6 REVERSION IN GUINEA-PIGS AND ITS EXPLANATION. 



to be conclusive, but suffices to show that the animals in question were not in 

 all respects homozygous, the condition it was desired to obtain in the experi- 

 ment. These animals were therefore promptly discarded as soon as evidence 

 was obtained that they produced other than agouti young. In the case of 

 animals 191, 6197, 7894, 8020, and 9939, the numbers of young are sufficiently 

 large to establish beyond question their position in Group III. 



The animals included in table 4 together have produced by red mates 51 

 black and 91 agouti young. If all the red mates used lacked the agouti factor, 

 we should expect equality of black and agouti young; but from this expecta- 

 tion we observe a considerable divergence. It is quite possible that in some 

 of these tests red animals were used which were descended from cf 2054 and 

 had inherited from him an unseen agouti factor, for in making the later tests 

 some such animals were employed, but it is doubtful whether this accounts 

 for the whole discrepancy, amounting to about 14 per cent. This doubt is 

 strengthened by the similar discrepancy observed in the tests with black mates, 

 where we expect equality of black and agouti, but obtain 5 and 12 respectively. 

 The case of animal 191 looks suspicious; he produced only one black young one 

 in a total of 22, yet the record is certainly accurate. We had provisionally pro- 

 nounced him a pure agouti, when the single black young one came which places 

 him in Group III. It seems doubtful whether the theoretical 50 per cent of 

 his gametes were entirely free from the agouti character. If so, such gametes 

 would seem to have been deficient in vitality. The deviation is, however, not 

 an impossible chance result, though it seems improbable. I regret that this 

 animal was not more extensively tested. The matings of animals of Group 

 III with other agouti animals produced 9 black and 46 agouti young, or 84 per 

 cent agouti, where we expect between 75 and 100 per cent. 



Agouti animals of Group IV are included in table 5. They represent the 

 "fully fixed" agouti type breeding exactly like a pure wild species as regards 

 color. Two thoroughly tested animals belong here beyond question, cri45 

 and cf 316; two others less fully tested belong here in all probability, viz, cf 160 

 and cri81, while 9 473 is included as a possibly pure animal. She was fully 

 tested qualitatively (by a mating with a brown-eyed yellow animal known to 

 lack both black and agouti), but the number of young obtained is not large 

 enough to make the test conclusive. Similar tests of other females might be 

 included in the table, but they would add nothing to the demonstrative case 

 of 0^145, which shows that the reversionary character obtained by crossing 

 can be obtained in a homozygous, pure condition, as it occurs in wild species. 

 The five animals enumerated in table 5 have produced only agouti young : (o) 

 by red mates, 52; (6) by agouti mates, 135; (c) by black mates, 26. 



The occurrence of four different types of F2 agoutis indicates that the Fi 

 agoutis, their parents, were heterozygous as regards two independent mendelian 

 unit-characters, in agreement with the interpretation already given. These 

 two characters are: (1) black-pigmented fur; (2) the agouti pattern (visible 

 only in black-pigmented animals). Designating these two units as B and A 

 respectively, the Fi agoutis are all of the formula AB (heterozygous in both 



