Neuroterus lenticularis. ii 



except the coxae and the base of the femora which are 

 brown. The antennae are fifteen jointed, the first two 

 joints being yellowish. 



Experimental breeding. My earliest attempts at 

 breeding on a large scale were made in 1875 with 

 Neuroterus lenticularis. These experiments are easy 

 enough with this species if only a sufficient number 

 of flies are available. As soon as they leave the gall 

 they begin to deposit their eggs in the oak buds. 



It had hitherto been taken for granted that this species 

 produced a gall resembling that from which it emerged ; 

 many points nevertheless remained obscure and puzzling. 

 It had been long known for example that the gall of 

 Neuroterus lenticularis did not appear until July, but as 

 the eggs were deposited by the flies in April, three 

 months must have passed away before any trace of gall 

 formation had become visible. It was assumed therefore 

 that the embryonic development of the larva demanded 

 this lengthened period ; and this was quite possible, 

 since other species seemed to require an mcubation of 

 even longer duration. For example Andricus curvator 

 emerges in June and lives for two or three weeks during 

 which it lays its eggs, but the galls do not appear 

 until early in the following year. This could only be 

 explained by supposing that the ^^;g remained dormant 

 during the winter, and did not develop until next spring, 

 as is known to be the case with the eggs of many butter- 

 flies. The three months' egg-rest of Neuroterus lenticu- 

 laris therefore was not unprecedented. More conclusive 

 evidence in support of this supposition was wanting ; 

 but even if it were correct it failed to explain another 



