1903.] IN FAXOY MICE AXD RATS. 79 



i7iter se. This point was elaboi'ately tested. Crampe states that 

 albinos true-bred for some generations behaved difi'erently from 

 extracted albinos, the former being, as he says, merely "absorbed," 

 i. e. recessive, on crossing with colour ; while extracted albinos 

 gave, as I understand him, a mixture of ancestral forms when 

 they were crossed with other types. This part of his paper 

 (10. pp. 573-5) is difficult to follow ; and I cannot find any example 

 showing precisely the nature of the distinction he means to 

 emphasise so far as albinos are concerned. We must here await 

 fresh expei'iments. We readily see, however, that though in 

 respect of its albinism we may regard the albino as always the 

 same, it may obviously be retaining other characters derived from 

 various progenitoi's. Accordingly we find, as will appear, albinos 

 apparently of the same species manifesting different properties in 

 crossing. I suspect, however, that Crampe is here extending to 

 the albino a generalisation i-eally based on a mistake arising from 

 misconception respecting the phenomenon of dominance. [See 

 note added p. 97.] 



We may now, though the evidence is imperfect, consider the 

 significance of the appearance of these many new forms in F,. 

 This phenomenon is a most usual result of a ci-oss between 

 distinct varieties. It is the source of the majority of our new 

 garden varieties, and of many at all events of the co^oitr- varieties 

 of domestic animals. In general terms we can declare that the 

 result of the cross — the "asymmetrical fertilisation," to speak 

 strictly — is the pi'oduction of a diversity of gametes. Pending 

 histological research, we cannot tell the origin of the chaiacters 

 borne by these gametes ; but from many circumstances it seems 

 inevitable that they must be regarded as created in such a case 

 partly by resolution of the charactei- brought in by the dominant — 

 which we therefore call a compound chaiacter, and partly by the 

 imperfect segregation of that compound or of its components fi-oni 

 the recessive character (and its components if it be also resoluble). 

 In most cases the process of resolution is not complete for all the 

 gametes ; and some of the gametes ai'e bearers of the wholly or 

 partly uni-esolved character, just as all the colour -bearing gametes 

 were in Cuenot's simpler case. The Mendelian hypothesis leads 

 us to believe that the actual numbers of each type of gamete 

 will be on the average definite, and that the union of any two of 

 them will give rise to a zygote of definite character. 



The number of types of gametes and their several properties 

 can only be determined on a minute analysis of each member of 

 the series of zygotes by exhaustive breeding. No such evidence 

 is yet complete in any one case, but we see already in cei'tairi 

 cases that some of the F, are homo- and some hetero-zygous, and 

 we are beginning to suspect the ratios of the gametic forms in a 

 few simple cases. 



Returning to Crampe's evidence, though the ratios are 

 quite uncertain, we find that the several types had different 

 properties. 



