80 MR. W. BAXESO>r ON COLOUR-HEREDITY [May 26, 



On breeding specimens of each type inter se he found the 

 following behaviour : — 



Type 1 might give Types 



9 



,, -J )) 5) " 



5, 3 ,, ,, r 



J5 4 ,, ,, ,, 



,: ^ 7) ;; " 



5J O J) !> " 



j; ' ?j >; " 



In other words, each type is dominant to albino and the wild 

 type is dominant to all. The grey forms are dominant to the 

 black. The black-and-white of type 5 is recessive to type 6, 

 (black with white markings), but the self-coloured iDlack does not 

 contain type 5. These are some of the deductions from the table. 

 The peculiarities of types 3 and 5 are especially noteworthy and 

 caU for fresh experimental study. 



It appears that types 3 and 5 could be ultimately bred true. 

 As to 6 and 7 the evidence is not very clear ; but as I understand 

 the account, neither was completely freed from throwing the 

 other. The breeding in these types was the least successful and 

 extensive. Possibly they are illustrations of the Mittel-rassen of 

 de Yries. It is especially noteworthy that the grey-and-white 

 type 3 and the black-and-white type 5 do not give rise to self- 

 grey gametes or to self-black gametes, a fact found again in 

 mice. We see therefore that there are gametes for black-and- 

 white and for grey-and-white, each of which may behave as a 

 single character and dominate over albino. 



Similarly when pure black-and-white was crossed with the wild 

 species, all the coloured types might appear in F^ but no albinos 

 (10. pp. 555-6). Therefore, in this very important case, when 

 black-and-white of type 5 appeared in the postei-ity of such a 

 cross, they were all homozygotes and produced only their own type 

 (p. 555). This fact may furnish a useful basis for a new experi- 

 ment. In strict accordance with our expectation, Crampe found 

 that black-and- whites which gave albinos when bred inter se, 

 gave albinos if crossed with albinos ; but when they did not 

 produce them themselves, they did not when bred with albinos. 



The similar vai-iety, grey-and-white (type 3), is always homo- 

 zygous except when it appears as a dominant containing types 

 4 or 5. But if I rightly understand Crampe's use of the word 

 " constant" type 5 is not produced by crossing type 3 with the 

 wild form, unless it was already brought in as recessive to type 3. 

 On the other hand, type 3 can be produced (in F„) by crossing 

 type 5 with the wild form. It is most deshable that the pro- 

 perties of these two types (3 and 5) should be fully explored. 

 They give a chance of investigating the resolving powers of a 

 recessive that is not albino, and free from several complications 

 attending the use of the latter. 



"When we try to pictm-e what is taking place in the resolution 



