Mr. G. A. Boulenger on H^-nienochirus. 123 



not even to belong to tlie family Dactylethridas, distinguished 

 from the Pipidas by the presence of teeth in the upper jaw. 

 The mouth is edentulous, and the structure of the vertebral 

 column and of the pectoral arch has more in common with 

 Ptj)a than with Xenopus, the presence of claws to the first 

 three toes being the only point of special affinity with the latter. 

 Hymenochirus should therefore enter the Pipid^i, if dentition 

 be deemed of sufficient taxonomic importance for family 

 distinction. We have then a very interesting example of 

 geographical distribution before us, and one which should 

 remove any doubts as to the natural affinity between 

 Pipa and Xenopus, which has been questioned by some 

 authors, whose doubts have, however, not received the endorse- 

 ment of recent investigators on the anatomy and development 

 (c/. Beddard, P. Z. S. 1895, pp. 827 & 841, and Ridewood, 

 J. Linn. Soc. xxvi. 1897, p. 53, and Anat. Anz. xiii. 1897, 

 p. 359). The presence of only six distinct pieces in the 

 vertebral column of Hymenochirus, as I have ascertained, is 

 so far unique among Batrachians, and is only approached by 

 the number (8) in Pipa and (7) in the fossil genus Pakeo- 

 hatrachus. I do not consider this as in any way adding to 

 the evidence in favour of the allocation of the latter to 

 the Aglossa, as advocated by some authors, since the reduc- 

 tion in the normal number is obtained in a quite different 

 manner. 1 think the discovery by Ridewood of distinctly 

 segmented ribs in the larvse of both Pipa and Xenopus tends 

 to further separate PalcBohatrachus, of which the larvae are 

 well known and show no distinct ribs, from the Aglossa, and 

 to justify their provisional retention among the Pelobatidai. 



In Xenopus there are 8 distinct prassacral vertebras, as 

 normal in Tailless Batrachians ; but in Pipa and in Palceo- 

 hatrachus the first vertebra is formed by the fusion of two, 

 as proved by the passage of the first spinal nerve through 

 the neural arch and the presence of a diapophysis, which is 

 invariably absent from the atlas ; this is also the case in 

 Hymenochirus, in which the sacro-coccygeal complex is seen 

 to be composed of the sacral vertebra and the urostyle, as in 

 Xenopus and Pipa, with only four vertebrte intervening- 

 bet ween it and the first, thus apparently ])ointing to excalation. 

 In PaheohatracJius, according to Wolterstorff (Jahrb. nat. 

 Ver. Magdeb. 1886, p. 31), 9 vcrtebrge may be recognized, 

 viz., 1st and 2nd fused, 3rd, 4th, 5th, 6th free, 7th, 8th, 9lh 

 lused to form a sacrum. 



On the whole the skeleton of Hymenochirus much more 

 resembles that of Pipa than that of Xenopus. The following 

 characters ore common to all three genera : — 



