observed the same thing in Nigella arvensis, he afterwards
found exactly the opposite arrangement in a species of Euphorbia, in which the stigmas can receive the pollen by the
aid of insects only from older flowers.
He goes on to say that he grounds his theory of flowers on these his six chief discoveries made in the course of five years; he then gives his theory at length, first of all explaining the nature of juice-secreting glands (nectaries), and organs for receiving or covering the nectar, and the contrivances for enabling insects to find their way readily to the juice. He calls attention to Koelreuter's excellent observations on the fertilisation of nectar-bearing flowers by insects, and notices that no one has hitherto shown that the whole structure of such flowers has this object in view, and can be fully explained by it. He finds the chief proof of this important proposition in dichogamy.
'After the flower,' he says, 'has opened in dichogamous plants, the filaments have or assume either all at once or one after another a definite position, in which the anthers open and offer their pollen for fertilisation. But at this time the stigma is at some distance from the anthers and is still small and closed. Hence the pollen cannot be conveyed to the stigma either by mechanical means or by insects, for there is as yet properly no stigma. This condition of things continues a certain time. When that time is elapsed, the anthers have no longer any pollen, and changes take place in the filaments the result of which is that the anthers no longer occupy their former position. Meanwhile the pistil has so changed that the stigma is now exactly in the place where the anthers were before, and as it now opens, or expands the parts of which it is composed, it often fills about the same space as the anthers filled before. Now the spot, which was at first occupied by the ripe anthers and is now occupied by the ripe stigma, is so chosen in each flower, that the insect for which the flower is intended cannot reach the juice without touching with a por-
