THE SUBFAMILY LIBELLITLBLE. 255 



base of the triangle, is much broader and shorter than on the front wings. In one or 

 two small genera with aberrant neuration (Nannodythemis &c.) the triangle of the hind 

 wings is replaced by a trapezium. The hind wings are nearly always broader than the 

 fore wings, and the neuration of their lower portion is dissimilar. There is nothing 

 properly corresponding to the subtriangular space \ but the lower sector of the triangle 

 is generally much better developed, and rises with the upper sector of the triangle from 

 the lower angle of the triangle. It runs subparallel with the upper sector, but is rather 

 more irregular. It varies slightly in position, and in some genera (Mesothemis, 

 Lep'themis, &c.) the sectors of the triangle are not united at their base, and in 

 Nannodythemis the nervure corresponding to the upper sector actually starts from 

 the middle of the triangle (or rather trapezium), and the lower starts from the lower 

 angle of the triangle 2 . In species with normal neuration two more sectors descend 

 from the lower basal cell, the innermost a little before the middle from or near the 

 normal cross nervure, and the outermost a cell beyond. These curve outwards below 

 the basal portion of the lower sector of the triangle, and then suddenly curve away to 

 the lower part of the hind margin, which they do not quite reach, uniting just above it 

 with a short, nearly perpendicular, nervure descending from the lower sector. These 

 may be called the subbasal sectors, and are not always distinctly present in the smaller 

 species. Towards the base the neuration becomes irregular. The membranule is usually 

 much larger on the hind wings than on the fore wings, and the adjacent space is 

 frequently tinged with yellow or brown to a greater or less extent. 



The lower basal cell, as I have said, is usually traversed by a cross nervure before 

 the middle. Sometimes there are two (Libellula &c.) and sometimes more. It 

 occasionally happens that a specimen of a species which has normally only one cross 

 nervure has two on one side or, more rarely, on both. 



Before dismissing the subject of neuration, I may say that, although this character 

 is subject to considerable variation, within both genera and species, yet the probable 

 amount of this can generally be estimated and allowed for. I have only met with a 

 single instance in which there appears to be a constant difference between the sexes of 

 a species in neuration, viz. Perithemis bella, sp. n. (postea, p. 324), in which the subtri- 

 angular space has three cells in the male, and only one in the female. 



f. Abdomen. 

 The shape of the abdomen sometimes varies so greatly in species which differ little 

 in any other salient character that it can only be used to a very limited extent in 



1 Mr. C. 0. Waterhouse suggests that the upper cells lying between the subtriangular and subbasal cells may 

 represent it. 



2 In such cases, cells lying beyond the triangle, but below the upper sector of the triangle, must be counted 

 as posttriangular cells. 



