Gametogeyiesis of the Gall-Fly, etc. 89 



pieces of the leaf containing them are easily cut out and fixed. Petrunke- 

 witsch's solution was used, Flemming and others that were tried proving 

 unsatisfactory. 



For spermatogenesis and the development of the egg before deposition the 

 larvte and pupte were removed from the galls at the end of May and early in 

 June, opened and fixed immediately in Flemming (strong) and Petrunke- 

 witsch. For this work Flemming's solution proved the better, but both 

 fixatives gave good results. The chief stain used in all cases was 

 Heidenhain's iron-hrematoxylin, safranin being used as a control. 



Since both males and females arise from the parthenogenetic eggs of the 

 spring generation, it seemed desirable to determine whether both sexes 

 arise from eggs laid by one female, and whether there are any differences in 

 the maturation of the eggs corresponding to the difference in sex. With 

 these objects in view, in collecting and preserving the eggs, in 1909 I kept 

 those laid by individual females separate ; and I also made some experiments 

 in letting the flies lay on growing oak branches, and rearing the galls. 

 In April I put single females in muslin bags or " sleeves " on the branches 

 of an oak tree ; removed the sleeves and marked the branches when the flies 

 had died, and searched the leaves for galls at the beginning of June. 

 In 1908 these experiments were wholly unsuccessful, perhaps because the 

 tree chosen had not suitable buds for the flies to lay in, for in 1909 I found 

 that they laid much more readily in buds from some trees than from others. 

 The more advanced buds were preferred. In 1909 I sleeved out 40 flies on 

 several different trees, including some quite young oaks from 1 to 4 feet in 

 height. On June 1, I found galls on seven branches of two large oaks, but 

 none on any of the small ones. The numbers of galls in the sleeved 

 branches were respectively 6, 7, 7, 2, 4, 3, 3, and from these I reared 

 respectively 5 ? 's, 7 c? 's, 6 's, 2 ? 's, 4 ? 's, 0, 0. 



In the galls which did not hatch the larvse either died or were parasitised. 



I searched the trees carefully for galls on branches which had not been 

 sleeved, and found none, so it may be concluded with confidence that these? 

 galls were produced by the flies sleeved out. Since in each case only one 

 sex emerged from the galls produced by one fly, although the numbers are 

 small I think it is justifiable to conclude that every fly of the agamic 

 generation produces eggs of only one sex, i.e. gives origin to either males or 

 females only, not to both. Cases of this kind are known in the Hemiptera 

 (Aphididte) and in Rotifers, but not hitherto, as far as I am aware, in the 

 Hymenoptera. 



