70 AMPHIBIA 



Absence of Urodela (except Amhly stoma persimile*), of Cystigna- 

 thidae, and practically of the Hylidae, only two of which occur 

 in the Himalayan district. But this great chain of mountains 

 should not he included within the region, while the outlying 

 spurs in Upper Burma (with Amhbjstomo^ are debatable ground. 

 The subdivision of this widely extended region is beset witii 

 difficulties, chietly on account of Madagascar and Papuasia. The 

 fauna of Madagascar is very remarkable. All its Amphibia are 

 Eirmisternal, a mixture of African and Indian forms. The island 

 agrees with Africa, in opposition to the Oriental countries, in no 

 special point ; all the Kaninae, except Megalixalus, Bcqiida, and 

 two rather common species of Bana, belong to different genera. 

 Madagascar differs from Africa by the absence of Apoda, of 

 Aglossa, and Bufonidae. On the other hand, it agrees with India 

 or with the Malay islands, in opposition to Africa, by the possession 

 of Dyscophinae, of the Kanine genus Bhaco2)horus, and the Engy- 

 stomatine genus Calophrymis. 



Africa and India agree with each other, and differ from 

 Madagascar by the possession of Apoda, the genera i>V/o and 

 Necto-pliryne, and by the close resemblance of several genera of 

 Eaninae. 



India, the Malay islands, and Papuasia witli Melanesia 

 possess Pelobatidae {LeptohracMtiin^^ Batrachopsis* Aderophn/s*), 

 and thereby differ consideralily from Africa and Madagascar. 

 Batrrichylodes* of the Solomon Islands has unmistakable affinities 

 with Bhrynodenna* of Karen, between Burma and Siam : Orco- 

 hatrachus* of Borneo mucli resembles Bhrynohatrach'us* (if West 

 Africa ; and Cornufer, typical of the Malay and Melanesian 

 islands, occurs also in West Africa. All these Eaninae indicate 

 that the Austro-^Ialayan and Melanesian islands belong to the 

 Palaeotropical region. Ceratohatrachus* type of a sub-family, is 

 peculiar to Melanesia. 



There are c(inse(|uently several possil)le modes of subdivision, 

 all with a different result, according to the group of Amphibia, 

 which we may select as of leading importance, e.g. .\poda or 

 Peloljatidae, <ir J3yscophinae and BIiaco2)horvs. The Engy- 

 stomatinae and Eaninae are to be eliminated, since tliey occur in 

 all the countries in question. We have either to leave the 

 whole region luidivided — and it is a significant tact that tlie 



'■ indicates Ampliibia wliieli are ]if<niliar to tln' r('S])C'etive regions or suli-regions. 



