44 
Let us now pass on to consider the nature of the seed- 
lings themselves. It was pointed out previously that 
Darwin had shown the beneficial effect of cross-pollination 
in the more vigorous development of the offspring. Another 
important feature of seed reproduction as compared with 
vegetative reproducton, particularly if the seeds are 
the result of cross-fertilisation is the occurrence of a con- 
siderable amount of variation. The more dissimilar the 
parents the more varied are the offspring, and the greater 
therefore the scope for the play of natural or artificial 
selection. 
The term variation has been used for two very different 
phenomena noticeable when examining a large number of 
plants or animals of the same species or kind. It is well 
lnown that the offspring of any two parents all have some 
different individual characteristics, and a close observation: 
of a number of seedling plants will show us that though 
they all have a oar resemblance, we find that - they 
differ slightly one from another in size,in the shape and 
texture of their leaves, and when they grow up in colour and 
conformation of their flowers. Indeed, if we had carefully 
examined the seeds from which they have grown we should 
have found that the latter showed already a considerable 
range of variation in shape and size, and possibly also in 
colour. Such slight individual variations are always 
found to fluctuate around a mean or average which we 
may look upon as the general character of the species or 
race. It is these slight individual variations which 
Darwin regarded as so important in the evolution of new 
forms, those least suited to the particular conditions of 
life gradually dying out and leaving those which were fit 
to survive. Natural selection operating in this way was 
thought to have produced the innumerable forms which we 
know as natural species. In the same way man, by making 
a choice of the plants most suitable to his purposes, has 
by artificial selection produced the strains and varieties 
now cultivated. 
Recently experiments have been carried out in which 
the seeds of certain plants have been carefully graded and 
the largest sown separately with a view to ascertaining 
whether by such constant selection the seeds could be in- 
definitely increased in size. This was, however, not found 
to be possible, for though the average size of the seeds was 
at first considerably raised, a limit was reached beyond 
