though I have watched the flowers during many years.) It is curious for how long a time the flowers of the heartsease and of some other plants may be watched without an insect being seen to visit them.  During the summer of 1841, I observed many times daily for more than a fortnight some large clumps of heartsease growing in my garden, before I saw a single humble-bee at work.  During another summer I did the same, but at last saw some dark-coloured humble-bees visiting on three successive days almost every flower in several clumps; and almost all these flowers quickly withered and produced fine capsules.  I presume that a certain state of the atmosphere is necessary for the secretion of nectar, and that as soon as this occurs the insects discover the fact by the odour emitted, and immediately frequent the flowers.

As the flowers require the aid of insects for their complete fertilisation, and as they are not visited by insects nearly so often as most other nectar-secreting flowers, we can understand the remarkable fact discovered by H. Muller and described by him in ‘Nature,’ namely, that this species exists under two forms.  One of these bears conspicuous flowers, which, as we have seen, require the aid of insects, and are adapted to be cross-fertilised by them; whilst the other form has much smaller and less conspicuously coloured flowers, which are constructed on a slightly different plan, favouring self-fertilisation, and are thus adapted to ensure the propagation of the species.  The self-fertile form, however, is occasionally visited, and may be crossed by insects, though this is rather doubtful.

In my first experiments on Viola tricolor I was unsuccessful in raising seedlings, and obtained only one full-grown crossed and self-fertilised plant.  The former was 12 1/2 inches and the latter 8 inches in height.  On the following year several flowers on a fresh plant were crossed with pollen from another plant, which was known to be a distinct seedling; and to this point it is important to attend.  Several other flowers on the same plant were fertilised with their own pollen.  The average number of seeds in the ten crossed capsules was 18.7, and in the twelve self-fertilised capsules 12.83; or as 100 to 69.  These seeds, after germinating on bare sand, were planted in pairs on the opposite sides of five pots.  They were first measured when about a third of their full size, and the crossed plants then averaged 3.87 inches, and the self-fertilised only 2.00 inches in height; or as 100 to 52.  They were kept in the greenhouse, and did not grow vigorously.  Whilst in flower they were again measured to the summits of their stems (see Table 4/41), with the following result:—­

Table 4/41.  Viola tricolor.

Heights of plants measured in inches.

Column 1:  Number (Name) of Pot.

Column 2:  Crossed Plants.

Column 3:  Self-fertilised Plants.

Pot 1 :  8 2/8 :  0 2/8. 
Pot 1 :  7 4/8 :  2 4/8. 
Pot 1 :  5 :  1 2/8.