12 



Britain, and has become so within the memory of man. There 

 must, of course, be a cause, but whatever it may be, it is probably 

 as liable to affect the insect in other countries as in our own. 



Notwithstanding the general resemblance between the members 

 of what I have termed the cinxia-gvonp, I am far from certain that 

 they have not been evolved by two quite separate lines of descent, 

 the one in connection with the auriniid, the other with the athaliid 

 group. The close apparent connection on the upperside between 

 cj/nthia and roiiianori, though the former is pronouncedly auriniid 

 and the latter at least as pronouncedly cinxiid on the underside, is 

 a startling, but perhaps quite deceptive, indication in this direction. 

 If, however, this is the case, the whole (lidi/uia-gvou]) must certainly 

 belong to the romanovid, while the piKcbe-gvoup more probably 

 belongs to the cinxiid line of descent. The connection between 

 atliene (so absurdly placed by Staudinger as a form of sa.votilis) and 

 roiiianori has been already mentioned, and the altaica, polaris, and 

 yet more strongly the chitndensis forms of didijina exhibit a decided 

 tendency towards the white and red coloration of the same species. 

 Too much stress, however, must not be laid on this, since the same 

 tendency exists in pincbe, which has the most athaliid underside of 

 the group. That the cinxia division is descended from the atlialia- 

 group through a not remote ancestor of vaiia seems to me, in the 

 face of the Asiatic species so closely connected with the latter, to be 

 almost beyond dispute, and if strong corroboration is needed it is to 

 be found m the fact that practically throughout the range of the 

 fl?/(o/m-group the most recent species, <leione,siilta)iensis, bellona and 

 their varieties, exhibit strong cinxiid tendencies on the underside ; 

 unless it be held m the latter case that bellona is a cinxiid species 

 having an atavistic tendency towards the athalia-growp, which (con- 

 sidering the apparent affinities of the one Melitaeid species, jezabella, 

 which has been found south, and only south, of the pal*arctic zone, 

 in south-west China, not only with the athaliid and cinxiid groups, 

 but also with the ancestral auriniids), is at least an arguable posi- 

 tion. If we are to divide the cinxiid and romanovid lines of descent, 

 cinxia, perse, arduinna, uabbii, pIta-be, and ictlierea (if it is a separate 

 species, which I regard as very doubtful) are on the former line, 

 roiiianori, atliene, didijuia, aeraina, infernalis, nibina, the saxatilis 

 group (if they are specifically distinct from didyuia and from each 

 other), !/iientij, triria, collina and ai/ar are on the latter. 



The North American species somewhat complicate the matter ; 

 perse appears to be so indubitably cinxiid, whilst miniita seems'' not 

 to be far removed from didi/ma : (jahbii also, which is certainly on 

 the cinxiid line, shows a distinct connection with the atlialia-gvon"^, 

 while })aUa, another species showing an approach to that group, is 



* This is, however, the only nearctic species which even seems to belong to 

 the romanovid section ; and if the two lines of descent are really different, 

 the latter may probably be confined to the Old World, and minuta really be 

 cinxiid. 



