XIV 



INTEODUCTION. 



is supposed to be supported by the occurrence in the Eocene of Wyoming and Utah 

 of a fish, Priscacara, which has been referred by some authors to the Cichlidiu. 

 But Priscacara has ireither the reduced number of branchiostegals nor the toothless 

 palate which characterize all living Cichlidiie, and it appears to me that it has no 

 bearing on the -problem of how two very closely allied genera, Acara and Paratilaina, 

 differing from it and agreeing with each other in some important features which are 

 obviously due to their near relationship, have come to inhabit South America and 

 Africa respectively. 



DlSTEIBUTIOK OP GlCHLIDJ!. 



The Indian Etrojjhis is an isolated type, more specialized than Parctroi^lus of 

 Madagascar, which appears to be its nearest relative, and, as has already been stated, 

 a study of the American Cichlida; leaves no possible doubt that the Mexican and 

 Central-American Cichlid fauna has originated with immigrants from South America; 

 consequently the hypothesis of the northern origin of the family should, I think, be 

 rejected. We are therefore led to believe that in early Eocene times Africa was 

 connected by land with South America on the one side and with India, vid Madagascar, 

 on the other. 



The Mexican and Central-American Cichlidic number about G5 species; they are 



