TRANSFORMISM 257 



organisms (the 5-one), differing from the ^-category in that some 

 advantageous character has changed so that it becomes more 

 advantageous, will have arisen. Now, as a rule, several characters 

 will vary in the same ways. Also an actual novelty of character 

 may have originated in one, or a few variants in A and since 

 this novelty, or mutation, is an advantageous one it will be largely 

 increased in frequency in the B'% (or progeny of the ^'s) because 

 the variants in A were enabled to survive and reproduce often. 

 Clearly the results of these processes will be that a new category 

 of organisms to which the original definition does not apply will 

 have evolved. And we may reverse the whole argument, as stated 

 above : assuming, instead of advantageous variations, some that 

 are disabilities to the organisms that display them. Then it is 

 easy to see that transformism will occur, but now the variants in 

 question will fail to survive, will not reproduce, or will survive 

 in smaller proportions and so there will be fewer progeny. In the 

 long run such disabled variants will tend to be eliminated from the 

 race. 



The question arises now — are the variations that we observe 

 in a naturally occurring population inheritable .' Experience 

 has proved that, in general, they are fluctuations and are not 

 inherited. If we select from a stock of wild animals some indivi- 

 duals that are, say, larger than the mean — those at a in the graph 

 A' of Fig. 35 for instance, and if w^e now breed from them the 

 progeny that we obtain will tend to regress towards the mean 

 of^'. 



Thus the children of exceptionally tall (human) parents will 

 tend to be taller than the mean of the whole population, but they 

 will not be so tall as their parents. And even when inbreeding 

 is practised this regression towards the mean value of stature 

 in the population will occur. Clearly the ordinary variations 

 that we see in a race-population are not, in the long run, inheritable 

 ones. A variant taken at random may display a variation that is 

 inheritable, but experience has shown that such variants occur 

 only exceptionally. And the result is even more clear when we 

 experiment upon a " pure race," say the classical beans. If we 

 rear plants from the heaviest beans taken from a plant of such a 

 pure race the mean weight of the beans produced by these 

 daughter-plants will be just the same as the mean weight of the 

 beans produced by the parent-plants. 



