several genera of insects, as, for instance, some Earwigs (Forficulæ) and Lamellicorn Beetles (Lucanus cervus), the Stag-beetle (Dynastes hercules), and Xylotrupes gideon, there are some males in which the antennæ, the secondary sexual characters by which they differ most markedly from the females, are extremely long, and others in which they are very short. Bateson, who has written most on this subject, distinguishes the two forms as "high males" and "low males." It is true that a continuous series of intermediate forms links the extreme types, but, none the less, the vast majority of the individuals are at one extreme or the other. Unfortunately, Bateson did not investigate the relations between these different types of males and the females, and so it is not known if there be female types with special sexual affinity for these male types. Thus these observa- tions can be taken only as a morphological parallel to heterostylism and not as cases of the law of complementary sexual attraction.
Heterostylous plants may possibly be the means of establishing my view that the law of sexual complements holds good for every kind of living thing. Darwin first, and after him many other investigators have proved that in heterostylous plants fertilisation has the best results, or, indeed, may be possible only when the pollen from a macrostylous flower (a flower with the shortest form of anthers and longest pistil) falls on the stigma of a microstylous blossom (one where the pistil is the shortest possible and the stamens at their greatest length), or vice versa. In other words, if the best result is to be attained by the cross-fertilisation of a pair of flowers, one flower with a long pistil, and therefore high degree of femaleness, and short stamens must be mated with another possessing a correspondingly short pistil, and so, with the amount of femaleness complementary to the first flower, and with long stamens complementary to the short stamens of the first flower. In the case of flowers where there are three pistil lengths, the best results may be expected when the pollen of one blossom is transmitted to another blossom in which the stigma is the nearest comple-
