SKELETON AND LAETKX OF XENOPUS AND PIPA. 119 



anything longer than the fourth, — a feature unparalleled among 

 living Anura (23. p. 178) ; and these facts acquire an exceptional 

 significance when taken in conjunction with the observations of 

 "Wolterstorff, that in PalceobatracJius the third and fourth digits 

 of the hind limb are either equal in length, or the third is but 

 slightly shorter than the fourth (43. II. p. 16 [96], and I. p. 42). 

 Further: — "Die Lange der Metacarpi, die stumpf'e Beschaffen- 

 heit des Pusses haben Palceolatrachus und Xenopus mit Pipa 

 gemeinsam" (43. II. p. 75). 



PalcdohatracTius is unique among Anura in having the diapo- 

 physes of the seventh, eighth, and ninth vertebrae confluent to 

 form a disc, perforated by two foramina which mark the limits of 

 the three diapophyses (43. p. 30). Among living Anura a multiple 

 sacrum is only met with in Pelohates and Pipa, although it may 

 occur exceptionally in the frog (Howes 21) and Bomhinator (Gotte 

 17; Camerano 5). In Pelohates and Pipa the expanded plate to 

 which the ilium is attached is formed by the united diapophyses 

 of the ninth and tenth vertebrae ; and in the late embryo of Pipa 

 (Stage V^I., anted) there is present near the posterior edge of the 

 plate a small triangular foramen which marks off the small trans- 

 verse process of the tenth vertebra from the expanded diapo- 

 physis of the true sacral (ninth) vertebra. The outlines of the 

 ninth and tenth vertebrae cannot be distinguished with certainty 

 in the adult. In both Pelohates and Pipa the urostyle is 

 immovably fused with the sacrum *. In Xenopus also the sacral 

 diapophyses are greatly expanded, and the urostyle is confluent 

 with the sacrum. The sacrum, however, is simple, and the lateral 

 plates consist of the diapophyses of the ninth vertebra only. 

 The tenth vertebra is not differentiated in Palceohatraclius, but 

 is included, as it is in most Anura, in the urostyle. This latter 

 is not confluent with the sacral vertebra, but articulates (43. 

 p. 29) on to a pair of condyles situated on the posterior surface 

 of its centrum. 



There are eight free presacral vertebrae in Xenopus^ but in 

 Pipa only seven, the first two vertebrae not being differentiated 

 from one another. In Palceohatraclius also the first two vertebrae 

 are confluent, and, since the seventh and eighth enter into the 

 formation of the sacrum, there are but five free vertebrae. The 



* Cope's statement (Q. p. 98 and 7- P- 252) that in Fipa the urostyle is 

 simple and attached to a single condyle is, if not misleading, liable to mis- 

 interpretation. 



