Meyrick. — Revision of the New Zealand Caradrinina. 89" 



character may be variable even within the limits of the same species ; 

 therefore we must not assume that if a character separates natural genera 

 in one instance it will also do the same in another. There is no scientific 

 reason why secondary sexual characters should not be used to define 

 wenera in those cases where they are found to indicate natural genera in 

 accordance with the above-mentioned principles ; tufts of hair (probably 

 scent-producing) in the male sex are, however, found in practice to be of 

 specific value only — at any rate, as a general rule. Hampson oddly and 

 inconsistently refuses to use any sexual characters for defining genera, 

 whilst invariably employing these same characters, even the specific tufts 

 of hairs, for forming sections of genera ; whereas these should in any case 

 be limited on exactly the same principles as genera, being of smaller value 

 but precisely the same nature. 



I will give one or two specific instances of the unsatisfactory nature 

 of Hampson's results, to illustrate my meaning. Hampson makes a new 

 genus Eriopygodes for two European species and the Hawaiian euclidias 

 Meyr. This could only be explained geographically by supposing that 

 at some former period a straggler of the genus from Europe had reached 

 the Hawaiian Islands, which is unlikely, but, of course, possible. But 

 euclidias is an insect of striking appearance, and two other Hawaiian 

 species, compsias Meyr. and niphadopa Meyr., are structurally and super- 

 ficially so close to it that it is impossible to doubt they are closely related. 

 These are placed about seventy pages off in the genus Hyssia, which con- 

 tains about fifteen North American, European, and New Zealand species. 

 and a separate origin from another straggler is required for them. The 

 difference stated is that Hyssia has the thorax clothed with scales mixed 

 with hair, the abdomen with dorsal crest on first segment ; Eriopygodes, 

 the thorax clothed with hair only, abdomen without crests. But euclidias 

 (of which I have a long series) certainly has a small abdominal crest, and 

 the difference in clothing of thorax is imperceptible. I conclude that 

 euclidias must be transferred to Hyssia. But the only distinction be- 

 tween Hyssia and the cosmopolitan genus Cirphis, with 140 species, is 

 that Hyssia has the thorax clothed chiefly with hair-like scales, and 

 Cirphis almost entirely with hair. This is a distinction without a 

 difference, and, in effect, I am quite unable to distinguish the species 

 assigned to these two genera by this or any other structural character, 

 though they are separated by two hundred pages, and placed in widely 

 remote branches of the phylogenetic tree. I am therefore obliged to unite 

 them, which makes the Hawaiian species a local group representative of 

 a cosmopolitan genus, and puts quite a different face on the matter. But 

 on examining Eriopyga, with 100 species (chiefly American, some European), 

 only stated to differ from the above by absence of abdominal crest, I find 

 that some at any rate (e.g., the European Turca L.) certainly possess a 

 small crest (no doubt the character is often difficult of observation, because 

 the base of abdomen is clothed with rough hairs, and the thoracic hairs tend 

 to conceal it also, but when present it is formed by scales of a different 

 character and different colour), and must be referred to Hyssia also. I 

 am not well supplied with the American species, but the genus at least 

 requires cleansing. And Bornlia, with forty species, mostly African and 

 Australian, is only stated to differ from Eriopyga exactly as Hyssia does 

 from Cirphis, a distinction found to be inappreciable, for the supposed 

 difference in form of wing (more oblique termen) cannot be seriously re- 

 garded as a generic character, and therefore this also needs reconsideration. 



