PHILIP P. CALVERT 351 



On siimmiiijo; up the data given here and by Dr. Ris it appears 

 that all species of Epigomphus thus far stucUed in both sexes 

 show a differentiation in the two sexes in the rear of the head, 

 this being least marked in obtusus and llama and most striking in 

 arrnatus. The occiput of the two sexes is similar in suhsunilis 

 but more or less different in other species, tumefadus female show- 

 ing the most comphcated structure of this part. The vertices of 

 male and of female are similar in siihohtusiis, ohtusiis and sub- 

 similis, dissimilar in the others, the maximum of specialization 

 being shown by tmnefactus female in one way and bj' the females 

 of verticicornis and arrnatus in another. 



A comparison of the appendages of the males of these species 

 with each other does not furnish any sure basis for deciding that 

 these organs are more specialized in one species than another, 

 since specialization, from a phylogenetic standpoint, may be 

 either by reduction (atrophy) or by comphcation (hypertrophy). 

 Perhaps we may take the testimony of the other sex on this 

 question and infer that those male appendages are most special- 

 ized in the species whose female shows the greatest number of 

 mating adaptations, irrespective of the degree of com]ilication 

 shown by the male appendages themselves. 



Judged by similarity in head structure of the two sexes, suh- 

 similis is the least specialized of the Costa Rican forms; it is a 

 species in which the second pale antehumeral stripe (of both 

 sexes) is represented only by a superior spot. Of the species 

 which show the greatest differentiation in head structure between 

 the two sexes, arrnatus has the second pale antehumeral stripe 

 complete, while tumefadus and verticicorii'is are like siibsimiUs 

 in this stripe. 



The data given in this paper would seem to indicate that if 

 the existing species of Epigomphus have come into existence 

 gradually, none of them can stand in an ancestral position to 

 any other, but that they represent end twigs of a fairly divergent 

 genetic tree. On the other hand, of course, is the possil)ility of 

 origin by mutation. By either hypothesis the most interesting 

 feature of their evolution is that the two sexes appear to have 

 been modified correlatively in structures concerned in the act of 

 mating so essential to the continuance of the race. It would be 

 desirable to ascertain whether these correlations are so exact as 



TRANS. A.M. ENT. SOC, XLVI. 



