56 ANNALS NEW YORK ACADEMY OF SCIENCES 



they are based on cenotelic characters. The generic characters are often 

 no better, because they, too, are usually cenotelic. 



Take, for example, the genus Cichlasoma, which is primarily distin- 

 guished by the presence of four to thirteen anal spines from the genus 

 JEquidens, which has only three anal spines. For all we know, a muta- 

 tion with four anal spines could have easily appeared from the three anal 

 forms, or vice versa. In fact, that is exactly what seems to be the case 

 in Eio Sao Francisco, where I found only two specimens with three anal 

 spines among thousands of the four-anal -spine form, Cichlasoma bimac- 

 ulatum. In southern Brazil, I found just the opposite, namely, only 

 three four-anal-spine forms of the same size, color, etc., as thousands of 

 three-anal-spine forms which are known as /Equidens portalegrensis. 

 Hence all such closely drawn genera must be used with great care in the 

 determination of the point of family origin and dispersal, because they 

 are based on such characters that a mutation or individual or local varia- 

 tion might readily establish a new genus and species in a very short time. 



The importance of having a clear idea of the exact status of the ich- 

 thyological taxonomy, and especially the value of generic and specific 

 character and which are paleotelic and which are cenotelic, are of prime 

 importance in the explanation of the distribution of the fishes. For ex- 

 ample, the writer described Geophagus brasiliensis iporangensis as a new 

 variety from the headwaters of Eio Eibeira de Iguape. This variety may 

 be a distinct species, and it may be only a somatic change due to the en- 

 vironment. The adult forms certainly look and measure like a distinct 

 species, but I observed that the young individuals, one and two inches 

 long, could not be told from the lowland young of Geophagus brasiliensis. 

 It is highly probable that this new variety would produce typical Geo- 

 phagus brasiliensis, if it were removed to the lower course of the same 

 river. The same may be true of Crenicichla iguassuensis, also described 

 by me from the Eio Iguassii. This species is closely related to C. lacus- 

 iris and would perhaps produce that species, if subjected to the same ex- 

 ternal conditions. These two examples illustrate the present status of 

 South American ichthyology. One can, I believe, readily separate out a 

 couple of hundred forms of the two thousand catalogued for South 

 America which are probably not species, but merely "ontogenetic species," 

 showing, for example, somatic changes of a color spot, a few more or less 

 scales and spines or some other trivial physiological difference. Many 

 of these species, in fact, are based on single characters, such as teeth, 

 which are variable structures. The same have been shown experimentally 

 with birds, beetles, butterflies, etc., living in different temperatures, etc., 

 iind until the fishes are better and experimentally known, their present 



