1899.] GROTE — SPECIALIZATIONS OF LKPIDOPTEROUS WING. 39 



be summed up as follows: The first radial branch may intersect 

 with subcosta {Idaides and Ancea) ; the cubital cross-vein or its 

 traces occurs in both series (in the Nymphalids in Li??inas, Heli- 

 ^o?iii(s, Moi'pho), while traces of the humeral cell of secondaries of 

 the Papilionides are found in the Limnadidai and Helicon iidai. 

 It is remarkable that just in these two latter families the resem- 

 blance is contradicted by the fact that both display the forked vein 

 viii, in contradistinction to the downwardly curved vein " ix " of 

 the Papilionides. It is this fact that allows me to consider the 

 characters of resemblance as residual and common to the lepidop- 

 terous wing, and as undecisive of questions of a nearer relationship. 

 These characters, the traces of the cubital cross-vein, etc., allow of 

 the view that the Limnadidse and Heliconiidce retain primitive 

 features and may more nearly represent early stages of the Nympha- 

 lids proper. Genera like Dione may supply a connection between 

 the Heliconians and Argynnince and render the view probable that 

 the brush-footed butterflies are monophyletic. 



Variability in the Veining. 



I have shown {Trans. Ent. Soc. Loiid. 1897, 342) that the varia- 

 tions in the neuration within the limits of a species take the same 

 direction with those used to define distinct species or genera. In 

 other words, they follow the chief directions which underlie most 

 changes in the neuration. These latter lie in the breaking up of 

 the median series and the suppression of the radial branches. For 

 instance, in the five-branched forms of the Anthocharini there will 

 be found a tendency to discard one vein in the direction of the 

 four-branched forms of Tetracharis. In a variable species standing 

 between Euchloe and Tetracharis the disappearing vein may be dis- 

 carded by anticipation (specialization) or retained by reversion 

 (generalization). In Pontia we have the three-branched form 

 definitely assumed, and we may consider that the ancestry of P. 

 daplidice has passed through the Etichloe and Tetracharis stages. 

 In the same way the passage of Mancipium to a three-branched form 

 is in process of accomplishment ; sufficient material has not yet 

 been examined of brassicce to determine whether the short apical 

 veinlet is discarded or not in the majority of individuals. In the 

 Zerynthiance we find individual variation in the direction of the 

 breaking up of the median series; sometimes vein ivj ascends the 

 radius ; again, in generalized individuals, it retains its older position 



