188 



Dr. L. Doncaster. Gametogenesis and 



of Series G-, sleeve 52, but it certainly occurred in other cases also. These 

 facts strongly suggested that the exceptions were due to the presence of eggs 

 of wild flies, which were remarkably numerous in 1914. It seemed probable 

 that, as the buds on the twigs of necessity pressed against the muslin in some 

 of the sleeves, wild flies might have been able to insert their ovipositors through 

 the meshes of the muslin and lay eggs in such buds. The sleeves used were 

 of muslin of two kinds, one considerably finer than the other. Unfortunately 

 no record was kept of which galls came out of sleeves of the two grades of 

 coarseness, but it seems probable, if the explanation suggested is correct, that 

 the sleeves containing no exceptions were of the finer mesh. I was prevented 

 in the summer of 1914 from repeating the whole experiment, as I wished to 

 do, and an attempt made in 1915 failed because the flies laid no eggs ; but, in 

 order to test whether flies can lay through the meshes of muslin, in the 

 spring of 1915 I sleeved a number of twigs, and then put a number of galls 

 in larger sleeves fastened outside the others. Some sleeves were also left on 

 the twigs with no outer sleeve, to see whether galls would be produced in 

 them by wild flies. As this was done before the end of February, two or 

 three weeks before the flies begin to emerge, there is no chance of eggs having 

 been laid in the buds before the sleeves were put on. The result of this 

 experiment was that out of 19 sleeves, each with galls in an outer sleeve 

 enclosing it, three contained galls of the sexual generation in May. In each 

 case the galls were on leaves derived from one bud only ; in one sleeve there 

 were 20, in the second 18, in the third 2. From the first two there were 

 reared 17 females in the one case, 15 females and 2 males in the other. 



Five twigs were covered with simple sleeves, to see whether galls might be 

 produced in them by wild flies, but none contained any. Although a much 

 smaller proportion of galls were produced inside the sleeves than in the pre- 

 ceding year, this experiment proves conclusively that it is possible for the flies 

 to oviposit through the meshes of the sleeves used, and the smaller proportion 

 is probably due to the fact that the season of 1915 was much less favourable 

 to the flies than that of 1914. Wild galls were extremely scarce at 

 Cambridge as contrasted with their great abundance in 1914, so that if it had 

 not been for the kindness of my friend Mr. E. E. Unwin, who sent me a 

 supply of galls from Eeading, I might have not been able to obtain enough to 

 supply sufficient material to complete the cytological work described below. 

 It is probable that the frost and snow of the last days of March, following a 

 mild early spring, killed many of the flies before they had laid. I think it 

 may, therefore, be concluded with certainty that the exceptions recorded in 

 about half the sleeves in the 1914 experiments were due to eggs laid by wild 

 flies. 



