1903.] IN FANCY MICE AND EATS. 95 



greater than the amount of the same pigment in the oiiginal 

 compound colour. But this consideration cannot be allowed much 

 weight, seeing that there may be an excess of pigment in hetero- 

 zygotes pi'oduced even from two gametes apparently bearing no 

 pigment elements at all {cf. p. 76). In the chemistry of pigment- 

 ation there may perhaps be interactions and cancellings so 

 complex as to make this particular problem as yet quite insoluble. 



Fuller analysis is especially needed also to determine the place 

 of the pied and diluted colour-bearing gametes in the series, but 

 it is fairly certain that they must be regarded as due to disinte- 

 gration and imperfection of resolution of the colour from the albino 

 character. 



Future experiment must decide the conditions determining- 

 resolution. Cuenot, as I understand his paper, got none in the 

 main expei'iment with wild mice ; but he states that he obtained 

 yellows, blacks, and pieds '■^ accessoirement" (perhaps by intro- 

 ducing some coloured fancy strain 'i). 



From this survey of evidence mostly ali'eady published, it is 

 clear that Mendelian analysis provides a means of elucidating a 

 large part of the phenomena. The majoi-ity of the obsei-vations 

 are in accord with the Mendelian hypothesis in a simple foi-m. 

 The true solution of several subordinate problems still remains 

 obscure. The value of the Mendelian analysis will be the more 

 appreciated when it is I'emembered that previously the whole body 

 of facts must have been regarded as a hopeless entanglement of 

 contradictions, as reference to any non-Mendelian discussion even 

 of these veiy phenomena will show. 



As I have elsewhere pointed ou.t, the central phenomenon in 

 Mendelian heredity is segregation. The characters in simplest 

 cases are treated as units in gameto- genesis. In more complex cases 

 there is resolution, sometimes also disintegration and imperfect 

 segregation, leading to the formation of fresh units. The gametes 

 bearing these units are produced in numerical propoi'tions which 

 on an average ai'e also definite, but as yet these proportions have 

 only been determined in the simple cases. There is no doubt that 

 fui-ther experiment will determine them in complex cases also. 



It is the object of Mendelian analysis to detei-mine 



(1) the constitution of the several types of gamete produced by 



each type of zygote ; 



(2) the numerical proportions in which each type of gamete is 



produced ; 



(3) the specific result of the union of any two of the types of 



gamete in fertilisation. 



Though for convenience we may still speak of inheritance as 

 being " Mendelian " or " non-Mendelian," we are rapidly passing- 

 out of the initial phase of the inquiry in which such expressions 

 are demanded. In our further investigations we are concerned 

 not so much with the question of the applicability of the simplest 



