78 ON THE ORIGIN OF SPECIES 



a great difference in cultivating the several varieties, 

 assuredly, in a state of nature, where the trees would 

 have to struggle with other trees and with a host of 

 enemies, such differences would effectually settle which 

 variety, whether a smooth or downy, a yellow or purple 

 fleshed fruit, should succeed. 



In looking at many small points of difference be- 

 tween species, which, as far as our ignorance permits 

 us to judge, seem quite unimportant, we must not forget 

 that climate, food, etc., probably produce some slight 

 and direct effect. It is, however, far more necessary 

 to bear in mind that there are many unknown laws of 

 correlation of growth, which, when one part of the 

 organisation is modified through variation, and the 

 modifications are accumulated by natural selection for 

 the good of the being, will cause other modifications, 

 often of the most unexpected nature. 



As we see that those variations which under domesti- 

 cation appear at any particular period of life, tend to 

 reappear in the offspring at the same period ; — for in- 

 stance, in the seeds of the many varieties of our culinary 

 and agricultural plants ; in the caterpillar and cocoon 

 stages of the varieties of the silkworm ; in the eggs of 

 poultry, and in the colour of the down of their chickens ; 

 in the horns of our sheep and cattle when nearly adult ; — 

 so in a state of nature, natural selection will be enabled 

 to act on and modify organic beings at any age, by the 

 accumulation of variations profitable at that age, and by 

 their inheritance at a corresponding age. If it profit a 

 plant to have its seeds more and more widely dissemi- 

 nated by the wind, I can see no greater difficulty in this 

 being effected through natural selection, than in the 

 cotton-planter increasing and improving by selection 

 the down in the pods on his cotton -trees. Natural 

 selection may modify and adapt the larva of an insect 

 to a score of contingencies, wholly different from those 

 which concern the mature insect. These modifications 

 will no doubt affect, through the laws of correlation, the 

 structure of the adult; and probably in the case of those 

 insects which live only for a few hours, and which never 



