Chap. IX. AND SELF-FERTILISED FLOWERS. 321 



were raised from seeds purchased from nursery -gardens, 

 or taken from plants growing in my garden, or 

 growing wild, and surrounded in every case by many 

 individuals of the same species. Plants thus circum- 

 stanced will commonly have been intercrossed by 

 insects; so that the seedlings which were first ex- 

 perimented on will generally have been the product 

 of a cross. Consequently any difference in the 

 fertility of their flowers, when crossed and self-fer- 

 tilised, will have been caused by the nature of the 

 pollen employed ; that is, whether it was taken from 

 a distinct plant or from the same flower. The de- 

 grees of fertility shown in the following table, F, 

 were determined in each case by the average number 

 of seeds per capsule, ascertained either by counting 

 or weighing. 



Another element ought properly to have been taken 

 into account, namely, the proportion of flowers whicli 

 yielded capsules when they were crossed and self-fer- 

 tilised ; and as crossed flowers generally produce a 

 larger proportion of capsules, their superiority in fer- 

 tility, if this element had been taken into account, 

 would have been much more strongly marked than 

 appears in Table F. But had I thus acted, there 

 would have been greater liability to error, as pollen 

 applied to the stigma at the wrong time fails to pro- 

 duce any effect, independently of its greater or less 

 potency. A good illustration of the great difference 

 in the results which sometimes follows, if the number 

 of capsules produced relatively to the number of flowers 

 fertilised be included in the calculation, was afforded 

 by Nolana prostrata. Thirty flowers on some plants 

 of this species were crossed and produced twenty-seven 

 capsules, each containing five seeds ; thirty-two flowers 

 on the same plants were self-fertilised and produced 



Y 



