IX THE DEBT OF SCIENCE TO DARWIN 465 
After a considerable amount of observation and experiment, 
he found that bees and moths visited the flowers, and that 
their probosces became covered with pollen while sucking up 
the nectar, and further, that the pollen of a long-stamened 
plant would be most surely deposited on the stigma of the 
long-styled plants, and vice versd. Now followed a long series 
of experiments, in which cowslips were fertilised either with 
pollen from the same kind or from a different kind of flower, 
and the invariable result was that the crosses between the 
two different kinds of flowers produced more good capsules, 
and more seeds in each capsule; and as these crosses would 
be most frequently effected by insects, it was clear that this 
curious arrangement directly served to increase the fertility 
of this common plant. 
The same thing was found to occur in the primrose, and 
in many other species of primulacee, as well as in flax (Linum 
perenne), lungworts (Pulmonaria), and a host of other plants, 
including the American partridge-berry (Mitchella repens). 
These are called dimorphic heterostyled plants. 
Still more extraordinary is the case of the common loose- 
strife (Lythrum salicaria), which has both stamens and styles 
of three distinct lengths, each flower having two sets of 
stamens and one style, all of different lengths, and arranged 
in three different ways: (1) a short style, with six medium 
and six long stamens; (2) a medium style, with six short and 
six long stamens; (3) a long style, with six medium and six 
short stamens. These flowers can be fertilised in eighteen 
distinct ways, necessitating a vast number of experiments, 
the result being, as in the case of the cowslip, that flowers 
fertilised by the pollen from stamens of the same length as 
the styles, gave on the average a larger number of capsules 
and a very much larger number of seeds than in any other 
case. The exact correspondence in the length of the style of 
each form with that of the stamens in the two other forms 
ensures that the pollen attached to any part of the body of 
an insect shall be applied to a style of the same length on 
another plant, and there is thus a triple chance of the maxi- 
mum of fertility. Some other species of lythrum, of oxalis, 
and pontederia, were also found to have three-formed stamens 
and styles; and in the case of the oxalis, experiments were 
2H 
