254 MR. W. P. KIRBT— A REVISION OF 



that the base rises in a right angle, with more or less equal sides, to meet it, thus 

 converting the triangle into a trapezium. 



In less aberrant genera the triangle generally forms nearly an isosceles triangle, 

 with its apex pointing downwards, but \aries considerably in size and shape. It is 

 either empty or crossed by a single nervure (sometimes by more), and in rare cases is 

 reticulated. 



From the apex of the triangle (or occasionally above it, as in Nannodijtheinis &c.) 

 another nervure curves to the hind margin. This is the upper sector of the triangle. 

 Below this is the lower sector, which is nearly always very irregular, and sometimes 

 very short. Beyond the triangle, between the lower sector of the arculus and the 

 upper sector of the triangle, is a series of cells, the posttriangular cells, which vary 

 from one to twelve at the base, and often increase considerably in number towards the 

 hind margin. Three may be considered the normal number towards the base, and the 

 middle cells are usually more or less hexagonal. It is not uncommon for the first row 

 of cells beyond the triangle to consist of one cell more than those which immediately 

 follow it. 



From the apex of the triangle, or from a point just below it, a line runs obliquely 

 upwards towards the base, which is united with the nervure above by one of the 

 straight or oblique lines descending from tlie lower basal cell. This may conveniently 

 be called the subtriangular space ; it is sometimes open, but is more frequently divided 

 into a number of cells, most often three, but frequently two, four, five, and upwards. 

 At the base of the inner margin is an opaque space called the membranule, which is 

 usually very small on the fore wings. 



I have not attempted to use the various slender and incomplete intermediate sectors 

 in various parts of the wings for classification, and I believe that the space between the 

 base and the subtriangular space presents no characters of any importance. 



The neuration in Odonata is so complicated that the real difficulty is rather to select 

 the characters which are of real value than to discover characters which might be 

 employed. 



e. The Hind Wings. 



In the main the neuration of the hind wings is a repetition of that of the fore wings, 

 as far as the upper basal cell. The cross nervures are generally less numerous. The 

 last antenodal cross nervure is always continuous, except as a rare anomaly in Lepte- 

 trum and one or two allied genera. When we come to the triangle, however, we find 

 that its base is usually formed by the continuation of the arculus, and that its apex is 

 directed outwards, lying at the angle formed by the lower sector of the arculus. 

 Occasionally {Agriono])tera &c.) the base lies far beyond the arculus, rendering the 

 triangle very small. The triangle is far less frequently traversed by a nervure than 

 that of the fore wings; and when it is traversed, a supratriangular nervure is most 

 frequently likewise present. The lower basal cell, the extremity of which forms the 



