54 [February, 



Personally, I shall not shrink from saying that, hypothesis or no, 

 I am very much inclined to believe that it is true. The evidence for 

 it is, no doubt, wholly circumstantial and inferential, but taken all 

 together it appears to me very strong. 



First, of course, there is the obvious argument from structure 

 and general appearance. The " facies " of Sphecodes is, beyond all 

 doubt, suggestive at least of parasitism. 



Then there is an argument from analogy, vt'hich might carry little 

 weight if it stood alone, but which, as corroborating other arguments, 

 is worth consideration. As a rule, an Aculeate genus known to be 

 parasitic appears from its structure to be a kind of offshoot from 

 some industrious genus, often that which it chiefly infests. Nomnda 

 is a disputed case, but Mr. Saunders has shown that its structure is 

 in some minute points accordant with that of Andrenci. Stelis and 

 Anthidium^ however, Ccelioxys and Megachile, Melectn and Antho- 

 pliora^ are all clear cases in which the parasite stands close to the 

 industrious genus in nearly all points of structure, except in the 

 pollinigerous apparatus. Supposing Sphecodes to be a parasitic off- 

 shoot of the Halictus stock, its structure is precisely that which the 

 above analogy would lead us to expect. 



Thirdly, there is the argument from the ordinary behaviour and 

 demeanour of the $ $ Sphecodes, which is unmistakeably that of 

 parasites. They visit but do not toil among the flowers ; they are 

 seldom seen even apparently at work on excavation ; they are never 

 found taking sole possession of an unoccupied place with a colony of 

 their own; but, almost wherever Halicti swarm, Sphecodes may be seen 

 dogging their movements, and sneaking about their burrows with the 

 same sort of vigilant and cautious activity which is so noticeable in 

 {e.g.) the Chrysids. They seem to be always reconnoitring an enemy's 

 position. They are "slim," mobile, inquisitive, restless, concentrating 

 their attention upon no visible operations of their own, while indus- 

 trious bees seem wholly absorbed in their own labours, and heedless 

 of the occupations of others. 



Fourthly, we have yet another argument, resting on the con- 

 current appearance — much too regular to be accidental — of particular 

 species of Sphecodes in and about the nesting places of particular 

 species of Halictus. It is true that the evidence as to these associa- 

 tions is somewhat perplexed at present, and, in fact, it seems likely 

 that few, if any, Sphecodes are wholly and solely attached to any one 

 particular host. But that as a rule Sphecodes of one species abound 

 in or are absent from particular localities, according as particular 



