328 SCIENCE PROGRESS 



eliminated. Thus we may find some cases in which the char- 

 acters of organisms have not changed materially during 

 geological epochs of time. Any considerable variation would 

 have been disadvantageous and so must have been eliminated. 

 Such cases are, however, as would be expected, comparatively 

 rare and occur chiefly among stationary or slowly moving 

 organisms. For the origin of this property of varying we 

 must therefore look back to the origin of life itself, and it 

 seems a work of supererogation to invent theories as to the 

 causes of variations during the later stages of evolution and 

 to treat them as though they had not been there all along. 



But there is one point about the variability of living 

 organisms which I do not think has received much attention, 

 and that is that it must obviously be the object of selection 

 just as much as any other character. Selection must increase 

 the variability among the individuals of a race just as it must 

 affect the length of a tail or the shape of a head. I shall have 

 more to say of this later. 



Prof. Thomson gives a number of theories as to the 

 causes of variation during the advanced stages of evolution, but 

 he assumes that in many cases variability does not already 

 exist. In explanation of this he says : " The cell which in the 

 embryo begins the germ-cell lineage may be identical with the 

 fertilised ovum, and the complete heritage may be continued 

 intact through successive cell divisions until the next genera- 

 tion is started and the process begins anew. The completeness 

 of hereditary resemblances depends, in Bateson's phrase, on 

 1 that qualitative symmetry characteristic of all non-differen- 

 tiating cell divisions.' " To me this appears to be a most 

 unwarrantable assumption. I have examined hundreds of thou- 

 sands of germ-cells destined to produce ova or sperms and I 

 have never seen two exactly alike even from the same indi- 

 vidual ; no one among the hundreds who have made similar 

 observations has ever done so either. Profs. Thomson and 

 Bateson must realise this themselves after due consideration. 

 Furthermore, the fertilised ovum cannot possibly be identical 

 with each of the germ cells which goes to form it. " That 

 qualitative symmetry characteristic of all non-differentiating 

 cell divisions " means no more in relation to Prof. Thomson's 

 "completeness of hereditary resemblance {i.e. the absence of 

 variation)" than that cells tend to produce cells more like 



