390 ♦ • JOHN H. GEROULD 



broods, shows at once that their parents all were heterozygous 

 for the new character. Moreover, since these heterozygous 

 grass-green-blooded parents were brothers and sisters (family 

 1920, i) with the exception of one male (s^") which was a 

 cousin, we must look to the previous generation of adult butter- 

 flies (1920, (j)) for the individuals in whose germ cells the 

 mutation may have occurred. 



The grass-green caterpillars that became this previous genera- 

 tion were the hibernating offspring of a white-winged female of 

 the preceding autumn. Of the seven individuals of this family 

 that were used in breeding, at least two (0^ 9, white-winged, 

 and either <j)^ 9 , yellow- winged, or her mate </>^cr) were heterozy- 

 gotes for blood color, but, mating in each case with normal 

 homozygous dominants for blood color, no recessives appeared 

 in their immediate offspring (1920, t, s). It seems probable, 

 furthermore, that other heterozygotes for the blue mutation 

 may have existed in this brood, 0, making up probably 50 

 per cent, but the offspring of only one pair (0^ 9 X (A^ = 

 were thoroughly tested (giving broods /S, a, 6, v). Two other 

 matings (0" 9 X ^-cf = X, and </)^ 9 X </>^ cf =0) were made, 

 one of which, X, was carried no further, while two couples from 

 the other produced only grass-green caterpillars. 



Since more than one and possibly 50 per cent of brood were 

 heterozygous for blue, it is probable that the mutation should 

 be traced back to the wild female of the autumn of 1919 or 

 to her mate, as the original heterozygous mutant, else we must 

 suppose that the similar mutation occurring by hypothesis in 

 two of her offspring originated quite independently of each other 

 and of the parental genotype, which would be quite contrary to 

 general principles in heredity. The case illustrates well how a 

 recessive mutation may be transmitted by heterozygotes, out- 

 crossing with normal dominants for many generations, without 

 coming to light. There is reason to suppose that this is often 

 the case under natural conditions and that close inbreeding of 

 the wild stock for a few generations is all that would be necessary 

 to demonstrate the existence of these latent recessive factors. 



