As a large number of seeds were placed on the sand to germinate, many remained after the pairs had been selected, some of which were in a state of germination and others not so; and these were sown crowded together on the opposite sides of one or two rather larger pots, or sometimes in two long rows out of doors. In these cases there was the most severe struggle for life among the crossed seedlings on one side of the pot, and the self-fertilised seedlings on the other side, and between the two lots which grew in competition in the same pot. A vast number soon perished, and the tallest of the survivors on both sides when fully grown were measured. Plants treated in this manner, were subjected to nearly the same conditions as those growing in a state of nature, which have to struggle to maturity in the midst of a host of competitors.
On other occasions, from the want of time, the seeds, instead of being allowed to germinate on damp sand, were sown on the opposite sides of pots, and the fully grown plants measured. But this plan is less accurate, as the seeds sometimes germinated more quickly on one side than on the other. It was however necessary to act in this manner with some few species, as certain kinds of seeds would not germinate well when exposed to the light; though the glasses containing them were kept on the chimney-piece on one side of a room, and some way from the two windows which faced the north-east. (1/7. This occurred in the plainest manner with the seeds of Papaver vagum and Delphinium consolida, and less plainly with those of Adonis aestivalis and Ononis minutissima. Rarely more than one or two of the seeds of these four species germinated on the bare sand, though left there for some weeks; but when these same seeds were placed on earth in pots, and covered with a thin layer of sand, they germinated immediately in large numbers.)
The soil in the pots in which the seedlings were planted, or the seeds sown, was well mixed, so as to be uniform in composition. The plants on the two sides were always watered at the same time and as equally as possible; and even if this had not been done, the water would have spread almost equally to both sides, as the pots were not large. The crossed and self-fertilised plants were separated by a superficial partition, which was always kept directed towards the chief source of the light, so that the plants on both sides were equally illuminated. I do not believe it possible that two sets of plants could have been subjected to more closely similar conditions, than were my crossed and self-fertilised seedlings, as grown in the above described manner.
In comparing the two sets, the eye alone was never trusted. Generally the height of every plant on both sides was carefully measured, often more than once, namely, whilst young, sometimes again when older, and finally when fully or almost fully grown. But in some cases, which are always specified, owing to the want of time, only one or two of the tallest plants on each side were measured. This plan, which is not a good one, was never followed (except with the crowded plants raised from the seeds remaining after the pairs had been planted) unless the tallest plants on each side seemed fairly to represent the average difference between those on both sides. It has, however, some great advantages, as sickly or accidentally injured plants, or the offspring of ill-ripened seeds, are thus eliminated. When the tallest plants alone on each side were measured, their average height of course exceeds that of all the plants on the same side taken together. But in the case of the much crowded plants raised from the remaining seeds, the average height of the tallest plants was less than that of the plants in pairs, owing to the unfavourable conditions to which they were subjected from being greatly crowded. For our purpose, however, of the comparison of the crossed and self-fertilised plants, their absolute height signifies little.
As the plants were measured by an ordinary English standard divided into inches and eighths of an inch, I have not thought it worth while to change the fractions into decimals.