THE STUDY OF HEREDITY 333 



aberration, 1 the same authors obtained Mendelian results. They 

 came to the conclusion that, in order to obtain Mendelian 

 segregation, variations occurring in a race occupying the same 

 geographical area must be crossed ; but that if characters in 

 geographically separated races are crossed, they blend. My 

 belief is that this happens simply because the variations in the 

 same locality are individual characters of recent origin, whilst 

 differences between two geographically separated races, which 

 are common to all the individuals of each race, are racial 

 characters and are comparatively ancient. 



Crosses between individuals belonging to different species 

 and even to different genera of fish, among the Salmonidae parti- 

 cularly, are common, and practically perfect blending of the 

 characters is almost invariable. 



The alternative transmission of personal or individual varia- 

 tions must be of enormous advantage in the process of evolution. 

 As even every cell is different from every other cell, the number of 

 variations round the mean of any character in the multicellular 

 organism must be incalculable. It is also obvious that most of 

 these variations must be useless and some actually injurious. 

 The rapid elimination of useless variations is of great importance, 

 and this rapidity is provided for by the alternative inheritance 

 of recent variations. Only 25 per cent, of the second generation 

 from the introduction of the variation can possess gametes which 

 all carry the character. Of the rest, 25 per cent, will not possess 

 the character at all and in 50 per cent, it will be present in only 

 half the gametes. If the variation be advantageous, it will thus 

 be more easily preserved ; if it be useless or injurious, it will be 

 more readily and rapidly eliminated. 



We have in certain constituents of the cell — the chromosomes 

 — and the mode in which they are alternatively distributed to the 

 gametes upon fertilisation, an exact parallel to the distribution 

 of the characters in Mendelian inheritance. I have elsewhere 

 suggested the probability of the intimate connection between 

 these phenomena. 2 



Sex is claimed as a Mendelian character, and with some 

 modifications I feel that this claim is justified. Leaving aside 

 the highly technical points in relation to chromosomes as deter- 



1 Entomologist's Record, xv. and xvi. ; Trans. Entomol. Soc. London, 1906, and 

 Proc. 1907. 



2 Hereditary Characters. 



