28 GENETIC STUDIES ON A CAVY SPECIES CROSS. 



gation of a separable belly factor, for the usual correlation between the 

 dorsal and ventral sides was observable. The zygotic class, Aa, occurs 

 as expected about as frequently as class AA'. It is represented by 

 9 356, 9 414, 9 481 and 9 2030, which were mated to recessive non- 

 agouti males and produced each two sorts of young, light-bellied tame 

 agouti and non-agouti, in equal numbers. The total number of young 

 which they produced was 10 agoutis and 10 non-agoutis, the most 

 probable expectation. These females have already been entered in 

 table 9, and it need only be added that their agouti young were in all 

 cases of the tame agouti type. 



Visible class A' a.— This class is distinguished from the zygotic class, 

 Aa, by its very dark ticked-bellied agouti, which was received from 

 the wild source. It should occur among the offspring of cross 2 as fre- 

 quently as the class AA' or the class Aa. The actual number produced 

 was 13, whereas the most probable expectation is 9. Like the class 

 Aa, it is heterozygous in agouti; and, when mated to non-agoutis, 

 produces equal numbers of agouti and non-agouti offspring; but the 

 agouti offspring are all of the dark, ticked-belly type. Of the 13 

 animals assigned to this class on the basis of visible characters, it was 

 possible to test 5 females and 1 male. Female 195, 9 421, 9 565, and 

 cf 41 2 were mated to non-agouti guinea-pigs, and produced 8 dark, ticked- 

 bellied agoutis and 14 non-agouti young. The most probable expec- 

 tation is a ratio of 11 : 11. The females have already been entered in 

 table 9, and it need only be added that the agouti produced were of 

 the modified wild agouti type. Female 357 and 9 484 were tested with 

 non-agouti males, but their young are so few that the test is inconclusive; 

 they produced 3 tick-bellied young and 1 albino, but no non-agoutis, 

 hence their supposed heterozygous character as regards agouti has not 

 been demonstrated. 



Visible class aa. — This non-agouti class consisted of 4 individuals, 

 whereas the most probable expectation is 9. With such relatively 

 small numbers as must content the breeder of mammals, it is sufficient 

 to know that the class occurs. The animals which made up this class 

 were sterile males or died prematurely. There is no reason to believe 

 that, if they had been tested, they would not have acted just as any 

 other non-agouti recessives. Having raised over 400 young from non- 

 agoutis bred to non-agoutis and observing no exception to the rule that 

 these recessives breed true, whether they are guinea-pigs or hybrids, 

 it is safe to assume that class aa is exactly what it appears to be, pure 

 non-agouti. 



"Pbesencb and Absence" Hypothesis Applied. 



In the foregoing discussion, wild agouti and tame agouti are con- 

 sidered allelomorphic to each other. This hypothesis seems unavoid- 

 able, for, if the two sorts of agouti behaved as units wholly independent 



