COLOR AND COAT CHARACTERS. 29 



in heredity, then the hybrids containing both sorts of agouti, each in a 

 single dose, should produce some non-agouti offspring when mated to 

 non-agoutis, but they do not. These hybrids (whether received from 

 Cross 1 or from Cross 2) would on that assumption have a formula 

 AA'aa' in place of AA' and should produce four kinds of gametes in 

 equal numbers, viz, AA', Aa', A'a, and aa'. One-fourth of the gametes 

 in that case should fail to transmit agouti; but the experimental evi- 

 dence given in tables 4 and 15 shows clearly that such is not the case, 

 and that, therefore, the hybrids produce only two kinds of gametes, 

 one of which carries tame agouti, while the other carries wild agouti. 

 Hence, the hypothesis that wild and tame agouti are not allelomorphic 

 is untenable, at least in the simple form stated. 



Nevertheless, it may be assumed that the dominant agouti of the 

 tame guinea-pig really contains the same agouti as C. rufescens, but 

 has an additional differential factor, D, firmly coupled with it. The 

 tame agouti in that case might be designated by the inseparable com- 

 bination AT), which is the equivalent of A in the foregoing pages. 

 The wild agouti would then be designated by A'd. The two would be 

 allelomorphic to each other and each to its absence, a/A. This expla- 

 nation does no violence to the observed facts, but is untenable without 

 the supplementary hypothesis of coupling. This line of explanation 

 does not simplify the statement that wild agouti and tame agouti 

 behave as allelomorphs to each other, although it allows one to account 

 for the origin of "wild" agouti from "tame" by a break in the supposed 

 coupling. It also has advantages from the standpoint of those who 

 consider unit characters unchangeable, except by the addition or sub- 

 traction of distinct factors likewise unchangeable. This method of 

 procedure, however, encounters difficulty in explaining how wild, light- 

 bellied, light agouti of the early hybrids might give dark, ticked-bellied 

 agouti and then these latter give the former again. 



Fewer suppositions make the first alternative explanation simpler, 

 for any sort of agouti is allelomorphic to any other sort of agouti in 

 guinea-pigs. The segregation of the modified dark, ticked-bellied 

 agouti from the light agouti is more apparent and striking on account 

 of the differences. That they do segregate may be due to their being 

 carried in homologous chromosomes. 



Morgan (1911) and Cufoot (1911) have described light-bellied agouti 

 mice which are dominant to the gray-bellied variety. The two forms 

 segregate. Hurst (1905), on the other hand, mated a very yellow 

 agouti rabbit to an albino, and got the darker wild gray type. He 

 reports that no segregation occurred. 



My experience with the transmission of the wild agouti factor to 

 the hybrids and its inheritance is as follows : 



(1) The wild agouti, when transmitted to the hybrids, may give a 

 very dark ticked-bellied coat. This modification may persist, become 



