( clxxviii ) 



But in the special nidificatory habits of Anthidium 9 I see 

 a very good reason why (in the larger species especially) 

 there should be a limit to the tnagnitude of the $ $. These 

 insects nidificate, not in holes excavated by themselves, which 

 might be of any size, but at the bottom of ready-made tubes 

 of more or less equal and fixed dimensions, — the empty shells 

 of certain particular snails, or the hollow stems of certain 

 particular plants (e. g. A. manicatum has been recorded as 

 utilising the hollow stems of Heracleimi for this purpose, and 

 lining their interior with down scraped from some woolly 

 plant). It would clearly be convenient for a $ with such 

 habits not to exceed a certain size : but for the ^ this would 

 matter less, since he need never enter such tubes again after 

 his first emergence from one as an imago. His possible size 

 is probably only limited by that of the particular stem in 

 which he has been reared ! If the species were naturally 

 a large one, a reduction in the average size of the 9 ? might 

 well arise by Variation and become fixed by- Selection. But 

 in the smaller species such reduction would probably be 

 needless, and would therefore not occur. And this is exactly 

 what seems to happen. (It may be noted, also, that in the 

 various spp. of Stelis, which seems to be a smallish parasitic 

 offshoot of the Anthidium stock, the $ $ are usually larger 

 than their ^ ^.) So I believe that here we have, not a 

 (^-character, but a 9-character to be explained, and that, as 

 usual, it is to be explained by the post-nuptial duties of 

 that sex. 



Pi'obably the case is not very different with the ? ? of 

 Methoca, etc. These insects generally occur running among 

 sand, broken ground, stones, roots of plants, etc., and dis- 

 appearing suddenly into any crack or crevice that they meet 

 with, as though in search of some hidden object. They are 

 believed to feed their young with larvae which they find 

 in their researches underground. Methoca, I am assured on 

 good evidence, attacks the larva of the Tiger-Beetle Cicindela. 

 Such habits would be favoured by the absence of wings and 

 somewhat diminutive size, as they are in the Worker Ants, 

 among which the Mutillidae were formerly reckoned. But 

 to the ^ (J, who visit flowers and need never traverse narrow 



