26 Rev. F. D. Morice on male terminal segments and 



been, that Colletes $ armatures have usually characters of 

 their own, by which the species can be certainly, and some- 

 times easily, recognized. But these characters do not 

 depend simply on the outlines assumed by certain portions 

 of the structure in certain positions. Those outlines are 

 often only the somewhat illusory and variable horizo7is of 

 solid objects viewed accidentally in this or that aspect, they 

 depend partly on the focussing adopted, partly on the 

 position of the armature as a whole, and partly on that of 

 each different jDart of it in relation to the others. This 

 last may depend on a variety of accidents, since the parts 

 are not all equally rigid, so that one or more of them may 

 easily be somewhat distorted or displaced by causes which 

 leave the rest of the structure unaffected. I do not think 

 that Radoszkowski sufficiently realized this fact, which in 

 itself seems to render his diagrammatic way of treating the 

 subject inapplicable. Besides this — or perhaps because of it 

 — he seems too ready to content himself with noting certain 

 arbitrarily selected details, which are often not sufficiently 

 distinctive to bear the weight he lays upon them. 

 Other parts of the structures he dismisses — one might 

 almost say, impatiently, giving the impression that he had 

 studied them not at all or only in certain species. Thus 

 he says that the volsellx " ne presentent aucune particular- 

 ite," which is quite curiously untrue in this genus, though 

 no doubt their position renders it difficult to make much 

 use of the characters exhibited in them. And in speaking 

 of the couverch genital, as he calls it, he defines it in the 

 introductory part of his paper as composed of " two pieces, 

 the sixth" (!) "and eighth segments"; but he figures it 

 both there and afterwards (in the only two cases where he 

 figures it at all) as consisting not of these two segments, 

 nor of either of them, but simply of the seventh ventral- 

 plate ! This is the more curious, because in his earlier 

 paper on the Mutillides (1885) he says quite correctly, that 

 the courerelc genitcd has been shown by Mr. E. Saunders to 

 consist of the seventh and eighth ventral plates ; and he 

 figures accordingl}^, as such, those two plates (clinging 

 together as they often do) both there, and also in a paper 

 published a year later (Hor. Soc. Ent. Ross., xx, 1886). It 

 would seem, then, that between 1886 and 1891 — the date 

 of the " Revision " — he must have somehow lost interest 

 in the couverch, or he would hardly have blundered about 

 its composition as stated above. He adds one further 



