THE AMERICAN TETRAGONOPTERINAE. 47 



festce and Bryconamericus peruanus of the Pacific slope of Ecuador are more intimately related 

 than festa is to Astyanax anterior of the upper Amazon. And in this case, Astyanax brevi- 

 rostris or Bryconamericus brevirostris whichever it may be, is intermediate between the two. 

 I am not competent to say whether brevirostris is moving from Bryconamericus to become 

 an Astyanax, or whether it has just completed the reverse process. Certainly festa and 

 brevirostris are more intimately related, have had a common ancestor at a less remote time, 

 than either of them with an Astyanax or Bryconamericus of southeastern Brazil. 



We recognize two types of genera, one a group of closely related species, descended 

 from a common ancestor and having certain distinguishing characters in common. Phena- 

 cogaster is such a genus. The peculiar scaling of the ventral surface has been developed 

 but once; and the species are all closely allied, differing from each other in but a few char- 

 acters. The other, a polyphyletic type, consists of species having a certain combination of 

 definite characters in common which easily distinguish members of the genus, but which, 

 instead of indicating a single ancestral line from which the species have diverged, are acquired 

 possibly one at a time along distinct lines converging to a common definition. Sometimes 

 the polyphyletic origin can be detected, sometimes not. Bryconamericus seems to me to be 

 such a genus; Hemibrycon, Deuterodon, and the larger genera are probably also polyphyletic. 



" Since it is difficult, or impossible, to say in any case which of the given characters has 

 appeared first, it is extremely difficult to point out lines of evolution leading to different gen- 

 era or species. We can only insist that certain innate possibilities may become actualities 

 anywhere along the line, possibly wherever they may prove advantageous, though the advan- 

 tage, to say the least, is not always obvious. 



" We may be permitted to assume that the more frequent character is the primitive one, 

 although this is certainly not always a safe assumption." 



Deuterodon with the character h is a genus of polyphyletic origin. Deu- 

 terodon iquape is found in southeastern Brazil, Deuterodon nasatus in Central 

 America. These two species technically belong to the same genus, genetically 

 they are most certainly not derived from an immediate common ancestor and 

 it is very probable that Deuterodon pinnatus and Deuterodon potaroensis from 

 Guiana, and Deuterodon acanthogaster from the Paraguay, are also independent 

 derivatives from the genus Astyanax. 



A somewhat similar case is presented by a character not mentioned in the 

 list, because I hesitate to propose a generic designation for it. It is this: — 

 In the vast majority of the species of Characins, the innominate bone is feeble 

 and entirely concealed. In Deuterodon acanthogaster and in Astijanax mucrona- 

 tus, this bone has become firm and the anterior end projects out of the body as 

 a distinct spine. There is no doubt whatever that this modification is arising 

 independently in the two species. 



It must be quite evident from the foregoing that the subfamily is a paradise 



for the student of divergent evolution. But the very conditions that make it 



of interest to the student of evolution make it the despair of the systematist 



whose object is to express relationship by grouping the species in an orderly 



■ array of genera and the individuals in an orderly array of species, always, if 



