458 GENEBAL RESULTS. CHAP. XTI. 



over, seeds are frequently exchanged or procured from 

 other gardens having a different kind of soil ; and the 

 individuals of the same cultivated species will thus be 

 subjected to a change of conditions. If the flowers are 

 not visited by our native insects, or very rarely so, 

 as in the case of the common and sweet pea, and 

 apparently in that of the tobacco when kept in a 

 hothouse, any differentiation in the sexual elements 

 caused by intercrosses will tend to disappear. This 

 appears to have occurred with the plants just 

 mentioned, for they were not benefited by being 

 crossed one with another, though they were greatly 

 benefited by a cross with a fresh stock. 



I have been led to the views just advanced with 

 respect to the causes of the differentiation of the sexual 

 elements and of the variability of our garden plants, 

 by the results of my various experiments, and more 

 especially by the four cases in which extremely incon- 

 stant species, after having been self-fertilised and 

 grown under closely similar conditions for several 

 generations, produced flowers of a uniform and constant 

 tint. These conditions were nearly the same as those 

 to which plants, growing in a garden clear of weeds, 

 are subjected, if they are propagated by self-fertilised 

 seeds on the same spot. The plants in pots were, 

 however, exposed to less severe fluctuations of climate 

 than those out of doors ; but their conditions, though 

 closely uniform for all the individuals of the same 

 generation, differed somewhat in the successive gene- 

 rations. Now, under these circumstances, the sexual 

 elements of the plants which were intercrossed in each 

 generation retained sufficient differentiation during 

 several years for their offspring to be superior to 

 the self-fertilised, but this superiority gradually and 

 manifestly decreased, as was shown by the difference 



