132 MB. G. A. BOULENGEU OK TltE AFRICAN AND [Mar. 1, 



2. A Revision of the African and Sja'ian Fishes of the 

 Family Cichlidce. — Part I. By G. A. Boulenger, F.R.S. 



[Received February 18, 1898.] 



(Plate XIX.) 



The CicJtlicld' form a very natural family of Perciform Acantho- 

 pterygiaus, which may be defined as follows : — 



No suborbital lamina of the suborbitals ; entopterygoid present, 

 small. Prsecaudal vertebrae with transverse processes from the 

 third to the last ; ribs, all but the last few nearly sessile, inserted 

 behind the transverse processes, not at their extremity, and 

 narroAvly separated if at all from the centra. Nostril single on 

 each side. Gill-membranes free from isthmus : five or six branchio- 

 stegals ; gills four, a slit behind the fourth ; pseudobranchiae 

 absent. Lower pharyngeal bones united, with persistent median 

 suture. 8oft portion of dorsal fin not more developed than the 

 anal : latter with three or more spines. 



The term " Chromides," by which this family is often designated, 

 is inadmissible, the name Chromis having been originally intended 

 for Sparus cliromis, h. = Chromis castanea, Eisso, a Heliastes of 

 Cuvier and Valenciennes. Cuvier, in the second edition of the 

 ' Regne Animal,' i-eferred to Lahrns niloticus only as a second 

 species of his genus Chromis, the definition of which is drafted 

 from Sjjarus chromis. tipariis niloticus can therefore not be retained 

 as the type of Chromis, and must be placed in the genus Tilapia of 

 Andrew Smith, in the family named Ciclilicla- by Bleeker (1859). 



This family includes a large number of bracJdsh- and freshwater 

 fishes from Africa, Syria, India, and Central and South America, 

 one species extending into Texas. 



Tlie nearest affinity of these Pishes is with the Centrarchiclce and 

 Fcrciche. A study of their skeleton sho\\s them to bear no special 

 relation to the Pomacentricla', Lahrida/, Scariclo;, or to any other 

 division of the ' Pharyngoguathi,' than which a more artificial 

 group has never been conceived. Yet, I regret to say, it is very 

 slow at being altogether abandoiied, remains of it lingering here 

 and there in the works of even the most advanced reformers — to 

 wit. Gill's superfamilies Pomacentroidea and Lahroidea (1893), 

 corresponding to the Suborders Chromides and Fharpu/oc/nathi of 

 Jordan and Evermann (1896). 



In the present paper I shall deal only \^'ith the genera repre- 

 sented in Africa and Syria, of which I am able to distinguish 

 nine. All agree in having the palate toothless, no supplemental 

 maxillary, a single, continuous dorsal fin, and two distinct lateral 

 lines. The number of vertebrae varies, in the specimens examined 

 by me, between 26 and 38 (13-19 + 12-19). 



I am under deep obligations to Professor Vaillant for his kind 

 assistance in communicating to me the types of several species 

 preserved in the Paris Museum. I have thus been enabled to 



