VII 
ON THE INFERTILITY OF CROSSES 
169 
which it is correlated; and though these differences are 
sometimes slight, and the amount of the infertility is not 
always, or even usually, proportionate to the external dif¬ 
ference between the two forms crossed, we must believe that 
there is some connection between the two classes of facts. 
This is especially the case as regards colour; and Mr. Darwin 
has collected a body of facts which go far to prove that 
colour, instead of being an altogether trifling and un¬ 
important character, as was supposed by the older natural¬ 
ists, is really one of great significance, since it is un¬ 
doubtedly often correlated with important constitutional 
differences. Now colour is one of the characters that most 
usually distinguishes closely allied species; and when we 
hear that the most closely allied species of plants are 
infertile together, while those more remote are fertile, the 
meaning usually is that the former differ chiefly in the colour 
of their flowers, while the latter differ in the form of the 
flowers or foliage, in habit, or in other structural characters. 
It is therefore a most curious and suggestive fact, that in 
all the recorded cases, in which a decided infertility occurs 
between varieties of the same species, those varieties are 
distinguished by a difference of colour. The infertile 
varieties of Verbascum were white and yellow flowered 
respectively; the infertile varieties of maize were red and 
yellow seeded; while the infertile pimpernels were the red 
and the blue flowered varieties. So, the differently coloured 
varieties of hollyhocks, though grown close together, each 
reproduce their own colour from seed, showing that they are 
not capable of freely intercrossing. Yet Mr. Darwin assures 
us that the agency of bees is necessary to carry the pollen 
from one plant to another, because in each flower the pollen 
is shed before the stigma is ready to receive it. We have 
here, therefore, either almost complete sterility between 
varieties of different colours, or a prepotent effect of pollen 
from a flower of the same colour, bringing about the same 
result. 
Similar phenomena have not been recorded among 
animals ; but this is not to be wondered at when we consider 
that most of our pure and valued domestic breeds are 
characterised by definite colours which constitute one of their 
