Triacantiiagyna and Gvnacantiia 23 



Thorax and abdomen as in Ihe male, to judge from preserved material, 

 the colors possibly slightly paler, especially those of the abdomen. Append- 

 ages similar to ditdcri in shape, basal narrowing slightly longer. 



Wings and legs as in the male. 



Of fourteen wings of males twelve had three cells in the most anterior row 

 in the anal loop, and two wings had two cells; of fourteen w-ings of females 

 all had three cells. In thirteen wings of males there were two cells just 

 posterior to A between the anal loop and the anal triangle, and in one wang 

 there were three cells; in the fourteen wings of females there were in every 

 case four cells posterior to A from wing base to anal loop. Twenty-two of 

 the twenty-eight front wings had one row of cells throughout between M. 

 and Rs, four wings of females had a single double cell, and one wing of 

 each sex had two double cells ; in all twenty-eight hind wings there were 

 invariably a single row of cells. 



In a dry brush-choked gully near Santa Marta we found slightly teneral 

 specimens of this species in December. Crawling about through the almost 

 impenetrable spiny vegetation the collector flushed occasional individuals, 

 a few of which it was possible to capture in his fingers. Other specimens 

 were taken soon after sunset flying along the road at Bolivar and other 

 specimens from various localities were flushed in forests. 



Caribbca, in the form of the superior appendages of the male, differs 

 from the other three species of the triUda group, but this difference is slight 

 and will be detected probably only by those having specimens of the several 

 species for comparison. The green color of the posterior surface of the second 

 femora, which thus resemble the first femora and differ markedly from the 

 third femora, is obvious in well colored specimens of both sexes. In ten- 

 eral individuals this green is replaced by gray and the color is less con- 

 spicious as the entire femur is paler in such specimens. But in these cases 

 the third femora are darker than the second, so the dissimilarity between 

 the second and third femora still holds. In trifida Rs forks more basally than 

 in other species of the group and this can be more readily detected in the 

 front wing (because of the greater proximity of the fork to the stigma) 

 than in the hind wing. In figure 146, Martin 25, t'.'ifida, Rs in the front 

 wing forks basal to the level of the last postnodal crossvein, and in the 

 hind wing basal to the level of the stigma a distance equal to or greater 

 than the length of the stigma. A comparison of Martin's figure 146 with 

 figure 3 of caribbea in this paper will show this difference, which however, 

 like the differences in appendages, is difficult of precise definition. See last 

 paragraph under T. ditdcri. 



Material examined : Mexico (i male, i female, U. S. N. M.) ; Guatemala, 

 Cayuga (Schaus and Barnes, forest, August 23, i female, A. N. S.) ; 

 Colombia, (Turbo, Mack, i male, M. C. Z.), Santa Marta, Dept. Magdalena 

 (December 13, 14 and 16, 1916, 9 males, 9 females, E. B. W.), Bolivar, 

 near Santa Marta (December 21, 23, 26, 1916, 2 males, t female, E. R. W.) ; 

 Puerto Berrio. Dept. Antioquia (February 8, 1917, i female, E. B. W.) ; 

 Venezuela, Caracas (Rolle. iQio, i female, Ris), San Esteban, Dept. Cara- 

 bobo (February 8, 1920, i female, E. R. W.), Palma Sola, Dept. Falcon 



