NATURAL SELECTION 393 



direct effects of the environment, and are not inherited. To-day, we 

 must throw these out and consider heritable variations only. Now we 

 find that these heritable variations mainly (at least) represent no more 

 than a shuffling of the stock-in-trade of the organism, and if any of 

 them involve absolutely new determiners, we do not know it. The 

 matter is complicated by the frequent appearance of new characters, 

 which experimental evidence shows to result from new combinations 

 of the old ones. Thus the pink-flowered and yellow-flowered stocks 

 (Matthiola) give white and red-cream flowers in the third (second 

 filial) generation, no matter if the two original strains had been bred 

 true and had remained constant from the beginning of time. Here 

 we seem to see something entirely new, but analysis shows that we have 

 no more than new combinations of certain of the grandparental 

 characters. 



What will natural selection do with such materials ? It can do no 

 more than favor certain characters or combinations of characters and 

 eliminate others. It can not even eliminate the recessives. The result 

 will be the production of a number of distinct types, without necessarily 

 any forward evolution — anything more than a shuffling and sorting of 

 determiners. So far as we can see, this is exactly what has happened 

 in the case of the oak leaves and many mollusc shells. 



The modern school of Mendelian experimenters, who have from neces- 

 sity confined themselves to determining the inheritance of relatively 

 simple characters, have come to think little of natural selection. They 

 have seen how various combinations can arise, greatly altering the 

 appearance of animals and plants, without selection having anything 

 to do in the matter. They have also seen how certain of these modified 

 types, or others like them, may multiply and spread, without being 

 obviously helped or hindered by selective agencies. Where the charac- 

 ters came from, they do not know; but neither do the selectionists. 

 Let it suffice that we have here an apparently mechanical arrangement, 

 which if left to itself will people the earth with diverse animals and 

 plants, a large proportion of which will get along well enough to sur- 

 vive. Possibly this description is unjust in its application to modern 

 experimenters generally, but it at least represents the attitude of some 

 of the more influential and at the same time marks the recognition of a 

 number of real and important facts. I think that while we shall gladly 

 incorporate the facts into our system, we shall in time come to wonder at 

 the limited view of nature implied by the attitude described. It is all 

 too simple and too easy, it does not take into account the real com- 

 plexities of life or of organization. It reminds us a little of the school 

 of zoogeographers who would bridge the oceans whenever it seems nec- 

 essary for some animal to cross, or to have crossed. The experimental 

 work itself is revealing this, as day by day new complications arise. 



VOL. lxxxii. — 27. 



