ODONATA OF BAJA CALIFORNIA. 521 



of variation which may be estimated at .15 of one per 

 cent. Kirby (I. c.) gives as a generic character of Or- 

 themis that the internal triangle of the front wings is 4- 

 celled* ("subtriangular space consisting of four cells"), 

 and four is indeed the more usual number, but that num- 

 ber varies too much to serve as a generic distinction. 

 Thus in the same lot of 208 6 , 119 ? , it was found that 

 the internal triangle was 3-celled in both wings of 9 indi- 

 viduals, in one wing of 25, and 5-celled in both wings of 

 23, in one wing of 55 individuals. Summarizing these 

 statistics in another way, the number of cells in the inter- 

 nal triancrle of the front wings is asvmmetrical with re- 

 spect to the right and left sides of the body in 80 individ- 

 uals out of 327, or 24>^%. Of these 80, 72 have 4-cells 

 in one front wing, 3 or 5 in the other, the 3 or 5-celled 

 wing being the right one in 43, the left one in 29. Among 

 those individuals in which the 3-celled condition exists in 

 one or both front wings, males and females are almost in 

 equal numbers, but where the 5-celled condition in one 

 or both front wings occurs the males are almost three 

 times as numerous as the females. 



One cross-vein in the discoidal triangle of the front 

 wings is a very constant feature; out of the same 327 in- 

 dividuals, two cross-veins were observed in the left wing 

 of a single male; a free condition of this triangle was not 

 found. 



Abdomen 5 32-35, ? 34-36. Hind wing 5 39-43' ? 4^- 

 44. 



Distribution. From Florida and Texas to southern 

 Brazil and Chili. 



'His figure, I. c, pi. Ivii, fig. 3, however, shows five cells. 



