478 Proceedings of the Royal Physical Society. 



ancestral group, and not from the more highly specialised, 

 would ensue from these generalised forms having retained a 

 greater degree of this potential variation. It would therefore 

 appear that the rate of actual variation is inverse to the 

 degree of potential variation. 



To go further, it is evident that in order to get continued 

 variation in one direction, true breeding must be adhered to. 

 This, indeed, would appear to be the reason of the tendency 

 of domestic productions to continue varying so rapidly in 

 one direction, namely, the extent to which selection is 

 brought into play. At each step of this process, however, 

 where a variation became fixed, it would seem that some 

 potential variation in other directions is lost. It then only 

 continues to vary within the limits mapped out by the 

 potential variation that remains to it. 



If we now turn to study what would be the effects of 

 amphymyxis on this process, we note in the first place that 

 as the specialisation advanced along a certain line, and con- 

 tinually lost potential variation, all the individuals of the 

 stock would share in this change, though in different degrees. 

 The consequence of this would be that the possible com- 

 binations and permutations of variation that could be pro- 

 duced through amphymyxis would be reduced within 

 narrower and narrower limits at each step of the process. 

 In fact, the range of unlike possibilities of variation in the 

 two uniting cells would become more and more limited. 



If we again look at amphymyxis in another light, namely, 

 as to what is its apparent use in Nature, we notice that it is 

 inevitable in most forms sooner or later to prevent exter- 

 mination of the stock, and that the result is what may be 

 termed rejuvenescence, and that it is usually followed by 

 a great and rapid development. It seems to me, therefore, 

 that there is a return of potential variation brought about 

 by this union, or rather a check on the too rapid loss of 

 this potential variation. If this be the case, the more 

 specialised the stock becomes, the more limited would be 

 this gain of potential variation, and the greater would be 

 the rapidity of the actual variation. Also, within the limits 

 of the stock, the more nearly related the individuals con- 



