5 IT) PAKENTAL (WTtE AMON(4 FRESH-WATER ETRHES. 



The liitest monographer of the family, J. Pellegrin, in 1904 recog- 

 nized 294 species of the family, and of these 161 were inhabitants 

 of Africa (including Syria) and 133 of America. The 294 were 

 ranged by him under 55 genera. The chief authority for African 

 fishes (Boulenger) in 1905 admitted 179 species for that continent, 

 and more recently Regan has enumerated 133 American species. 

 Including subsequently introduced or fortified species, there are now 

 known nearly, if not quite, 350 species. 



It is especially noteworthy that each genus is strictly limited to one 

 continent or at least continental area, not a single African having 

 American representatives, nor any American an African one. There is 

 not, however, a natural aggregate of American types to l)e contrasted 

 with another of African. Nevertheless most of the American types are 

 closely related to each other, and there is not the diversity that is 

 manifested by the African. In one respect, nevertheless, the varia- 

 tion is nnich greater. All the African species have three anal spines, 

 and three only, while in the American they range from three to as 

 many as fourteen. 



Most of the American species at least are very much like the famil- 

 iar sunfishes of the North American streams and lakes, and when the 

 writer first saw a small school of coscorobs {CicJilaHoma) in a river 

 of Trinidad lie thought it was one of the familiar sunfishes. Not 

 until he had carefully examined some that were caught was he un- 

 deceived. A couple of writers of an illustrated book on the Fishes of 

 Guiana (1841) were not only impressed with the likeness, but per- 

 petuated their impressions in the nomenclature, actually referring 

 species of the family to the genera " Pomotis " and " C entrdrchua /" 

 and so also had two very great ichthyologists (Cuvier and Valen- 

 ciennes) been misled by a colored figure of a Cuban fresh-water fish, 

 which they referred to the genus CentrarehuH^ but which proved to 

 l)e a member of the characteristic middle American Cichloid genus 

 Heros or Clelddsoma. Nevertheless the likeness of Cichlids to Cen- 

 trarcliids is entirely superficial, for they differ in the single nostrils, 

 toothless palate, the lateral line, the number and arrangement of 

 l)ranchiostegal rays, the single lower pharyngeal bone, and various 

 other anatomical characters. Still they may have originated from 

 not very remote common progenitors, which were, however, neither 

 specialized Centrarchids nor Cichlids. 



The Cichlids exhibit remarkable diversity of oviposition and care 

 of the eggs and embryonic young. Some " lay their eggs in the 

 sand," or, like the northern sunfishes, " build a kind of nest in the 

 sand or mud in which they deposit their eggs, hovering over them 

 until the young are hatched." Among such are the Flygrogonus 

 {Astronotus) and Chcetohranchus^ according to Agassiz. One 



