78 MR. W. BATESON ON COLOUR-HEEEDITY [May 26, 



these qualitative observations are of great value and will provide 

 a basis for future work. 



In rats the Mendelian rules, in their simplest form, are plainly 

 inadequate to express the facts, and we soon meet a, number of 

 deductions of specific application, each needing full investigation. 

 Crampe's account is long and difiicult to follow. At first sight 

 also it seems not wholly consistent in certain particulars, but the 

 conclusions here summarised seem well established. 



Breeding albino rats with wild M. decumanus, Crampe found 

 that F^ might have one of lioo forms, being either a self -grey like 

 the wild type, or gi'ey with white marks. Unfortunately no precise 

 description of this and of the other broken-coloured rats is given, 

 and we do not know the precise extent and distribution of the 

 white*. According as F^ presented the first or second form, the 

 subsequent offspring produced from F^ bred inter se, differed. 

 The whole series of colours presented by such offspring is arranged 

 by Crampe in seven types, thus : — 



1. Self-grey. 



2. Grey with white marks. 



3. White and grey. 



4. White (albmo). 



5. Black-and-white. 



6. Black with white marks. 



7. Black without marks. 



The self-coloured grey in F, gave in their posterity all the 

 types except 3 and 5, but Fj of type 2 gave all seven types. 



The nature and cause of the heterogeneity in F^ is as yet 

 unexplained. Such an occurrence is, however, not rare. In my 

 own poultry experiments for example, the dark feathers scattered 

 in the white F^ raised between a bi-own and a white breed may 

 be either chequered or plain black. According as one or tlie 

 other form appears in F^ the posterity probably differ, though 

 this point is not yet established in the case of poultry. 



The existence of two classes in F^ indicates in all probability 

 the existence of two classes of gametes, either in the wild 

 decimianus or in the albinos, but in which we cannot say. From 

 the evidence, it seems that both forms of F^ could be produced 

 by the same paii- of parents, but I cannot find the fact explicitly 

 stated. Both foi-ms occurred in F^ not only when decumanviS 

 was crossed with albino, but also when it was crossed with type 3 

 and with type 5. 



Only the albino could cause all seven types to appear in progeny 

 (F.^ &c.) raised from a cross with the wild type. 



The albino was recessive to all the other six types, and albinos 

 of whatever parentage gave nothing but albinos when bred 



* First crosses shown me by Miss Douglas were grey except for an irregular but 

 small amount of white on the chest and belly. I take this to be Crampe's type 2. 



