A SOMAIIC MUTATION IN DKOSOl'HII.A MKLANOGASTKM 101 



The question was now, whether in the mosaic hoth testes, one 

 testis or perhaps none of them were derived from the cell in which 

 a somatic sex-linked mutation was supposed to have taken place. 

 This could he ascertained hy outcrossing his daughters and controlling 

 whether all of them, half of them, or none of them were heterozygous 

 for the gene which in tiie mosaic caused the hristle and hair alteration. 

 Ten bl pr cii daughters from C. 2702 were accordingly outcrossed 

 singly to wild-type males. The result of these outcrosses is presented 

 in Tahle 1. 



Two additional mass cultures were at the same time made up. 

 In one of them (C. 2714) 14 bl pr cu females from C. 2702 were mated 

 to Sk cii hrothers. In the other (C. 2713) 16 Sk cu females from the 

 same culture were simultaneously crossed to w" v f males from stock. 



The result of these tests proved the correctness of the somatic 

 mutation hypothesis. The singed character did not manifest itself in 

 the daughters of the exceptional male, but reappeared unchanged in 

 his grandsons. The curling of the hairs and bristles in the mutant 

 part of the mosaic was accordingly produced by an ordinär sex-linked 

 recessive. 



Moreover, of the 10 daughters separately tested 7 had received a 

 paternal X carrying the mutant gene, w-hile 3, giving only non-singed 

 sons, must have received from the father an Ä'-chromosome without 

 this gene (Table 1). This indicates that only one of the testes of the 

 mosaic can be derived from the cell which contained the mutated Ä''- 

 chromosome. The other testis, — or at any rate part of it — , has 

 descended from a cell which carried the original unmntated gene for 

 wild-type bristles. 



This conclusion is also confirmed by the results obtained in the 

 mass cultures (Tables 2 and 3). Here we would on the latter assumption 

 expect about a quarter of the sons to be singed, while if both testes car- 

 ried the mutated A', we would get singed and non-singed sons in equal 

 numbers. The total output (ten days count) of the two cultures w^as 215 

 non-singed daughters, 144 non-singed and 64 singed sons. This is a fairly 

 good 3 : 1 ratio, when it is remembered that the numbers are small 

 and derived from mass cultures. It is true that the somewhat lowered 

 viability of the singed individuals, which is apparent from the data 

 presented in Table 1, may in a degree be responsible for the difference 

 in number between the non-singed and the singed classes. But the 

 effect of this differential viability is counterbalanced to a certain extent 

 by the marked falling back of the bl pr cu class in C. 2714. When 



