VII ON" THE IXFERTILITY OF CROSSES 171 



to insects, for it is found that silkworms which produce white 

 cocoons resist the fungus disease much better than do those 

 which produce yellow cocoons. ^ Among plants, Ave have in 

 North America green and yellow-fruited plums not affected by 

 a disease that attacked the purple-fruited varieties. Yellow- 

 fleshed peaches suffer more fi'om disease than white-fleshed 

 kinds. In Mauritius, white sugar-canes were attacked by a 

 disease from Avliich the red canes were free. White onions 

 and verbenas are most liable to mildew ; and red-flowered 

 hyacinths Avere more injured hj the cold during a severe 

 Avinter in Holland than any other kinds.- 



These curious and inexjDlicable correlations of colour Avith 

 constitutional peculiarities, both in animals and plants, render 

 it probable that the correlation of colour AAath infertility, 

 Avhich has been detected in several cases in jDlants, may also 

 extend to animals in a state of nature ; and if so, the fact 

 is of the highest importance as throAving light on the origin 

 of the infertility of many allied species. This aahII be better 

 understood after considering the facts Avhich Avill be now 

 described. 



The Isolation of Varieties by Selective Association. 



In the last chapter I have shoAvn that the importance of 

 geographical isolation for the formation of new species by 

 natural selection has been greatly exaggerated, because the 



1 In the Medico- Chirurgical Transactions, vol. liii. (1870), Dr. Ogle has ' 

 adduced some curious physiological facts bearing on the presence or absence 

 of white colours in the higher animals. He states that a dark pigment in the J 

 olfactory region of the nostrils is essential to perfect smell, and that this 

 pigment is rarely deficient except when the whole animal is pure white, and 

 the creature is then almost Avithout smell or taste. He observes that there is 

 no proof that, in any of the cases given above, the black animals actually eat 

 the poisonous root or plant ; and that the facts are readily understood if the 

 senses of smell and taste are dependent on a pigment which is absent in the 

 white animals, who therefore eat what those gifted with normal senses avoid. 

 This explanation however hardly seems to cover the facts. We cannot sup- 

 pose that almost all the sheep in the world (which are mostly white) are 

 without smell or taste. The cutaneous disease on the white patches of hair 

 on horses, the special liability of white terriers to distemper, of white chickens 

 to the gapes, and of silkworms which produce yellow silk to the fungus, are 

 not explained by it. The analogous facts in iilants also indicate a real con- 

 stitutional relation with colour, not an aflTection of the sense of smell and 

 taste only. 



- For all these facts, see Animals and Plants under Dmnesfication, vol. ii. 

 pp. 335-338. 



