lY. — A Revision of the American Cichlidae.^ 



BY CARL H. EIGENMANN AND WILLIAM L. BRAY. 



Read Dec. 4, 1893. 



While examining the specimens of Cichlidse, of Cornell Univer- 

 sity, the need of a revision of the generic definitions became evi- 

 dent. We have attempted this revision in the present paper. The 

 synonymy of the genera has been added, and where we have had a 

 sufficient number of species, or where the genus has been small, we 

 have added keys to the species. Several genera, as Astronotus, 

 Crenicichla, and Greophagus, need revision ; but with the limited 

 amount of material at our disposal, such a revision would necessarily 

 have been a compilation and have failed entirely in its object. 



The specimens were largely collected by the late Frederick C. 

 Hartt. We have been able to examine them through the courtesy 

 of Prof. B. G. Wilder, of Cornell University. 



Analysis of the Genera and Subgenera of American Cichlid.^. 



a. Spinous and soft portions of the dorsal of equal extent, or the former the 

 longer. 

 h. First gill arch normal. (Without additional lobe above.) 

 c. Gill-rakers long. 



d. Crill-rakers close set and very long, setiform, numerous (about 85). 

 e. Anal spines three, alternating. Body compressed, oblong, covered 

 with scales of moderate size. Dorsal spines numerous (13 w 14). 

 Each jaw with a front series of small awl-shaped teeth, behind 

 which is one or more series of smaller teeth. Cleft of month of 

 moderate width. Scales on cheeks in five or more series. Soft 



dorsal and anal naked 1. CliaetobrailcllUS 



ee. Anal spines six. Dorsal spines 15 or 16. Preorbital less than 

 orbit in width. Scales on cheek in 3 or 4 series. Soft dorsal and 

 anal scaled 2. CliaBtolvrailCllopsis 



1 Contributions from the Zoological Laboratory of the Indiana University, 

 No. V. 



Annals N. Y. Acad. Sci., VII, Jan. 1894.— 40 



