68 COAT CHARACTERS IN GUINEA-PIGS AND RABBITS. 



and long-coat, for all long-haired albinos are pure in these characters, 

 and as the result of experiment it may be stated that he was likewise 

 pure as regards the third or dominant character, rough coat. He was 

 mated with nine different pure pigmented, short-haired, and smooth 

 females, producing 29 young, all pigmented, short-haired, and rough 

 (see PI. 2, figs. 3 and 4), but, presumably, all containing the three cor- 

 responding recessive characters albinism, long coat, and smooth coat. 

 He was likewise mated with three other pigmented short-haired, 

 smooth females, which differed from those previously mentioned only 

 in this, that they contained recessive albinism. They produced by 

 this mating 15 young, of which all, of course, were short-haired and 

 rough, but 9 were albinos and 6 pigmented. The latter should be 

 similar in nature to the 29 young of the matings previously described, 

 making in all, from matings of c? 2002, 35 young bearing the three 

 dominant characters, but supposed to contain the corresponding reces- 

 sives, i. e., to be hybrid in all three pairs of characters. A number of 

 these young which were tested by suitable matings proved to be of the 

 character expected. They have produced by different matings all the 

 eight possible visibly different combinations of characters, viz : 



(i.) Pigmented rough short, P. R. Sh. (compare PI. 2, fig. 4). 

 (2.) Pigmented rough long, P. R. L. (compare PI. 5, fig. 9). 

 (3-) Pigmented smooth short, P. Sni. Sh. (compare PI. 3, fig. 6). 

 (4.) Pigmented smooth long, P. Sm. L. (compare PI. 5, fig. 10). 

 (5.) Albino rough short, A. R. Sh. (compare PL 2, fig. 3). 

 (6.) Albino rough long, A. R. L. (compare PI. I, fig. i). 

 (7.) Albino smooth short, A. Sm. Sh. (compare PI. 4, fig. 7). 

 (8.) Albino smooth long, A. Sm. L. (compare PI. 4, fig. 8). 



This result shows clearly that there is no necessary correlation be- 

 tween either character of one pair and either character of another pair. 

 On the hypothesis that each pair of characters is, in its transmission, 

 wholly independent of the others, i. e., that no correlation whatever 

 exists among the pairs, we should expect triple hybrids, like the 35 

 young of d 1 2002 already described, to produce, when mated inter se, 

 young visibly of the eight different sorts just enumerated in the following 

 proportions : 27 P. R. Sh., 9 P. R. L., 9 P. Sm. Sh., 3 P. Sm. L., 

 9 A. R. Sh., 3 A. R. L., 3 A. Sm. Sh., i A. Sm. L. It will be 

 noticed that at least 64 young must be reared to an age at which all 

 three coat-characters can be determined before we may expect all these 

 classes to be represented among the offspring. As yet only S young 

 have been reared to such an age, so that the results are scarcely signifi- 

 cant quantitatively, yet they indicate that the classes which are expected 

 to be largest will really be so, for the eight young are distributed thus : 

 4 P. R. Sh., 3 P. R. L., i P. Sm. Sh. 



Certain of the triple-hybrid offspring of c? 2002 were mated in ways 



