220 Mr Doncaster, On the Inheritance of Goat-colour in Rats. 



ancestry of the "Irish" used and my experience is in accord with this, 

 although in Exp. 63 there is only 1 black ancestor three generations 

 back. Crampe does not separate the two forms of " Irish," but in 

 all his matings black-and-white appeared, so that when blacks 

 were produced the parents may have been heterozygotes black 

 black-and-white. But in my mating 63 blacks appeared from two 

 parents of type 6 a, so that this explanation is impossible, and we 

 therefore find that blacks bred together can throw Irish, and Irish 

 can give black. There is further no sharp line of distinction 

 between black and " Irish a," for the latter may have only about 

 half-a-dozen white hairs on the chest, or may have a considerable 

 white patch. 



It is practically certain, therefore, that black is only an extreme 

 form of " Irish a," and not really a separate variety ; either may 

 be bred true by continued selection, but they are not allelomorphic 

 to one another. Both when paired with black-and-white give the 

 heterozygous form " Irish b." That this is the true explanation is 

 indicated by Exp. 55, where the $ parent was "Irish 6" made 

 by crossing black-and-white by black, the " Irish a " derived from 

 2 "Irish" parents. If black were dominant over "Irish a" the 

 offspring should have consisted of black and "Irish b" in equal 

 numbers, but actually 4 " Irish 6," 2 " Irish a " and 2 black were 

 produced. The conclusion must be that as far as pattern is 

 concerned there are only two allelomorphs in rats, viz. " pied " 

 and "self," the pied consisting of coloured hood and stripe, the 

 self either completely coloured or having a small white mark on 

 the ventral surface. When crossed together there is no complete 

 dominance, and a heterozygous form is produced having the whole 

 ventral surface white. ^T 



Of the other experiments not much remains to be said. I.jyfpst 

 of them were begun in a rather indiscriminate fashion before I 

 realized that albinos were not all alike in constitution, and many 

 of the albinos used were dead before they could be tested. Pair- 

 ings were made between " half-wilds " (children of albino by wild) 

 and albinos in Exps. 14, 21 a, 44, 24, and 26, giving a total of 

 87 young, 44 of which were albino, a close approximation to the 

 Mendelian equality. If the albinos used all bore the black and 

 piebald determinant, the expectation is that the coloured off- 

 spring should consist of equal numbers of Crampe's types 2, 3, 5, 6 ; 

 the actual numbers are 8, 9, 10, 16 respectively. Of the five 

 albinos used in these experiments, 3 were bought, and in my 

 experience albinos obtained from dealers generally carry the black 

 and piebald characters. Unfortunately, however, most of them 

 died before they could be tested in this respect. 



In only one experiment (No. 43) did I succeed in getting 

 young from " half- wilds " paired together, 7 out of 25 being 



