250 APPENDIX III LETTERS PUBLISHED IN NATURE. 



the causes of divergence are all causes of segregation ; while the causes 

 of unification, whether of functions or of structures, are causes of 

 intergeneration. If the environments which surround the isolated 

 portions are the same, the use of the environment, and, therefore, 

 the forms of selection, may become divergent ; if the use continues 

 unchanged, some useless divergence in the method of securing the 

 use may appear; or, if all the relations to the environment, whether 

 useful or useless, remain unchanged, "the adjustment of the male and 

 female elements to each other" are liable to become slightly diver- 

 gent, producing mutual infertility, or the preference of the sexes for 

 certain shades or arrangements of color in their mates may become 

 slightly different, or, through some slight difference in the hereditary 

 elements distributed in each separated portion at the first, one, or all 

 of these causes of accumulated divergence may be introduced. I 

 think it is evident that we have here a general principle which is as 

 applicable to a wide range of divergences as it is to the divergence 

 that produces mutual infertility and sterility. 



The context shows that the prominent idea in Mr. Wallace's mind 

 was divergence in the adjustment of the male and female elements, 

 through correlation with "some diversity of form or color," resulting 

 from divergent forms of natural selection, which had been induced by 

 exposure to "somewhat different conditions of life." But if the rea- 

 soning is correct in the sentences I have quoted above it gives an 

 explanation of similar divergences when the separated portions are 

 exposed to the same environment and where there is no possible 

 advantage to be gained by divergence. This is one of the principles 

 I have used in the explanation of the divergences of Sandwich Island 

 land mollusks ; and I think that in the earlier stages of the develop- 

 ment of infertility between allied forms it is often the only explana- 

 tion that is applicable. It should, however, be remembered that, for 

 divergence of this kind, it is not always necessary that the isolation 

 should be either complete or very long continued, and that, when the 

 forms that are not fully fertile with each other meet and more or less 

 commingle, there is, through the very laws of propagation, without 

 any aid from natural selection, a constant increase in the ratio of the 

 pure breeds to the mongrels, and an accumulating intensity in the 

 segregative instincts and the physiological incompatibilities. As this 

 point has been fully discussed in my paper on "Divergent Evolu- 

 tion," I do not need to enlarge on it here. 



There is, however, another phase of the subject which is indicated 

 by Mr. Wallace's suggestion that infertility depends on "such a 

 delicate adjustment" that it is more easily affected by isolation than 



