COLOR AND COAT CHARACTERS. 15 



between dorsal and ventral pigmentation; for the darker the dorsal 

 surface, the darker is the ventral surface, and in any one animal of 

 either species the dorsal surface is always darker than the ventral. 

 Two facts may here be emphasized: 



(1) The agouti of the wild rufescens has less power to exclude black 

 and brown from the hair than has the agouti of the tame. 



(2) Agouti, from whatever source derived, produces a more striking 

 effect on the belly than on the back, restricting black or brown more 

 completely in the former region. It is one and the same agouti factor, 

 but it causes a different reaction in these two parts of the coat. Breed- 

 ing many agoutis has shown that there are not two factors, one for 

 restriction of black on the ventral side and one for the restriction of 

 black on the back. If this were true, the two factors could be dissoci- 

 ated and transmitted independently, but this has never been accom- 

 plished. It may be objected that, even with this evidence, we can not 

 be sure that two or more factors do not exist in complete coupling. 

 The objection, in a measure, answers the argument, for if the coupling 

 of factors is complete, we can only deal with them as one unit character. 



The surmise that the wild race would be found homozygous in agouti 

 proved true (see table 1). The original wild male was father of 27 

 young, all agouti, like himself, while his sons and grandson sired 19 

 other agoutis. Had either one of the original parents, cf 1 or 9 3, been 

 heterozygous, it would have been possible to extract recessive non- 

 agouti individuals from the matings of their offspring inter se, for half 

 of their offspring would likewise have been heterozygous. The con- 

 viction that the wild race must be homozygous in agouti is furthermore 

 strengthened beyond a reasonable doubt by the matings of 4 wild males 

 with 10 different non-agouti guinea-pig females (table 2); 37 agouti 

 young were thus produced. If the wild parent in any of these matings 

 had been heterozygous, these matings must siu-ely have produced some 

 non-agouti offspring, but such was not the case. Our point is therefore 

 well established by these 83 offspring. A second, point, in a measure 

 dependent on the first, may be stated thus: the wild agouti character 

 dominates its absence, even though the absent condition is presented 

 by the tame female parent.^ In both particulars the results agree with 

 similar matings among guinea-pigs. Such was Nehring's experience, 

 also, with Cavia aperea; for, though he did not understand alternative 

 inheritance at the time of his experiments in 1893, he nevertheless gives 

 data which show conclusively that Cavia aperea, a different wild 

 Brazilian species, is likewise homozygous in agouti and dominant to 

 its absence in a mating with Cavia porcellus. It is rather surprising 

 that no one has studied Nehring's data and referred to them or pre- 



'Castle (1906) has reported on the dominance of the wild agouti when mated to non-agouti. 

 The wild stock at that time was supposed to be the common Brazilian Cavia aperea. The fact 

 of dominance reported was correct; the error of classification is corrected in this paper. 



