14 



(in the restricted sense), since their larva agree with those of the 

 Pyrameids in possessing a row of central spines, whilst the Bren- 

 thids and Argynnids agree with the other Vanessids in being with- 

 out it. Possibly this may not be a uiatcer of so great importance 

 as others, and I am not suggesting it as probable that the Melitieids 

 are older than the Brenthids and Argynnids, since the closed cell of 

 the hindwing in the latter is a more ancestral character, still I am 

 not at all certain that a comparatively recently lost nervule might 

 not re-appear, especially since a part of this nervule is present in 

 many of the MelitiBas, and I have come across a specimen of varia 

 with the cell completely closed. On the other hand it is possible 

 that the Pj^rameids may not be the oldest branch of the Vanessid 

 stirps. In any case the Brenthids are closer to the Melitceids than 

 are the Argynnids, and it is at least possible that they, or something 

 like them, may be the source of both, since there are some grounds, 

 such as geographical distribution, for doubting the theory that the 

 Argynnids are older than the Brenthids. It is a remarkable circum- 

 stance that the species of the MelittBids that most closely approach 

 their nearest neighbours, the Brenthids, on the upperside, all belong 

 to the most recent group, viz., ardiihma, tiiria, and hellnna : this 

 group is on the other hand the farthest removed from the Brenthids 

 ^n the underside, and this Brenthid upperside cannot be regarded 

 as indicating a closer relationship than exists between the latter 

 and the other groups, but must have developed quite independently. 

 If we take either the upper or underside of the three British 

 species we shall at once be struck by the difference between them, 

 yet an examination will show that the wing pattern is essen- 

 tially the same in all. Different again as the upperside is from the 

 under in each species, the same pattern is essentially reproduced on 

 the two sides, though there is generally considerable obsolescence 

 on tbe underside of the forewing, and the sufl'usion at the base of 

 the iiindwing on the upperside usually effaces some part of the 

 wing pattern, which is further concealed by the fact that the diver- 

 sified colouring of the underside hindwing is never entirely repro- 

 duced on the upperside and frequently not even indicated. The 

 pattern of the forewing is therefore best studied on the upper-, that 

 of the hindwing on the underside. That of the forewing consists 

 essentially of the following parts : — Black border. Two black lines 

 of varying width and conspicuousness, nearly parallel to the border; 

 these we will call the '^ outer and inner .siibterminal lines." Between 

 the border and the outer subterminal line, the ground colour shows 

 more or less in the form of lunules, the third of which, counting 

 upwards from the anal angle, often projects further towards the 

 disc of the wing than the others ; the direction and curve of the 

 inner subterminal line is a somewhat valuable character in deter- 

 mining the different species. Further towards the base is a sharply 

 elbowed almost sickle-shaped black line of very variable breadth, 

 curving sharply out from the costa towards the outer margin, then 

 inwards towards the base, and again somewhat outwards, spreading 



