242 HABITS AND INSTINCTS IN LfiPTINOTARSA. 



cies. It is just the kind of a process which would be best adapted for the 

 splitting up of a polymorphic species through the limitation of reproduction to 

 the more mediocre of the individuals in each group, and the tendency to 

 eliminate intermediates, and also perhaps to prevent the crossing of closely 

 related forms. Assortive mating is no doubt equivalent to some one of 

 Gulick's (1905) divisions of segregation, but I can not see the need of 

 the extremely finely divided divisions of segregation, and prefer to use for 

 the present the simplest terminology. Although assortive mating (or seg- 

 regation or isolation in reproduction) is undoubtedly a valuable aid in the for- 

 mation of species, it must nevertheless find characters already developed to the 

 extent of having a selective value before it can begin to exert any influence. 

 It can not bring into existence anything; it can only preserve that which is 

 already in existence. Moreover, in the genus Leptinotarsa at least, a great 

 many characters, as, for example, color and ornamentation, are not influenced 

 at all by this form of selection, so that if, perchance, these non-selective char- 

 acters should exist in combination with characters which have a selective 

 value, they would be carried along with the selective characters in the segre- 

 gations which assortive mating brings about. In this way non-utilitarian 

 characters might be preserved by selection through being associated with 

 others which are perhaps less conspicuous, but which have selective value. 

 However, in order to explain all nature and evolution through selection, one 

 must not assert that all apparently non-selective characters are correlated with 

 selective characters which have not as yet been discovered, and perhaps never 

 can be. Assortive mating has been invaluable in the production of new 

 characters in experiment and in the perpetuation of the new races which have 

 arisen in my experiments, and I can see no valid reason why it should not be 

 an equally important factor in nature, although this is not as capable of 

 demonstration or of experimental proof. It is possible, however, as I shall 

 show later, that there is the best of circumstantial evidence that this selective 

 process has been a powerful factor in the evolution of this genus of beetles ; 

 and if it is a potent factor here, why should it not also be one in other cases? 

 Although in these beetles, as far as we can determine, color and ornamenta- 

 tion seem to play no part in selective mating, there is no reason apparent why 

 these factors may not be of selective value in other animals. It is well to re- 

 member that we are viewing these beetles as outsiders, and are not, therefore, 

 able to appreciate the minute individual differences which they exhibit and 

 which may play important roles in the mating of the species. It is quite pos- 

 sible that there are in animals characters of a selective value which we are 

 unable to recognize, but which the animals themselves appreciate. Even in 

 our own species, the individuals when in a crowd appear so much alike that 

 we are unable to recognize persons whom we know from strangers. In 

 foreign races of our own species we recognize individual differences and 



