BY R. J. TILL YARD. 745 



base of the wing, whereas, in the latter, it is always more or less 

 removed from the base. The change has clearly been brought 

 about by a gradual shifting of the point of origin of Rs distad 

 along R. Evidence of this is furnished by Glenoleon and allies, 

 in which Rs still arises, in the hindwing, so close to the wing- 

 base that only one crossvein is interposed in the radial space 

 before the origin of Rs. N. Banks has, indeed, divided the 

 Myrmeleonidce. into two subfamilies on this very character; the 

 MyrmeleonincH having several crossveins before Rs in the radial 

 space in the hindwing, the Dendroleonince only one (rarely two). 

 On phylogenetic grounds, this subdivision is now seen to be 

 clearly justified. Moreover, as might be expected from their 

 zoogeographical distribution, we can say definitely that the 

 Dendroleonince is the more archaic subfamily of the two. 



We may conclude, then, that the Myrmeleonidct are a special- 

 ised and highly successful offshoot from Nymphid-like ancestors, 

 the course of evolution being marked by the following changes : 



(1). In the venation, attachment of Cuj, in the forewing, to 

 M2, to form a single strong vein, the cuhifomedian; progressive 

 movement distad of the origin of Rs; gradual reduction in length 

 and distinctness of Cu^,; also gradual reduction in the general 

 density of venation and in the size and prominence of the ptero- 

 stigma. 



(2). In the antennce, a gradual reduction from a moderately 

 thickened and fairly long filiform antenna of numerous joints, as 

 in Nymphid(B (itself probably a specialisation from the longer 

 and slender Osmylid-form), to a shorter and stouter form with a 

 gradual tendency to the formation of a club at the tip. 



(3). In the larval life-history, a change from a wandering (pro- 

 bably nocturnal), carnivorous larva, .with an oval abdomen and 

 omnivorous tastes, to a sedentary, pit-dwelling, ant-feeding form 

 with a more rounded and specialised abdomen. 



True Myrmeleonidce being unknown from even the latest of the 

 Mesozoic strata, we shall probably be right in regarding this 

 family as being of early Tertiary origin. 



10 



