R. C. PUNNETT 309 



Castle's Tan Dutch for example were almost all of grades 3-4, and 

 their variation did not extend beyond grades 2 and 5. Again the 

 members of his White Dutch strain all fell into grades 15-17, and 

 in the course of ray experiments I have also established a strain in 

 which the variation is about the same. I think we may fairly regard 

 these as homozygous strains. But when we turn to the Dark Dutch, 

 varying from grades 1-7, I feel that the position is different, ^nd I am 

 not prepared to classify together under the same head animals which 

 are almost self-coloured and others which might pass for the typical 

 Dutch of the fancy. It is quite possible, as will appear later, that 

 among Castle's Dark Dutch several strains could be got to breed 

 approximately true to diflferent grades of pigmentation. My own data 

 point to the fact that it is possible to obtain a strain of Dutch breeding 

 true to about grades 7 and 8. If this is substantiated by further 

 experiment it introduces a new factor into the case. Further, I am 

 inclined to believe that it is possible to obtain a strain which breeds 

 true to about grades 11-13. Here again my experiments are not yet 

 complete, for larger numbers are required before this point can be 

 regarded as established. For the present I look upon the following as 

 weak points in Castle's interpretation, viz. (a) the inclusion of grades 

 1-7 in. a single category, (b) the non-recognition of a true breeding 

 strain of typical Dutch (grades 7-8)S and (c) the non-recognition of 

 a true breeding strain of Spotted Dutch varying about grades 11-13. 



The interpretation that I wish to put forward provisionally is as 

 follows. We may suppose that there are several pattern forms with 

 reduced pigmentation which breed approximately true, viz. 



(a) White Dutch, comprising animals of grades 15-17. 



(6) Spotted Dutch, varying about grades 11-13. ( 



(c) Typical Dutch, varying about grades 7-8. 



1 Castle regards these grades as generally, if not always, heterozygotes between Dark 

 and White Dutch (op. cit. p. 7), and this is doubtless very often true. Nevertheless I think 

 it equally beyond a doubt that such animals may be homozygous. Indeed Castle probably 

 used such an animal in his experiments. Of the original two Dutch bucks he started with 

 one ( (? 3036) was from his breeding behaviour obviously heterozygous. But the evidence 

 given by Castle strongly suggests that the other ( <? 3037) was not carrying White Dutch. 

 There is no record of his being mated directly to White Dutch, but five of his Fi daughters 

 from self-coloured mothers were mated with <? 3036 (op. cit. Tables 14 and 16) and produced 

 50 offspring, all of grades 1-9. Had <f 3037 been heterozygous it is most likely that some 

 of his tested daughters would have carried White Dutch, and would, when mated with the 

 heterozygous <? 3036, have produced some offspring with a grade of pigmentation 

 approximating to White Dutch. 



