424 On Freshicater Fishes from Madajascar. 



])arasplienoicl only; articular surface broadly ovate, almost 

 iieart-.sliaf)ed. Vertebrae 34 (17 + 17); fourth witii a pair ot" 

 very small inferior apophyses ; praecauelals with parapopliyae3 

 from the fourth ; ribs subsessile. 



]\Iaclagascar ; two sj^ecies. 



This genus is quite distinct from any of the African genera, 

 but is closely related to the Indian Etroplus, wliich dilFers 

 from Paretrophis in its more generalized dentition, the jaws 

 with 2 or 3 series of tricas))id teeth, those of tiie outernjost 

 series enlarged, in the adult truncate, often without lateral 

 cusps, and the lower ))haryngeal with most of the teeth 

 slender, uni- or bicuspid, only the two middle rows being 

 formed of large blunt teeth. 



The Madagascar Cichlidse belong to three endemic genera, 

 two of which appear to be related to West-African genera, 

 whilst the third is closely related to, but more specialized 

 than, the only Indian genus of the family. Except the 

 Oichlida;, none of the families of fishes characteristic of the 

 fresh waters of Africa occurs in Madagascar, which is poj)U- 

 lated chiefly by freshwater genera or species of marine 

 families (Kuhliidw, Atheiinidae, Eleotridaj). The Ostario- 

 physi, which are dominant in the freshwater fauna of :dl 

 other parts of the world except the Australian llegion, are 

 absent from ]\radrtgascar, except for two species of the endemic 

 genus Anc/iarliis^ which belongs to the Ariidae, one of the 

 two families of Siluroids that form a!i exception to the rule 

 that the Ostariopliysi are strictly freshwater fishes. 



The presence of Cichlidre in Madjigascaris probably due to 

 the fact that some fishes of this family are found in waters of 

 fairly high salinity. Species of each of the three Madagascar 

 genera have been found in brackish lagoons on the coast, 

 whilst Ktroplus suratensis of India and Ceylon is charac- 

 teristically an estuarine fish, and, according to Day, " extends 

 its range into brackish or even saline water." It is evident 

 that Madagascar has not been connected during the Tertiary 

 with either Africa or India to an extent that sufficed for the 

 passage of true freshwater fishes, but it may have receiveii its 

 Cichlidre from Africa at a time when it was only narrowly sepa- 

 rated from or even temporarily connected with that continent, 

 and pi'rlia|)S from India when the islands of the Indian Ocean 

 were more extensive and a brackish-water fish might |)ass 

 from one to another ; this time can hardly have been later 

 than the beginning of the Miocene. 



