements, which on the left side extended backwards to join the first sacral rib. ‘The sacral vertebree are two in number and have large diapophyses. They lie more caudad than those of Caswarius, resting against the hinder border of the downward ischiadic or post-acetabular process of the ilium, and therefore conspicuously caudad vf the acetabulum. In Caswarius the sacral ribs lie immediately behind the acetabulum. ‘The outer free ends of the sacral ribs give rise to a large swollen process from the dorsal surface, which extends upwards and backwards to reach the diapophysis of the 2nd sacral, sometimes extending forward to that of the Ist. In one specimen in the Museum collection the lumbo-sacral has developed a parapophysial process on the left side. PHYLOGENY OF THE PALZOGNATHA AND NEOGNATHA. ies Sometimes the former fuse with The Ist caudal bears both para- and diapophyses. those of the sacral. Behind this lie seven caudals without parapophyses, before the first free caudal is encountered. The neural spines of these are not transversely expanded as in Casuarius. In Fhea there are ten pre- and five post-sacral vertebrae. The pre-sacral comprise two thoracic, four lumbar, and four lumbo-sacral; the last. of these may develop parapophysial processes. The sacral ribs fuse with the ilium more dorsad than in Struthio, Caswarius, or Dromeus. In the last three they abut more or less extensively against the post-acetabular ascending process of the ischium. In Rhea they rise dorsad, so as to reach the level of the centre of the acetabular are. The Ist sacral is intermediate in position, with regard to the acetabulum, between Casuarius and Struthio. In the latter, it will be remembered, it lies considerably behind the acetabulum, in Caswarius level with its hinder border. The sacral ribs ave very short and almost indistinguishably fused with their diapophyses. In the possession of diapophyses these vertebre differ from those of Casuarius and Dromeus, and resemble Struthio, in which, however, they are much more extensively developed. The first post-sacral, or lst caudal, bears both para- and diapophysial processes, but they are almost indistinguishably fused. The 2nd candal bears diapophyses only. The 3rd to 5th caudals are somewhat remarkable in that they are represented only by atrophied vertebre. The centra of these are reduced to thin delicate rods, supporting slender diapopbyses and feeble neural spines. ‘The 6th vertebra resembles the oth, but has a stronger neural spine. It is the first vertebra to escape from the clasp of the post-acetabular ilium. The 7th has a much stronger neural spine and slightly thicker centrum, but like the 6th has lost its diapophyses. The 8th and 9th have inflated cancellous centra and short neural spines; they fuse by their ventral aspects with the ischia. The 10th is the first free caudal, of which there are eight in all: these have short thick centra and short neural spines. In Apteryx there are nine pre- and five or six post-sacrals, with the true sacrals sixteen in all. The pre-sacrals represent two thoracic, four lumbar, and three lumbo-sacral. The first post-sacral (1st caudal) bears short para- and diapophysial processes. The post-sacral are chiefly remarkable for the great lateral compression which they have undergone. The synsacrum of the Dinornithide differs in many respects from that of all the It resembles in many ways that of Dromeus and Caswarius The pre-sacrals are apparently There is an other Paleognathe. on the one hand, and that of Struthio on the other. eleven in number, viz., two thoracic, six lumbar, and three lumbo-sacral. indication of a fourth lumbo-sacral bearing parapophysial processes, which have almost completely fused with those of the 1st sacral. There are five post-sacrals. VOL. XV.—PART v. No. 14.—December, 1900. 2H 214 MR. W. P. PYCRAFT ON THE MORPHOLOGY AND The pre-sacrals resemble those of Dromeus and Casuarius in that they descend below the level of the ventral border of the pre-ilium, but not greatly so. In Struthio this feature is much more marked, the whole centrum and base of the neural spine being exposed in side view. ‘The parapophysial processes of the first caudal are directed backwards instead of forwards. From this cause the true sacral vertebre are completely isolated. In Struthio, Casuarius, Dromeus, and Apteryx, it will be remembered, the para pophysial process of the first or even first and second caudals may combine with those of the true sacrals to form a sacral mass, the true nature of which is difficult to make out in the adult. The post-sacrals (caudals) resemble those of Struthio rather than any other Ratite in the great development of the diapophysial process, which extends upwards, outwards, and backwards to separate the post-acetabular ium. In Dinornis, however, this separation of the post-acetabular ilia is carried to a greater length than in any other Ratite. The neural spines of the Dinornithide are peculiar in that they combine dorsally to form a broad lozenge-shaped superficial plate of bone which serves as a cover to the huge fossa that otherwise exists between the upwardly directed transverse processes and the post-ilia. The synsacrum of the Crypturi differs from that of all the other forms herein described. Moreover, in the points wherein it differs from these it resembles the Neognathe. One of the most noticeable differences is the fact that the post-sacrals have lost their neura] spines. In all the flightless Palwognathe these are of great length. ‘The diapophyses of the post-sacral region rise from the top of the centrum, at the base of a long, faint, median ridge marking the position of the neural spines in other forms. These diapophyses develop thin plates of bone which fuse one with another to form a lozenge-shaped plate, protecting the inter-diapophysial fosse, as is done by the similar plate from the neural spines in Dinornis. ‘The synsacrum of the Tinamous includes one thoracic, four lumbar, five lumbo-sacral, and five post-sacral. There are no parapophysial processes abutting against the pectineal process as in the flightless Palwognathe. ‘This process, it should be mentioned, has shifted further backward relatively to the sacral vertebra, so that a line drawn across from the base of the anterior concave border of the one would pass immediately in front of the first sacral. The sacral ribs are very long, and articulate with the ilium immediately behind the acetabulum by a long vertical roughened surface. The rib of the second sacral is vestigial or wanting. The free caudal vertebree in the Palwognathe do not exceed ten in number. Some- times the 9th and 10th are more or less fused. In Struthio the 10th is laterally compressed to form a kind of pygostyle. PHYLOGENY OF.THE PALZOGNATHA AND NEOGNATHA, 215 In Rhea there are but seven free caudals. This is due to the great backward extension of the ischia. Counting from the free end of the post-acetabular ilium backwards, as in other forms, there are eleven post-iliac caudals. Riss. The anterior cervical ribs (pleurosteites) have already been described. It is probable that the free cervical ribs (cervico-dorsal, Mivart) and the vertebre to which they articulate really belong to the thoracic series. ‘This is indicated by the fact that these ribs have lost their connection with the sternum, and have lost their sternal ribs through the backward shifting of the sternal plate. In Casuarius and Dromewus there may be three or four free cervicals. In Struthio and Rhea, two. In Apterya one, which bears a large uncinate (in A. oweni there is a shorter and more style-like rib anterior to this and wanting the uncinate). In the Crypturi there are two free cervicals. The thoracic ribs in Casuarius and Dromeus agree with Struthio, and differ from Rhea, in the fewness and slight development of the uncinates. In Casuarius the Ist five pairs of thoracic ribs articulate with the sternum by means of sternal ribs. The 6th and 7th are long, but have no sternal segments. The 7th pair are more or less overlapped by the pre-ilium. The 8th pair are very short and underlie the pre-ilium. The 9th pair are represented by a scale-like vestige repre- senting little more than the tuberculum and capitulum, which in the young bird can still be made out (e. g. 1874 a Mus. Coll.) closely applied to the pre-ilium. In C. bennetti there are but three free cervical ribs and only four pair of thoracic ribs articulating with the sternum by means of sternal segments. ‘The 5th pair have very long sternal segments, but they fail to reach the sternum, We may gather from this, however, that they have not long lost their connection therewith. The uncinates of the ribs of Caswariws are never large. Generally only two or three appear to be present, and these may anchylose with their respective ribs. In a half- grown C salvadorii I find five uncinates. These have been probably lost in maceration in most of the skeletons which I have examined. In Dromeus there are two free cervicals; the second is very long, and almost articulates with the anterior lateral process of the sternum. Following these come four pairs of thoracic ribs connected with the sternum by sternal ribs. These are succeeded by a pair (7th) with very long sternal segments, but which fail to reach the sternum. ‘The next two pairs have no sternal segments, and are overlapped by the pre-ilium. In Struthio there are two free cervicals and five pairs of thoracic ribs articulating with the sternum by means of sternal ribs. The pair next behind these has long sternal segments which do not reach the sternum. Jn a mounted specimen in the collection of S. camelus this 6th sternal segment actually articulates with the sternum. 2H 2 216 MR. W. P. PYCRAFT ON THE MORPHOLOGY AND The following pair (7th thoracic) are long, but without sternal segments; the 8th pair are short and fuse with the pre-ilium. In Rhea there are two free cervicals, three pairs of thoracic articulating with the sternum by sternal ribs, and provided with very long uncinates. ‘The 4th and 5th pairs are long, but have no sternal segments. The 6th pair are very short and fuse with the pre-ilium. ; Amongst Dinornithide there appears to be some variation. In Dinornis maximus there are three free cervicals, two thoracic connected with the sternum by sternal segments, one long free rib, and three free ribs behind this overlapped by the pre-ilium. In Pachyornis there seems to have been but one free cervical or cervico-thoracic, three thoracic articulating with the sternum, four long free ribs, and short free ribs, these last two being overlapped by the ilium. The uncinates were very long. In Apterya the ribs are remarkable for the great breadth and flatness of their vertebral segments. ‘There may be one or two pairs of cervico-thoracic ; there are four thoracic articulating with the sternum (in a skeleton of A. australis I found five on one side, and in a skeleton of A. oweni, both in the Rothschild Coll., I found but three thoracic vertebra connected with the sternum) and four pairs of free ribs behind these, all of which are overlapped by the pre-ilium. ‘The last pair are often exceedingly short. Six pairs of ribs bear uncinates, the series commencing in the last cervico-thoracic. The uncinates are long and broad, extending backwards to the hinder border of the rib next behind. In the ribs of all the Palwognathe, save Apterya and some Tinamous, there is a large pneumatic aperture lying in the cleft between the tuberculum and capitulum. THE STERNUM AND PEcTORAL ARCH. a. The Sternum. As with the other portions of the skeleton so far examined, so with the sternum, Dromeus possesses the most generalized type. The sternum of the young Vromeus and that of the young Caswarius are very similar in appearance; in the adult they are readily distinguishable. In the adult Dromeus the anterior lateral processes are very long, slender, and directed upwards and backwards. There is no spina externa nor interna, but a deep emargination in their place. The ventral lips of the coracoid grooves are well developed. ‘The articular surfaces for the coracoids are short, not exceeding in length the distance between the articulation for the first and third sternal ribs. These grooves differ from those of all the other Palwognathe in that they overlap, causing the bases of the coracoids to cross one another in the middle line as in many Carinate. PHYLOGENY OF THE PALZOGNATHEZ AND NEOGNATH. 217 Fig. 4. A.—Ventral aspect of the sternum of Dromceus nove-hollandie, to show the form and position of the coraciod grooves and anterior lateral processes. B.—Ventral aspect of the sternum of Dromeus ater. Note the degenerate condition of the coracoid grooves and the well-marked protuberantia sterni. C.—Ventral aspect of the sternum of Rhea macrorhyncha for comparison with that of Dromeus. Note the difference in the position of the coracoid grooves, of the articulation for the sternal ribs, the large protuberantia sterni, and the thin places indicated by shading denoting the position of former fenestre or fissures. D,.—Ventral aspect of the sternum of an embryo Apteryx after Parker, to show the possible vestigial keel. a.s. = articulation for sternal rib. c.g. = coracoid groove. a.l.p. = anterior external process. ps. = protuberantia sterni. p.lp. = posterior lateral process. k, = keel. v.c.g. = ventral lip of coracoid groove. 218 MR. W. P. PYCRAFT ON THE MORPHOLOGY AND The sternum of Dromeus ater is remarkable, possessing as it does characters in common with both Casuarius and Rhea, but is at the same time Dromeine. As in D. nove-hollandie, the anterior lateral processes are conspicuously large and have the same form. The outline of the free edge of the corpus sterni from the articulation of the last thoracic rib backwards is the same. ‘The coracoid grooves are obsolete. That of the right side is represented by a portion of the inner angle of the lower lip of the groove. ‘This, as is seen in the figure (4 B), persists in the form of a median asymmetrical tubercle. The obsolescence of these grooves indicates that the pectoral arch and wing were much more degenerate than in Dromeus nove- hollandie or D. ivroratus. The corpus sterni differs from that of D. nove-hollandie in the relatively much greater development of the protuberantia sterni, which resembles that of Rhea, and in the relatively narrower width across the region of the incipient posterior lateral processes. Four ribs articulate with the sternum, in D. nove- hollandie only three. The following measurements in millimetres are of interest :— D. nove-hollandie. D. ater. Greatestjlen thyme amie ee eer 160 95 Greatest width (across ant. lat. proc.) . . 140 90 Least width (across post. lat. proc.). . . 105 55 It should be noted that the metasternum of D. afer bears traces of fracture, indicating a somewhat greater length than the above measurements. ‘The breadth of the sternum nearly equals its length. In Caswarius it is much longer than broad. The articular surfaces for the sternal ribs are divided one from another by pneumatic apertures forming pockets of great depth, extending from without inwards towards the middle line. ‘The hinder border, both in Dromeus and Casuarius, is shield-shaped. No external lateral processes are present. The sternal plate is of great thickness and highly pneumatic. In the young Dromeus the anterior lateral processes are very short, and the ventral margins of the coracoid grooves are almost confluent and not separated by a deep median notch. In Casuarius the sternum is relatively much longer than in Dromeus, the increase in length being most noticeable from the articulation for the last thoracic rib backwards. The coracoid grooves are much longer, and the anterior lateral process is almost obsolete. In the middle line between the ventral lips of the coracoid groove in Casuarius there opens forward a large pneumatic aperture; this runs backwards into the body of the bone for some considerable distance, penetrating a low median ridge or protuberance resembling a vestigial keel. This is wanting in CO. unappendiculatus, a shallow groove taking its place, the sides of which further back become conspicuously raised to form a pair of long low ridges. PHYLOGENY OF THE PALZAOGNATHA AND NEOGNATHA, 219 The sternal plate of Caswarius is of great thickness and highly pneumatic. Such specific differences as I have been able to find will appear in the “ Key.” The sternum of Rhea resembles that of Dromeus and Casuarius in several respects, but is yet very readily distinguishable from either. The coracoid groove, though shallow, is sharply defined. The outer and inner lips, though but feebly developed, are subequal in length. The groove of the left side is separated from its fellow of the right by a smooth-lipped and gently hollowed plate of bone representing the anterior border of the sternum, equal in length to the groove itself. ‘The anterior lateral processes are large. The articular surfaces for the ribs are closely crowded, and follow immediately behind the anterior lateral process. Thus the distance from behind the articular surface for the last rib to the hinder border is more than twice the distance of the space separating the coracoid grooves; or, in other words, the articular surfaces of the ribs occupy only one-sixth of the whole lateral sternal border, in Caswarius it occupies about one-third. The body of the sternum is notched posteriorly and tapers from before backwards. It presents a deep basin-like hollow dorsad and is produced ventrad into a long, swollen, ridge-like prominence. ‘The corpus sterni is much thinner than in Dromeus or Casuarius. The articular surfaces for the sternal ribs are each more or less distinctly divided into two, the outer being the larger. The pocket-like interarticular spaces (incisure intercostales) are of great depth. Large pneumatic apertures open below the base of the anterior lateral process. For further remarks on the sternum of /hea, see p. 224. In Struthio the sternum is roughly about as broad as long. It resembles that of Dromeus in some respects and hea in others. The form and position of its coracoid grooves resemble that of Dromeus. ‘The ventral lip, however, is wider than the dorsal. Both dorsal and ventral lips converge mesiad. The right and left grooves are separated by a slight hollowing out of the anterior sternal border. The anterior lateral processes are moderately well developed and project outward and forward. The articular surfaces for the sternal ribs are of great width and occupy nearly the whole lateral border of the sternum, thus differing from both Dromeus and Rhea. The pocket-like pneumatic incisure intercostales are relatively less deep than in the foregoing forms. The median portion of the bar serving for the articulation of the ribs has lost its glenoid surface, which is preserved only at either end. In this point Struthio and Rhea agree. There are large posterior lateral and in some sterna incipient median lateral processes. The ventral prominence of the sternum lies beliind the centre of the sternal plate, and is broader and more flattened than in Ahea; in which, by the way, it lies in the middle of the sternal plate. 220 ; MR. W. P. PYCRAFT ON THE MORPHOLOGY AND Apteryx, in the form and position of its coracoid grooves, resembles Rhea more closely than any other group. Like Rhea they are widely separated by a deeply concave median portion of the anterior border of the sternal plate. The groove lies at the base of the anterior lateral process. It is very short, and has a conspicuously projecting lower lip. The sternal plate differs from that of all the foregoing in being non-pneumatic. Posterior lateral processes are well marked, and Its breadth exceeds its length. The articular divided one from another and a median process by a pair of notches. surfaces for the ribs, as in Struthio, extend along the greater part of the lateral sternal wall. Occasionally, ¢.g.,in an A. australis described by Owen, the sternum was fensterated. Beddard describes a pair in A. australis mantelli. In another of this species were a pair “ of pin-hole size.” Parker directs attention to a distinct keel on the sterna of A. australis mantelli. As Beddard remarks, though not easy to see, this is readily felt. He found traces in A. haastii, and I think to these we might add A. oweni. | * 6 | 8 2B 2 7 Casuarius unappendiculatus, 9.1... | a | 10 43 5 6 casuarius beccarit, Q «1... 4 8 | 1 0 42 55 POELCHOS, Ci b0000000000008 a0) | 10 4 o bicurunculatus ........-+..| a. OU | 0 7 3 UDOIGHON G+ piaamanenooeolos | 3 103 113 32 33 | Dromeeus nove-hollandia ............ 10 «6 0) 5 | Apterya australis 20.004. esses esses AA 47 7 a OW CRIS WII a TL oe 3. 4} 43 7 | Tue Liver. The liver, in all the Palwognathe, save Struthio, agrees with that of the Neognathe in that it is divided into two lobes. In Casuarius and Dromeus the right lobe is larger than the left. A Spigelian lobe is represented. Both Casuarius and Dromeus possess a gall-bladder. In the former the gall opens, in common with the pancreatic ducts, into a diverticulum on the distal loop of the duodenum. In Apteryx the left liver-lobe is larger than the right; the gall-bladder is occasion- ally absent. In Rhea the right and left liver-lobes are equal in size. ‘The gall-bladder is generally wanting. 252 MR. W. P. PYCRAFT ON THE MORPHOLOGY AND Struthio is remarkable in that, according to Beddard, the right and left lobes are fused into a single heart-shaped lobe. There is a Spigelian lobe. The gall-bladder is wanting; and there is but a single bile-duct, which opens near the pylorus. Mr. Beddard found the free edge of the liver in Rhynchotus rufescens pierced by three small vessels of the venal portal system—a fact which, as he points out, has a very lizard-like appearance. RESPIRATORY ORGANS. a. THE Lunes (Pl. XLV. fig. 9). The costo-pulmonary muscles in Caswarius are short, thick, and powerful, five pairs in all; they arise from the thoraco-sternal articulation of the ribs and extended fanwise on to the aponeurosis covering the lung. Those of Rhea are relatively longer than in Casuarius. In Apteryx, according to Huxley [41], they form “ broad flat bands,” which ‘take their origin from the vertebral ribs, at some distance below the attachment of the pulmonary aponeurosis, and, proceeding obliquely dorsad and forwards, spread out and are inserted into that part of the aponeurosis which covers the posterior inferior facet. In the relative size and distribution of the ecto-, ento-, and para-bronchia, and in the size of the vestibule, the flightless members of the Palwognathe do not appear to differ much one from another or from the Neognathe. In the lungs of a Casuarius unappendiculatus the bronchial rings extended backwards as far as the 4th ento-bronchium. ‘The inner border of the anterior and posterior inferior facets of the lung are deeply sculptured by a system of numerous parallel grooves of open tubes, closely packed and running along the surface of the lung from the mesial border outwards so as to converge towards, and open into, a spacious chamber communicating posteriorly and ventrally with the Ist ento-bronchium (Pl. XLV. fig. 9), further forwards and also ventrally into the anterior end of the pre-bronchial ostium. These grooves give the ventral face of the lung, after the removal of the aponeurosis, the appearance of having been bored by 'Teredos, the groove representing a hemisection. ‘They are crossed and recrossed by transverse strands of tissue and punctured by minute apertures. They may be divided into two groups according to whether they are fed by a trunk from the Ist or the 3rd ento-bronchium. Thus, as will be seen in fig. 9, all the secondary grooves cephalad of the main trunk marked “ent. 1.” are fed from this, whilst ento-bronchium 11. supplies all the grooves caudad thereof. These grooves are converted into tubes by the pulmonary aponeurosis. b. Tur Arr-sacs (Pl. XLV. fig. 8). In Casuarius the pre-bronchial sac is paired, and extends forward on each side of PHYLOGENY OF THE PALZOGNATHE AND NEOGNATHA. 253 the cesophagus for a considerable distance beyond the level of the Ist cervico- thoracic rib. The sub-bronchial sac is divided into right and left loculi, enclosing a median tunnel-like space between them, the roof of which is formed by the trachea. This tunnel can be traced into a large pneumatic cavity in the sternum Opening on its anterior border in the middle line between the coracoid grooves. The intermediate sacs were three in number. A similar subdivision has been de- scribed by Beddard in Podargus and “one or two Accipitres.” In the present instance it would seem that this third sac is due to a subdivision of the anterior intermediate. The posterior intermediate was large. Its roof was pressed inwards by the forward extension of the posterior abdominal. The posterior abdominal was of considerable size. That of the right side enclosed the right kidney; on the left side the anterior lobe of the kidney was found to lie without the sac and to be enclosed in a separate sac. The intermediate and posterior abdominal sacs of the right side were markedly smaller than those of the left. The sacs are fed as follows:—The pre-bronchial and sub-bronchial by a common ostium from ento-bronchium I. ; the anterior intermediate (a) from ento-bronchium II., and (6) from the anterior of two canals arising one immediately behind the other from the ventral floor of the meso-bronchium immediately below the apertures of the mn-Iv. ecto-bronchia. The ento-brenchial ostium opens into the anterior, the ecto-bronchial into the posterior loculus. ‘The posterior intermediate by the hinder of the two hinder ecto-bronchial canals. The posterior abdominal by the terminal end of the meso-bronchium—the ostiwm posterius. The 2nd ento-bronchium does not open, as described by Huxley in Apterya, into the sub-bronchial ostium, but turns upwards and backwards, and terminates cecally like the 4th. The air-sacs in Rhea differ in no essential respect from those of Casuarius. The sub-bronchial sac is unpaired. In one Rhea americana which I dissected the anterior intermediate sac was fed by two ostia widely separated one from another. In a second example the ostium of the anterior intermediate sac of the left side and the septum dividing this sac from the posterior intermediate were both wanting. The posterior abdominal sacs are much larger than those of the Cassowary. Apteryx differs from all other birds, as was first pointed out by Huxley, in that, owing to their smaller size, the posterior intermediate and abdominal sacs are enclosed within the sub-pulmonary chamber. In all other birds a great increase in the size of these two has driven out the posterior abdominal sac, so that it projects, like a hernial sac, between the peritoneum and the parietes into the abdominal cavity. The respiratory organs of the Palwognathe are singularly uniform in character inter se and with the Neognathe. VoL. XV.—PaRT v. No. 19.— December, 1900. 2N 254 MR. W. P. PYCRAFT ON THE MORPHOLOGY AND c. THe WINDPIPE. The most valuable of the numerous treatises on the windpipe of the Palwognathe —from a systematic point of view—is that by Forbes [20]. We are concerned here only with the thoracic end of the trachea, and the degree of specialization which it has reached in the direction of the formation of a “ syrinx.” Struthio and Apteryx, up to the time of Mr. Beddard’s paper [8], have been held to represent the least specialized forms in this particular. In Struthio there are no intrinsic muscles. There is a membrana tympaniformis interna, but only a trace of a membrana semilunaris. ‘ Internally,” according to Forbes, ‘the mucous membrane of the interior is greatly thickened, forming a vocal cord, in the region of the last three tracheal rings and first two bronchial semirings. The lateral tracheal muscles stop at the point of insertion of the sterno-tracheales.” In Apteryx there is a broad membrana tympaniformis interna and a feeble membrana semilunaris. As in Struthio there is no pessulus and, it was believed, no intrinsic muscles. Mr. Beddard’s dissections, however, showed that these were ‘especially plainly developed in Apteryx australis.” In this species he found “a thin band of muscle continuous above with the extrinsic muscle; this passes down the side of the trachea and gradually thins out into fibrous tissue, which seems to die away upon the last tracheal ring, hardly reaching the bronchi. Its state, in fact, is clearly one of degenera- tion, and very much on a par with the state of the corresponding muscle in the Hoatzin (Opisthocomus). In A. mantelli the same muscle was visible, but it did not Seems ir to be anything like so well developed as it is in the species first referred to.” It appears to be entirely wanting in A. haastii and A. oweni. There is a feeble vocal cord, and the lateral muscles stop at the insertion of the sterno-tracheales. Casuarius, as Forbes remarks, shows one interesting peculiarity, the lower end of the trachea becoming slightly dilated above the bronchi. Furthermore, the last tracheal rings are incomplete behind, the median space to which this gives rise being filled up by fibrous and elastic tissue arranged transversely. By the contraction of this, after death, the ends of the rings are drawn inwards, so forming within the tube a low longitudinal ridge, resembling the tracheal septum of certain Petrels and Penguins. There is no pessulus, the membrana tympaniformis interna passing over from one bronchus into the other. A very slight thickening of this membrane in the middle line between the two bronchi suggests a membrana semilunaris, but, as Forbes remarks, there is nothing that can properly be called by this name. The outer vocal cord is well developed. Forbes describes, but so far I am not able to confirm this, “On the internal wall of PHYLOGENY OF THE PALZOGNATHA AND NEOGNATHA. 255 the! bronchus) 75.0 a second much slighter and less prominent fold, slightly concave forwards, running somewhat obliquely backwards and downwards, and supported by the anterior ends of the third and sixth bronchial semirings.” There are no intrinsic muscles. The sterno-tracheales in Casuarius and in Dromeus expand so as to meet in the middle line in front. The tracheo-bronchial region of the trachea in the adult Dromwus has not yet been described. In a young specimen dissected by Forbes there seemed to be no great differences between it and Casuarius. The membrana semilunaris seemed to be better developed. The trachea of Dromwus is peculiar, in that its front wall is pierced by a long, vertical, slit-like median aperture, caused by the deficiency of certain of the tracheal rings in this region. The lining membrane of the tube projects from this opening and forms a large dilatable sac, which can be filled with air at will. It is probably concerned with the production of the drumming sound which this bird occasionally makes. It has been described in detail and figured by Murie [63]; but seems to have been first discovered and described by Peter Fremery [21] in 1819. In Apterya, it will be remembered, the syrinx is but feebly developed, and the intrinsic muscles are degenerate. The intrinsic muscles of Rhea were first correctly described by Forbes [20]. The last four or five, or even six, tracheal rings fuse more or less completely to form a box. The number of rings involved in this fusion can generally be traced by the interannular sutures, which appear never to be completely obliterated. There is a complete and cartilaginous pessulus and a membrana semilunaris. Forbes describes, and I am able to confirm, ‘‘on the middle part of the membrana tympani- formis interna a thickened fibro-cartilaginous area.” ‘This is triangular in form, its base resting on the pessulus, and its apex terminating near the 4th bronchial semiring. It would be interesting to compare this thickened tissue with certain fibrous pea- shaped pads found in this region in Sula bassana, first described by Garrod, and Catarrhactes chrysocome, described and figured by Watson in the ‘ Challenger’ Reports. The broncho-trachealis (intrinsic syringeal) muscle in my dissections, as in Forbes’s, arises from the last ten tracheal rings, and passes downwards to be inserted into the anterior ends of the last four or five bronchial semirings and the membrane between. The sterno-trachealis terminates just pre-axiad of the origin of the intrinsic muscle. In the Casuariida, as Forbes has pointed out, the “ terminal part of the trachea tends to develop into an expanded tympanum”’; the tracheal rings are softer and become imperfect posteriorly, the junction of the trachea and bronchi is less marked, and the arrangement of the sterno-tracheales is peculiar. In Struthio and Apteryx the tendency of the trachea is to become narrower at its termination, and the distinction between this and the bronchi is much sharper than in Caswarius. 2Nn 2 256 MR. W. P. PYCRAFT ON THE MORPHOLOGY AND d. THe Cnoana (Pl. XLV. fig. 1). The choane of the Palwognathe, it is interesting to remark, differ in form from the Neognathe in their larger size. In all the Neognathe the aperture is slit-like, whilst, as will be seen in the figure, in the Palwognathe it is triangular or cordiform in shape. CIRCULATORY SYSTEM. Tue Heart. The heart of the Palewognathe does not differ materially in structure from that of the Neognatiw. It has been already ably described and its most important features pointed out by Owen [67], Lankester [45, 46], Beddard [5], Rolleston [87], and others, so that little remains for me to say. Fig. 8. A. ace. t. SCp. w weet SS A.—Dissection of the heart of Causuarius wnappendculatus, to show the right auriculo-ventricular valve and the ‘‘ moderator band.” B.—--Dissection of the heart of Rhea americana, to show the right auriculo-ventricular band and the abnormal septal valve. ace.t. = accessory tendon. sep.v. = septal portion of right auriculo-ventricular valve. s.v. = semilunar valves. r.du.v. = right auriculo-ventricular valve. p. = papillary muscle. mod.b. = moderator band. PHYLOGENY OF THE PALZOGNATHA AND NEOGNATHA. 257 My only point concerns the moderator band first described in Casuarius by Rolleston. It seems to have been taken for granted by subsequent writers that this was always present, if not in all the Palwognathe, at least in the Casuariide. I have only succeeded in finding it once, and in the heart of a Casuarius casuarius salvadorti, though I have examined specimens of C. wnappendiculatus, C. casuarius heccarit, C. casuarius violicollis, C. casuarius sclateri, and Casuarius casuarius, In addition to the large papillary muscle binding the right auriculo-ventricular valve to the anterior wall of the ventricle, 1 find two or more very slender “ chorde tendinez’’ extending from the outer surface of the auriculo-ventricular valve to the anterior ventricular wall, recalling the similar structures described and figured by Lankester [45] in the heart of Ornithorhynchus. In Rhea and Dromeus only one of these threads was present. I have, however, only had the opportunity of examining one Dromeus for this purpose. The septal half of the valve in Casuarius is thick and fleshy. In Dromeus it is relatively somewhat larger, and, in the only adult heart which I have had an opportunity of examining, differs from Caswarius in that its free ventral or hinder border is curved forwards at its insertion towards the floor of the left auricle, forming by this means a sort of pocket-like flap on the wall of the inter-ventricular septum. In Rhea this same arrangement occurs again. In one adult which I dissected (fig. 8B) this was remarkably well developed. The valve was very thin, and the forward pocket-shaped extension into the anterior end of the ventricle was strongly developed and was continued yet further in the form of a narrow slip of muscle attached only at its ends. In Dromeus, in addition to one or two delicate tendinous threads from the outer surface of the auriculo-ventricular valve to the anterior ventricular wall, there were numerous fleshy strands of muscle, interlacing one another, running parallel with the valve, and from it to the anterior wall of the ventricle. - Trabecule carne were well developed both in Rhea and Dromeus. CAROTIDS. All the Palwognathe, save Apterya, have normally two carotids. Occasionally, in Rhea, the left only persists. In Apteryx there is only one carotid, the left. CLOACA, BURSA FABRICI, AND URINO-GENITAL SYSTEM. a. CLOACA. The terminal end of the alimentary canal in the Palwognathe, as inthe Neognathe, is divisible into a copro-, wro-, and proctodewm, as defined by Dr. Gadow [24]. In Casuarius, Rhea, and Apteryx the coprodeum passes gradually into the rectum, but in Struthio and some Neognathe the coprodeum is preceded by an additional 258 MR. W. P. PYCRAFT ON THE MORPHOLOGY AND chamber, though this is but small and not very sharply defined. In this succession of chambers therefore, as Dr. Gadow has pointed out, Struthio resembles some Saurians and young Crocodiles. In Struthio, defecation and micturition are separate acts, and the urine, instead of passing from the uro- to the coprodeum, is transferred to the bursa Fabricii, which, unlike that of all other birds, is of considerable size and remains permanently open. Its function, acting as a physiological dorsal bladder, is unique. A true allantoic bladder, communicating directly with the ureters, is wanting in the Aves. The cloacal aperture of the Palwognathe, as in Neognathe, is round, not transverse as in Reptilia. Fig. 9. The terminal end of the alimentary canal of Caswarizs violicollis, to show the cloaca and its divisions into copro-, uro-, and proctodeum and the form and position of the bursa Fabricii. b.f. = bursa Fabricii. c.d, = coprodeum. u.d. = urodeum. p.d. = proctodzeum. cl, = clitoris. ovid, = aperture of oviduct. u. = aperture of ureter. 6. Tue Bursa. The bursa Fabricii of the Palwognathw has been very carefully studied by Forbes [20] and Wenckebach [97]. In the Paleognathe—save only in Apteryx—as Forbes [20 4] has shown, the bursa in young birds is of considerable size. Furthermore, he held that it differed PHYLOGENY OF THE PALZOGNATHA AND NBEOGNATHA. 259 fundamentally from that of the Neognathw, since in these last it opens by a more or less constricted aperture into the general cavity of the cloaca, whilst in the Struthiones the exact opposite was the case, the cloaca opening into the bursa. oH rompthisewaryer it will be at once evident that the cloaca does not open directly to the outer surface, but indirectly through the bursa by means of its large posterior and inferior aperture.” I find myself unable to confirm this point, and I think the accompanying figure will justify me. The only difference which I can see between the Palwo- and Neognathe in this particular is that in the latter the bursa opens into the roof of the proctodeum by means of a minute aperture, in the former it opens by a very large one, so large as to absorb nearly the whole roof of the proctodeum. When these parts are dissected from spirit-specimens, in which hardening has taken place in situ, Forbes’s description seems to be amply verified, as I can testify. The accompanying figure was made from a specimen in which the hinder end of the gut had been removed soon after death, the cloacal aperture set widely open, and the whole plunged into formol and left for two or three days. The wall of one side was then removed and the true relations revealed. The function of the bursa is yet unknown. c. CopuULATORY ORGANS. The Paleognathe all possess a well-developed copulatory organ. That of Struthio is a solid grooved organ containing erectile tissue and resembling very nearly that of the Chelonia and Crocodilia. It can be retracted into a pouch in the floor of the proctodeum. The m. protractor penis receives a slip from the transversus analis, which descends from the pelvic bones. A pair of retractores penis arise from the pelvis and are attached to the ventral aspect of the middle portion of the organ. In the remaining Palwognathe the penis resembles that of the Anseres amongst the Carinatwe, being evertible and more or less markedly spirally twisted and grooved dorsally. The copulatory muscles of the Palwognathe, Dr. Gadow [24] points out, differ from those of the Neognathe, inasmuch as the latter are derived from the sphincter ani solely, whilst in the Palwognathe they are partly differentiated from muscles which are still attached to the pelvis, and are therefore skeleto-genital. DEVELOPMENT. The egg of the Palwognathw, save in Apteryx and perhaps the Crypturi, differs from that of the Neognathw in that the pore-canals are branched. In Apteryx, as in the Neognathe, they are unbranched. In Struthio, according to Nassonoy [64], and Apteryx, according to Parker [71], an 260 MR. W. P. PYCRAFT ON THE MORPHOLOGY AND opercular fold grows backwards from the hyoid arch so as to cover the first and second branchial clefts. This has been noticed in some Neognathe. Sushkin [94] has figured and described it in Falco tinnunculus. ‘The first discovery of this operculum appears to have been made by Rathke [84] in 1832. PHYLOGENY OF THE PALAOGNATH A. I enter upon the discussion of this subject with very considerable trepidation—not only on account of the difficulty of the problem, but also because it brings me into direct conflict with the opinions of those who speak after deliberation and long experience, and are, for this reason, and rightly, accounted authorities amongst us. It is now generally admitted that the Palwognathe must be descendants of an Avian proto-carinate form capable of flight. Parker [71] has given a number of characters which lend the strongest support to this view. ‘To many they will be convincing. ‘The most important of these are :— ‘The fact that the skeleton of the fore limb is that of a true wing.” “The presence of an alar membrane or patagium.” “The presence of remiges and tectrices majores.” The alar membrane is a structure of great significance. According to Parker, it is the remnant of a once extensive scale-covered patagium. As these scales gradually became transformed into feathers the size of the patagium decreased. As to the mono- or polyphyletic origin of the group there is much divergence of opinion. Newton and Beddard consider the Palewognathe to be monophyletic; Fiirbringer, Gadow, and Parker polyphyletic. With these last I also agree, but not entirely so. There can be no doubt but that the Palwognathe represent a lower grade than the Neognathe. Nevertheless the latter are closely related thereto. It is probable that they have been derived as a diverging branch from that stock which gave rise to the Rhetde, Dinornithide, and Atpyormithide. The diagram (fig. 10, p. 265) is intended to make this clear. The very close resemblance of the palate of the Crypturt to that of Rhea and Dinornithide and the general resemblance in the form of the pelvis of Crypturi and Dinornithide lend strong support to this view. The structure of the pelvic arch in the Palewognathe is of considerable interest, inasmuch as it bears upon the relation of this group to the Neognathe. It retains its most primitive form in Archwopteryx, Dinornis, Aipyornis, Hesperornis, Ichthyornis, Apteryx, and in the Crypturi. In all these the ischiadic fissure and free pubes are retained throughout life. In Meognathe the ischiadic fissure becomes converted into a foramen by synchondrosis between the ilium and ischium. In Struthio the ischiadic fissure is permanent, but the pubes fuse in the middle line. PHYLOGENY OF THE PALAOGNATHA AND NEOGNATH. 261 In Rhea the ischiadic fissure is closed; the ischia, furthermore, meet one another below the vertebral column, and the pubes fuse with them posteriorly. Thus Séruthio and Rhea are each unique in the form of the pelvic arch. Struthio seems to be a stumbling-block. Thus Prof. Newton writes [65]: “Some systematists think there can be little question of the Struthiones being the most specialized and therefore probably the highest type of these Orders” (‘ Zatite”). Others, in spite of the great amount of specialization which Struthio has undoubtedly undergone, would regard this form as the most primitive, not only of living birds, but of the Palwognathe living and extinct. According to Furbringer, on account of the great number of its primaries, its coracoid, and many of its shoulder- and arm-muscles, the form of its pelvis, two-toed foot, ‘‘gewisse Eingeweide, etc., entfernen sie weiter von allen anderen lebenden Vogeln, als diese in der Regel unter einander divergiren. Zugleich bietet Struthio neben verschiedenen mehr vorgeschritten Specialisirungen progressiver und retrograder Natur eine Anzahl héchst primitiver Charaktere dar, die ihre tiefer als die anderen Ratiten stellen und durch welche diese sehr alte Form Blicke bis in die friihesten Vorzeiten des Vogelstammes thun liisst.” Beddard considers ‘‘that Struthio is removed far from the Dinornithide, as well as from other Ratites, by the structure of its palate, which diverges much.” He does not, however, go as far as Fiirbringer, for he continues further on: “There are really, however, not a large series of characters in which they may be fairly said to be more primitive than some other groups..... ” This last statement is evoked in answer to Fiirbringer’s contention that Struthio is really very primitive. I find myself in agreement with Beddard: more than this, I believe that we have to look to Dromeus and Not to Struthio for the most primitive of living birds. This opinion is based mainly upon a study of the bones of the palate in this and the allied forms. These in Dromeus seem to represent the ground type from which, or from some closely allied form, the palates of Struthio, Rhea, and Dinornis may be derived. ‘The palate of Apterya must be considered by itself. The palate of Struthio compared with that of Dromeus will be found by no means so fundamentally different as seems to have been supposed. On the contrary, it is contended that the palate of Struthio is a specialized and, in some respects, retrograde form of that of Dromeus. The palato-pterygoid connection is similar in both. The palatine and the forward extension of the maxillo-palatine processes are both considerably longer, both actually and relatively, in Séruthio than in Dromeus. Struthio lacks palatine processes to the premaxilla, and has a degenerate, one might almost say vestigial, vomer. Like all degenerate structures, however, it gives some certain signs of a former perfection. Thus, in a skull which in this paper is attributed to Struthio meridionalis the vomerine limbs extend backwards to within a very short VOL. XV.—ParRT y. No. 20.—December, 1900. 20 262 MR. W. P. PYCRAFT ON THE MORPHOLOGY AND distance of the pterygoid. If this proximal extension were carried back sufficiently far to enable the vomer to join the pterygoid the resemblance to the Dromeine type would be really striking. If we turn now to the palate of Rhea, and compare this with that of Dromeus, we shall find the differences more decidedly marked. The peculiar and almost complex relations which obtain between the palato-pterygo-vomerine articulations have already been described in these pages (p. 206), and are sufficiently distinct to render further comparisons unnecessary. But the palate of Rhea is curiously similar to that of Dinornis and the Tinamous. So similar as to render it probable that these are related more closely one to another than to Dromeus or Struthio. The affinity of Rhea to the Dinornithide has already been hinted at by Beddard, though upon different grounds. In discussing the position of Struthio, he remarks that though “ this is removed far from the Dinornithide, as well as from other Ratites, by the structure of its palate, which diverges much,..... it is not clear that Rhea is so remote; the existence of an apparent homologue of the maxillo-nasal bone... .. isa point of somewhat striking likeness to Hmeus.” Nathusius, again, in studying the egg-shells of Rhea and Dinornis, was so impressed with the likeness between them that he proposed to unite them in the same genus. Parker [73], however, strangely enough, remarks: ‘I know of no character in the skull of Rhea by which it definitely approaches the Moas..... ” T say strangely enough advisedly, for I feel sure that had Parker’s attention been drawn to the points to which attention is directed in this paper he would have grasped its significance and written quite otherwise. About the palate of Apyornis we as yet know nothing. In the roof of its skull it resembles the Dinornithide, inasmuch as, just as in this group, there are no out- standing supra-orbital processes to the lachrymal. In some other points it resembles Struthio; but since in its pelvis and other points it closely resembles Dénornithida, I am inclined to place it near this group. The palate of the Tinamous is truly Dromwognathous, but approaches the Neo- gnathine type, tending towards schizognathism. In its pterygo-palatine articulation, as I have already pointed out (p. 208), it closely approaches the Meognathe, the palatine having shifted from contact with the body of the pterygoid to form a connection more or less intimate with its tip. The palate of Apteryx in the peculiar forked pterygoid and the complexity of the relations between it and the palatine and vomer seems to differ markedly from all the other Palwognathe, or, indeed, from all living birds. Here, again, is a conclusion diametrically opposed to that of Parker. Whether or not I failed to appreciate the points of resemblance Parker claims must remain to be seen. This discrepancy is certainly remarkable, and will seem to cast doubt upon the PHYLOGENY OF THE PALZOGNATHEH AND NEOGNATHA, 263 4 validity of my conclusions rather than his, for those who know his monographs on the development of Apteryx and the skull of the Dinornithide will agree that they are monuments which mark an epoch in the history of this subject. In his work on the skull of the Dinornithide he wrote: ‘The marked differences between the Moas and Kiwis are certainly for the most part adaptive... .. The real affinities underlying these differences are, however, shown by the striking similarity of the bones of the palate in the two forms.” The skull of Caswartus, it may be remarked here, differs in no essential respect from that of Dromaus. In whatsoever it differs from Dromeus point to specialization, as, for instance, the development of the casque upon the mesethmoid (p. 199, Pl. XLIV. fig. 3) and the small quadrato-jugal fossa. The skull of Dromeus, it would seem, must be regarded as the most generalized of living birds, at least in so far as the bones of the palate are concerned. Struthio somewhat nearly approaches Dromeus in the arrangement of these bones, fthea, the Tinamous, the Dinornithide, and probably the Apyornithide agree more closely one with another than with Dromeus. Apteryx differs from all in the direction of increased complexity of these parts. We have now Struthio and Dromeus with Casuarius opposed to Rhea, Dinornithide, and Crypturi, and possibly Apyornis. ‘These last we may further subdivide by means of the pelvis. This will separate AKhea from the remaining forms; inasmuch as in Rhea the pelvis is long and narrow, with the post-acetabular ilia meeting in the middle line, as in Apteryx. In the Dinornithide, Mpyornis, and the 'Tinamous the post-acetabular ilia are separated by the long transverse processes of the synsacrum. It may seem that this character of the pelvis is a somewhat artificial one, but reflection will show, I think, that it is probably not so. It does not seem to belong to the category of adaptive characters, since Rhea and Apteryx in the general confor- mation of the pelvis agree, though their habits are very different, and both bear some _ similarity to that of Struthio—similar in so far as the great length and transverse width are concerned. In the Tinamide, Dinornithide, and Afpyornis the pelvis owes its great breadth to (1) elongated transverse processes, and (2) the broad dorsal plane of the post-acetabular ilium. The pelvis of Struthio bears an undoubted resemblance to the Dinornithine form, inasmuch as, like Dinornis or Atpyornis, the transverse processes bear the post- acetabular ilia away from all contact with the neural spines of the synsacrum. Fiirbringer’s view with regard to the Dromeide is not exactly in harmony with the views adopted here, but it lends some support thereto nevertheless. He says Dromeus and Casuarius “bilden zwei sehr nahe verwandte Familien..... welche etwas hoher als die Struthionidw, aber tiefer als die Rheide stehen und im Ubrigen sowohl von den anderen Ratiten ..... als von den Carinaten eine entfernte 202 264 MR. W. P. PYCRAFT ON THE MORPHOLOGY AND Stellung einnehmen,” and adds: “Ihre Vorgeschichte kommt mir am meisten dunkel By this time the tendency of the foregoing will no doubt have revealed itself. Its intention is to suggest that Dromeus and Casuarius—the Dromeide—are to be regarded as the most primitive of the Palwognathe ; and that the Struthionide are most probably the nearest approach thereto, for, in spite of the great specialization which they have undergone, they undoubtedly retain many archaic characters. The Rheide, Dinornithide, Aypornithide, and Crypturi are all probably more nearly related one to another than to the Dromeidw. The Apterygide are the most aberrant of all, but have affinities with the Dinornithide. This implies that the Palwognathe are to be regarded as polyphyletic—probably triphyietic. An attempt to express this relationship is given in the diagram on p. 265. I propose to discuss the details of this “tree” in so far as concerns the Neognathe elsewhere. Of the Palwognathe I need only say I recognize 7 Orders, viz. :— Casuaril. Struthiones. Rhee. Crypturi. Dinornithes. Epyornithes. Apteryges. Their relation to the other members of the Class may be expressed as follows :— Class AVES. Subclass I. ARCHAORNITHES. » Il. NEoRNITHEs. Grade A. Palwognathe. » B. Neognathe. It would have been helpful here to discuss the further evolution of this group to its final end in the various genera and species, but this is unfortunately impossible in the present contribution owing to lack of material. The diagram will give a slight indication of the writer’s views on this subject. PHYLOGENY OF THE PALZOGNATHE AND NEOGNATH &, 265 Fig. 10. Pici f f f Passeres Trogones’ f f Colii Coraciss Bucerotes Accipitres Cesena i Columbs H | # Pterocles Steganopodes Af i y Aisas Musophagi if Cathartes 5 = Limicola, Palamides Gruidss = FS Rallida S Hurypygide Mesitides V Y Y VY Y, 47 iY | LF Yf WA Tubinares ‘GM A UY ( A == Colymbt (including Hesperornis) ¢} 4 H H ESS ) A Yj Sphenisci Z ~ IA, LF Dinornithes »)) 4 pyornithes Z A A“ Za A LY Za tity LS” 4 Apteryges Phylogenetic diagram showing the probable Relationships of the Palwo- and Neognathe. |||| == Palewognathine stock. 38e = Eutaxic forms. Z = Diastataxic forms Blank branches indicate eutaxic forms amongst a diastataxic group; blank spaces in a branch that both eu- and diastataxic forms occur in the same genus or family, 266 MR. W. P. PYCRAFT ON THE MORPHOLOGY AND SUMMARY. The group of birds which we collectively designate the “ Ratita” is the same as that to which this name was originally given’ by Merrem in 1813 (with the addition of Apteryx, then unknown), on account of the fact that all agreed in the absence of a keel to the sternum. They were “raft-breasted,” in contradistinction to the Carinate, or keel-breasted !. Nitzsch, Huxley, Newton, Sclater, and Sharpe have successively adopted Merrem’s subdivision of the class Aves—adding the Saururee, unknown to Merrem. Other workers, both before and since Merrem, have accorded to the “ Ratite” no greater prominence than that of an order or suborder of the class Aves. Sometimes even this was considered too great a distinction, and they were reduced to the rank of a Family, including such forms as Bustards, Plovers, &c. The “ Ratitee” of Merrem were apparently regarded by that author, as well as by Huxley, Newton, Sclater, Sharpe, and other systematists, as a monophyletic group. Fiirbringer, Gadow, and Parker stand prominently forward as advocates for a poly- phyletic origin. Gadow’s views are admirably set forth in his contribution to Bronu’s ‘'Thier- Reich ’ [26]. Briefly, he regards the “ Ratitee” as a morphologically monophyletic group, standing in the same relation to the class Aves as the Anthropomorphe to the rest of the Mammalia. / With this view we entirely agree. We venture further, and include the Crypturt within the group. This was done long since by Garrod [30] and Stejneger. Beddard [7], later, has shown a strong inclination to follow suit. Thus, he says:—“ It is unquestionably to the Struthiones that they showed the greatest number of important likenesses, so much so, indeed, that their inclusion in one great group with them would be by no means an unreasonable way of disposing of them.” The Dromeognathous palate is sufficient justification for such an association. Consent to this, however, renders inoperative the old term “ Ratitz,” which at best is but a makeshift ; for many of the Carinate are Ratite, whilst the Tinamous have a large keel to the sternum. We propose instead the terms Palwognathey (Ratite + Crypturi) and Meognathe (Carinatee—Crypturi). The adoption of these surmounts the difficulty indicated above. Gadow regards the Struthiones as a group of primitive forms more nearly repre- senting Proto-Carinate than any other living birds. Their retention under the common name “ Ratitee” he regarded as convenient rather than an indication of close affinity. Similarly we regard the Palwognathe as the unsuccessful descendants of proto- carinate forms, but retaining probably but few of their characteristics unaltered. It would be too much to regard any of the existing Meognathw as the direct descendants of any of the existing or extinct Palwognathe known to us, but it seems not * Blanchard in 1859 substituted for Merrem’s the terms Tropido- and Homalosterni. PHYLOGENY OF THE PALZOGNATHE AND NEOGNATHA. 267 improbable that their source may be traced to that stock which gave rise to the Rheo- Dinornithine type of palate and pelvis. Casuarius, which forms the nucleus of this research, is here regarded as only generically distinct from Dromwus. These two forms are, it is here contended, closely allied. Struthio is perhaps derived from the same ancestral stock. Hitherto, how- ever, these two forms, Caswarius and Dromeus, have been regarded as more closely allied to Dinornis and Apterya than any other forms. Apteryx we regard as quite distinct. The interrelationships of this group are indicated in the diagram (p. 265). The skeleton of Dromeus ater, procured at the last moment, at great trouble, by Mr. Rothschild, has proved of considerable interest in one or two respects. It does not differ much, however, from that of the larger species: wherein it differs it appears to approach the Cassowaries. I have seen no authenticated skeleton of D. irroratus. Skulls of adult Caswarius and Dromwus are badly needed; as also is the adult skeleton of Rhea darwini. It is interesting to note that the characteristic feathered tarso-metatarsus of &. darwini is temporarily reproduced in the embryo of R. americana (fig. 2 D, p. 155). Whether the four species of Ostrich indicated in this paper will be further confirmed is a point which time alone will show. It seems probable. I hope to be able to supply much that is Jacking in this paper by a further contri- bution at no distant date in the shape of an Appendix. Since this was written the British Museum has acquired a hind-limb of Genyornis. This is remarkable chiefly on account of the fact that digit 1. was becoming absorbed. The proximal phalanx was of great length and slenderness. In this we may trace the early stage in the evolution of a second two-toed Struthious bird. As Messrs. Stirling and Zeitz have shown, Genyornis was undoubtedly a near ally of Dromeus. Key To THE OSTEOLOGY OF THE PALHOGNATHA, based on the Characters of the Adult Skeleton. A. SkuLt. (Plates XLIL—XLIV.) The pterygoid free anteriorly or resting on the paired posterior ends of a large voiser ; connected with the palatine either by the opposition of its outer border with the mesial border of the palatine or by squamous suture, or by complete fusion, but never by means of opposed glenoid surfaces. The vomer is always present, and paired, and is either fused with or tightly wedged in between the maxillo-palatine processes ; basipterygoid processes always present, of great size, seated at the base of the parasphenoidal rostrum, and articulating with the extreme proximal end of the pterygoid; nasal processes of the premaxilla unpaired; parasphenoidal rostrum and mesethmoid continued forward beyond the level of the lachrymal. A. Beak short or of medium Jength, never probe-shaped. a. Vomer deeply cleft posteriorly into widely expanded right and left limbs running directly backwards beneath the pterygoids ; palatine articulating 268 MR. W. P. PYCRAFT ON THE MORPHOLOGY AND posteriorly by its mesial border with external border of the combined vomer and pterygoid; voraer of great length, continued forwards beyond the free end of the parasphenoidal rostrum; pterygoid laminate; orbital process of quadrate decurved and articulating with the pterygoid at its base ; anterior palatine foramen very small, and lying near the tip of the beak beyond the vomer; anterior narial and lachrymo-nasal fossz, in dried skull, confluent ; mesethmoid with a truncated anterior border ; nasal process of premaxilla rod-shaped ; nasal with a vestigial external process. a’, Vomer expanded into a broad flattened plate between the maxillo- palatines, and separated from the parasphenoidal rostrum; beak dorso-ventrally depressed ; palatine fenestrated posteriorly, not anchy- losed with pterygoid; horizontal plate of mesethmoid appearing on the surface of the skull between nasals, but never inflated to form a casque ; lachrymals fusing with nasals and frontals ; with long supra- orbital processes '. Vomer only very slightly enhonded between the arate palabiviesl Beal laterally compressed ; palatine unfenestrated and anchylosed with the pterygoid and vomer; horizontal plate of mesethmoid, nasals, and more or less of frontals inflated to form a casque, sometimes of great size ; quadrato-jugal fossa not extending forwards to level of antorbital plate 6. Vomer large, ata teciieeal aeceies investing the par ee esoitel rostrum ; distal end of pterygoid and hinder end of vomer hidden (in ventral aspect of skull) by the palatine which underlies them; pterygoid rod-shaped. c', Vomer deeply cleft anteriorly, exposing the parasphenoidal rostrum between a pair of long spurs; anterior palatine foramen long and slit-like. a", Lachrymo-nasal fossa bounded anteriorly by a bony spike rising upwards from the middle of the external border of maxillo-palatine ; with large post-orbital processes ; orbital process of quadrate very short, not extending beyond level of articulation with pterygoid ; lachrymal with large orbital processes; mesethmoid visible as a lozenge-shaped plate between nasals . 5 '. Lachrymo-nasal fossa bounded anteriorly by a deine rade articu- lating above with the nasal and below with the hinder border of the maxillo-palatine plate ; no post-orbital processes ; orbital process of quadrate moderately long, with a free end projecting beyond articu- lation with pterygoid; lachrymal without orbital processes ; nasal schizorhinal; inter-orbital septum perforate ; mesethmoid not visible on surface of skull d', Vomer truncated anteriorly, not extending so oe ere as the aa of the parasphenoidal rostrum; anterior palatine foramen small, closed by anterior end of parasphenoidal rostrum ; large post-orbital processes ; lachrymals fused with frontals and not readily traceable; orbital Dromeus. Casuarius. Rhea. Tinamide. PHYLOGENY OF THE PALZOGNATHE AND NEOGNATHA. 269 process of quadrate long, articulating with whole inner border with pterygoid ; olfactory cavity of great size . . . . . . . . . . Dinornithide. c. Vomer deficient posteriorly, not extending backwards to reach the pterygoid, anteriorly extending forwards, as in Dromeus, to conceal the parasphe- noidal rostrum ; the body of vomer deeply hollowed, fitting closely on to the rostrum, and embraced on either side by the maxillo-palatine pro- cesses ; maxillo-palatine processes hastate; anterior palatine processes of great size; pterygoid laminate; palatine rod-shaped, fused proximally with external border of pterygoid ; premaxilla wanting a palatine process ; lachrymals continued backwards to fuse with a chain of supra-orbitals and with the frontals, so as to enclose a supra-orbital foramen ; lachrymal with an osiculum lachrymo-palatinum ; nostrils holorhinal ; anterior narial and Jlachrymo-nasal fossee confluent; quadrate with long recurved process bearing a well-defined projecting facet for the pterygoid in the middle of its internal ventral border . . . . . . . . . . . Struthio. B. Beak elongated, decurved, probe-shaped ; anterior narial fossa minute, con- fined to tip of beak ; lachrymo-nasal fossa enormous, extending forwards to within a short distance of tip of beak ; skull rounded, without post-orbital processes ; lachrymal vestigial ; olfactory chamber of great size, extending backwards to optic foramen; vomer passing backwards to fuse with the pterygoid and palatine, and, with these, helping to form the floor of a tubularmecessropemingcimward sia mcnurny amie i) eens ane nnnEEA penta: Key to the Species. STRUTHIO. a. Maxillo-palatine processes not rising dorsad, the anterior border of the maxillo-palatine process at right angles to the long axis of the skull. a’. Palatines projecting distad beyond anterior border of maxillo-palatine process ; maxillo-palatine processes without any external lateral fenes- trated plate é Porth, gti ener Ua BG as ety ae en S. australis, 6'. Palatines not reaching as far forward as the level of the anterior border of the maxillo-palatine process. a’, Vomer very short; parasphenoidal rostrum thick; maxillo-palatine processes with an external lateral fenestrated plate. . . . . . . S.molybdophanes. 6’. Vomer very long ; maxillo-palatine processes without an external lateral fenestrated plate; parasphenoidal rostrum slender ; roof of the skull with deep depressions mesiad of the supra-orbital ledges; nasals meeting in the middle line proximad of the nasal process of pre- med 5 o 6 oo 0 6 © 9 6 6 6 6 6 BH 6 bh 8 0 (Sh aRHOAaOS”. b. Maxillo-palatine processes rising dorsad ; anterior border of maxillo-palatine processes sloping obliquely backwards ; vomer broad, with short, pointed, posterior limbs; palatines projecting distad beyond maxillo-palatine pro- CESSESP a sbucrie hint sad OO. wae veh SS ey th el then cea ke The tS ee Seecamelus: 1 We believe that Struthio massaicus (Neum.) is identical with this bird—W. R. & W. P. P. VOL. XV.—Part v. No. 21.—WDecember, 1900. 2eP 270 MR. W. P. PYCRAFT ON THE MORPHOLOGY AND APTERYX. a. Vomer long, equal to or longer than the width across the paroccipital processes at widest, notched anteriorly, aud with the median cleft not exteading forwards beyond the middle - . . A, australis. b. Vomer less than width across paroccipital processes. a’, Vomer notched anteriorly, the median anterior portion nearly as long as the paired posterior limbs; paroccipital processes well defined; basi- temporal plate bounded antero-laterally by a pair of ridges passing forwards into parasphenoidal rostrum . : 6 6 0 8 6 0-0 Zils AMIRI, bo’. Vomer imperfect anteriorly, median portion much shorter than the paired jee limbs. = Size not less than.7 inches) 2) 953, 3 2 ee wl le el ee Aa mantelle: . Size not less than 6} inches . . . . . . - . . . A. owen, Raea. A. Temporal fossa more or less sharply defined and extending backwards to meet the lambdoidal ridge; lachrymal with elongated dorsal processes arching backwards over orbit. a. Posterior nares markedly wider anteriorly than posteriorly ; anterior vome- rine cleft not extending backwards to the level of the hinder angle of the lachrymo-nasal fossa; length of quadrato-jugal fossa less than width across beak measured across the maxillary pillars bounding the lachrymo- nasal fossa anteriorly . . . . . : 5 6 0 6 0 6 6 Li COAG b. Posterior nares not wider anteriorly chan Heeiecionley anterior vomerine cleft extending backwards to the level of the hinder angle of the lachrymo- nasal fossa; length of quadrato-jugal fossa equal to width across beak at the level of the maxillary processes bounding the lachrymo-nasal fossa. .macrorhyncha'. B. Temporal fossa not sharply defined and not extending backwards to lambdoidal ridge; lachrymals with short orbital processes; anterior vomerine cleft very SINOTialins Meme Ge Seno W hol wot Moni dg lecd vaecoila Byetodia tego sie okt a. ye dali COORG Dromzvs. a. ae not exceeding 7 inches. . Post-orbital processes short . . . . . .. . - . . . « .« D.nove-hollandie. Q Post- orbital processes long, extending downwards to ihe a short distance ES) Grecnechemosmeaal ime ss 5 5 6 5 6 oo 6 ce om 6 0 Jb Garena b. Size not exceeding 5 inches; roof of skull with a very conspicuous cerebral COM 6 6 6 so 6 8 o 6 0 6 59 0 0 2D, QuaP 1 The specific characters of this species have been drawn from a single adult skull, kindly lent me by Mr. Beddard from the Zoological Society’s Collection. 2 This skull is from a skeleton in the Florence Museum. The bones of the palate are missing. The cranium closely resembles that of D. nove-hollundi, differing therefrom chiefly on account of its very small size and the prominence of the cerebral dome, whilst in the peculiar form of the inferior limbs of the lachrymal it PHYLOGENY OF THE PALHZOGNATHA AND NEOGNATHA. B. VerTEBR# (not including Crypturi) }. 2 ‘ 1 All the pre-synsacral vertebre are heterocelous and free; all the synsacral vertebre develop high neural spines ; intercentra are wanting. A. Hyperapophyses of axis well developed; cervical vertebre short; cervical diapophysial lamelle large and conspicuous; at least one lumbo-sacral vertebra ; sacro-caudal (post-sacral) vertebrze well ossified. a. Hyperapophyses of atlas moderate; neural spine of axis rather high; hyperapophyses of axis well developed ; cervical vertebra rather short ; neural spines of cervical vertebrae well developed and greatly excavated postaxially ; no neural spines laterally bifureating ; catapophyses com- mence at about the sixth vertebra, and do not form a ring; hypapophyses well developed, not much produced postaxiad; diapophysial lamella extending towards postzygapophysis, large and conspicuous; cervical ribs moderately long and styliform; thoracic and lumbar neural spines not much axially extended, and not subequal in height; post-sacral vertebre raised in part, more or less, above ventral margins of acetabula ; sacral vertebre not plainly distinct from post-sacrals ; sacral and post- sacral neural spines laterally expanded and separating the post-acetabular ilium ; caudal vertebrie eight or nine; pygostyle cylindrical. a’. Total number of vertebre fifty-four; cervical vertebra seventeen or eighteen ; vertebre with ribs nine; post-sacral vertebre nine or ten ; pre-axial articular surface of atlas moderately notched dorsally ; hypa- pophysis of axis much elongated; hypapophyses developed from the eighteenth or nineteenth to the twenty-third vertebra; a conspicuous fossa on each side of the pre-axial part of the neural spines of dorsal vertebrz ; post-sacral vertebra raised decidedly dorsad of ventral Tinian OF acim 5s 56 vw 6 6 0 6 0 09 16 6 @ 5 4 b/. Total number of vertebre from fifty-five to fifty-nine ; cervical vertebra fifteen or fourteen ; vertebre with (free) ribs eleven or twelve; sacro- caudal (post-sacral) vertebre eight or nine; pre-axial articular surface of atlas moderate, may become a foramen; hypapophysis of axis extremely long; hypapophyses developed from eighteenth to twentieth vertebra ; fossze beside pre-axial part of neural spines very small ; sacro- caudal (post-sacral) vertebra raised scarcely at all dorsad of acetabula . b. Hyperapophyses of atlas large ; neural spine of axis lofty ; hyperapophysis of axis moderate or extremely large; cervical vertebrz short ; neural spines of cervical vertebra generally more or less well developed ; cata- Dromeus. Casuarius. resembles D. trroratus. species. In the conspicuous development of the cerebral dome it differs from both existing I would remark here that the specific characters attributed to D. irroratus are based upon an examination of a single unnamed skull in the British Museum Collection ; but these appeared to me to be sufficiently marked to justify my determination of this skull as of D. wroratus. 1 This Key is compiled, with modifications, from Mivart’s memoir, 2P 9) ) bo bo MR. W. P. PYCRAFT ON THE MORPHOLOGY AND pophyses may form a ring ; diapophysial lamella extending towards postzygapophysis, large and conspicuous or extremely so; cervical ribs moderate or small; sacro-caudal (post-sacral) slightly raised dorsad of ventral margin of acetabulum ; cervical vertebre fifteen. ce’. Total number of vertebre from forty-five to forty-seven ; post-synsacral vertebre three ; pre-axial articular surface of atlas scarcely at all notched dorsally ; hyperapophyses very long ; a minute hypapophysis to atlas; no hypapophysis to axis; neural spine of axis very high ; hyperapophysis of axis extremely large ; catapophyses commence at the fifth vertebra, in the tenth and eleventh vertebre they form a ring ; hypapophyses developed from the twelfth cervical to the third thoracic vertebra; diapophysial lamella extremely large and con- spicuous, developing conspicuous postaxially extending processes ; cervical styliform ribs small; thoracic and lumbar neural spines exceedingly extended axially and subequal in height; synsacrum with sacral vertebrz plainly distinct from post-sacral ; post-sacral bent much ventrad of ilia ; free caudal vertebr seven or nine; pygostyle conical. ’, Hyperapophyses of axis moderate ; neural spines of cervical vertebrae well developed and deeply excavated postaxially ; neural spines of fourth, fifth, sixth, and seventh cervical vertebre bifurcating laterally ; catapophyses commencing at the fourth vertebra; hypapophyses commence at fifteenth (last cervical) vertebra; diapophysial lamella extending towards postzygapophysis, large and conspicuous; cervical ribs probably moderately long and rather styliform; thoracic and lumbar neural spines not much axially extended, and probably not sub- equal in height ; synsacrum with post-sacrals not bent ventrad of ilia. B. Axis vertebra with a hypapophysis, with or without well-developed hyper- apophyses ; cervical vertebrz greatly elongated ; neural spines of cervical vertebree small or almost obsolete; catapophyses commencing about the sixth vertebra; catapophysis never forming a ring; diapophysial lamella extending towards posterior zygapophysis, small and inconspicuous; a conspicuous fossa on either side of pre-axial part of neural spines either of cervical or thoracic vertebre; synsacrum with sacral vertebre plainly distinct from post-sacral vertebre. e’. Total number of vertebre fifty-six ; cervical vertebrae seventeen ; vertebre with ribs ten; no lumbo-sacral vertebree ; synsacrum with eight post- sacral vertebre ; pre-axial articular surface of atlas with a large dorsal notch; hyperapophyses of atlas moderate ; hypapophysis of axis rather small; neural spine of axis moderate ; neural spines of cervical vertebree small, but not obsolete; no neural spines bifurcating laterally; no neural spines aborted ; hypapophyses developed from the eighteenth (first thoracico-dorsal) to the twenty-first (first thoracic) vertebre ; hypapophyses but little produced pre-axially ; a conspicuous fossa on each side of pre-axial part of neural spines of thoracic vertebr, but not of cervical; cervical ribs long and styliform; thoracic and thoracico- Apteryz. Dinornithide. PHYLOGENY OF THE PALZOGNATHA AND NEOGNATHA. lumbar neural spines but little axially extended, but considerably extended dorso-ventrally and not subequal in height ; synsacrum with post-sacrals not at all raised dorsad of acetabula, but rather depressed ventrad ; post-sacrals perfectly ossified ; no lumbo-sacral vertebre; pre-caudal vertebre sacral plainly distinct from post-sacral vertebree ; ten; pygostyle a laterally compressed plate . UM SON Ys Struthio. f’. Total number of vertebre forty-nine; cervical vere fourteen ; pre- axial articular dorsal surface of atlas with only a minute dorsal notch ; no hyperapophysis to atlas; hypapophysis of axis not very large; neural spine of axis not very high; hyperapophyses of axis well deve- loped; neural spines of cervical vertebre almost obsolete, those of fourth to seventh vertebre abort; neural spines of thirteenth to sixteenth vertebre bifurcate laterally; hypapophyses developed from last cervical to second dorsal vertebra; hypapophyses narrow and pre-axially produced ; a conspicuous fossa on each side of pre-axial part of neural spmes from twelfth to the fifteenth inclusive ; cervical ribs short and not styliform; free cervical ribs nine ; synsacrum with post- sacral vertebre raised quite dorsad of acetabula, also very imperfectly ossified and much reduced; sacral plainly distinct from post-sacral vertebre ; post-sacrals nine ; free caudal vertebre five ; pygostyle small amGl q@ylbnchaenll g 5 6 6 o 540 0 Oo 6 Rhea. The vertebral formula is :— Synsac. eS IDOE 65065000000 Cy. 18: Oy. Th. 3: Th.3+1+4+2: Lb.6: Lb.se.2: Se. 2: Cd. 949=55. — eH 6 18 Synsac. Casucriusteeieia Cy.15: Cv. Th. 4: Th.5+1+3: Lb.5: Lb.sce.2: Se. 2: Cd. 94+9=55. —S SS — 9 18 Synsac. poe Struthto 28s secs selene Cv. 17: Cy. Th.3: Th. 542: Lb. 8: Se. 2: Cd.8+10 = 5). a) nd 7 18 Synsac. TOG go 009660000000 Cy. 14: Cy. Th.3: Th. 34241: Lb.3: Lb.se.4: Se.2: Cd. 94+8=49. S— Sue 6 17 Synsac. a i ZVOTBED 650 0000606 . Cy.15:; Oy. Th.1: Th.44+341: Lb.4: Lb.se.4; Se. 2: Cd. 348=45. (A. a. mantelli.) ame na fal = 274 MR. W. P. PYCRAFT ON THE MORPHOLOGY AND C. Srernvum (fig. 4, p. 217) anp Pectorat Arca}. The sternal plate is keelless; the coracoid anchyloses with the scapula, and lacks the acro- coracoid, the coraco-scapular angle is very wide; furcula vestigial or absent. A. Coracoid grooves not widely separated. a. Metasternum pointed, no posterior lateral processes; pro-coracoid not forming a distinct bar. a’. Coracoid grooves and bases of coracoids overlapping ; anterior lateral process long, directed upwards and backwards ; a pair of short recurved processes lying on the anterior sternal border between the anterior lateral processes ; with vestigial clavicles; sternal plate nearly as broad as long; coracoid longer than broad . . . . . . . . . . )~Dromeus. b’. Coracoid grooves not overlapping ; anterior lateral processes of sternum short or obsolete; no clavicles; sternal plate much longer than broad; @ormconel AS lownel as lkime 6 5 6 3 6 oo 6 6 0 6 oo fl lo 6(CRRITRINS. 6. Metasternum more or less pointed, with large posterior lateral processes ; pro-coracoid of great size, forming a distinct bar of bone ; body of the sternum with a conspicuous ventral prominence ; no clavicles ; articular surfaces for ribs of great'width . . . . . . ..... =. . «. Struthio. B. Coracoid grooves widely separated. a. Corpus sterni basin-shaped with a marked ventral prominence ; posterior lateral processes absent ; anterior lateral processes conspicuous, each with a large pneumatic aperture at the base; articular surfaces for ribs crowded together immediately behind the anterior lateral processes ; coracoid grooves at the base of the anterior lateral processes and separated one from another by the wide emarginate body of the anterior border of the sternum ; coracoids long and slender ; pre-coracoid relatively small. Rhea. b. Sternal plate flat, slightly or not at all pneumatic ; metasternal and posterior lateral processes always present ; facets for ribs not separated by deep pneumatic pits. a’. Articular facets for ribs widely spaced, and lying entirely behind the anterior lateral processes, and not separated by deep pneumatic pits ; coracoid nearly as broad as long; pre-coracoid not conspicuous. . . 0 6 6 9 AYiaIER B. Pubes anchylosed candied mth one another or Beh ‘g@linn, c. Post-acetabular ilia widely separated from ischia, and from one another in the mid-dorsal line; pubes fused caudad (a) with the ischia and (d) with one another in the mid-ventral line; the synsacral vertebre appear- ing below the level of the ventral border of the ilium . . . oer ey SENULIIOS d. Post-acetabular ilia anchylosed with ischia, and with one ETotiier in the mid-dorsal line; pubes anchylosed with ischia caudad; ischia closely united in the middle line one with another throughout almost entire length (unique) ; the post-sacral vertebrae of the synsacrum atrophied . Rhea. Key to the Species. Ruea. a, Pre-ilium not projecting forward beyond the last thoracic vertebra and rib, which forms the limit of the synsacrum cephalad . . . . . . . . . WR. americana. 6, Pre-ilium projecting forward conspicuously beyond the last thoracic vertebra pln) ¢ 5 6.6) Gg oS ob 6 0 6p a 59 6 6 5 6 6 o 6 Oo 5 JenmomeAahD. APTERYX. A. Width across anti-trochanter equal to or less than the width of pre-acetabular ilium at widest. a. Pre-acetabular ilium of great width, with a strongly arched dorsal border, and large laterally expanded antero-ventral angles; width across pre- acetabular ilia at widest part=to length of ischium, and far exceeding length of post-acetabular ilium; post-acetabular illum with a sudden widening at its free end; breadth of pre-ilium at widest=breadth across AMT 5 o 0 0 0.0 0 6 0 6 6 0 0 9 6D . - A. australis. B. Width across anti-trochanter always greater than width of pre- acetal ilium at widest; post-acetabular tapering caudad; breadth of pre-ilium at widest less than breadth of anti-trochanter. 6. Pectineal process large ; five lumbar vertebree. a’. Post-acetabular ilium with a concave ventral border, and a slight notch immediately behind the anti-trochanter . . . . . . . . . . A. haasii. o'. Post-acetabular ilium with straight or convex ventral border ; no notch behind the anti-trochanter . . . SD ee ey cig hind Lao hier EAR OWWCNE® c. Pectoral process small ; four lumbar ver ayes Je Gh soy ideo oe echo oo) neh Chai VoL. XV.—PAR?T v. No. 22.—December, 1900. 2a 278 MR. W. P. PYCRAFT ON THE MORPHOLOGY AND Drom us. a. Length 17 in.; pectineal process well developed ; dorsal border of pre-ilium much arched ; obturator fissure nearly closed ; pubo-ischiadic fissure nearly Gogetil 5 o 6 6 5 0 6 ob oo eo ol ll 68 ll ol oO Di pipeiolianite, 6. Length 93 in.; pectineal process obsolete; dorsal border of pre-ilium not arched ; obturator fissure wide ; pubo-ischiadic fissure very wide. . . . Dz ater. EK. Prcrorat Lime. A. Extended forearm and hand longer than the humerus; ulnare resembling that of the typical Neognathe ; radius and ulna long, with but a narrow chink between them; distal end of Me. II. flabelliform . . . . . . . Rhea. B. Extended forearm and hand much shorter than the humerus. a. Humerus with a shallow, but blind subtrochanteric fossa; a very strongly developed linea aspera for the triceps, forming a conspicuous triangular ridge; radius and ulna short and bowed; ulnare a simple nodule; distal end of Me. II. not markedly expanded; manus as in Rhea, resembling typical carimate form, though too small for purposes of flight . . . . Struthio. 6. Humerus much bowed, without subtrochanteric fossa or indication of pectoral crest ; radius and ulna with but a very narrow chink between them; a vestigial radiale and ulnare; a vestigial pollex bearing an unequal phalanx, with the Me. III. incomplete and free distally . . . Apteryax. C. Extended forearm and hand about equal to the length of the humerus; a subtrochanteric fossa ; radiale and ulnare anchylosed with metacarpals. a, Subtrochanteric fossa large ; ulnare indistinguishable ; radiale anchylosed with radius and carpo-metacarpus, but distinguishable. . . . . . . Cuaswarius. 6. Subtrochanteric fossa very small; no carpal elements distinguishable ; neither pollex nor third digit represented. . . . . . =. =. ..: . Dromeus. F, Pretvic Lie. When the bones are pneumatic only the femur is conspicuously so; save in the Dinornithide the extensor bridge is wanting; the hallux, when present, is never large. a. Toes three in number ; tarso-metatarsus far exceeding femur in length. a’. Ectocnemial crest large, rivalling the procnemial in size ; ento-condyle with a deep lateral pit; distal extremity of shaft above entocondyle (lateral surface) with a deep groove. a, Procnemial crest extending down the shaft; femur with a large (SATTEAG CNOA GG 0 6 6 6 Go oo 5 ol ee ol LDR UD, b!. Procnemial crest not extending down the shaft ; femur non-pneumatic; ungual phalanx of inner toe greatly elongated . . . . . . . . Casuarius. 4!. Ketocnemial crest small, a mere tubercle of the procnemial; no ento- condylar pit or groove on the lateral surface of the shaft ; a moderately conspicuous extensor groove; popliteal fossa of the femur very deep; ungual phalanges laterally compressed. «9. . . » = » 2 1-2.) Ried, PHYLOGENY OF THE PALAOGNATHEX AND NEOGNATEHA. 279 6. Toes two in number ; tarso-metatarsus far exceeding the femur in length. ce’. Ectoenemial crest small, a mere tubercle on procnemial ; fibula extending to level of proximal tarsal mass; extensor groove obsolete; a deep ento- condylar pit continued backwards into a groove; the lateral ventral border of ento- and ecto-condyles continued backward and upward to project spur-wise from the shaft . ce. Toes four; tarso-metatarsus short. d', Extensor bridge wanting. ce", Femur long and slender; tibio-tarsus with a deep intercondylar gorge ; entocondyle very large, projecting conspicuously in front of shaft in lateral view ; none of the phalanges conspicuously shortened d", Femur very short and thick, with enormous great trochanter; tibio- tarsus without intercondylar gorge, ento- and ecto-condyles not being clearly distinguishable: Ph. 2. D. I. short; Ph. 2, 3, 4. D. IV. short. e’, Extensor bridge present ; femur short ; ento-condyle of tibio-tarsus large, resembling that of Apterya List oF WORKS REFERRED TO OR CONSULTED. 1, Aux, P. H. E.—Syrinx of Rhea. Bull. Soc. Philomat. 1874, p. 38. Struthio, Apteryx. Aipyornis. Dinornithide. 2. Autis, T.—On the Skeleton of the Apteryx. Journ. Linn. Soc., Zool. xi. 1873, p. 523. 3. Anprews, C. W.—Note on a complete Skeleton of Apyornis from Madagascar. Geol. Mag. vol. iv. 1897, p. 241. 4. Anprews, C. W.—On the Skull, Sternum, and Shoulder-girdle of Apyornis. Ibis, 1896 p. 376. 5. Bepparp, F. E.—On the Heart of Apteryx. Proc. Zool. Soc. 1885, p. 188. 6. Bepparp, F. E.—On the Alimentary Canal of the Martineta Tinamou. pp- 61-66. Bepparp, F, E.—The Structure and Classification of Birds. 1898. comY . . vol. vi. 1899, pp. 386-402. eo] Ibis, 1890, vol. iit. Bepparp, F. E.—Notes on the Anatomy of the Genus Apteryx. Novitates Zoologice, g. Beppaxp, F. E. and Mrrcnetx, P. C.—On the Structure of the Heart of the Alligator. Proc. Zool. Soc. 1895, pp. 343-349. 10, Buumensacn, J. F.—Abbildung. naturhistor. vol. vii, 11. Borcxinc, A.—De Rhea americana. Diss. Inaug. Bonn, 1863. 12. Burcxuarpt, R.—Ueber Apyornis. Paleeontol. Abhandl. neue Folge, Bd. ii. Heft 2 (Jena, 1893). 13. Cunnincuam, R. O.—Notes on some Points in the Osteology of Rhea americana and Rhea darwini. Proc. Zool. Soc. 1871, p. 105. ZQy2, 280 14. 15s 16. C7 18, 1g. 20. MR. W. P. PYCRAFT ON THE MORPHOLOGY AND Cuvier, Baron G.—Lecons d’Anat. Comp. t. ii. 1799-1800 (Ann. vin.), p. 69. Datuas, W. S.—On the Feathers of Dinornis robustus. Proc. Zool. Soc. 1865, p. 265. Ducuamp, G.—Observations sur ’ Anatomie du Dromeus. Ann. Sci. Nat. xvi. 1873. Epwarps, A. Mitnr- and Oustater, E.—Notice sur quelques Espéces d’Oiseaux du Muséum WHistoire Naturelle. Centen. Mus. d’Hist. Nat. Paris, 1893 (l’Emeu ou Hmou noir, p-. 246). See also Mitwn-Epwarps. Evans, F. P.—Note on the Carotids of Rhea americana. Ann. & Mag. Nat. Hist. xi. 1884, p. 458. Frower, Sir W. H.—On the Skeleton of the Australian Cassowary, Casuarius australis. Proc. Zool. Soc. 1871, p. 32. Forses, W. A.—On the Conformation of the Thoracic End of the Trachea in the Ratite Birds. Proc. Zool. Soc. 1881, p. 778. 20A. Forpes, W. A.—On the Bursa Fabricii in Birds. Proc. Zool. Soc. 1877, pp. 804-318. 21. 22. 2s 24. 25. 26. Dy). 28. 29. 30. 3I. 32. 33: 34: 35: 36. 37: 38. 39: 40. Fremery, P. J.—Specimen Zoologicum sistens observationes, presertim Osteologicas de Casuario nove-hollandie. Fixsrincer, M.—Untersuch. zur Morphol. und System der Végel. Amsterdam, 1888. Gapow, H.—Zur vergleich. Anat. der Muskulatur des Beckens und hinteren Gledmasse der Ratiten. Jena, 1880. Gapvow, H.—Remarks on the Cloaca and on the Copulatory Organs of the Amniota. Phil. Traus. Roy. Soc. 1887, pp. 5-37. Gapow, H.—On the Anatomical Differences in the Three Species of Rhea. Proc. Zool. Soc. 1885, p. 308. Gapow, H.—Bronn’s Thier-Reich. Bd. vi. Vogel, 1891, Anatom. Theil. Gavow, H.—Ibid. Syst. Theil, 1893. Garrop, A. H., and Darwin, F.—Note on an Ostrich lately living in the Society’s Collection. Proce. Zool. Soc. 1872, pp. 356-362. Garrop, A. H.—On the Carotid Arteries of Birds. Proc. Zool. Soc. 1873, p. 457. Garrop, A. H.—On certain Muscles of the Thigh of Birds and on their Value in Classifica- tion: Part I. Proc. Zool. Soc. 1873, pp. 626-644. Garrop, A. H.—Ditto: Part II. Op. cit. 1874, pp. 111=128. Gnorrroy St. Hitarrn, E.—Compos. des Appareils génitaux, urinaires et intestinaux, d leurs points de rencontre dans l’Autruche et dans le Casoar. Mém. d. Hist. Nat. tome ix. 1822, p. 438. Gruser, W.—Ueber das Thranenbein der Straussartigen Végel. Bull. Ac. Sci. St. Pétersb. 1855, p. 161. Haventon, S.—On the Muscular Mechanism of the Leg of the Ostrich. Proc. Roy. Irish Acad. vol. ix. 1866. ! Havenron, 8S.—On the Muscular Anatomy of the Emu, Dromeus nove-hollandie. Ibid. 1868, pp. 487-497. Haventon, S.—On Muscular Anatomy of the Rhea. Ibid. p. 497. Hitprsranor, G. F.—Bemerkung iiber den Embryo des Strausses im Hie. Voigt’s Magaz. d. Naturkunde, Bd. xi. 1806, pp. 97-126. Hurron, IF. W.—The Moas of New Zealand. Trans. N. Z. Instit. vol. xxiv. Houxtey, 'T’. H.—On the Classification of Birds. Proc. Zool. Soc. 1867. Huxtey, I’. H—The Anatomy of Vertebrated Animals. 187 iy PHYLOGENY OF THE PALZOGNATHA AND NEOGNATH®. 281 - Huxiuy, T. H.—On the Respiratory Organs of Apteryx. Proc. Zool. Soc. 1882, p. 560. - Kyox, R.—Observations on the Anatomical Structure of the Cassowary of New Holland. Edinb. Philos. Journ. vol. x. 1824, p. 132. - Kwox, R.—Additional Observations on the Structure of the Trachea in the Cassowary Emeu of New Holland. Edinb. Philos. Journ. vol. x. 1823-24, p. 137. . Layxester, E. R.—On the Right Cardiac Valve of Echidna and of Ornithorhynchus. Proc. Zool. Soc. 1883, pp. 8-14, pls. iii., iv. . Lanxesrer, E. R.—On the Heart described by Prof. Owen in 1841 as that of Apteryx. Proc. Zool. Soe. 1885, pp. 289, 240. » Lanxuster, E. R.—On the Right Cardiac Valve of the Specimens of Apteryx dissected by Sir Richard Owen in 1841. Proc. Zool. Soc. 1885, pp. 477-482. . Livpsay, B.—On tiie Avian Sternum. Proc. Zool. Soc. 1885, pp. 684-716, pls. xlii—xly. . LypsekKker, R.—Cat. Foss. Birds, 1891. - Macazister, A.—On the Anatomy of the Ostrich. Proc. Roy. Irish Acad. 1867. . Macartney, J—An Account of the Organ of Voice in the New Holland Ostrich. Brit. Assoc. Rep. Bristol, 1836. . Mau, A. W.—Om Luftrér-sacken. (fv. K. Vet.-Ak. Férh. 1880, no. 7, p. 33. . Marsnatt, W.—Beobachtungen iiber das Verhiltniss der Federn. Zool. Gart. vol. xvi. 1875, p. 121. . Maver, A. F. J. C.—Analecten fiir vergleich. Anatom. 1839. . Mecxet, J. F.—System der vergleich. Anatom., Abtheil. 11. 1825, p. 202. - Mecxet, J. F.—Beitrige zur Anatomie des Indischen Casuars. Archiv f. Anat. u. Phys. 1830, p. 200; 1832, p. 273. . Merere, J. H. ne.—Ueber die Federn der Végel. Morphol. Jahrb. xxiii. Bd. 4, 1895, pp. 562-591. . Munzpier, M.—Vergleichende Osteol. des Penguine. Bull. Soc. Imp. Nat. Moscou, 1887. . Mitnn-Epwarps and Granp1pimr.—Observations sur les Aipyornis de Madagascar. Comptes Rendus Acad. Sci. t. exviti. 1894, p. 122. . Mitnz-Epwarps and Granpipier.—Sur les Ossements d’Oiseaux provenant des Terrains récents de Madagascar. Bull. Mus. Hist. Nat. pt. 1895, p. 9. See also Epwarps. . Mircuxztz, P. C.—On the Intestinal Tract of Birds. Proc. Zool. Soc. 1896, pp. 186-159. . Mircnett, P. C.—On so-called Quintocubitalism. Journ. Linn. Soc., Zool. vol. xxvii. . Mrvart, Sr. G.—On the Axial Skeleton of the Struthionide. Trans. Zool. Soe. x. p. 1. . Morte, J.—On the Tracheal Pouch of the Emu. Proc. Zool. Soc. 1867, p. 405. . Nasonov, N. V.—K istorii razvitiya Afrikanskagho Strausa. [On the history of the deve- lopment of the African Ostrich.] 3 Pt. [Published by the Zoological Section of the Warsaw University.] 1894-96. . Newton, A.—Dictionary of Birds (Introduction), 1896, p. 109. . Owen, Sir R.—Mewmoir on Anatomy of the Southern Apteryx. Trans. Zool. Soe. vol. 11. 1838. . Owen, Sir R.—On Dinornis. Trans. Zool. Soc. vu. p. 145. . Owen, Sir R.—Art. “ Aves,” Cyclopedia Anat. & Physiol, 1835-86. . Patmer, W.—Avifauna of the Pribilof Islands. Rep. Jur Seal Investigations, 1896-7, part 3. . Panper, C. H. and v’Atton, ..—Vergleich. Osteologie, Abtheil. 11. Lief. i. 1827. . Parker, T. J.—Observations on the Anatomy and Development of Apferyz. Phil. Trans. Roy. Soc. 1891, vol. 182. MR. W. P. PYCRAFT ON THE MORPHOLOGY AND . Parker, T. J.—Additional Observations on the Development of Apteryx. Phil. Trans. Roy. Soc. 1892, vol. 183. . Parker, T. J.—On the Cranial Osteology, Classification, and Phylogeny of the Dinornithide. Trans. Zool. Soc. vol. xiii. 1895. . Parser, T. J.—Text-Book of Zoology, vol. ii. 1897. . Parker, W. K.—On the Osteology of the Gallinaceous Birds and Tinamous. ‘Trans. Zool. Soe. vol. v. 1862. . Parker, W. K.—On the Shoulder-Girdle and Sternum. Ray Soc. Publications, 1868. . Parxer, W. K.—On the Structure and Development of the Skull in the Ostrich Tribe. Phil. Trans. vol. 156, 1866, p. 113. . Parxer, W. K.—On the Remuants or Vestiges of Amphibian and Reptilian Structure found in the Skull of Birds, both Carinate and Ratitz. Proc. Roy. Soc. vol. xlii. 1888. . Parker, W. K.—On the Morphology of a Reptilian Bird, Opisthocomus cristatus. Trans. Zool. Soc. vol. xii. 1891. . Parxer, W. N.—Note on the Respiratory Organs of Rhea. Proc. Zool. Soc. 1883. . Pycrart, W. P.—A Contribution to the Pterylography of the Tinamiformes. Ibis, 1895, p. lL. . Pycrart, W. P.—Contributions to the Osteology of Birds. Part II. Proc. Zool. Soc. 1898, pp. 958-989. . Pycrarr, W. P.—Some Facts concerning the so-called ‘‘ Aquintocubitalism ” of the Bird’s. Wing. Journ. Linn. Soc., Zool. vol. xxvii. pp. 236-256. . Raruxe, H.—Anatomisch-philosophische Untersuch. tiber den Kiemenapparat und das Zung- enbein der Wirbelthiere. Riga, 1832. . Remoucnames, E.—Sur la Glande gastrique du Nandou. Bull. Ac. Belg. t. 1. 1880, p. 114. . Removucuames, E.—Archives de Biologie, t. i. 1880, pp. 583-594. . Rotieston, G.—The Harveian Oration, 1873. Collected Scientific Papers and Addresses, vol. 11. p. 740. . Rotieston, G.—On Homologies of certain Muscles connected with Shoulder-Joint. Trans. Linn. Soe. xxvi. (1870), p. 609. . Rorascuitp, W.—The Genus Apteryxr. Nov. Zool. vol. vi. 1899, pp. 861-386. . Sappey, P. C.—Recherches sur l|’Appareil respiratoire de Oiseaux. 1847. . Scrater, P. L.—List of Vertebrated Animals. . Srrruine, HE. C. and Zurrz, A. H. C.—Preliminary Notes on Genyornis newtoni, &c. Tr. Roy. Soc. S. Australia, xx. 1896, p. 171. . Srrewine, i, C. and Zertz, A. H. C.—Description of the Bones of the Leg and Foot of Genyornis newtoni. Trans. Roy. Soc. 8. Australia, xx. 1896, p. 191. . Susuxin, P. P.—Zur Morphologie des Vogelskelets. Nouv. Mém. Soc. Impér. des Natu- ralistes, Moscou, xvi. liv. 2, 1899, pp. 1-163. . Van Bemmeten, J. F.—Onderzoek van een Rhea-embryo. Tijd. Ned. Dierk. Ver. 1886-87, p- cev. . Wepemeyer, G. H. L. C.—Der mit der Luftréhre des Emu Casoars in Verbindung stehende muskulose Beutel. Froriep’s Notizen, 1824-25, p. 7. . Wencxesacn, V. F.—De Ontwikkeling en de Bouw der Bursa Fabric. Inaug. Diss. Leyden, 1888. PHYLOGENY OF THE PALZOGNATHA AND NEOGNATHA. 283 98. Woopwaxp, A. S.— Vertebrate Paleontology, 1898. 99. Wray, R. S.—On some Points in the Morphology of the Wings of Birds. Proc. Zool. Soc. 1887, pp. 343-356. 100, Wunperticu, C.—Beitriige zur vergleichenden Anatomie und Entwickelungsgeschichte des unteren Kehlkopfes der Vogel. Nov. Act. Ac. Leop.-Carol. xlviii. 1884. 1o1. Zanprer.—Ueber das Gefieder des amerikanischen Strausses. Schrift. phys.-dkonom. Gesell. K@nigsb. xxix. 1889, p. 31. EXPLANATION OF THE PLATES. PLATE XXII. Casuarius casuarius. Common Cassowary (p. 113). (Drawn from life from a bird formerly at Tring, now living in Dublin.) PLATE XXIII. Casuarius caswartus beccarti. Beccari’s Cassowary (p. 116). (Drawn from life from a bird formerly at Tring, now in the Society’s Gardens.) PLATE XXIV. Casuarius casuarius salvadorii. Salvadori’s Cassowary (p. 120). (Drawn from life from a bird formerly alive at ‘Tring, now mounted in the Tring Museum.) PLATE XXV. Casuarius caswarius australis. Australian Cassowary (p. 123). (Drawn from life from a bird that lived ten years at Tring, and is now living in the Society’s Gardens.) PLATE XXVI. Casuarius casuarius violicollis. Mauve-necked Cassowary (p. 122). (Drawn from life from the type specimen formerly at Tring, now living in the Society's Gardens.) PLATE XXVIL. Casuarius casuarius intensus. Blue-necked Cassowary (p. 121). (Drawn from lite from the type specimen formerly at Tring, now mounted in the Tring Museum.) PLATE XXVIII. Casuarius bicarunculatus. 'Two-wattled Cassowary (p. 129). (Drawn from life from a bird formerly at Tring, now living in the Society’s Gardens.) 284 MR. W. P. PYCRAFT ON THE MORPHOLOGY AND PLATE XXIX. Casuarius unappendiculatus. One-wattled Cassowary (p. 132). (Drawn from life from a bird in the possession of Mr. Blaauw.) PLATE XXX. Casuarius unappendiculatus. One-wattled Cassowary (p. 132). Whole birds, adult and young. (The figure of the adult was made from Mr. Blaauw’s bird; the young was drawn from a bird at Tring, the skin of which is preserved in the Tring Museum.) PLATE XXXI. Casuarius unappendiculatus occipitalis. Jobi Island Cassowary (p. 135). (The head, neck, and naked parts were drawn from a bird now living in the Society’s Gardens, the casque from the mounted specimen in the Paris Museum.) PLATE XXXII. Casuarius unappendiculatus aurantiacus. Kastern One-wattled Cassowary (p. 136). (Drawn from life from the type now living in the Berlin Zoological Gardens.) PLATE XXXII. Casuarius philipi. Sclater’s Cassowary (p. 138). (Drawn from life from the type now living in the Society’s Gardens. ) PLATE XXXIV. Casuarius papuanus. Westermann’s Cassowary (p. 139). (Drawn from life from two specimens formerly at Tring and the Society's Gardens, now preserved in the Tring Museum.) PLATE XXXV. Casuarius papuanus edwardsi. Milne-Kdwards’s Cassowary (p. 141). (Drawn from life from a specimen formerly at Tring, now mounted in the Tring Museum.) PLATE XXXVI. Casuarius picticollis, Painted-necked Cassowary (p. 143). (Drawn from the plate in Gould’s ‘ Birds of New Guinea.’) PLATE XXXVILI. Casuarius picticollis hecki. Heck’s Cassowary (p. 144). (Drawn from life from the type specimen living in the Berlin Zoological Gardens.) PHYLOGENY OF THE PALZOGNATHA AND NEOGNATHA. 285 PLATE XXXVIII. Casuarius lorie. Loria’s Cassowary (p. 142). (Drawn from a sketch made from a fresh-killed bird by Dr. Loria.) PLATE XXXIX. Casuarius bennetti. Bennett's Cassowary (p. 145). (Drawn from life from a bird formerly at Tring, now in the Zoological Society’s Gardens.) PLATE XL. Map showing the geographical distribution of the six subspecies of Casuwarius casuarius and of Casuarius bicarunculatus. PLATE XLI. Map showing the geographical distribution of the three subspecies of Caswarius unappendiculatus, two of Casuarius papuanus, two of Casuarius picticollis, Casuarius lorie, and Casuarius bennetti. PLATE XLII. The dorsal aspect of the Skull (figs. 1-4). Fig. 1. The skull of Rhea americana (p. 270), to show the form of the temporal fossa, the partly fused mesethmoid plate, the form of the lachrymals and _nasals, and the broad nasal process of the premaxilla. . The skull of Dromeus nove-hollandie (p. 270), showing the rod-like nasal process of the premaxilla, the fused lachrymal, broad interorbital region, and Les) ore | w) widely separated temporal fossze. . The skull of Struthio molybdophanes (p. 269), showing the supra-orbital ossicles running backwards from the lachrymal to the frontal, and the broad nasal process of the premaxilla. The mesethmoidal suture is obliterated. Fig. 4. The skull of Apterya australis (p. 269), showing the small size of the lachrymal and the absence of its supra-orbital processes, the complete fusion of the = or gs [Se nasal process of the premaxilla, and the rounded form of the frontals, which do not protect the orbit above. Ventral aspects (figs. 5-8) Fig. 5. The ventral aspect of fig. 1, showing the large size of the palatine processes of the premaxilla and of the maxillo-palatine processes. The peculiar relations between the vomer, palatine, and pterygoid (described on p- 207), and the anterior end of the parasphenoidal rostrum between the anterior cleft of the vomer. VoL. XV.—PaRT v. No. 23.—December, 1900. QR 286 MR. W. P. PYCRAFT ON THE MORPHOLOGY AND Fig. 6. Ventral aspect of fig. 2, showing the broad vomer and its continuance backwards on to the pterygoid. The palatines articulate with the external border of the foot of the vomer and external border of the pterygoid. In Rhea the articulation between pterygoid and vomer is concealed by the palatine. articulate with the pterygoid. 7. Ventral aspect of fig. 3, showing the vestigial vomer, which does not now Note the absence of the palatine processes of the premaxille. The palatine has fused with the external pterygoid border. Fig. 8. Ventral aspect of fig. 4, showing the complete fusion of the vomer, pterygoid, and palatine. ADDITIONAL LETTERS. ant.b,f. =anterior basi-cranial fontanelle. ant.h.=antrum. b.6.p.= basitemporal platform. b.p.=basipterygoid processes. hy for.= hypoglossal foramina, 1, 2, 3. !.=lachrymal. Lp.=lachrymo-nasal pillar. ir. =lambdoidal ridge. m.e.=mesethmoid. mz.p.=maxillo-palatine process. n.pmx.—=nasal process of premaxilla. n.=nasal. 0.c.=occipita] condyle. pa.=palatine. pa.r.=parasphenoidal rostrum. p.pmx.=palatine process of premaxilla. p-p.=paroccipital process. pt.=pterygoid. s.=supra-orbital bones from lachrymal to frontal. v.=vomer. vag for.=vagus foramen. PLATE XLITI. The immature Skull. Fig. 1. Dorsal aspect of the palatine bones of Apteryx mantelli (p. 204), to show the sutures, as yet unclosed, between the vomer, pterygoid, and palatine. Fig. 1a. Ventral aspect of the same, to show similar sutures. Fig. 2. Sagittal section of skull of Apteryx australis mantelli, after Parker, for the purpose of comparison with Dromeus. Note the appearance of the squamosal on the inside of the skull in Apteryz. i) ey oe) . 8. Ventral aspect of the skull of Emeus (p. 270), for the purpose of comparison with Rhea. Note the similarity in the relations of the pterygoid and vomer. Fig. 4. Sagittal section of cranium of embryo Dromewus nove-hollandie, to show the separate bones. Fig. 4a. Outer view of fig. 4. Note how completely the squamosal conceals and protects the pro- and opisthotic bones. In the nestling Penguin the squamosal is deeply emarginate posteriorly and exposes the otic bones. Fig, 5. Ventral aspect of base of skull of Ahea americana, to show sutures, PHYLOGENY OF THE PALZOGNATHA AND NEOGNATHA. 287 g. 5a. Lateral aspect of skull of Rhea americana, to show sutures, Note the vertical spine (lachrymo-nasal pillar) from the maxillo-palatine process, which replaces the maxillary process of the nasal; the remarkably small size of the quadrato-jugal; and the outer and inner views of the lower jaw. . 56. Lateral aspect of pterygoid of Rhea americana (p. 191), to show its relations with the vomer. . 5c. Outer aspect of mandible of fig. 5, to show sutures. . 5d. Inner aspect of mandible of fig. 5, to show sutures. g. 6. Quadrate of Bubo maximus, to show double, widely separate, squamosal articular head. Nat. size. Fig. 7. Quadrate of Dromeus irroratus?, posterior aspect, to show single articular head. Nat. size. Fig. 8. Quadrate of Apteryx owent, posterior aspect, to show double squamosal head. ADDITIONAL LETTERS. a.=angulare. o.h.=otic head of quadrate. als.=alisphenoid. o.r.=orbital process of quadrate. ar, =articulare. p.=parietal, ant.=antrum of Highmore. pf.=precondylar fossa. b.s. = basisphenoid. pro. =pro-otie. d,=dentary. $.a.=supra-angular. cor.=coronoid. $.0.=supra-occipital. ep.o.=epiotic. sp.=splenial. ex. =exoccipital. sq.=squamosal. ff.=floccular fossa. sq.h.=squamosal head of quadrate. fr.=frontal. vag.f.=vagus foramen. m.=meatus internus. vo.pr.=vomerine process of premaxilla, mes. =mesethmoid. l.p.=lachrymo-nasal process of premaxilla. op. = opisthotic. PLATE XLIV. Fig. 1. Dorsal aspect of the skull of an adult Dromeus ater (p. 270), to show the conspicuous, embryonic, cerebral prominence and the fused lachrymals. . la. Lateral view of the same skull (fig. 1), showing the cerebral prominence and temporal fossze. Fig. 2. Dissection of a skull of a young Dromeus nove-hollandia, lateral view. to show the form and position of the mesethmoid for comparison with fig. 3 et seq. Fig. 3, The mesethmoid of a young Casuarius unappendiculatus (p. 199), to show the mesethmoid and the large median superposed element forming the centre of the casque of the adult. ‘ig, 4a, MR. W. P. PYCRAFT ON THE MORPHOLOGY AND ig. 4. Lateral aspect of a portion of the skull of Casuarius sclatert, showing the frontal, nasal, and median elements of the casque. Dorsal aspect of fig. 3 a. ‘ig. 4b. Dissection of a skull of Casuarius sclateri=fig. 3 a, showing the mesethmoid and a portion of the ventral border of the median ossification of the casque interpolated between the mesial border of the anterior ends of the frontal. Fig. 5. Lateral view of the synsacrum of a young Casuarius wnappendiculatus after removal of the right innominate, to show the great height of the neural spines of the vertebree. Fig. 6. Lateral view of the synsacrum of an embryo Dromeus nove-hollandie (p. 212), prepared for comparison with fig. 5. Note that at this stage the neural spines have not yet risen above the centra of the vertebre, but form only a series of lateral wings. ‘ Fig. 6a. Dorsal aspect of a portion of the synsacrum (fig. 6), to show more clearly that the neural spines do not rise above the centra as in the adult condition (fig. 5). Fig. 6 6. Lateral view of the same pelvis (fig. 6), to show the separate elements of the innominate. ADDITIONAL LETTERS. ant. = anti-trochanter. o.p.fr.= orbital process of frontal. jf.c.=frontal moiety of casque. | p-=pubis. i. =ilium. p.p.=pectineal process. il.is.f.=ilio-ischiadic foramen. pubo-isch.f.= pubo-ischiadic fissure. is. =ischium. s.c.=sacral vertebra 1-2. /b. \-6=lumbar vertebra 1-6. s.cd. =sacro-caudal vertebree. pyaar lb, \-2=lumbo-sacral vertebrie 1-2. s.th.=sacro-thoracic ,, ; EESTI me.c.= median element of casque. v. = vestigial thoracic rib. n.c.=nasal element of casque. v.me.=veutral border of mesethmoid. vb.=obturator fissure, PLATE XLV. Fig. 1. The buccal cavity of Rhea americana (p. 248), to show the large choane characteristic of the Palwoguathe. Fig. 2. A neossoptile of Calodromas elegans (p. 162), after Pycraft, to show the large aftershaft, equalling the main shaft in size, the complex rami of the main shaft, and their production forwards into horny filaments resembling the grooved ribbon-shaped bands of the young Ostrich (fig. 7). Fig. 2a. A radius from the distal end of the main shaft. PHYLOGENY OF THE PALAOGNATHAZ AND NEOGNATH &. 289 Fig. 8. A neossoptile of Rhea americana (p. 164): neither the main nor the aftershaft ig. 10. ealull, , te, have a distinct rhachis. The main shaft is represented by three stiffened rami, produced into filaments as in Tinamous and Ostrich, the aftershaft by - numerous slender rami. ig. 3a. A portion of a radius from fig. 3 (p. 164). Note the absence of fila. e. 4, A neossoptile of Dromeus nove-hollandiev (p. 162). The aftershaft is barely distinguishable ; the main shaft well developed: compare with figs. 3-7. . 4a. A portion of a radius of fig. 4. Fila are yet represented but are small. ge. 5. A portion of a radius of Casuarius casuarius, showing the still further suppression of the fila. . 6. A neossoptile of Apterya mantelli (p. 163). There is no aftershaft. e. 6a. A portion of a radius of fig. 6. Note that fila are numerous and large, indicating a less degenerate condition than figs. 4a and 5. .7. A neossoptile of Struthio camelus, showing the three peculiarly modified rami described on p. 164, and representing the main shaft, and the numerous downy rami representing the aftershaft. iad . Ta. A portion of a radius of fig. 7. These are flattened as in Crypfuri (fig. 2). Note the “ eye-spots.” . 8. Dissection of the air-sacs of the left side of a Casuarius unappendiculatus (p. 252), showing the three divisions of the anterior intermediate sacs and the positions of the ostia throughout. Note the position of the anterior end of the posterior abdominal air-sac—forcing down the roof of the posterior inter- mediate. The posterior abdominal, furthermore, is seen to take the form of a‘more or less sausage-shaped sac adherent to the inner wall of the posterior intermediate. It encloses a part of the kidney. g. 9. Dissection of the left lung of Casuarius unappendiculatus (p. 252), to show the costo-pulmonary muscle, meso-, ecto-, and ento-bronchia and their ramifications. Note that the ento-bronchia 1 and 3 between them supply all the tubule of the ventral surface of the lung described on p. 252. The systems fed by ento-bronchium 1 have been exposed by dissection, those fed by ento-bronchium 2 lie caudad of these and are concealed by the aponeurosis. Ventral aspect of a dissection of the wing of an adult Casuarius casuarius, to show the hypertrophied calami described on p. 165. Dorsal aspect of a preparation of the wing of an Ostrich. After Wray. . Dorsal aspect of a portion of the wing of Rhea americana (p. 156), to show the remarkable distribution of the coverts in transversely oblique bundles. Ventral aspect of the wing of Apteryx mantelli (p. 161), to show the number of remiges and their remarkable resemblance to the young Cassowary described on p. 165. VOL. Xv.— PART v. No. 24.—December, 1900. 25 29) ON THE MORPHOLOGY ETC. OF THE PALZOGNATHA AND NEOGNATHA. Fig. 14. A remex of a nestling Casuarius australis, dorsal aspect, before the loss of the rhachis described on p. 165. Fig. 14a. Ventral aspect of fig. 14, to show the peculiar rhachis. Fig. 14}. Section of the rhachis of figs. 14-14 a (p. 165). Fig. 15. A remex of Apteryax mantelli (p. 166), for comparison with that of the young Cassowary described p. 165. ADDITIONAL LETTERS. a. =aftershatft. a.int.1-3 =anterior intermediate air-sacs 1-3. a.s.=ala spuria. ap.=aponeurosis. c.=calamus. c.c,=carpal covert. ch.=choane. cl.=claw. ¢.p.m. = costo-pulmonary muscle, e.r.= carpal remex. c.r.=cubital remiges. *—=hypothetical carpal covert. OER a REELEINC Xs e.=eye-spots. ect.=ecto-bronchia. ent. =ento-bronchia. eu. = eustachian aperture. f.=filum. h.=heart. k.=\kidney. mb. =meso-bronchium. m.d.r.=metacarpo-digital remiges. o.=ostium posterius. obl.sep.= oblique septa. o.ent. 3=ostia of ento-bronchium 3. o.mb, 1-2=ostia of outer and inferior lateral branches (1-2) of meso-bronchium. o.sub.=ostia leading into sub-bronchial sac. DIMI 5p pre-bronchial sae. p.0.=pre-bronchial air-sac. 2? d2 p.int.= posterior intermediate air-sac. p.abd. = posterior abdominal air-sac. p.a.=pulmonary artery. para.b.=parabronchia. p.v.=pulmonary vein. rd.=yradius. rh.=rhachis. rh.""=rhachis of aftershaft. rm. =ramus. r.r.=ribbon-shaped prolongations of rami. r.g.=rima glottidis. t.=tongue. i.f.=terminal filaments of the rami. t.mj.=tectrices majores. tmd= *,, medi. i — minores. tes — sae marginales. 3~——==passage from ento-bronchium 1 to vestibule. 4.2 =traverses entire length ento- bronchium 1 to vestibule. 5» =passes from ento-bronchium 1 to vestibule. 6=—» =traverses whole length ento- bronchium 4. traverse lateral meso-bronchia into inter- anterior and posterior mediate air-sacs. Troamd,Look. Soo Wb XV GC ALY J.G.Keulemans del. et lth. Mintern Bros.imp. CASUARIUS CASUARIUS Trans. look. foc Vol, XV.GC_ AMM. J.G.Keulemams del.et hth. NGntern Bros.imy CASUARIUS -CASUARIUS 1 CU T) ie) O i rr U a re 1 c ben real clases? ay Trans. Look, See Vol AV IG ARXIV J.G.Kewlemans delat lith. MinternRros.ime CASUARIUS CASUARIUS SALVADORILI TS ra Tranb Lock, Soo Vl AV GICAL. Mintern Bros.imp. J.G.Keulemans del.et hth RIUS CASUARIUS AUSTRALIS ee aay Peery ey. Trans, 006, fo VE, XV. FE, AXVI. Mintern Bros.imp J.G.Keulemans del.et lth dp) CASUARIUS CASUARIUS VIOLICOLLI J.G.Keulemans del. et ith CASUARIUS RIUS INT! } Ve vi et Trans. Lock, Soe Met NV GG ANV UM. J.G.Keulemans delet lth Mintern Bros.imp CASUARIUS BICARUNCULATUS ic Pee fi ek Lae { mel Le an fen ant i am? Ch ER : i m % i Ate 5.G.Keulemans del,et hth. Trams. Toot foe Vol XV..G0. KATE. CASUARIUS UNIAP ENDICULATUS VV B Mintern Bros.imp J.G-Keulemans del, et ith CASUARIUS UNIAP EN NG Na Ltern Trans.Lool. See VL XV. FO_AXAT. J.G.Keulemans del et hth Mintern Bros.imp CASUARIUS UNIAPPENDICULATUS OCCIPITALIS Trans. Lool Soo. Ul IVIG AXED. 9 wy 5 J.G.Keulemams del.et kth : Mintern Bros.imp CASUARIUS UNIAPPENDICULATUS AURANTIACUS. Trans. Lol. Sco Vb XV. JC. XXXM. J.G.Keulemans del et hth Mintern Bros.imp CASUARIUS PHTEDPT: Trand.Locl. Soc Vil, XV. FC. NAAT. J.G Keulemans asl.et lth MmternBros.imp. CASUARIUS PAPUANUS. Lob, Soo Nob NV. FC. ANY. J.G Keulemans del. et hth CASUAT ROUSING) S) EDWA Mintern Bros.imp Trans.oot. SooM AV FU ANLV Oo) J 5 CASUARIUS Sle Le Oi Rs. J.G.Keulemans delet hth Mintern Bros. imp bi tet ‘ ta meee Ne ! ie its Trans. Lo0b, Soc Vol XV.IG, XAXVE. Mintern Bros.imp JG Keulemans del. et lith CASUARIUS PICTICOLLIS HECK. Trond A lool, ; Se LC MEL. Zi i AY AAAVT J.G Keulemans ) az aio oe 5 del et hth Mintern Bros.imp > pan) Sr ens 00b, Soo WH XV. FO NANTY T’SHIYVMOSSVO HHL AO NOILNEAIYLSIa *STYFOYMIUNADAG QF) ST O-ISTD STLLONSD) “) [) “WAP PS STONED) “9 [ TOP DATOS STILEONSO * [Ta] *81]007]010 sTn.LONS -) ALDODdG STLIDNSP) -) [Fa STL) a UOPUOT GUST 60a] SPI1OfuOze Wo muvesy JO SLT SEL NE Al NEE A ES JIPD UOOH \ pn e a uopuoT qn s7, boag spuopunje YImMuvess) JO Seq AHAE Ted O° ¢ | eg, o & oA Ge ‘ 4 : YUuIAle “sq Speccmpy : —— es ost TIX to AK OVOP PPP Le PMOL Srans. Zoot. Soc. Vol. XV Fl. XL. mxp. Se db Daniel Ltd Uneee STRUCTURE OF THE PALMOGNATHE, Photoprint by Bale Danielsson Spans Fok. Soe. Vol XV PE NLL | a | : ag Ve Me A | y by hy. | ( A j / J. eth.py.sphi | | ge vt. vag f. 3 Py. AERESPE oe 1 Z | | spl. _ss.duig. prsphif’ \ Sa ones ear ae meth.’ “ Cor . | o Cae aN bon sy \i7l.17t. ) Lab = — a ee a | ‘nV0.0. = | par. 4 ob.s. bo. a | | Sq. Photoprint by Bale a Danielsson Ltd! H.Grénvold. del STRUCTURE OF THE PALMOGNATHSA . | Srands. Loot. Soe. Vol. XV. Fl. XLV. $.c. a A hae vie ql ‘ z Photoprint by Bale «Danielsson Lté T H.Grénvold. del SMMENOT GAMO MSE (Ost ARISE Ish PALMOGNATHA. / Trams. Look. Soc. Vok. XV Fé XIV OLN. | ~h obl.sep. 0.mb.1-2. par.b ; t.ms. im. tmd. STRUCTURE OF THE PALAOGNATH., Photoprint by Bale «Danielsson Lid E H.Gronvold. del. TRANSACTIONS OF THE ZOOLOGICAL SOCIETY OF LONDON (continued). VOLUME XIII. (1891-1895, containing 62 Plates) . Partl. (1891, containing VOLUME XIV. (1896-1898, containing 47 Plates) 2. (1891, containing 3. (1891, containing 4. (1892, containing 5. (1893, containing 6. (1893, containing 7. (1898, containing 8. (1894, containing 9. (1894, containing 10. (1895, containing 11. (1895, containing Part l. (1896, containing 2. (1896, containing . (1897, containing 6 Plates) 6 Plates) 6 Plates) 1 Plate) . 9 Plates) 4: Plates) 6 Plates) 6 Plates) 6 Plates) 5 Plates) ° ° . . ” 7 Plates and Title and Index) ,, 2 Plates) 6 Plates) 9 Plates) . (1897, containing 10 Plates) . (1898, containing 3 4, 5. (1898, containing 6 7. (1898, containing 8 . (1898, containing VOLUME XV. Part 1. (1898, containing ” »? 2a %% 2. (1899, containing 3. (1899, containing 4. (1899, containmg 2 Plates) 7 Plates) . 6 Plates) . 5 Plates and 8 Plates) 5 Plates) 4 Plates) 4 Plates) 5. (1900, containing 24 Plates) 9 2 22 eb} Title and Index) ,, To Fellows. Gass Price 6 8 0 15 ll CS) MOO Or SL ta OS eS 15 Price 5 0 0 1 0 6 0 0 0 Price 0 15 22 22 2 2 (0) 09 (0) 2) 2 12 d 3 9 9 0 6 3 0 0 0 3 0 9 (aa) (8S) fe) (Ss) Aol ey) XS) Ooocaes To the Public. 3 1 (ey ee) (0) Mel (oe eye Oooo eH Ss. Il 1 a) ee COU OT NS OO Col Siete OO Re 1) a we — ~w © | oS d. Se) tea (SY ehn E) a Qe o Seo OOM Omori Onmicre, fe) SWISS sy, fm) CONTENTS. VI. A Monograph of the Genus Casvarius. By the Hon. WALTER Rorascuitp, Ph.D., PZS. With a Dissertation on the Morphology and Phylogeny of the Pauao- anaroa (Rarira and Crypturi) and NwoenaTia (Cartnata). By W. P. Pronamt.. (Plates*XXIE-XLV.). - 2 ee ee page 109 THE PUBLICATIONS OF THE ZOOLOGICAL SOCIETY OF LONDON. {ue scientific publications of the Zoological Society of London are of two kinds—“ Proceedings,” published in an octavo form, and “ Transactions,” in quarto. According to the present arrangements, the “ Proceedings” contain not only notices of all busi- ness transacted at the scientific meetings, but also all the papers read at such meetings and recom- mended to be published in the “ Proceedings” by the Committee of Publication. A large number of coloured plates and engravings are attached to each annual volume of the “ Proceedings,” to illustrate the new or otherwise remarkable species of animals described in them. Among such illustrations, figures of the new or rare species acquired in a living state for the Society’s Gardens are often given. ; The “ Proceedings ” for each year are issued in four parts, on the first of the months of June, August, October, and April, the part published in April completing the volume for the preceding year. — The “Transactions ” contain such of the more important communications made to the scientific meetings of the Society as, on account of the nature of the plates required to illustrate them, are better adapted for publication in the quarto form. They are issued at irregular intervals. Fellows and Corresponding Members, upon payment of a Subscription of £1 1s. before the day of the Anniversary Meeting in each year, are entitled to receive all the Society’s Publications for the year. They are likewise entitled to purchase the Publications of the Society at 25 per cent. less than the price charged for them to the Public. A further reduction of 25 per cent. 1s made upon purchases of Publications issued prior to 1871, if they exceed the value of five pounds. Tellows also have the privilege of subscribing to tle Annual Volume of the ‘ Zoological Record ” for a sum of £1 (which includes delivery in the United Kingdom only), but this privilege only holds good if the subseription is paid before the First of December in each year. Such of these publications as are in stock may be obtained at the Society’s Office (8 Hanover Square, W.), at Messrs. Longmans’, the Society’s publishers (Paternoster Row, H.C.), or through any bookseller. Pecember 1900. 1 Le SCLATER, secretary, TRANSACTIONS OF THE ZOOLOGICAL SOCIETY OF LONDON. VoL. XV.—Parr 6. LONDON: PRINTED FOR THE SOCIETY, SOLD AT THEIR HOUSE IN HANOVER-SQUARE ; AND BY MESSRS. LONGMANS, GREEN, AND 0O,, PATERNOSTER-ROW. April 1901. Price 16s. Taylor and Francis Printers, ] [Red Lion Court, Fleet Street. TRANSACTIONS OF THE ZOOLOGICAL SOCIETY OF LONDON. To Fellows. To the Public. L's. ds i AS. ds VOLUME I. (1833-1835, containing 59 Plates). . Price 313 6 . . . 418 O* VOLUME | II. (1835-1841, containing 71 Plates). . 5 400... 5 6 6* VOLUME III. (1842-1849, containing 63 Plates). . , 38 8 3 . . . 4 11 O* VOLUME IV. (1851-1862, containing 77 Plates). . ,, 6 2 0... 8 2 6+ VOLUME _ V. (1862-1866, containing 67 Plates). . , 5 4 3 . . . 619 OF VOLUME VI. (1866-1869, containing 92 Plates). . ,, ll 5 O .-15 0 OF VOLUME VII. (1869-1872, containing 73 Plates). . ,, 10 4 0. . .1812 OF VOLUME VIII. (1872-1874, containing 82 Plates). .,, DB 8 os ar ce eee Os ae VOLUME IX. (1875-1877, containing 99 Piates)iti, srs lite) Oo Oe 2 Ort VOLUME X. (1877-1879, containing 95 Plates) . 3 10 0 8 ed 7- 207 GENERAL INDEX, Vols. I. to X. (1833-1879) 6 Ol cog to nO) 0) (0) VOLUME XI. (1880-1885, containing 97 Plates). ._,, DAs Ora ele 6. 0 VOLUME XII. (1886-1890, containing 65 Plates). . ,, OS Oe ey. 4 0 * No copies of these volumes remain in stock. + Only complete copies of these volumes are left in stock. Continued on page 3 of Wrapper. feo) VII. Notes on the Transformations of some South-African Lepidoptera. By Lieut.-Colonel J. Maucotm Fawcerr }. Received October 17, 1899, read February 6, 1900. [Puates XLVI.-XLIX.] THE following paper is the result of a collection and observations made during a residence of a year at Ladysmith and Pietermaritzburg, Natal, during which particular attention was paid to the habits and early stages of the Lepidoptera of the district. CONTENTS. I. Remarks on the Climatic and Natural Conditions of Natal: p. 291. II. Observations on the Rhopalocera: p. 292. III. Observations on the Heterocera: p. 302. IV. Explanation of the Plates: p. 319, T.—RemMarkKS on THE Curmmatic AND Natural Conpitions or NATAL. The climate may be described in short as one marked by extreme and rapid changes of temperature, except in what is known as the “Coast Belt.” On the higher plateaux, about 3000 feet, a change of wind from west to east has, in the experience of the writer, caused a sudden drop of 22° of the thermometer in one hour; while in the colder months (June to September) the range between day- and night-temperatures is frequently as much as 70°. Such rigorous conditions, taken in conjunction with the want of covert, may account for the extremely poor butterfly-fauna of the higher levels. The rainy season, here coincident with the hotter months, usually begins in November and lasts till the middle of March, and is characterized by exceptionally severe thunderstorms and heavy sea-fogs and rains from the south-east ; during the remainder of the year the weather is dry, the days being warm and the nights very cold. Natal may be divided, for the purpose of these Notes, into three well-defined regions or zones, as the country rises in successive plateaux inland :— 1. The Coast Belt or Zone, from sea-level up to 1000 feet elevation. (Tropical Zone.) 2. The Middle Zone, from 1000 to 5000 feet elevation. (Subtropical Zone.) 3. The High Levels; 5000 feet elevation and over, (Temperate Zone.) ? Communicated by R. Trimen, F.R.S., F.Z.8. VOL. XV.—Part VI. No. 1.—April, 1901. 20 prri, 292 LIEUT.-COL. J. M. FAWCETT ON THE 1. The Coast Zone, in which Durban is situated, has a tropical climate, and abounds in thick jungles or “ bush ” (as it is called in South Africa) of virgin forest, the open spaces being mainly under cultivation for tea, pine-apples, sugar-cane, and bananas. In this district Bellevue, some 12 miles out of Durban, and Pinetown are good localities for Butterflies, while a fine tract of virgin forest, called the ‘* Berea,” extends to the very outskirts of the city of Durban; but, as a rule, all the bush in this zone is full of Butterflies and Moths of species characteristic of the tropics. 2. ‘The Middle Zone comprises (for the purpose of these Notes) the Maritzburg district. The country here is much wore open, and consists of grassy downs with here and there a little mealie cultivation, and a few plantations of wattle (Llephantorrhiza burchellii) and blue gum (/ucalyptus globulus), imported from Australia. About 70 miles inland a plateau runs north and south, parallel to the sea, and attaining a height in some places of 5000 feet : on the eastern slopes of this range are the only coverts of virgin forest in this zone; they are known as the “Town Bush ” (above Maritzburg) and “ Karkloof” (about twenty miles north of it). In these coverts alone are the rarer species of Rhopalocera to be found: the commoner species, the larvee of which feed on more or less cultivated plants, such as orange, lemon, passion- flower, vine, oak, &c., being found in Maritzburg and its vicinity. It is remarkable, however, how many larve in Natal feed on imported plants. 3. The High Levels comprise the country above the 5000 feet plateau above alluded to up to the ‘Transvaal and Orange Free State. It mainly consists of open grassy downs (called “ veldt” by the Boers), with here and there rocky knolls (‘‘ kopjes” or ‘“‘koppies ”) covered with large boulders, entirely devoid of timber, with the exception of a few imported blue gums planted round the farm homesteads ; from this district, however, came but few of the forms mentioned below. 1I1.—OBSERVATIONS ON THE RHOPALOCERA. Most of the species which have been observed are double-brooded, the larvee of the first brood appearing in November and December, changing to pupe in January, and the imagines appearing early in February; while the second brood is met with in March, and remains in the pupal stage till the following October. ‘Those which are single-brooded generally appear as imagines in October. In the case of species which have already been figured and described, merely the writer’s experiences and observations in rearing them are recorded, and- the reader is referred to the published descriptions and figures. Mr. J. F. Quekett, Curator of Durban Museum, and Mr. W. Haygarth, of Natal Government Railways, have rendered much valuable assistance in the preparation of these Notes, and have presented many of the larve described to the writer. TRANSFORMATIONS OF SOME SOUTH-AFRICAN LEPIDOPTERA. 293 Mr. J. Medley Wood, Curator of the Durban Botanic Gardens, has kindly identified the various food-plants. The larve are figured natural size, except where stated otherwise. RHOPALOCERA. Family NYMPHALIDE. Subfamily DaNaAIN& (Bates). 1. Danais cHrysippus (Linneus). The larva and pupa of this well-known species have been already figured by Mr. Trimen, in ‘ Rhopalocera Africe Australis,’ plate i., 1862. The South-African larva seems to differ considerably from De Nicéville’s description of the Indian form in ‘ Butterflies of India, Burma, and Ceylon,’ where only “ two large yellow patches” on the second and third somites are mentioned, and nothing is said of the crimson bases to the “ dark retractile (?) horns.” The larve reared in Maritzburg fed on “ Gomphocarpus fruticosus,’ a common plant in marshy places, and also on the ‘ veldt” in Natal; on being cut the plant exudes a bitter white milky substance, which may account for the bitter taste of the imago, which is said to be its protection, ‘This larva is much infested by the maggots of a very small hymenopterous insect, and out of many examples taken, only two survived to hecome imagines. I have heard similar experiences detailed by other collectors. The imago is, however, common all over Natal, and is the only butterfly to be seen on the high levels during the colder months. I have noticed the alcippus, but not the dorippus form in Natal. Subfamily AcR a#1N & (Bates). 2. AcR#A ACARA (Hewitson). The larva and pupa of this species have been already figured by Mr. Trimen in ‘South African Butterflies,’ i. plate i. figs. 1, 1 a. Numerous examples of this larva were reared in Maritzburg on a species of Passi- flora, which is a common plant on the verandahs of houses. ‘The period passed in the pupal stage in January was only 11 days. Some of these larve were placed in a box in which some of the same species had already changed into pups, and the Jarvie 272 294 LIEUT,-COL. J. M. FAWCETT ON THE proceeded to devour the pup, although there was plenty of the food-plant in the box. This is the only species in which I have noticed this carnivorous habit. I have not obtained this species in the coast districts. 3. ACREA PETR#HA (Boisduval). (Plate XLVI. fig. 1, larva; figs. 2, 38, pupa.) Larva. Ground-colour pale golden brown, with a dorsal and lateral black lines, and a black transverse line on each segment bearing two largish white spots and six long branched black spines, those on 3rd, 4th, and 5th segments being longer than the remainder. Head large proportionately to body, black with a white bifid mark on front. ‘Thoracic legs and claspers yellowish. The young larve reared were all blackish in colour, and fed in companies on Oncoba kraussiana (Planch.). Pupa appears to be dichromatic, some being pale brown and others ferruginous ; in both forms the fine black lines and spots peculiar to Acrwa pup are much reduced. The pupal stage lasted 15 days in January. ‘The imago is common in the Durban and Maritzburg districts. 4. ACR@A ENCEDON (Linneus) = A. lycra (Fabricius). (Plate XLVI. figs. 4, 4, larva ; fig. 6, pupa.) Larva, Slaty black, with a yellow lateral line above prolegs and claspers. On each segment three deep black fine transverse lines enclosing two white patches dorsally and two yellow patches laterally. On the centre black line of each segment are placed six black spines (branched). Head, thoracic legs, and claspers black. Pupa waxy white, with the usual fine black lines on the wing-covers and black spots with orange centres on the abdominal segments. The larva figured was not fully grown. It became one-fifth longer and propor- tionately thicker before changing into a pupa. It feeds on a species of Commelina, a common weed in gardens at Maritzburg. The imago is one of the commonest butterflies in Durban, both on the Berea and in the Back Beach Bush ; but is not so often seen in Maritzburg. 5. ACRHA RAHIRA (Boisduval). (Plate XLVI. fig. 7, larva; figs. 8, 9, pupa.) Larva. Back and sides blackish; thoracic legs, claspers, and a line above them chrome-yellow. A dorsal white stripe, and on each segment four yellow spots from which spring four branched yellow spines, the lower pair springing from the yellow spiracular line. These spines are shorter than in the majority of Acrwa larve. Head yellow. Feeds on a species of groundsel, Erigeron canadense (Linneus). TRANSFORMATIONS OF SOME SOUTH-AFRICAN LEPIDOPTERA. 295 Two figures of the pupa are given: one pupa is waxy white and similar to the pup of other Acrew, the other ferruginous. The ferruginous pupe had nearly always been attacked by ichneumons, with which the larve were much infested. The imago is a common butterfly in the coast district. 6. Acr#a BUxTONI (Butler). (Plate XLVI. fig. 10, larva; figs. 11, 12, pupa.) Larva. Pale buff dorsally, deepening to pale green on the sides, with a buff lateral spiracular line above thoracic legs and claspers, which are also buff. ‘Two dorsal pale green stripes, interrupted on every segment by a pale yellow transverse stripe bearing four black branched spines; below these are two buff-coloured spines springing from the buff spiracular line. Head yellowish. Feeds on a species of nettle, locally called “ pink hibiscus” (although it is not a hibiscus at all). It is a common plant on the Berea, Durban, where I found the larva, and has been identified for me by Mr. Medley Wood as Triumfetta rhomboidea (Jacq.). The larva is here figured, x $, to exhibit the detail. Pupa waxy white, with the usual fine black lines and spots with orange centres, beautifully gilded ; pupe formed in the dark, however, inside a box, are slaty black. Imago fairly common in Durban district. 7. PLANEMA ESEBRIA (Hewitson). This larva and pupa have been figured by Trimen, ‘South African Butterflies,’ i. pl. i. figs. 2& 2a. I took it ona species of nettle (Urtica sp.), in the covert known as the “‘ Back Beach Bush,” near Durban. My specimens resulted in two females of the white variety A. The imago is a very common butterfly in the Durban district. Subfamily NyMPHALIN& (Bates). 8. JUNONIA CEBRENE (Trimen). (Plate XLVI. fig. 15, pupa.) This larva has been already figured by Trimen, ‘South African Butterflies,’ pl. 1. fig. 4. 1 reared the larva from an egg which I saw deposited by the female on Justicia natalensis, a common plant amongst the grass of the veldt. Pupa chocolate-brown, with four rows of small yellow spots on each segment, and a submarginal row of white spots round the edge of the wing-covers, with a row of four white spots inside them, ‘The specimens reared passed 13 days in the pupal stage 296 LIEUT.-COL. J. M. FAWCETT ON THE (23rd Jan. to 4th Feb.). I reared one specimen from the egg which I saw deposited by the female, but regret that I made no notes of the earlier stages, save that the young larve were, during the first two moults, lighter and more violaceous in colour, and the spines were proportionately much smaller. The imago is common everywhere in Natal. 9. JUNONIA cLELIA (Cramer). (Plate XLVI. fig. 14, larva; fig. 15, pupa.) Larva. Dark slaty black, covered with minute yellowish-white atoms, and each segment bearing black branched spines as in the larva of J. cebrene, from which it only differs in being lighter coloured and brownish underneath, and in having an interrupted white lateral line above spiracles. Feeds on Asystasia coromandeliana. Pupa. Dark chocolate-brown ; wing-covers lighter ; yellowish-white spots on thorax and each segment and wing-covers, as in pupa of J. cebrene. Duration of pupal stage 11 days (March 1 to 12). Imago very common in Durban, more rarely seen in Maritzburg. 10. Pyramers carpur (Linneus). The larva of this well-known species has been so often described that any description here would be superfluous; 1 will therefore merely state that I reared it ona large species of thistle named Stobawa discolor (Harv.). The larva was very similar to that of Junonia clelia, but differed in being longer in proportion to its width ; the branched spines were yellow instead of black, and the black dorsal line was more defined ; the pupa was beautifully gilded. 11. Cuaraxes canprope (Godart). (Plate XLVI. fig. 16, larva; fig. 17, pupa.) Larva. Grass-green, irrorated with minute yellow spots, which coalesce into a lateral line below the spiracles, which ends on the 12th segment in a bifid tail. On the 6th and 8th segments are placed three yellow spots with orange centres, one dorsal and two subdorsal, the dorsal spot being larger than the subdorsal spots and more oval in shape. Head dark green in front, brownish to pale yellow at the sides, with four light brown serrated horns, two springing divergently from the top of the head and two from the sides. Feeds on a species of Croton, probably Croton sylvaticus (Hochst). Pupa. Grass-green ; very short and thick, suspended by tail to stem of food-plant. I am indebted to Mr. Walter Butcher for a specimen of this larva from the neighbour- hood of Durban, where the imago is common on the Berea in the summer months. TRANSFORMATIONS OF SOME SOUTH-AFRICAN LEPIDOPTERA. 297 Family PAPILIONID. Subfamily Prerin & (Bates). 12, Prmris severtNA (Cramer). (Plate XLVI. fig. 18, larva; figs. 19, 20, pupa.) Larva. Dark green on back, covered all over with very fine whitish hairs. ‘Two lighter subdorsal lines, and below them a row of yellow spots, one on each segment, placed in a darker green line. A light yellow lateral line along spiracles above a light green line just above thoracic legs and claspers, which, with under surface, are white. Head dark brown with minute white spots; 1st segment light yellow anteriorly. Feeds on Capparis corymbifera. Pupa. Bluish green with two yellow spots on each segment subdorsally, and four on each side of thorax. A fine black line along dorsal side of thorax, and two small white tubercles with black bases at angle of wing-covers. Imago common in Durban and Maritzburg. This larva is described in ‘South African Butterflies’ as ‘Dull reddish sandy on back, with a median longitudinal streak of violaceous grey.” It may be dimorphic ; but I have never met with a specimen of this colour (‘S. African Butterflies,’ vol. iii. p. 69) 13. Pipris HELLICA (Linneus). (Plate XLVI. fig. 21, larva; fig. 22, pupa.) Larva. Yellow above, bluish green underneath ; a light blue dorsal stripe, and on each side of it a broader greenish stripe, and below these a pale yellow spiracular stripe; body covered by minute black papilla and some sparse short hairs; head bluish, covered with black dots. Feeds on a common weed in the grass of the veldt, of which JI did not get the name. Pupa. Yellow above and green underneath, covered with minute black spots. Pupal stage lasted only 7 days in February. Common everywhere in Natal. 1“ My description (0. c.) of this larva was made from numerous Natalian specimens sent alive to me by Colonel Bowker. I observed, on the page quoted, that in its earlier stages the larva is dorsally much tinged with greenish, and it is of interest to find that the green tint commonly persists till the full-grown stage.”-— t. TRIMEN. 298 LIEUT.-COL. J. M. FAWCETT ON THE Subfamily PaPILIONIN& (Bates). 14. Papitto poticenes (Cramer). (Plate XLVI. figs. 23-28, larva; figs. 29, 50, pupa; fig. 31, head of larva with tentacles.) Larva. 1st stage. Chocolate-brown, each segment with a transverse yellow stripe ; three yellow spines on thoracic segments, and a yellow bifid tail consisting of two yellow divergent spines. The larva does not exceed 4 inch in length in this stage. 2nd stage. Ground-colour ferruginous red, with transverse black stripes as follows :-— one between each two segments, and two on each segment enclosing a greyish-blue streak hetween them. Ist and 2nd segments and bifid tail yellow with black points. A black spiracular line divides the ferruginous upper portion of the body from the lower portion, which is violaceous grey. ‘Three thin black spines with scarlet bases on thoracic segments. Head ferruginous, thoracic legs and claspers pale grey. 3rd stage. Ground-colour has become green, fading to yellow at sides above the black lateral line; the black line dividing the segments has become obsolescent, and the two black lines on the segments enclosing the blue streak have become ferruginous. Otherwise as in 2nd stage. In this stage larva attains its full size. 4th (and final) stage. Entirely green above, with the exception of the three black spines on the thoracic segmeuts, which still retain their scarlet bases, and are joined by transverse black streaks; under surface violaceous grey. Head and Y-shaped tentacular organ pale green. Larva feeds on Uvaria caffra, called by the natives “ Maswinda,” a common bush in the coverts of the coast-belt. Pupa green, very broad, and very much flattened, with a very prominent thoracic dorsal projection pointing forward; the lateral projections pale ferruginous; two subdorsal lines of black spots on each side. One pupa formed ina box was pale pinkish, with the subdorsal region strongly marked with black. Only the 2nd stage of this larva is described in ‘South African Butterflies’ (vol. iii. p. 203). The imago is common in the Durban district. 15. Papitio BRasipAs (Felder). (Plate XLVI. figs. 32-34, larva; figs. 35, 36, pupa.) Larva. st stage. Pale yellowish ferruginous, with reddish diagonal streaks on sides; back of 3rd, 4th, and 5th segments occupied by an olive-green area, with a TRANSFORMATIONS OF SOME SOUTH-AFRICAN LEPIDOPTERA. 299 median dorsal pink stripe from head to tail, bounded by an olive-green stripe on each side. Ist segment reddish brown superiorly, with short tubercles on each side; a pair of short pink tubercles dorso-laterally, bounded inwardly by dark olive-green, and joined by a fine pink transverse line on 2nd and 35rd segments; tail bifid. 2nd stage. Pale green above, darker green underneath ; all the markings described in Ist stage defined by slightly darker green, except the olive-green area on the back of the thoracic segments, which has become dark green in this stage, while the pink median dorsal line and transverse stripes have become a rather startling whitish cross. 3rd (and final) stage. Shortly before changing into a pupa the larva loses all its distinctive marks, and becomes very pale yellow all over, almost diaphanous. Feeds on Popovia caffra (Kaffir name, “small Maswinda”), a common shrub in the coast district. Pupa. Bright green. On each side of back, from apex of dorso-thoracic projection to anal extremity, a narrow yellow elevated stripe; on summit of back a pair of similar stripes commencing considerably apart near base of dorso-thoracic projection, and converging till they meet at anal extremity. A series of diamond-shaped lozenges down the back, medially, and a series of diagonal streaks subdorsally between the two yellow stripes, and extending beyond them to the cephalic area. Duration of pupal stage 21 days (24th January to 16th February). Common in Durban and the coast-zone. I have taken the allied species Papilio leonidas in the same coverts and at the same time. It is, I should say, doubtful if they are distinct species. P. brasidas is the commoner species in the Durban district. I have not met with either in the vicinity of Maritzburg. 16. Papitio morania (Angas). (Plate XLVI. figs. 37-39, larva; fig. 40, pupa.) Larva (early stage). General colour grass-green on back, incisions of segments pale blue; two dark chocolate-brown lateral stripes above spiracles (which are white), Head, thoracic legs, claspers, and bifid tail brownish yellow. Ist segment yellow dorsally, bearing two short tubercles on each side; 2nd and 5rd segments bearing dorso-laterally a pair of black spines, short, with orange bases, and joined by two broad transverse streaks of chocolate-brown; that on the 2nd segment having three brown spots posteriorly, and that on 3rd segment five brown spots anteriorly, and being much larger than the other, extending posteriorly over part of the next segment; from this three dark brown longitudinal lines, one dorsal and two subdorsal, extend to the 11th segment, where the two subdorsal lines become two chocolate-brown spots joined by fine transverse lines, while the dorsal line (which is much finer than the others) extends to the tail. A double lateral line of black spots along the segments, two on each segment along the upper line, and one on each segment in the lower line. 2nd stage. On casting its last skin, the larva emerges exactly the same as in the VOL. XV.—Part vi. No. 2.—April, 1901. 2U 300 LIBUT.-COL. J. M. FAWCETT ON THE early stage ; but in a very short time (perhaps half a day) it loses all its distinguishing characters in that stage, and becomes uniform pale green; a yellow lateral line above a darker green line being all that remains to represent the two very distinct brown lateral lines of the early stage, and the under surface, thoracic legs, claspers, and head have all become pale green. ‘The black spines on the thoracic segments, however, remain the same, except that their bases become more reddish. In this stage the larva is very similar to the larva of P. policenes in its green stage; but the latter always retains its brown lateral streak and its blue-grey under surface. 3rd (and final) stage. Like the larva of P. brasidas, this larva also becomes pale yellow and almost diaphanous just before pupating. It feeds on Uvaria caffra (the large “ Maswinda”), like the larva of P. policenes. Pupa. Grass-green and more rounded than in pupa of P. brasidas. Dorso-thoracic prominence thicker and pointing more upward; a pair of raised yellow lines from thoracic prominence along the sides to anal extremity, also a second pair of similar lines run subdorsally from base of thoracic prominence to tail, where they converge and meet. Pupe formed on Ist February remain over the winter months in this stage. The imago is common in the vicinity of Durban. 17. Papitio pEMopocus, Esper. (Plate XLVI. figs. 46-48, larva; figs. 49, 50, pupa; fig. 51, head of larva with tentacles.) Larva (early stage). Ground-colour black, with short yeilow tubercles on the thoracic and 10th, 11th, and 12th segments. Top of Ist segment yellow between the two tubercles ; a dorsal line of diamond-shaped lozenges commences grey on the 4th segment, becomes white on 6th and 7th segments, and darkens again to grey and brown on the remaining segments to the tail, where it is obsolescent. On 4th, 5th, and 6th segments are white lateral spots, which converge and meet the dorsal line of lozenges on 5th and 6th segments; similar white lateral spots on 10th, 11th, and 12th segments. ‘ail bifid and yellow, head reddish. ‘This larva, which apparently mimics the droppings of birds, varies but little in this stage. Full-grown stage. The ground-colour in this stage varies according to the colour of the lemon-leaves on which the larva is feeding, from pale green almost yellow to dark green; the transverse streaks on the sides also vary from light brown almost ferruginous in the pale specimens to deep velvety black in the dark green examples. The most common form has a grass-green ground-colour, with velvety black transverse streaks on the three thoracic segments, and black diagonal streaks on the 7th, 8th, and 9th segments, converging till they almost meet dorsally on Sth and 9th segments. Under surface and claspers greyish white, and a broad white lateral line above spiracles. Head and thoracic legs reddish ; head surmounted by two reddish tubercles, between TRANSFORMATIONS OF SOME SOUTH-AFRICAN LEPIDOPTERA. 301 which the ¥-shaped tentacular process is situated. The transverse black streaks on the 2nd and 3rd segments have a line of ferruginous ocellate spots variegated with light blue and yellow. The black diagonal streaks on 7th, 8th, and 9th segments are mottled with rather indistinct purple ocellate markings, and edged outwardly with fine lines of pale yellow. There are also short diagonal streak-like spots on the 9th and 12th segments just above the white spiracular line. Two yellow tubercles over the anal extremity. The pupa also varies in colour according to its surroundings: the most common form is light green, paling to two yellow subdorsal streaks; head and cephalic prominences (which are well-marked) brown, as is also the thoracic prominence posteriorly, from which a broad brown streak extends almost to the anal extremity. Several dark green and brown spots on the segments. I first bred this larva at Ladysmith, in the high level zone, in April 1898, on a plant named Clausena inequalis; the larve began pupating on the 7th April, and the first imago emerged on 12th October, 1898. I noticed that these specimens were smaller than those which I subsequently reared on lemon-leaves at Maritzburg in December and January following; these individuais only averaged a fortnight in the pupal stage. This is the commonest and most widely distributed Papilio in S. Africa. 18. Papinio nireus (Cramer). (Plate XLVI. figs. 41-44, larva; fig. 45, head of larva with tentacles.) Larva. 1st stage. Ground-colour black, paling to greenish on the sides, the under surface, thoracic legs, and claspers being greenish yellow; head, thoracic, 10th, 11th, and 12th segments surmounted by orange-coloured tubercles in pairs ; the pairs on the three thoracic segments being joined by transverse orange lines bearing two to three tubercles each, smaller than those at the extremities. 7th, 8th, and 9th segments ferruginous ; 5th, 6th, and 10th segments black superiorly with white sides. In this stage the larva somewhat resembles that of P. demodocus. In some individuals the black markings are replaced by very dark green. 2nd stage. Ground-colour grass-green superiorly, under surface greyish white. No ferruginous markings on 7th, 8th, and 9th segments, and no lateral white spots on dth, 6th, and 10th segments, these being replaced by a pale lemon-coloured lateral streak. ‘The tubercles, which have become light lemon-coloured, are confined to the thoracic and 12th segments. 3rd stage (final). Ground-colour generally dark green, but, like P. demodocus, varying according to the colour of the leaves of the food-plant. A wide pale greenish-blue oval area with a thin whitish line across its centre, on the summit of the 3rd and 4th segments, bounded anteriorly and posteriorly by a transverse line of ocellate spots :— the anterior series outlined in black on an ochreous transverse band, the ocelli at the 2u2 302 LIEUT.-COL. J. M. FAWCETT ON THE extremity of the band being larger than the rest; the posterior series indistinct: four ocelli in each series have light blue centres. A lateral white stripe along the spiracles from 4th to 12th segments: tubercles on anal segment yellow, well-defined, and joined by a raised yellow pale streak; tubercles on back of Ist segment pale and incon- spicuous. ‘The Y-shaped tentacles in this species are bright crimson and much longer than those of P. demodocus; the scent emitted is, however, similar. Pupa. Bright green; point of dorso-thoracic prominence, two spots below it (at abdominal base), and edge of lateral abdominal angles creamy reddish. In shape somewhat attenuated anteriorly, cephalic processes short and directed laterally outward, so that the frontal line of head is widened and but slightly concave. Thoracic lateral angles moderately acute ; dorsal prominence also elevated acutely but not inclined forward. Sides of abdomen widely flattened, and so extended as to form a very marked angle on each side of third abdominal segment; whence abdomen narrows very rapidly and greatly to the extremity. Infra-pectoral region, where wing- covers meet, very strongly convex. A marked constriction dorsally at junction of thorax and abdomen. In its natural position, attached vertically or nearly so, head uppermost, the anterior portion of this pupa is seen to be very much more bent backward than it is in P. demodocus. J have taken this description of the pupa, with some alterations, from ‘South African Butterflies,’ as I thought it so well describes the peculiarities in shape of the pupa. The imago is not so often met with as P. demodocus, and seems to prefer wooded country more. In its haunts, however, and in the coast district it is one of the commonest species. III.—OBsERVATIONS ON THE HETEROCERA. Family SATURNIIDZ. 1. Acrias Mimos&. (Plate XLVII. figs. 1, 2, larva; fig. 3, cocoon.) Larva. Ground-colour grass-green, with paired dorsal series of long conical humps with yellow apices, surmounted by three or four short black hairs, and the same number of longer yellow hairs, from 2nd to 10th somites inclusive ; the 11th somite has only one similar dorsal hump, and the Ist somite no hump, with the black and yellow hairs planted just above the head ; a subspiracular line of small tubercles with similar hairs. Between each somite, from 3rd to 11th, a yellow transverse streak TRANSFORMATIONS OF SOME SOUTH-AFRICAN LEPIDOPTERA. 305 folding over a blue transverse streak at the incisions of the somites. Head and thoracic legs ferruginous, spiracles white, with ferruginous centre. In the early moults this larva is ferruginons, the head and thoracic legs being black. Feeds on Sclerocarya caffra (Sond.), called in Natal “the wild mango tree”—a common tree in the coast districts, but not found higher up. When about to change the larva spins a greyish-white silky cocoon, with a line of ventilation-holes round the exit end, which it attaches longitudinally to a twig of the food-plant. The pupa is chocolate-brown and of the usual shape. This moth appears to be double-brooded, the larva appearing in November and February, and the perfect insects in January and the following October. Common at Sydenham, near Durban. 2. NUDAURELIA WAHLBERGI (Boisduval). (Plate XLVII. fig. 4, larva; fig. 5, pupa.) Larva. Ground-colour deep velvety black, each somite bearing four branched ferruginous spines with reddish bases, two subdorsally and two laterally, from 2nd to 12th somite inclusive, the spines on the 2nd somite having black bases. Between the subdorsal and the lateral row of spines are placed a collection of small yellow spots on each somite; the red bases of the subdorsal and lateral spines being joined on 10th and 11th somites. Spiracles white: a subspiracular row of small yellowish tubercles bearing a few whitish hairs, one on each somite. Head and legs concolorous with body. Feeds on English oak, but has also been taken in large numbers on peach-trees, in gardens in Maritzburg—both these trees being imported species in Natal. When full- fed it burrows underground, where it makes a sort of web !, inside which it undergoes its transformation. Individuals which changed about 14th March reappeared as imagines on 6th May and subsequent days. Pupa dark chocolate-brown. Common in Maritzburg. 3. Buna#A cAFFRARIA (Stoll). (Plate XLVII. fig. 8, larva.) Larva. Ground-colour deep velvety black ; each somite, from 4th to 12th, bearing eight yellow tubercular processes, two subdorsally, two laterally, and four (in two rows) on each side subspiracularly. The 2nd somite bears four black processes, two sub- dorsally and two laterally ; the 3rd somite bears four black processes, as in the 2nd, «The group of Antherwa to which this species belongs—Nudaurelia—is noted for forming no cocoon, and the closely allied species 4. (N.) cytherea, menippe, and tyrrhea, as well as Buncea caffraria, as far as I have observed, simply bury themselves without any attempt at forming even a ‘ web.’”—R, Tren. 304 LIBUT.-COL. J. M. FAWCETT ON THE and two small yellow processes on each side, in line with the subspiracular processes on the other somites. Spiracles red; those on the 4th to 11th somites being sur- rounded by an irregularly shaped red area. Head and legs concolorous with body. Larva figured has not attained its full size. Mr. Quekett writes: ‘‘The food-plants of this Bunea are, at Durban, Celtis kraus- siana (Bernh.), and Ekebergia meyeri (Presl), and at and near Maritzburg a species of Cussonia, on which I have taken the Jarve in the Botanic Gardens.” This larva, which much resembles that of NV. wahlbergi, undergoes its transforma- tion underground in a similar manner to that larva, and is double-brooded, the larvae appearing in November and December, and the second brood in March and April; these latter individuals remain pupe until the following October, my first specimen emerging on 12th of that month. Pupa dark reddish brown; very similar to that of Gynanisa maia (Klug). 4, GYNANISA MAIA (Klug). (Plate XLVII. fig. 6, larva; fig. 7, pupa.) Larva. Ground-colour pale green, with paired dorsal series of humps ; each somite, from 2nd to 12th, bearing four silver spikes with yellow points inclined backwards, one subdorsally and one laterally on each side; also a number of small silver spots placed irregularly over each somite; the first somite has no spikes, but is raised to a sharp ridge, with a black edge, which may be of assistance to the larva in forcing its way through the ground. Spiracles purple, and immediately beneath them a purple lateral line having on its lower edge a yellow raised lateral line bearing a small orange-coloured tubercle on each somite, and thickening considerably above anal claspers, where it has a series of small black tubercles superiorly. Head green, with black side-streaks defining the eye. Under surface darker green with minute white spots, and a row of small red tubercles, one on each somite just above the claspers, which are green. Thoracic legs pale brown banded with black. Feeds on the common wattle (Elephantorrhiza burchellii, Bth.) and on several species of thorny acacia. When full-fed the larva burrows underground, and there undergoes its transformation, making little or no preparation in the way of a web. The pupa is dark reddish brown, and is remarkable for the large size of the antenne covers. ‘The moth is double-brooded, the larve appearing in November, and the imagines emerging end of January ; the second brood of larvee appearing in March and emerging in the following October. At Fort Napier, Maritzburg, the wattle-trees were completely cleared of their leaves by the March brood this year. If the collector possesses one female, any number of males can be secured by placing the female in a box on the verandah. The writer has seen as many as twenty males at one time fluttering round the box about 10 p.w. Next morning the verandah was strewn with their wings, a cat, as was subsequently discovered, having caught and devoured them. TRANSFORMATIONS OF SOME SOUTH-AFRICAN LEPIDOPTERA. 305 5. PSEUDAPHELIA APOLLINARIS (Westwood). (Plate XLVII. figs. 11, 12, larva; fig. 13, pupa.) Larva. Ground-colour bluish grey, each somite with a broad transverse indented black streak, thickest on 2nd somite, and two finer black transverse lines across the upper part of the body between the somites. Body broader in the middle than anteriorly or posteriorly. Above anal extremity a sharp-pointed black horn—a fine black spiracular line, and immediately below it a broad raised ferruginous line bearing a small black tubercle crowned with some short yellow hairs on each somite. Under surface and abdominal claspers pale yellow. Head, thoracic legs, and anal claspers black. A tuft of short hairs on the summit of the 2nd somite. Feeds on Jurrea heterophylla (Smith), and undergoes its transformation underground ; the period passed in pupa state (in March) was 17 days. Pupa dark red-brown, abdominal somites strongly marked ; chiefly remarkable for the long sharply-pointed tail resembling the point of a thorn. The imago is a common species, flying by day, in the Berea Bush, Durban. 6. Urota stvope (Westwood). (Plate XLVII. fig. 9, larva; fig. 10, pupa.) Larva. Ground-colour velvety black, with a series of paired pale yellow humps formed by a thick raised transverse yellow streak, bearing a pair of short tubercles crowned with short yellow hairs dorsally on each somite, and a series of similar short tubercles laterally at the lower end of each streak. Spiracles pale yellow, and below them an interrupted, raised, subspiracular line covered with short yellow hairs. Head black, surmounted by a yellow transverse band where it joins Ist somite, and crowned with some yellow hairs. Thoracic legs black, abdominal claspers yellow banded with black, anal claspers and extremity yellow, all covered with yellow hairs. Feeds in companies on the “ Kaffir boom ” tree (EKrythrina caffra, Thb.). Pupa dark red-brown and of the usual form, with a sharpish spur or horn at the anal extremity. ‘The transformation is carried out underground. Fairly common in Durban district. 7. Lupra sminax (Westwood). (Plate XLIX. fig. 6, larva; fig. 7, cocoon.) Larva. Ground-colour rufous, with irregular pale blue spots surrounded by a thin white line. On each somite a short tuft of black hairs, sarmounted by a tuft of longer and finer hairs of same colour, and the first five and last somites covered with short yellow hairs. Head, legs, and claspers brown. This larva, which looks, through a microscope, like a piece of old china ware in colour, is one of the most unpleasant larvee to handle that I have ever met with. The short black hairs on each somite possess poisonous qualities, which produce on the hand a white rash akin to that produced by a bad stinging from nettles. It forms a 306 LIEUT.-COL. J. M. FAWCETT ON THE cocoon round itself, with its hairs, on the side of the box. Food-plant oak; but I have also reared the larva on Jasmine (Jasminium pubigerum). On quitting South Africa on a short leave of absence home in May 1899, I left two specimens of this larva, in the pupal stage, together with all my larvee which had not completed their transformations, in the care of a friend who was remaining on in Pietermaritzburg, as I was afraid that, during the voyage, the heat of the tropics would cause them to emerge prematurely ; but to prevent, as I thought, the possibility of any mistake in identifying the imagines, I placed each larva or pupa in a different box, with a small ticket, with a reference to the plate and figure of the larva in my sketch-book, which I requested my friend to attach to the pin of the imago when it emerged. I subsequently returned to South Africa when war broke out, and on the relief of © Ladysmith was invalided home ; on passing through Pietermaritzburg in March 1900, my box containing the emerged imagines (which had all been placed in a large cork box which I had left with my friend) was returned to me, when I found my ticket referring to the drawing of this larva pinned to an imago of Ludia smilax (Westw.). On reaching England I forwarded a figure of the imago to Mr. R. Trimen, and asked him for the name of the moth, as I was unacquainted with it. Iwas therefore much surprised to receive the following reply from him :—‘I know the moth you figure in your note of yesterday very well, it is Ludia smilax (Westw.); but I am much surprised to hear that this species resulted from the ‘ old china’ caterpillar figured by you, because I have by me figures of no fewer than three quite different larvee attributed to ‘ smilav’—I mean different from each other, as well as entirely different from yours. If you are quite certain about the moth resulting from the old china larva, then this stage of L. smilav presents amazing variability. W. D. Gooch and Dr. J. KE. Seaman made drawings of this caterpillar ; they are rather rough, but those of the apparently more prevalent variety agree in representing a pale greenish larva with broad black rings; Gooch thought that this was the male, but he does not seem to have proved this, though he bred a male from one of this pattern. Gooch’s second form of larva is quite like the other in shape, tubercles, hairs, &c., but the black rings are almost everywhere broken up into mere scattered spots, and the ground-colour, generally, is dull yellow; he thinks this is the female. **Much more amazing is Dr. Seaman’s second form of smilax larva; it is black, marbled with white, and with vivid scarlet rings! There is no doubt that some larve seem to vary almost indefinitely, but if ths is a true case all through, I think it is unsurpassed.” it is therefore with some doubt that I append the description of this larva as the larva of Ludia smilax (Westw.), inasmuch as I cannot vouch myself, from personal observation, for its accuracy, though I have no reason to doubt the accuracy of my friend, who is certain that no mistake has been made. TRANSFORMATIONS OF SOME SOUTH-AFRICAN LEPIDOPTERA. 307 Family SPHINGID. Subfamily ACHERONTIIN A, 8. ACHERONTIA ATROPOS (Linneus). No description is necessary of the larva of this well-known species; and 1 will merely remark that I reared it on a species of Spathodia, an imported tree which is often met with in Berea Bush, Durban. The larva underwent its transformation on 14th February, and the imago emerged on 17th March after a pupation of 38 days. This is one of the instances of larvee feeding on imported trees in Natal. I also reared the dark form (body uniformly fuscous, first three somites pink subdorsally) at Maritzburg, on Jasminium pubigerum (also an imported plant in Natal). This larva underwent its transformation on 5th April, and hibernated in the pupal stage ; this form is much rarer than the green and yellow form. Subfamily SMERINTHINA. 9. LopHostetaus puMoLINIT (Latreille)!. (Plate XLVIII. fig. 7, larva; fig. 8, pupa.) Larva. Ground-colour very pale green, a pair of blue-black steely branched spines with pale yellow bases and basal areas subdorsally on each somite from 2nd to 10th. The 1st somite has no spines, and the 11th has only one spine, thicker than the others and replacing the horn in other species. A lateral row of smaller black spines springing from the upper edge of a spiracular row of large yellow spots; a subspiracular row of small black spines springing from lower edge of above-mentioned spots, and, below these, two spines placed diagonally on the 4th, 6th, 7th, 8th, and 9th somites just above the claspers ; the 5th somite having three spines, and the 2nd, 5rd, and 10th somites one spine each in this series. Abdominal claspers yellow, with black extremities, each extremity bearing three short black divergent spines ; anal extremity and claspers horny and rufous, with a broad black edging. Head pale green superiorly, pale ferruginous inferiorly; two black vertical stripes on the face, ending with a detached black spot above them. Sides of head black, as in the larva of Acherontia atropos. ‘Thoracic legs pale ferruginous, banded with black. 1 «A brief description of the extraordinary spiny larva of this Hawk-moth—drawn up by me from a coloured drawing by Mr. W. D. Gooch, and a coloured photograph by the late Dr. J. KE. Seaman—was published by Prof. Meldola in his English edition of Weismann’s ‘ Studies in the Theory of Descent,’ vol. ii, pp. 527, 528 (1882).”—R. Trimen. VOL. XV.—PART V1. No. 3.—April, 1901. 2x 308 LIEUT.-COL. J. M. FAWCETT ON THE Feeds on Hibiscus tiliaceus (Linneus). When full-fed the larva burrows under ground, and forms a sort of chamber with a web, in which it undergoes its trans- formation. The pupa is dark reddish brown, and only distinguishable from that of A. atropos by its greater thickness and the abdominal somites being more horny. The specimens reared remained in the pupal state from February till the following October. The imago is not uncommon in Durban and the coast districts. Subfamily CH HROCAMPIN A. 10. CHaRocaMPA CAPENSIS (Linneus). (Plate XLVIT. figs. 17, 18, larva; fig. 19, pupa.) Larva. Ground-colour pale green, thickly irrorated subdorsally with darker green diamond-shaped spots, from 5th to 11th somites; these spots coalesce into a series of diagonal streaks along the somites subdorsally and spiracularly. A paler green lateral stripe from 5th to 11th somite, with a dark green stripe along its upper edge; horn very short and pink ; a reddish “eye ”-like spot edged with white superiorly on 4th somite. Head and claspers green, thoracic legs pink ; spiracles red. When ready for its transformation the larva becomes dull pink flesh-coloured, the lateral line only remaining green; the eye-like spot and diamond-shaped irrorations become black, and the ground-colour of the dorsal region becomes dull yellowish. The figure is taken from a larva in this final stage. Feeds on common vine. Pupa. Head, wing-covers, and dorsal region dark brown, with black markings ; abdomen pale pink flesh-coloured. Time passed in pupa state about five weeks. When about to change the larva spins a web amongst leaves, &c., on the surface of the ground in some sheltered place, and undergoes its transformation inside. The imago is fairly common in the evening in flower-gardens in Maritzburg. I have also reared a red form of this larva, which may be described as follows :— Ground-colour uniformly ferruginous; a pale subdorsal stripe from 2nd somite to horn, white above, darkening to yellow beneath, bounded inferiorly by dark red diagonal stripes on each somite from 4th to 10th. On 4th somite a conspicuous ‘eye ”-like spot, black superiorly, white inferiorly. Mr. R. Trimen writes to me of this larva:—* The red form of C. capensis larva is, at Cape Town (where the species feeds on the cultivated vine), very much rarer than the green one; the colour of the latter is decidedly protective on the cultivated vine, but that of the red one was not so (except slightly, perhaps, when the larva is on TRANSFORMATIONS OF SOME SOUTH-AFRICAN LEPIDOPTERA, 309 the older woody stems). I imagine, however, from one of the red ones which I found on its native food-plant, the wild Cape vine (Cissus capensis), that this tint was probably acquired in relation to the latter plant, which is densely clothed with red down on the younger shoots and underside of the leaves. I have noticed the common Cape Butcher-bird (fiscus collaris) taking capensis-larve from the cultivated vines at Cape Town and spiking them on thorns; such persecution would seem to show that these larve are in need of protective resemblance to their food- plants.” 11. CuHarocampa osiris (Dalman). (Plate XLVIII. figs. 3, 4, larva; figs. 5, 6, pupa.) Larva. Ground-colour pale green, thickly irrorated with darker green diamond- shaped spots as in C. capensis. A paler green lateral stripe from 5th to 11th somites, ending at a very small yellow horn. ‘Two large “ eye ’-like spots (green with a light blue centre, on which are placed six minute white spots), surrounded by a narrow light yellow iris, on 4th somite; two smaller round yellow spots on 5th somite. Head, spiracles, and claspers green ; mandibles and thoracic legs pink. When ready for transformation the larva becomes pale reddish brown with a violaceous suffusion; the intersections of the somites and lateral stripe become white, while the diamond-shaped spots become black where they are most thickly placed near the junction of the somites; the large “eye-like ” spots become brown with a green iris, while the yellow spot becomes black. The effect of these changes is to give the larva, when viewed dorsally, a very startling resemblance to a snake of the banded Krait species. Feeds on the common vine. Pupa \ight brown dorsally, variegated with black spots and streaks defining head, eyes, antenn, wing-covers, nervures, and abdominal somites; breast, wing-covers, and underside of abdomen pale pink flesh-coloured. This pupa is chiefly remarkable for its abnormally long snout, the palpi-covers being produced to form a shield or covering for the haustellum. The transformation takes place amongst leaves on the surface of the ground, without much attempt at making a chamber with a web, the leaves being loosely strung together by means of a silken thread. This larva was received from Bellevue, a suburb of Durban ; the imago seems to be rather uncommon. 12. CHHROCAMPA BALSAMIN® (Boisduval). (Plate XLVIII. fig. 1, larva ; fig. 2, pupa.) Larva. Ground-colour bright grass-green ; from 6th to 11th somites a brown inter- 2x2 310 LIEUT.-COL. J. M. FAWCETT ON THE rupted dorsal streak, and traces of two double subdorsal streaks mainly formed by small brown spots arranged in transverse rows across the somites. On the 4th somite a pair of black “ eye ”-like spots with a white iris; on the 5th somite a pair of red “eye ”-like spots with a white iris, and between these the dorsal streak above mentioned is continued, green in colour, to the head; thoracic legs pink; head and claspers green; horn light brown and straight, ending in a point. Feeds on Jussiwa repens (Linneus). Pupa. Head, thorax, and wing-covers yellowish brown, remainder of body reddish brown, paler dorsally and abdominally; palpi-covers slightly produced and prominent. The pupa is formed in a light web amongst leaves on the surface of the ground, in the usual manner of the larvee of Chwrocampine. The imago emerges in about a fortnight (in February). Found in Durban district. 13. Cuarocampa ipricus (Drury). (Plate XLIX. fig. 1, larva; fig. 2, pupa.) Larva. Ground-colour pale green; each somite, from 4th to 10th, with a pair of triangular spots, brown superiorly, yellow inferiorly, at its anterior edge. The pair of spots on 11th somite are lengthened into a brown and yellow streak, and end in the horn, which is brown, and sharp-pointed like a thorn. The pair of spots on the 4th somite are larger than the remainder, and bear on their yellow area an oval, black, “eye ”-like spot with a white pupil near its upper edge ; each spot has three minute white points on its brown area; spiracles white; a pinky-white subspiracular line, and above it a collection of brown points at each intersection of the somites. Head and claspers green, thoracic legs light brown. Feeds on Spermacoce natalensis (Hochst.), a common herb among the grass of the veldt. Pupa. Pale golden brown, with small black spots defining the nervures on the wing- covers, and sprinkled over the abdominal somites ; spiracles black. Period of pupation about 23 days. Habits and transformation similar to other species of the genus. The imago is very common, hovering over flowers in gardens in Maritzburg, from the early afternoon till dusk. The larva and pupa are here figured x $ in size on nature. 14. Dapayis neru (Linneus). The transformations of this insect are so well known that figures and detailed descriptions are unnecessary here; but it may be as well to note that numerous specimens were reared from the larva on oleander, which plant is a common shrub in gardens in Maritzburg. ‘The colour of the larva in Natal is very pale green (almost TRANSFORMATIONS OF SOME SOUTH-AFRICAN LEPIDOPTERA. 311 yellow) dorsally, darker below; a white lateral line with numerous small silver-white spots sprinkled over, above, and below it, the lower edge of the line is bounded by a broad light blue area extending downwards as far as the spiracles from 6th to 10th somites. On 3rd somite a pair of large, black, double-pupilled, eye-like spots, the pupils being white, surrounded by a blue and a crimson iris. Pupa pale ferruginous ; habits and transformation similar to those of C. capensis. Duration of pupa state 10th or 11th February to 4th March. Subfamily SPHINGINA. 15. ProroparcE Mauvriti, Butler. (Plate XLVIII. figs. 9, 10, larva; me, IL, pupa ; fig. 12, larva, dorsal view of 1st four somites.) I am informed at the British Museum that the name MJacrosila solani, by which this species is usually known in South Africa, should be applied to the Mascarene form. Larva. Ground-colour grass-green, under surface darker. Paired humps on Ist and 2nd somites ; a purple dorsal stripe from 4th somite to horn; lateral oblique purple stripes from 4th to 10th somites ; these stripes join the dorsal stripe on every somite, and are defined inferiorly by parallel narrow white oblique stripes; horn ferruginous, long, and beset with yellowish tubercles; spiracles small, red, with black centres. Head green, with vertical black stripes on the face and sides, as in larva of A. atropos. Thoracic legs black, claspers green. Feeds on Duranta plumieri, a common shrub in gardens in Natal; but which, according to Mr. Medley Wood, “is included in the flora of Natal by mistake, and is most certainly not indigenous.” I have also reared a dark form of this larva, which may be described as follows :— Ground-colour very pale brown, the oblique purple stripes of the green form being represented in this form by oblique stripes of dark violaceous grey covered with blackish atoms. Head pink, with lateral and frontal vertical black stripes; first three somites pink subdorsally, surmounted by paired dorsal diamond-shaped fuscous spots. It will be observed from this description that the head and first three somites of this larva bear a striking resemblance to the dark form of the larva of Acherontia atropos. Spiracles black; prolegs and claspers pale brown with fuscous bands; horn yellow and serrated. The specimen figured fed on Dahlia variabilis (an imported plant in Natal), Pupa. Dark reddish brown, with a long external sheath for the proboscis. ‘The transformation is effected underground, where, as in the case of A. atropos, the larva constructs a sort of chamber for the pupa. Mr. Trimen states “ the case of M/. solani seems to me especially interesting, because not only is the larva imitative of that of A. atropos in both forms, but the moth also 312 LIEUT.-COL. J. M. FAWCETT ON THE is decidedly imitative of atropos-moth, so much so that I have taken it for atropos when at rest on a tree-trunk. At Cape Town both the dark forms (of solani and atropos \arve) are rare in comparison with the green ones.” Subfamily MACROGLOSSIN&. 16. CEPHONODES HYLAS (Linneus). (Plate XLVIII. figs. 13-18, larva; fig. 19, pupa.) This well-known insect has been reared from six different forms of larva in Natal, which show almost every gradation from an almost wholly green larva with white subdorsal stripes to an almost wholly black one, in which the subdorsal stripes are replaced by very dark grey ones. The various forms will now be described in detail. Larva. 1st form. Ground-colour pale green, a white-bordered light blue dorsal line, and a thin white subdorsal line bordered superiorly by a thin carmine line, the combined lines ending in a yellow streak at the base of the horn. Spiracles white, bordered with red edges, that on 11th somite surrounded by a rufous area. Head green, Ist somite beset with yellow tubercles superiorly ; thoracic legs and claspers brownish ; horn green, curved and pointed, and beset with yellow tubercles. Plate XLVIII. fig. 13. This seems to be the Indian form described by Hampson in ‘ Fauna of British India.” 2nd form. Ground-colour pale green; a white-bordered blue dorsal line ; a subdorsal white line defined in places superiorly and inferiorly by small black oval spots ; under surface, legs, and claspers pale brown. Spiracles white, surrounded by oval red areas ; a yellow subspiracular line, interrupted and defined inferiorly with black, which curves up vertically on the posterior half of each somite from 6th to Yth, otherwise as in Ist form. Plate XLVIII. fig. 14. 3rd form. Similar to 2nd form, but differing in the following characters :—The dorsal line is dark grey ; the black oval spots defining the white subdorsal line inferiorly are much more numerous and extend downwards between the spiracles till they meet a black subspiracular line. The red areas surrounding the spiracles more enlarged. Plate XLVIII. fig. 15. 4th form. Similar to 3rd form, but differing as follows :—Dorsal line darker, the black oval spots defining the white subdorsal line inferiorly are coalesced into a con- tinuous black lateral line; and the whole body below it is pale brown, with only an isolated patch of green on each somite round the red spiracular areas. Ist somite yellow, covered with small yellow tubercles. Plate XLVIII. fig. 16. 5th form. Dorsal and subdorsal stripes as in 4th form. Ground-colour black ; a rufous area round each spiracle, and a few irregular yellow marks on each somite ; a few traces of the yellow subspiracular line, and a yellow patch at base of horn; claspers and anal extremity brown; head and Ist somite yellowish. Plate XLVIII. Mes lis TRANSFORMATIONS OF SOME SOUTH-AFRICAN LEPIDOPTERA. 315 6th form. Entirely black, with the exception of the red areas surrounding the spiracles, and the head, base of horn, and anal extremity and claspers, which are dull ferruginous ; dorsal line black; subdorsal line and Ist somite dark grey. Plate XLVITI. fig. 18. All these forms of larva were found at the same time feeding on Gardenia and also on Kraussia lanceolata (Sond.). When ready for their transformations the larve burrowed underground, and there underwent their change to pupe; the perfect insects emerged in from 3 to 4 weeks. In order to test whether all the different forms of larva produced the same imago, each form of larva was placed in a separate box and carefully labelled, but when the imagines emerged it was impossible to detect any difference between them. Pupa. Dark reddish brown. ‘The imago seems common all over Natal. Family ZYGANID. 17. Zycaya (AnteRIs) aMPpLA (Walker). (Plate XLIX. figs. 27, 28, larva; fig. 29, cocoon.) Larva. Ground-colour white, with pink stripes and black spots. A white dorsal line bordered by two subdorsal rows of black spots, two on each somite; beneath these a pink lateral line bordered inferiorly by a white spiracular line, on which is situated a row of smaller black spots in groups of two on each somite. A subspiracular pale fulvous line, and below it an interrupted line of black lunules, one on each somite ; under surface and claspers white ; head black, with a white bifid frontal stripe. General form of larva thick in middle, tapering towards extremities, and very similar to larvee of Lyceenide in shape. Feeds on a bush with very large leaves with spiny sedges (like holly) of which 1 have not been able to obtain the name. The pupa is formed in a strongly-woven oval cocoon, usually pale fulvous, and affixed to a leaf of the food-plant; time passed in pupal stage about 25 days (March 3rd or 4th to March 29th, 1899). Found in large numbers on the food-plant in the Back Beach Bush, Durban. Family LASIOCAMPID &. 18. Gonomera Postica (Walker). (Plate XLIX. figs. 3, 4, larva; fig. 5, cocoon.) Larva. There are two common forms of this larva in Natal, one with long grey hair and one with long fulvous hair. 1st form. Ground-colour deep velvety black, thickly covered on sides with long grey hair ; a thin black dorsal line and paired black subdorsal patches of short black hairs on each somite from 4th to 12th. Head black, covered with brownish-grey hairs ; 314 LIEUT.-COL. J. M. FAWCETT ON THE spiracles, thoracic legs, and claspers red. The 2nd form is similar to the Ist, but with the long hairs bright fulvous. Both forms feed on the same food-plant at the same time, grow to the same size, and produce male and female imagines irrespectively. This larva should not be handled without gloves, as its hair penetrates the skin of the hand and causes much irritation. It appears in great numbers in November and early December, feeding on common wattle (Elephantorrhiza burchellii) and also on a species of thorny acacia ; the second brood appearing in March and April. ‘The pupa is formed inside a cocoon beset with the hairs of the larva, which cause a very irritating effect to the fingers if it 1s grasped carelessly, and is affixed to a branch of the food-plant, as shown in the figure. Pup formed in April hybernate in this state, and emerge as imagines in the middle of October. In common with species of Saturniidz, the male of this moth seems to possess in a high degree the power of discovering the whereabouts of the female, and the possession of one female will enable a collector to secure a large number of males. Common and widely distributed in Natal. Family LYMANTRID A. 19. Duicuta FascrAta (Walker). (Plate XLIX. figs. 13, 14, larva.) Larva. Ground-colour velvety black ; dorsal area white, thickly covered with white hairs, except on the 4th, 5th, 11th, and 12th somites, which are surmounted by tufts of long black hairs; a dorsal red spot on the 9th and 10th somites; a spiracular line of red spots, and below this an interrupted white line bearing long white hairs. Head black; thoracic legs and claspers red. Feeds on species of Bauhinia, and also on rose-bushes and oak, the two latter being imported plants in Natal. When ready for its transformation, the larva forms a silky cocoon with its hairs, inside which it changes and remains as a pupa for about 24 days (Dec. 30th to Jan. 25rd). Common in gardens in Maritzburg. 20. DASYCHIRA GEORGIANA, sp.n. (Plate XLIX. fig. 19, imago; fig. 20, larva ; fig. 21, pupa.) Allied to D. horsfieldii. Head and thorax white; branches of antenne fulvous. Fore wing white, irrorated with black scales; the antemedial lines more angled than in D. horsfieldii; the hind wing white, with some darker hairs on the internal area. Abdomen orange, with a dorsal line of black spots on first four somites. Underside white, no cell-spots. TRANSFORMATIONS OF SOME SOUTH-AFRICAN LEPIDOPTERA. 315 Larva. Ground-colour velvety black, with very long pale yellow hairs; each somite with two red and four greyish tubercles, from which spring the long silky yellow hairs, except the 4th, 5th, 6th, and 7th somites, on which the two dorsal tubercles coalesce, and from these spring four thick dorsal tufts of shorter grey hairs, and 11th somite, which bears a tuft of shorter blackish hairs. Head red above, black beneath ; thoracic legs and claspers red, and a red transverse band on 12th somite. Feeds on common wattle (Hlephantorrhiza burchelli). The pupa is formed in a slight web constructed from the hairs of the larva, and is pale fulvous with black markings; wing-covers greyish, with veins delineated by thin black lines. The descriptions are made from two specimens reared in Maritzburg in February and March. There is a female specimen of the imago in the British Museum unnamed. Family HYPSID A. 21. Hypsa apuipas (Hopffer)=H. subretracta (Walker). (Plate XLVII. fig. 16, larva.) Larva. Chocolate-brown ; under surface greyish ; an indistinct dark dorsal line. Three red tubercles, one dorsally and two subdorsally, on Ist to 3rd somites; paired red subdorsal tubercles, four on each somite, on a black subdorsal line, from 4th to 12th somite ; a broad white spiracular streak on 5th, 6th, 10th, and 11th somites. Head dark brown ; thoracic legs and claspers fuscous. A few sparse brown hairs on the body. Feeds on the common fig, both cultivated and wild; the specimen figured fed on the cultivated fig, but I have reared others which I found as pupe in hollows on the trunks of the wild fig-trees round the gardens in the square at Durban. When about to change, the larva rolls up a fig-leaf in which it spins a web, and the same process is gone through in the hollows of fig-tree trunks. ‘The specimen figured remained as a pupa from 18th February to 42nd March. Pupa black, with polished surface. Imago fairly common in Durban and Maritzburg. Family ARCTIIDA. 22. SPILOSOMA PUELLA (Druce). (Plate XLIX. fig. 15, larva.) Larva. Ground-colour dark fuscous, each somite with six large and two small black tubercles surmounted by tufts of black hairs. A pale ferruginous dorsal line, darker on the summit of each somite; under surface paler; head and thoracic legs black, spiracles ferruginous. ; Feeds on the Grenadilla passion-flower. vol. XV.— Part vi. No. 4.— April, 1901. 2Y¥ 316 ’ LIEUT.-COL. J. M. FAWCETT ON THE Pupa black, with polished surface, formed inside a cocoon constructed out of the hairs of the larva. ‘The imago emerged in about 17 days (March 8-26). Common in Maritzburg. Subfamily ARCTIIN &. 23, PHISSAMA FLAVA (Walker). (Plate XLIX. fig. 8, larva.) Larva. Ground-colour pale fuscous, each somite with a red-brown area superiorly, on which is situated a dorsal line of white spots surrounded by black, one on each. somite; also a series of black tubercles bearing tufts of black and reddish hairs. Head, thoracic legs, and claspers ferruginous. Feeds on Ornithogalum eckloni, a species of lily, and other plants. When ready for its transformation the larva weaves a cocoon with its hairs and remains a pupa for about 17 days (Ist to 17th December in the case of the specimen figured). The imago is common in Maritzburg district. 24, PHISSAMA SCREABILE (Walker). (Plate XLIX. fig. 9, larva.) Larva. Ground-colour greyish white, the somites broadly black transversely and beset with long black and white hairs; the white spaces between the somites traversed by fine black transverse lines ; under surface black. Head, thoracic legs, and claspers bright red. Feeds on Ornithogalum eckloni, also on mealie: a larva reared on the latter plant differed from the individual figured in its lighter parts being buff instead of white ; this larva produced a female imago. The pupa is formed inside a cocoon constructed with the hairs of the larva. Duration of the pupal stage about three weeks (9th to 29th December, specimen figured). Imago is common in Maritzburg district. Subfamily NycTEOLIN &. 25. Farias INSULANA (Boisduval). (Plate XLIX. fig. 25, larva; fig. 26, cocoon.) Larva. Ground-colour pale bluish-black, with deep black markings; Ist to 3rd somites pale brown; 4th somite with a white transverse streak; an interrupted white dorsal line from 4th to 8th somite; 9th to 12th somite pale brown dorsally. A sub- dorsal line of paired short brownish spines with scarlet bases; a lateral line of similar paler spines. Head and legs pale brown. Feeds on Hibiscus. ‘The pupa is formed in a stout thickly-woven cocoon attached to stem of food-plant. The larva is figured x $, to get in detail. Common in gardens in Maritzburg. TRANSFORMATIONS OF SOME SOUTH-AFRICAN LEPIDOPTERA. 517 Family AGARISTID. 26. EUSEMIA BUTLERI (Walker). (Plate XLVII. fig. 14, larva; fig. 15, pupa.) Larva. Ground-colour light bluish-grey, with fine black transverse stripes and spots on each somite; a reddish spiracular line. Head, thoracic legs, and Ist and 11th somites superiorly reddish ochreous; a few sparse hairs distributed over the somites ; thoracic legs fuscous; ventral claspers blue-grey, anal claspers reddish ochreous. Feeds on common vine (cultivated). The larva undergoes its transformation underground, and remains a pupa about one month. Pupa reddish fuscous, very rugged in appearance, the abdominal somites very strongly defined. The imago has a very strong penetrating scent. Not uncommon in Maritzburg. Family NOCTUIDA. Subfamily TRIFIN &. 27. GLoTTULA PANCRaTI (Cyr.). (Plate XLIX. fig. 17, larva; fig. 18, stem of food- plant showing hole.) Larva. Reddish brown, each somite with three small black warty spots, one dorsally, two subdorsally, and two small pale yellow subdorsal spots; between each two somites a transverse row of five oblong pale yellow spots, divided mesially by the intersection of the somites. Head, legs, and claspers ferruginous; a few short fulvous hairs on the somites. This larva feeds on a species of lily, eating the flowers as well as the stems; also making a burrow inside the stem (as shown in fig. 4¢). It can eat its way out of any box, except tin. When full-grown it undergoes its transformation underground. Pupa red-brown, with strongly-marked somites. The tmago is noteworthy owing to its exceedingly small and short pale green haustellum, which would seem to be of little use to it. 28. AGRoTIS SEGETIS (Schiff.). (Plate XLIX. fig. 16, larva.) Larva, Ground-colour pale brown, with small black spots placed irregularly on the somites ; fuscous dorsal and lateral lines, the former narrow, the latter rather broad and bounded inferiorly by a narrow white spiracular line. Under surface, thoracic legs, and claspers pale fulvons. Head ferruginous, the eyes defined by two crescentic black frontal lines. Feeds on the cultivated vine, and undergoes its transformation underground. The pupal stage lasted about one month, in March. Imago common in Maritzburg in February and March. 22 318 LIEUT.-COL. J. M. FAWCETT ON THE Subfamily QUADRIFIN &. 99, PoLYDESMA GLAUCINANS (Guen.). (Plate XLIX. fig. 22, larva; fig. 23, pupa in web.) Larva. Ground-colour bright grass-green, with a broad white dorsal line from head to 12th somite, which bears on its centre a narrow pale blue dorsal line ; an interrupted white lunular lateral line, bordered with black superiorly. Head, thoracic legs, and claspers green; the first pair of abdominal claspers aborted; the 12th somite tapering to a point above the anal claspers. Feeds on wattle (Elephantorrhiza burchellii). Pupa pale brown, formed in a cocoon among the stems of the food-plant. Imago very common in Maritzburg in March. 30. Tanropyea sytvina (Stoll), (Plate XLIX. figs. 10, 11, larva; fig. 12, pupa in cocoon.) Larva (early moults). Ground-colour buff; on each somite, except Ist and 11th, a broad very dark crimson transverse band; on Ist and 11th somites he band is pale ferruginous and bears some small black spots. In the final moult the above-mentioned bands become deep velvety black, with the exception of those on the Ist and 11th somites, which remain the same, the ground-colour becoming pale greenish. Head, legs, and claspers pale ferruginous. Feeds on Ornithogalum eckloni (Sch.), a species of lily common in swampy places. When full-fed the larva burrows underground and constructs a chamber for the pupa, which is surrounded by a strong hard envelope, of a consistency similar to cement, with a small hole at one end for the exit of the imago. Pupa deep red, with black antemedial and postmedial lines on the wing-covers. The period of pupation lasts about five weeks. The imago is not uncommon in Maritzburg in January and February. Family GEOMETRID A. 31. Boarmra acactarta (Boisduval), (Plate XLIX. fig. 24, larva.) Larva. Ground-colour pale reddish brown, inclined to olive-green subdorsally ; a pair of dorsal processes with whitish centres on 5th somite, and a pair of dorsal white spots on 11th somite. Head, legs, and claspers reddish. Feeds on a small species of sunflower, which is a common plant in gardens at Maritzburg. The transformation to pupa is effected underground. TRANSFORMATIONS OF SOME SOUTH-AFRICAN LEPIDOPTERA, 319 My specimens of the imago were identified as B. acaciaria at the British Museum, but they are much paler and have fewer transverse bands than the examples in that collection, or the specimen figured by Hampson in ‘Fauna of British India,’ vol. iii. p. 265. IV. EXPLANATION OF THE PLATES. PLATE XLVI. Fig. 1. Larva of Acrea petrea (Boisd.), p. 294. Figs. 2, 5. Pupa of ditto. Figs. 4, 5. Larva of Acrwa encedon (Linn.), p. 294. Fi . Pupa of ditto. cao Fig. 7. Larva of Acrwa rahira (Boisd.), p. 294. “Figs. 8, 9. Pupa of ditto. Figs. 11, 12. Pupa of ditto, Fig. 13. Pupa of Junonia cebrene (Trimen), p. 295. Fig. 14. Larva of Junonia clelia (Cramer), p. 296. Fig. 15. Pupa of ditto. Fig. 16. Larva of Charaxes candiope (Godart), p. 296. Fig. 17. Pupa of ditto. Fig. 18. Larva of Pieris severina (Cramer), p. 297. Figs. 19, 20. Pupa of ditto. Fig. 21. Larva of Pieris hellica (Linn.), p. 297. Fig. 22. Pupa of ditto. Fig. 23, Larva of Papilio policenes (Cramer), 1st stage, p. 298. Figs. 24, 25. Ditto, 2nd stage (dorsal and lateral views). Fig. 26. Ditto, 3rd stage. Figs. 27, 28. Ditto, final stage (dorsal and lateral views). Figs. 29, 30. Pupa of ditto (dorsal and lateral views). Fig. 31. Head of larva with V-shaped tentacles protruded. Fig. 32. Larva of Papilio brasidas (Felder), early moults, p. 298. 320 LIEUT.-COL. J. M. FAWCETT ON THE Figs. 35, 54. Ditto, final moult (dorsal and lateral views). Figs. 35, 36. Pupa of ditto (lateral and dorsal views). Figs. 37, 38. Larva of Papilio morania (Angas), early moults (lateral and dorsal views), p. 299. Fig. 59. Ditto, final moult. Fig. 40. Pupa of ditto. Fig. 41. Larva of Papilio nireus (var. lyeus) (Cramer), Ist stage, p. 501. Fig. 42. Ditto, 2nd stage. Figs. 43, 44. Larva of Papilio nireus (var. lywus) (Cramer), final stage, p. 301. Fig. 45. Head of larva with V-shaped tentacles protruded. Fig. 46. Larva of Papilio demodocus (Esper), early moults, p. 300. Figs. 47, 48. Ditto, final moult. Fig. 48 shows the attitude adopted by the larva when the V-shaped tentacles are protruded. Figs. 49, 50. Pupa of ditto (dorsal and lateral views). Fig. 51, Head of larva with V-shaped tentacles protruded. PLATE XLVII. Fig. 1. Larva of Actias mimose, early moults, p. 302. Fig. 2. Ditto, final moult. Fig. 5. Cocoon of ditto, showing ventilating holes. Fig. 4. Larva of Nudaurelia wahlbergi (Wallengren), p. 303. Fig. 5. Pupa of ditto. Fig. 6. Larva of Gynanisa maia (Klug), p. 304. Fig. 7. Pupa of ditto. Fig. 8. Larva of Bunea caffraria (Stoll), p. 305. Fig. 9. Larva of Urota sinope, p. 805. Fig. 10. Pupa of ditto. Figs. 11, 12. Larva of Pseudaphelia apollinaris (Westwood) (dorsal and lateral views), p. 305. . 13. Pupa of ditto. . 14. Larva of Eusemia butlert (Walker), p. 317. . 15. Pupa of ditto. Fig. 16. Larva of Hypsa aphidas (Hopff.), p. 315. Fig. 17. Larva of Cherocampa capensis (Linneus) (green form, final stage), p. 308. Fig. 18. Ditto, red form. Fig. 19. Pupa of ditto. eee Q aq’ 0q” ee de Tm Ot BR oo LO Cc oo TRANSFORMATIONS OF SOME SOUTH-AFRICAN LEPIDOPTERA. PLATE XLVIII. Larva of Cherocampa balsamine (Boisduval), p. 309. Pupa of ditto. Larva of Cherocampa osiris (Dalman), p. 309. Ditto, final change before pupation. 6. Pupa of ditto (ventral and lateral views). . Larva of Lophostethus dumolinii (Latreille), p. 307. . Pupa of ditto. . Larva of Protoparce nauritt (Butler), green form, p. 511. . Ditto, dark form. . Pupa of ditto. . Dorsal view of first four somites of larva, dark form. . Larva of Cephonodes hylas (Linn.), Ist form, p. 512. . Ditto, 2nd form. . Ditto, 3rd form. . Ditto, 4th form. . Ditto, dth form. . Ditto, 6th form. . Pupa of ditto. PLATE XLIX. . Larva of Cherocampa idricus (Drury), X $, p. 310. Pupa of ditto, x ¢. . Larva of Gonometa postica (Walker), red form, p. 315. Ditto, grey form. Cocoon of ditto. . Larva of Ludia smilax (Westwood), p. 305. . Cocoon of ditto. . Larva of Phissama flava (Walker), p. 516. . 9. Larva of Phissama screabile (Walker), p. 316. re, 1G), . Ditto, final moult. 2. Pupa of ditto, in underground chamber. . Larva of Dulichia fasciata (Walker), p. 514. . Ditto, dorsal view. . Larva of Spilosoma puella (Druce), p. 315. . Larva of Agrotis segetis (Schiff.), p. 317. . Larva of Glottula pancratii (Cyr.), p. 317. Larva of Teniopyga sylvina (Stoll), early moults, p. 318. os 322 TRANSFORMATIONS OF SOME SOUTH-AFRICAN LEPIDOPTERA, Fig. 18. Stalk of food-plant, showing entrance to chamber made by larva. Fig. 19. Dasychira georgiana, sp. n., p. 314. Fig. 20. Larva of ditto. Fig. 21. Pupa of ditto in web. Fig. 22. Larva of Polydesma glaucinans (Guen.), p. 318. Fig. 23. Pupa of ditto in web. Fig. 24. Larva of Boarmia acaciaria (Boisduval), p. 318. . 25. Larva of Earias insulana (Boisduval), x $, p. 316. . Cocoon of ditto. s. 27, 28. Larva of Zygena (Anteris) ampla (Walker), p. 313. . 29. Cocoon of ditto. ] Leal] a gq og dg dq 99 bo i=) 1 aransbolbte, Vol AVTENINT JM. del : F West,Newman chr RCKnight lith TRANSFORMATIONS OF SOUTH-AF RICAN RHOPALOCERA . ple UE JMF. del j West,Newman chr. ai C Knight lith TRANSFORMATIONS OF SOUTH-AFRICAN HETEROCKRA Sranaloo Soc. Vol VLE ALUM. ANS go” wy € ar Qi B — PANS tiem ue 1 Ona m Wiss It fe \ 4 WUUDROLY \AVBLRYe sabaye sett ) y I NP 0h aay Heres i West Newman chr E.C.Knight lith \ i e i . iy 2 - t c f < = t 1 ‘ fs iv Z oe ‘ s ; . Xe 5 # ; = a F ¥ coy \ par rs = - Bap bol ee ULLAL A By del t Knight lth, JM 1a West,Newman chr TRANSFORMATIONS OF SOUTH-AFRICAN HE TEROCERA. TRANSACTIONS OF THE ZOOLOGICAL SOCIETY OF LONDON (continued). To Fellows. To the Public. &. Ss. d. £ os. d, VOLUME XIII. (1891-1895, containing 62 Plates). . Price6 8 38 811 0 Parrl. (1891, contaming 6 Plates) . . .. . 0) Toro atest’ (0) » 2. (1891, containing 6 Plates) . . ... Soe 0 loee9 Ieee » 9 (1891, containing 6 Plates) . . . . . Ol oa0) 1 4 0 » 4. (1892, containing 1 Plate) . » O 4 6 0 6 0 » 9. (1893, contaiming 9 Plates) se gO ly mone. 015 0 » 6, (1893, containing 4 Plates) 5 Oe Ok Ore torrets O Deer », 7%. (1898, containing 6 Plates) . ... . Ota) 012 0 », 8. (1894, containing 6 Plates) . . . . . » 0 9 O 012 0 » 9. (1894, containing 6 Plates) . . .. . pe ORES 015 0 » 10. (1895, containing 5 Plates) . . .. . BO OO 012 0 », Ll. (1895, containing 7 Plates and Title and Index),, 015 9 Terle0 VOLUME XIV. (1896-1898, containing 47 Plates) . Priced 5 O oO) 0) Part l. (1896, contaming 2 Plates) . ... . 5s iO. aoe Oe: 0 8 0 » 2. (1896, containing 6 Plates) . . , . . 0 tO ON. 012 0 » 9% (1897, containing 9 Plates) AOC | Oras We rah 240) » 4 (1897, containing 10 Plates) . . ... Ti aleel Ores Users 2 0 0 oe 5e (S98, contami, «2 Plates)s 0+. mente » 09 6 0 0 8 0 » 6. (1898, containmg 7 Plates) . ... . » O18 O bad 0 » 7. (1898, containmg 6 Plates) . . .. . Haws ao 015 0 » 8. (1898, containing 5 Platesand Titleand Index),, 0 9 O 012 0 VOLUME XV. Parr 1. (1898, containing 8 Plates) . . . . . PriceO 15 9 Tspeles'O » 2 (1899, contaming 5 Plates)... . . SOOM IKe O One 7=0 » 9. (1899, containing 4 Plates) . . . . . Sst Ones O 012 0 » 4. (1899, containmg 4 Plates) . . .. . Ono) 20 012 0 3, ». (1900, containing 24 Plates) . . . . . eee 310 0 7 won (LoOl, contaiming: “4 Plates) ).1) 2a. ea Oe 015 0 VOLUME XVI. Part 1, (1901, containing 6 Plates) . . . . . PriceO 18 0... . 1 4 0 CONTENTS. VII. Notes on the Transformations of some South-African Lepidoptera. By Lieut.- Colonel J. Macoum Fawcerr. (Plates XLVI-XLIX.) . . . - page 291 THE PUBLICATIONS OF THE ZOOLOGICAL SOCIETY OF LONDON. Tux scientific publications of the Zoological Society of London are of two kmds—* Proceedings,” published in an octavo form, and “‘ Transactions,” in quarto. According to the present arrangements, the “ Proceedings” contain not only notices of all busi- ness transacted at the scientific meetings, but also all the papers read at such meetings and recom- mended to be published in the “ Proceedings”? by the Committee of Publication. A large number of coloured plates and engravings are attached to each annual volume of the “ Proceedings,” to illustrate the new or otherwise remarkable species of animals described in them, Among such illustrations, figures of the new or rare species acquired in a living state for the Society’s Gardens are often given. The “ Proceedings” for each year are issued in four parts, on the first of the months of June, August, October, and April, the part published in’ April completing the volume for the preceding year. The “Transactions ” contain such of the more important communications made to the scientific meetings of the Society as, on account of the nature of the plates required to illustrate them, are better adapted for publication in the quarto form. They are issued at irregular intervals. Fellows and Corresponding Members, upon payment of a Subscription of £1 1s. before the day of the Anniversary Meeting in each year, are entitled to receive all the Society’s Publications for the year. They are likewise entitled to purchase the Publications of the Society at 25 per cent. less than the price charged for them to the Public. Al, 5 all ig. 22. ie. 26, THE ANATOMY OF POLYPTERUS, 337 kidney. ‘The section shows also the similar appearance of the testis-duct and ureter. The uriniferous tubules of the kidney are embedded in a mass of lymphoid tissue. A similar section passing behind the testis (p. 325). When compared with fig. 15 the figure shows the relatively great development of the main duct of the testis-ridge, the small development of the ductules of the ridge in number and size, and the small relative size of the ureter. A similar section in the region of the urinogenital sinus (p. 525), showing the opening of the vas deferens into the latter and its lining of large columnar cells. The large relative size of the vas deferens is seen on comparing with fig. 14. PLATE LIT. . A vertical section through the testis and ureters of a very young male P. senegalus, 9 cm. in length (p. 325), showing the precocious development of the testis-duct, the tubules of the testis not being yet formed. A similar section through the posterior termination of the testis-ducts (p. 325), showing that they are only just acquiring their connection with the urino- genital sinus. A vertical section through the developing ovary and ureter of a very young female P. senegalus, 9 cm. in length (pp. 325, 326, 327), showing the ova developing from the external wall of the ovary, the ovary divided into a number of loculi, and the artery and vein in the median wall. A similar section behind the ovary passing through the peritoneal opening of the oviduct (pp. 325, 326). . A similar section behind the peritoneal opening of the oviduct (pp. 325, 326), showing the oviduct lying between the mesoarium or female genital ridge and the ureter. ‘To be compared with fig. 17. . A-similar section through the posterior termination of the oviduct (p. 526), showing that they have not yet acquired their connection with the urino- genital sinus. To be compared with fig. 20. . Left operculum bearing external gill of Polypterus lapradei, 30 cm. in length (pp. 831, 332), showing the blood-supply to the gill and the efferent artery giving off a branch which runs parallel with the afferent artery. The muscles of the operculum and the gill are also shown. An enlarged drawing of the tip of the external gill (pp. 331, 352), showing its minute structure and the continuation of the afferent into the efferent artery at the end of the gill. . A pinnule of the external gill highly magnified (p. 331). 338 ON SOME POINTS IN THE ANATOMY OF POLYPTERUS. XIV. EXxXpLaANATION OF THE LETTERING OF THE FIGURES. abp. Abdominal pore. ovd.ep. Epithelium of oviduct. ab.p.c. Abdominal pore-canal. p.ep. Peritoneal epithelium. pl. Pinuule. aff.art. Afferent artery. pn. Pinna. an.f. Anal fin. an.f.m. Anal-fin muscle. ab.v. Abdominal vein. r. Rectum. sp. Spermatozoa. ao. Aorta. sp.art. Spermatic artery. 6.c. Blood-corpuscle. sp.ep. Spermatic epithelium. b.v. Blood-vessel. sp.v. Spermatic vein. b.w.m. Body-wall muscle. sw.bl. Swim-bladder. cv. Cardinal vein. t. Testis. cd.v. Caudal vein. t.d. Testis-duct. con.tis. Connective-tissue. t.d.ep. Epithelium of duct of testis. d.m. Dorsal muscle. t.d.sh. Testis-duct sheath. e.ov.w. External wall of ovary. tr. Testis-ridge. eff.art. Efferent artery. t.r.tbs. Testis-ridge tubules. fb. Fat-body. t.tbs. Testis-tubules. gap. Genital aperture. u. Ureter. g-p. Genital papilla. m.ov.w. Median wall of ovary. k, Kidney. kv. Kidney-vein. u.ep. Epithelium of ureter. u.g-.ap. Urinogenital aperture. u.g.s. Urinogenital sinus. u.s. Urinary sinus. lt. Lymphoid tissue. u.sh. Sheath of ureter. m.cps. Malpighian corpuscles. u.ths mes.o. Mesoarium. mes.t. Mesorchium. . Uriniferous tubules. v.d. Vas deferens. v.d.ep. Epithelium of vas deferens. n.o. Nucleus of ovum. v.m. Ventral muscle. o. Ovum. w.o.u. Point where oviduct ends in wall of o.at. Attachment of ovum. mmctes op.m. Opercular muscle. w.v.u. Point where vas deferens ends in wall ov. Ovary. of ureter. ov.d. Oviduct. a | | Trani Loc Soc Vol, AV FLL. 9p. v 3 Wilson Cambridge. ANATOMY OF POLYPTERUS. Trans.Loot. Soc Vol AV LLL gio ekinees EB ee eEere assesses RSs Gases setsee E Wilson ,Cambri ANATOMY OF POLYPTERUS Sry ae i Pane ok : ' hate } Vv : ; ‘ ‘ t . 2 : H ‘ ' ' n ‘ . 5 . ¥ (he f Frans, Loot, Soo Vol. SV, HOLL. S eo ot oe ted : at ef art. " gfart (SSS SGA ES SES O85 eae Seve Be, wou 24 ; 1 Wilson , Cambridge ANATOMY OF P OLYPTERUS . : ihiae 39 LIST OF THE PAPERS CONTAINED IN VOL. XV. Awnprews, C. W., B.Sc., F.Z.8., Assistant in the British Museum (Natural History). On the Extinct Birds of Patagonia.—I. The Skull and Skeleton of Phororhacos inflatus Ameghino Bounencer, G. A., F.R.S., F.Z.8. Report on the Collection of Fishes made by Mr. J. E. S. Moore in Lake Tangan- yika during his Expedition, 1895-96. With an Appendix by J. E. 8. Moors, A.R.C.S. Second Contribution to the Ichthyology of Lake Tanganyika.—On the Fishes ob- tained by the Congo Free State Expedi- tion under ‘Lieut. Lemaire in 1898 .... Brapy, G. Srewarpson, M.D., D.Sc., LL.D., F.R.S., C.M.Z.8. On the Marine Copepoda of New Zealand . Boneerr, J. S., M.A., F.Z.8., Trinity College, Cambridge. On some Points in the Anatomy of Poly- pterus Fawcerr, Lieut.-Colonel J. Matcorm. Notes on the Transformations of some South-African Lepidoptera .......... VOL xv.—Part vil. No. 3.—Apri/, 1901. Harmer, Srpnny F., Sc.D., F.R.S., Superinten- dent of the University Museum of Zoology, Cambridge. On a Specimen of Cervus belgrandi Lart. (C. verticornis Dawk.) from the Forest- Bed of HastyAneliagyancrierercie tie Pycrart, W. P. A Monograph of the Genus Casuarius. By The Hon. Watrer Roruscurc, Ph.D., F.Z.8. the Morphology and Phylogeny of the With a Dissertation on Paleognathe (Ratite and Crypturi) and Neognathe (Carinate). By W. P. Pycrarr Roruscuitp, Hon. Watrer, Ph.D., F.Z.8.. A Monograph of the Genus Casuarius. By The Hon. Watrer RornscHrcp, Ph.D., F.ZS. the Morphology and Phylogeny of the With a Dissertation on Paleognathe (Ratite and Crypturi) and Neognathe (Carinate). By W. P. PycraFrr Page 97 109 INDEX OF SPECIES, Abdimia, 246. Acartia denticornis, 34. ensifera, 33, 34, 49. Acherontia atropos, 307. Aerea acara, 293. buxtoni, 295, 319. encedon, 294, 319. lycia, 294. petraa, 294, 319. rahira, 294, 319. Actias mimose, 302, 320. Aipyornis, 171, 172, 173, 174, 175, 176, 179, 180, 183, 195, 221, 232, 235, 237, 260, 2638, 274, 76, 279, 281. Africa, on the Transformations of some Lepidoptera, of, 291-322. | _ Agrotis segetis, BG, Gell, Alestes imber?, 4. macrolepidotus, 5, 25, 95. Pp ; macrophthalmus, 5, 25, 95. rueppellir, 5, 6. Amymone claus, 40. Andrews, C. W. On the Extinct Birds of Patagonia. —I. The Skull and Skeleton of Phororhacos | inflatus, 55-86. Anomalopteryx, 179, 221. Anoplopterus platychir, +, 5, 24. Antherwa (Nudaurelia) cytherea, 303. —— (——-) menippe, 303. ( ) tyrrhea, 308. Apteryx, 76, 150, 154, 159-290. australis, 159, 163, 220, 225, 232, 234, 251, 254, 270, 277, 285. mantel, 155, 161, 162, 163, 169, 220, 223, 232, 234, 270, 273, 277, 286. haastu, 163, 220, 232, 234, 244, 254, 270, 277. 244, 196, EIC., IN VOL. XV. Apterya mantelli, 254, 286, 289, 290. owent, 168, 215, 220, 223, 225, 232, 234, 251, 254, 270, 277, 287. Aptornis, 69, 71, 72. Archeopteryx, 153, 260. Arius commersonii, 18. Artotrogus brevicaudatus, 48, 53, 54. ovatus, 48. | Auchenaspis biscutata, 5, 24, 95. Bagrus meridionalis, 4. Barbus sp., 6. pagenstechert, 5. trimaculatus, 4, 5. Barilius quentheri, 4. Bathybates, gen. noy., 15. ferow, 4, 15, 27, 30, 89. Boarmia acaciaria, 318, 322. Boulenger, G. A. Report on the Collection of Fishes made by Mr. J. E. 8. Moore in Lake Tan- Expedition, 1895-96, eanyika during his 1-30. Second Contribution to the Ichthyology of Lake Tanganyika.—On the Fishes obtained by the Congo Free State Expedition under Lieut. Lemaire in 1898, 87-96. Brady, G. Stewardson. On the Marine Copepoda of New Zealand, 31-54. Bubo maximus, 287. Budgett, J. S. On some Points in the Anatomy of Polypterus, 323-838. Bunea caffraria, 303, 320. Calamoichthys, 327, 330, 335. calabaricus, 334. Calanus arewicornis, 32. jfimmarchicus, 32. Caligus longicaudatus, 49, 53. Calodromas elegans, 154, 162, v41, 250, 288. 33B2 342 Canthocamptus furcatus, 45. Cariama, 60, 65, 68, 71, 72, 76, 77, 79, 80, 82, 83, 84, 85. Carinatze, on the Morphology and Phylogeny of the, 149-290. Oasuarius, a Monograph of the Genus, 109-290. sp., 117, 118, 130, 152, 139. altyugus, 120. —— uaruensis, 130. aurantiacus, 137. —— australis, 117, 118, 123, 124, 125, 128, 280, 290. becearti, 114, 116, 117, 118. — bennetti, 111, 112, 128, 188, 139, 140, 145, 146, 147, 148, 188, 227, 228, 251, 276, 285. maculatus, 112, 148. bicarunculatus, 111, 112, 116, 117, 122, 130) WSik Qi 25h 283285. casuartus, 111, 112, 113, 114, 121, 122, 124, 180, 181, 134, 189, 140, 147, 151, 152, 209, 227, 228, 257, 276, 283, 285, 289. australis, 110, 112, 122, 123, 209, 283. beccan lil 2 UG SIA ile SONS i 251, 257, 276, 283. intensus, 112, 121, 283. salvadorit, 112, 120, 121, 122, 187, 257, 2838. selatenv, V2> Ws, 119; 127, 124) 152) 155, 187, 257. violicollis, 112, 117, 122, 153, 257, 283. — edwardsi, 141, 142. —— emeu, 113, 130. galeatus, 118, 117, 130, 1381, 133, 146. javanensis, 114. johnson, 123, 124, 125. kaupi, 130, 132, 139, 143, laglaizei, 136. —— lorie, 111, 112, 142, 148, 144, 155, 169, 285. —— nowe-hollandice, 280. oceipitalis, 185, 136. orientalis, 114. -—— papuanus, 111, 112, 135, 136, 139, 140, 141, 142, 145, 147, 284, 285. edwardsi, 112, 141, 145, 284. philipi, 111, 112, 137, 138, 284, 129, 129, 188, 152, 153, INDEX. Casuarius picticollis, 111, 112, 142, 143, 144, 145, 251, 276, 284, 285. heckt, 112, 114, 148, 284. regalis, 124, salvadorii, 120, 121, 215. sclatert, 117, 118, 288. tricarunculatus, 120, 121. unappendiculatus, 111, 112, 130, 1382, 138, 134, 1385, 136, 137, 1388, 1389, 140, 153, 218, 228, 238, 248, 251, 252, 256, 257, 275, 284, 285, 287, 288, 289. aurantiacus, 112, 136, 284. occipitalis, 112, 135, 186, 137, 284. laglaizer, 136. rufotinctus, 112, 137. violicollis, 258. westermanni, 135, 189, 140, 141, 144. Catarrhactes chrysocome, 255, * Centronma thomsoni, 47, 54. Centropages discaudatus, 35, 50. pectinatus, 36, 50. typrcus, 30. aucklandicus, 35. Cephonodes hylas, 312, 321. Cervus belgrandi, on a specimen of, from the Forest- Bed of East Anglia, 97-108. —— browni, 104. carnutorun, LOG. dama, 99, 100, 104, 105, 107, 108. curyceros germane, 106. haiberniw, 106. giganteus, 98, 100, 101, 103, 104, 105, 107, 108. belgrandi, 106. carnutorum, 106. italice, 106, 107. , 106. —— typicus, 106. TU megaceros ruffiz, 107. verticornis, 97-108. Cherocampa balsamine, 309, 321, capensis, 308, 320. idvicus, 310, 321. osiris, 309, 321. Charaxes candiope, 296, 319. Chauna, 250. Chlamydoselachus, 333. INDEX. 543 Chlamydoselachus anguineus, 334. Chromis diagramma, 19. paterfamilias, 18. Chrysichthys cranchit, 95. Chaunga, 65, 84, 85. Ciconia, 71, 72, 80, 83, 246. Citharinus geoffroy?, 6 Clarias sp., 5. anguillarvis, 5, 24. liocephalus, 5, 24, 30. Clausocalanus arcuicornis, fureatus, 32, 33, 49. Oleta serrata, 40. Congo Free State Expedition, on the Fishes obtained in Lake Tanganyika by the, 87-96. Copepoda, on the New-Zealandian Marine, 31-54. Oorematodus shiranus, 4. 215 32, 33. Coryceus robusta, 4:6. Crypturi, on the Morphology and Phylogeny of the, 149-290. Ctenochromis nuchisquamulatus, 17. Cyclops ewarti, 38, 51. Dactylopus cinctus, 41. hanseni, 41, 52. strom, 41. tishoides, 41, 53. Danuis chrysippus, 293. aleippus, 293. dorippus, 293. Daphnis nerii, 310. Dasychira georgiana, 314, 322. horsfieldvi, 314. Diaphorapteryx, 71, 72. Dinornis, 168, 171, 173, ae 180, 181, 182, 183, 186, 193, 194, 210, 214, 221, 260, 262 267. maximus, 216. — robustus, 280. (Pachyornis) elephantopus, 138, 139. Diomedea, 62, 64, 72, 79, 80, 82, 84. Distichodus rudolphi, 6 5 263, | Drepanopus furcatus, 32, 33. pectinatus, 33. Dromeus, 76, 118, 131, 188, 153-288. ater, 172, 173, 175, 179, 217, 218, 222, 22 7, 234, 235, 267, 270, 275, 278, 287. — irroratus, 179, 267, 270, 271, 287. Dromeus novee-hollandic, 146, 153, 155, 175, 179, 217, 218, 929 297, 934. 935, 251, 270, 275; 278, 280, 285, 286, 287, 288, 289. Dulichia fasciata, 314, 321. Earias insulana, 316, 322. Echidna, 281. Ectinosoma australe, 39, 51. Eetodus, gen. nov., 21. descampst, 4, 90, 92, 96. longianalis, 91, 92, 96. melanogenys, 4, 21, 91, 92, 96. Emeus, 183, 186, 190, 191, 192, 193, 262, 286. DAenbone pinguis, 4. Entomolepis, gen. nov., 48. ovalis, 48, 54. Eretmodus, gen. noy., 16. cyanostictus, +, 16, 30. Eusemia butleri, 317, 320. Euterpe ucutifrons, 39. gracilis, 39. Extinet Birds of Patagonia, Andrews on the, 55-86. Falco tinnunculus, 260. Fawcett, Lt.-Col. J. Malcolm. Notes on the Trans- formations of some South-African Lepidoptera, 291-322. Flavia, gen. nov., +1. crassicoriis, +2, 51. Francolinus, 77. Fulica, Fundulus teniopygus, 5. 75, 76. Gasterosteus spinachia, 18. Genyornis, 267. newtoni, 282. Glottula pancratii, 317, 321. Glyptodon, 85. Fonometa postica, 313, 321. Grammatotria, gen. nov., 90. lemairri, 90, 96. Grus, 68, 71, 72, 79, 80, 83. 12, Gynanisa maia, 304, 320. Haplochilus johnstoni, +. -— tanganicanus, 5, 19, 25, 30, 96. On a Specimen of Cervus bel- grandi Lart. (C. verticornis Dawk.) from the Forest-Bed of East Anglia, 97-108. Harpacticus acutifrons, 39. chelifer, 44. Harmer, Sidney F. 544 INDEX. Harpacticus glaber, 44, 58. Mastacembelus flavomarginatus, 23. Hesperornis, 76, 150, 226, 260. —— liberiensis, 23. Hippateetryo, 109. —— loennbergii, 23. indicus, 114. —— marchit, 23. Hydrocyon forskalii, 5, 25. —— marmoratus, 22, 23. Hypsa aphidas, 315, 320. ——. moorit, 5, 22, 23, 30. subretracta, 315. —— niger, 23. Ichthyology of Lake Tanganyika, 1-30, 87-96. | —— nigromarginatus, 23. Tchthyornis, 260. —— ophidium, 5, 23. Idya furcata, 45. —— shiranus, 4, 23. Julidochromis, gen. nov., 11. tanganice, 23. ornatus, 4, 12, 30. Megalapteryx, 202. Junonia cebrene, 295, 319. Moore, J. E. S., on the Fishes collected in Lake clelia, 296, 319. Labeo sp., 5, 25. Tanganyika in 1895-96 by, 1-30. Mormyrops anguilloides, 6. forskalia, 5. zambanenje, 4, 6. —— mesops, 4. | Mormurus catostoma, 4. rueppellia, 5. discorhynchus, 4, 6. Labidocera cervi, 37, 50. longibarbis, 5. nerti, 38. oxyrhynchus, 5. Lamprologus compressiceps, 4, 7, 10, 29. Neognathe, on the Morphology and Phylogeny of congoensis, 10, 88. the, 149-290. elongatus, 4, 9, 10, 29. | New Zealand, on the Marine Copepoda of, 31-54. fasciatus, 4, 7, 10, 29. Nudaurelia wahlbergi, 303, 320. furcifer, 4, 9, 10, 29. | Nyeticorax, 245. lemairei, 88, 96. | Ocydromus, 72. —— modestus, 4, 8, 10, 29. | Otthona spinifrons, 38. moorii, 4, 8, 10, 29, 88. | Opisthocomus, 153, 239, 248. Laophonte meinerti, 40, 51. | cristatus, 241, 282. serrata, 40. Ornithorhynchus, 257, 281. Lates microlepis, 4, 6, 29, 88. | Osmerus, 334. niloticus, 7. | Pachyornis, 216. Lemaire, Lieut., on the Fishes obtained in Lake | Palaognathe, on the Morphology and Phylogeny of Tanganyika by the Congo Free State Expedi- the, 149-290. tion under, 87-96. Panochthus, 85. Lepidoptera, on the Transformations of some South- | Papilio brasidas, 298, 319. African, 291-322. | —— demodocus, 300, 320. Lepidosiren paradoawa, 334. —— leonidas, 299. Lepidosteus, 328, 330, 334. —— morania, 299, 320. Leptoptilus, 239. | —— nireus, 301. Lophostethus dumolinii, 307, 321. —— , var. lywus, 320. Tudia smilax, 305, 321. | —— policenes, 298, 319. Macrosila solani, 311. Paracalanus parvus, 32, 49. Malapterurus electricus, 5, 25. Paratilapia afra, 4. Mastacembelus sp., 5. cavifrons, 5. dewindti, 88, 96. dimidiata, 4. congicus, 23. eryptacanthus, 23. | INDEX. Paratilapia furcifer, 4, 14, 30. intermedia, 4, 13. leptosoma, 4, 14, 30. ——- livingstonii, 4. —— longiceps, 4, 15. longirostris, 5. macrops, +, 138, 30. modesta, 4. —— pfefferi, 4, 12, 30. retrodens, 5. —— robusta, 4. ventralis, 4, 13, 15, 30, 89. Patagonia, on the Extinct Birds of, 55-86. Paurocope, gen. nov., 46. robusta, 46, 53. Pelotrophus microcephalus, 4. microlepis, +. Peltidium interruptum, 44. novee-zealandiw, 44, 50, 53. Perissodus, gen. nov., 20. microlepis, 5, 21, 30. Petrochromis, gen. nov., 20. polyodon, 5, 20, 30. Phasianus, 77. Phissama flava, 316, 321. screabile, 316, 321. Pheenicopterus, 246. Phororhacos inflatus, on the Skull and Skeleton of, 55-86. Phroso, gen. noy., 42. gracilis, 42, 52. Pieris hellica, 297, 319. severina, 297, 319. Planema esebria, 295. Plecodus, gen. nov., 22. paradowus, 5, 22. Podargus, 253. Polydesma glaucinans, 318, 322. Polypterus, on some Points in the Anatomy of, 323— 338. bichir, 5, 26, 28, 334. laprade, 324-338. senegalus, 6, 324-338. Porcellidium fuluum, 46. Protoparce mauritii, 311, 321. Protopterus avnectens, 5. Pseudaphelia apollinaris, 305, 320. Pycraft, W. P. vo He Or Psophia, 71, 72, 79, 80, 83, 84. On the Morphology and Phylogeny of the Palseognaths (Ratitee and Crypturi) and Neognathe (Carinate), 149-290. Pyrameis cardui, 296. Ratite, on the Morphology 149-290. Rhea, 109, 150-289. and Phylogeny of the, americana, 154, 155, 1é : 245, 249, 250, 251, 253, 256, 267, 270, 219, 2, 279, 280) 285; 286; 287, 288; 289. darwim, 155, 153, 245, 246, 267, 270, 279. macrorhyncha, 158, 206, 217, 221, 245, 249, 251, 270, 275, 277. | Rhynchotus rufescens, 252. Rothschild, Hon. Walter. A Monograph of the Genus Casuarius, 109-148. Scutellidium plumosum, 45, 53. tishoides, 45. Serpentarius, 59, 60, 65, 77, 79, 83. Simochromis, gen. noy., 19. diagranma, 5, 19. Spilosoma puella, 315, 321. Struthio, 76, 150, 159-289. australis, 269, 275. camelus, 159, 215, 251, 269, 275, 289. casuarius, 113, 116. meridionalis, 206, 261, 269, molybdophanes, 223, 269, 27 7). 285. 2 5; Sula bassana, 255. Synodontis afrofischeri, 5. multipunctatus, 5, 24, 30, 95. —— schal, 6. smithii, 6. zambesensis, +. Teniopyga sylvina, 318, 321. Tanganyika (Lake), on the Fishes collected by Mr. J. E. S. Moore in, 1-30. , on the Fishes obtained by the Congo Free State Expedition in, 87-96. Telmatochromis, gen. noy., 10. temporalis, 4, 11, 30, 88. vittatus, 4, 10, 29. Temora longicornis, 35. tenuicauda, 34, 50. 346 Thalestris australis, +3, 52. ciliata, 43, 52. forficula, 43. Tilapia aurata, 4. burtoni, 5. — callipterus, 4. —— dardennii, 93, 96. —— desfontainesi, 94. grandoculis, 94, 96. johnstoni, 4. kirkti, 4. — luteristriga, 4. — lethrinus, 4. —— mucrolepis, 94, 96. —— mossambica, 4. ——— milotrca, 5, 6, 18. —— nuchisquamulata, 5. —— obliquidens, 5. — rendalli, 4. labiata, 5, 17, 30, 95. INDEX. Tilapia rubropunctata, 93, 96. —— sauvagii, 5. ——- shirana, 4. ——- simonis, 18. —— squamipinnis, 4. —— subocularis, 4. —— tanganice, 5. —— tetrastagma, +. —— trisirami, 6. —— williamsi, 4. Tisbe furcata, 45. Trematocara, gen. nov., 89. marginatum, 89, 96. Tropheus, gen. nov., 17. moorit, 5, 18, 30. Urota sinope, 305, 320. Xenorhynchus, 246. Xenotilapia, gen. nov., 92. sima, 92, 96. Zygena (Anteris) ampla, 318, 322. END OF VOLUME XV. PRINTED BY TAYLOR AND FRANCIS, RED LION COURT, FLEET STREET. TRANSACTIONS OF THE ZOOLOGICAL SOCIETY OF LONDON (continued). To Fellows. To the Public. AGE Ge Lars dg VOLUME XIII. (1891-1895, contaming 62 Plates). . Price6 8 3 811 0 Parrl. (1891, contaming 6 Plates) . ... . re Oblong 9, Ea) » 2. (1891, containing 6 Plates) . . . . . Be Wh lao) a eee ed) » 98 (1891, containing 6 Plates) . ... . 53) O18) 0 1 4 0 » 4 (1892, contaming 1 Plate). . ... . » O 4 6 0 6 0 » 5. (1898, contaming 9 Plates) .... . Ob 6) 015 O » 6. (1898, containing 4 Plates) ped) 0 012 0 », 7. (1898, containing 6 Plates) . . .. . FO) 0 012 0 » 8 (1894, contamimg 6 Plates) .... . eS) 0 012 0 » 9. (1894, contaiming 6 Plates) . . .. .- Pen OL es OzTa AO » 10. (1895, containing 5 Plates) . . .. - oe Om OREO aire 012 0 ,», 11. (1895, containing 7 PlatesandTitleandIndex),, 015 9 . Leola: VOLUME XIV. (1896-1898, containing 47 Plates) . Price 5 0 7 0EK0 Parr]. (1896, containmg 2 Plates) . . .. . OP mound, On 8= 20, » 2. (1896, containing 6 Plates) . . . . - Ae) 0 012 0 » 8. (1897, containing 9 Plates) . . . . . » O15 9 RAO) » 4. (1897, containing 10 Plates) . . . . . eeleel es O) 2 0 0 » 5. (1898, containing 2 Plates) . . . - - OOO 0 0 » 6. (1898, containing 7 Plates) . . .. . aye Op Ee) 1 » 7. (1898, containing 6 Plates) : j Ops: OnlbrO » 8. (1898, containing 5 Plates and Title and aos 0 9 O OnI2270 VOLUME XV. (1898-1901, containing 52 Plates) . Price5 15 6 714 0 Parr 1. (1898, containing 8 Plates) . . . . - fy Orbe EO » 2. (1899, containing 5 Plates) . . . - - sie Onto me O OTe O », 93» (1899, containing 4 Plates) . . .. - pa RRO SO) 0 12°.0 » 4. (1899, containing 4 Plates) . . . =. - Ohh 0) O26 ,, 5. (1900, containing 24 Plates) . . . . . ee LS 16 310 O » 6. (1901, containing 4 Plates) : : je Ogi) Rens: 0152-0 » 7. (1901, containing 3 Plates and Title and Cimaagie 0 9 0 012 0 VOLUME XVI. Parr 1. (1901, containing 6 Plates) . . . . . PriceO 18 O.. . 1,-4.0 CONTENTS. VILL. On some Points in the Anatomy of Polypterus. By J.S. Bupexntr, M.A., P.Z.8., Trinity College, Cambridge. (Plates L.--LIl.). . . . . - . « page 323 List of the Papers contained in Vol. XV. . . . . . . - . +. . . 339 Index ofSpecies zen Vol MVe eo Titlepage and Contents to Vol. XV. up scientific publications of the Zoological Society of London are of two kinds—“ Proceedings,” published in an octavo form, and “ Transactions,” in quarto. According to the present arrangements, the “ Proceedings” contain not only notices of all busi- ness transacted at the scientific meetings, but also all the papers read at such meetings and recom- ‘mended to be published in the “ Proceedings” by the Committee of Publication. A large number of coloured plates and engravings are attached to each annual volume of the “ Proceedings,” to illustrate the new or otherwise remarkable species of animals described in them. Among such illustrations, figures of the new or rare species acquired in a living state for the Society’s Garden are often given. The “ Proceedings” for each year are issued in four parts, on the first of the months of June, August, October, and April, the part published in April completing the volume for the preceding year. Commencing with January 1901 they will form two half-yearly volumes. The “‘ Transactions” contain such of the more important communications made to the scientific meetings of the Society as, on account of the nature of the plates required to illustrate them, are better adapted for publication in the quarto form. They are issued at irregular intervals. Fellows and Corresponding Members, upon payment of a Subscription of £1 1s. before the day of the Anniversary Meeting in each year, are entitled to receive all the Society’s Publications for the year. ‘They are likewise entitled to purchase the Publications of the Society at 25 per cent. less than the price charged for them to the Public. A further reduction of 25 per cent. is made upon purchases of Publications issued prior to 1871, if they exceed the value of five pounds. Fellows also have the privilege of subscribing to the Annual Volume of the ‘ Zoological Record’ for a sum of £1 (which includes delivery in the United Kingdom only), but this privilege only holds gocd if the subscription is paid before the First of December in each year. Such of these publications as are in stock may be obtained at the Society’s Office (3 Hanover Square, W.), at Messrs. Longmans’, the Society’s publishers (Paternoster Row, E.C.), or through any bookseller. April, 1901, pokes SCLATER, Secretary, oe poe ECPI i | | i | | Ey en Sawonsie Pees nen bent . eS Sepa reetn Sate eterno Bey Scae eeaNS, creed ae : poe tes Aastha Seay cirweea ASU Yeti tal, EPS Henn five