faction to Darwin to think he had been the chief agent in resuscitating him.
His first publications on the fertilisation of plants were two short communications to the 'Gardeners' Chronicle' in 1857 and 1858, on the fertilisation of the kidney-bean, and here it is evident that his chief interest lay in the question how far the different varieties of the bean are liable to natural intercrossing by insects. In 1860 and 1861 he worked at the fertilisation of the British orchids. This was, to a great extent, a rest to him amid the severer work entailed by the 'Variation of Animals and Plants,' and was considered by him as culpable idleness. During the whole of the latter part of 1861 and the spring of 1862 he gave himself up to the work, and the book on 'The Fertilisation of Orchids' was finished at the end of April 1862. His letters show the keen pleasure he felt in making out the complex relations between insects and orchids—a pleasure which he contrived to convey to his readers. The principles worked out in the 'Orchid' book for a single group have been accepted for flowers in general, and thus a new department of botanical research has been founded. This new field of work, which has been so largely extended by Hermann Müller and others, has reacted in a way especially satisfactory to its founder, namely, in showing how points of importance to the welfare of an organism may be hidden in apparently unimportant peculiarities, and it thus gives a basis of solid experience to the often-repeated caution as to our ignorance of the relations of organisms to their environment. No one with a knowledge of the wonderful mutual relations between flowers and insects will be inclined to dogmatise rashly as to the uselessness of any structure, or as to the consequent impossibility of its having been modified by means of natural selection.
A book which Darwin has described as the 'complement' of the 'Orchid' book was published fourteen years afterwards, in 1876. 'The Effects of Cross and Self Fertilisation,' which is hardly known except to professed naturalists, was the result of eleven years of experimental work, and contains conclusions of the highest theoretical interest. It is the complement of the 'Orchid' book, because, while that work showed how perfect are the means for insuring cross-fertilisation, the later book showed why cross-fertilisation is important. At the time of the publication of the 'Fertilisation of Orchids' no one could positively assert that a plant which is adapted for cross-fertilisation has an advantage over others not so adapted. 'The Effects of Cross and Self Fertilisation' supplied this want, and showed therefore, that each variation affecting the capabilities of a flower for cross-fertilisation must be severely tested in the struggle for life. Formerly we could only surmise that such variations were sifted out by a selective agency of unknown character; now we can show that a selective agency of a definite kind and of measurable strength must be ever at work. He showed, too, how the advantages of cross-fertilisation are in some unknown way connected with the advantages arising from changed conditions of life, and he was thus enabled to throw more direct light on the philosophy of the existence of sex than any previous writer. It is characteristic of Darwin's mode of work that the whole of this important research originated in an accidental observation. He noticed, in the course of experiments directed to another object, that the offspring of a cross were superior in vigour, even in the first generation, to seedlings of self-fertilised parentage. It is not so characteristic of him that it should have required, as he has recorded, a repetition of the accident before his attention was thoroughly roused.His next important botanical work—on heterostyled plants—was the investigation and ultimate solution of a problem at first sight of trivial significance, but really of an extremely complex character. As early as 1838 he noticed what seemed to be an unmeaning variability in the length of the style, or organ through which the influence of the pollen is conveyed to the ovules. But when he found that the primrose presents two sets of individuals, differing in a constant and striking degree in the length of the style and in other characters, he became convinced that his first idea was erroneous. Even after he had given up the variability hypothesis, he started with quite a wrong idea as to the meaning of the facts, and only attained the solution of the problem through the destruction of his preconceived ideas by a rigorous course of experiment. He showed that the two forms in the primrose, or the three forms in Lythrum, although each is a perfect hermaphrodite, are nevertheless connected with each other in a manner resembling to a certain extent the relationship between the sexes of animals. The working out of this curious result gave him, as he has mentioned, more pleasure than almost any other research. Besides giving the explanation of hitherto neglected facts, the work on heterostyled plants is of importance in a way that has not been always recognised, namely, in throwing light on the question of hybridisation. He found that in the primrose, for instance, a 'long-styled' individual crossed by another long-styled flower is comparatively infertile. So that