My dear Sir

The cases amongst Leguminosae where the apetalous flowers almost (but not quite) without anthers produce more seed than the perfect flowers are certainly two South of France Ononises O. minutissima and parviflora the North American and some of the Asiatic Lespedezas, a South American Clitoria and one of the closely allied genus Neurocarpum—of which the apetalous flowers have been described as Martia the North American Amphicarpaea monoica.—² I had formerly thought that a much more curious instance of bisexuality was in Arachis, and Stylosanthes and I may have given you those names, but in those cases I was mistaken—deceived by the sudden changes, that take place after fecundation—so take care how you quote me for them.³

The most curious instance of this kind of bisexuality in British Plants, and one in which the mode of impregnation has not been observed is in our Violets—V. odorata V. canina and V. palustris.⁴ If you look at any of them in July you will find numbers of minute flowers—a calyx, small abortive petals and stamens shorter than the calyx or even none at all, a perfect ovary which always ripens its seed whilst the showy spring flowers scarcely ever do   The variety called Russian violet is very remarkable for the great number of these summer flowers which I have seen go on till September when they gradually begin to have petals and stamens again.⁵

In Ononis parviflora the apetalous flowers are very early, in spring, the perfect ones come later   In Viola the perfect ones are early the apetalous ones later

Similar unisexuality has been observed in North American Cistineae (Helianthemum & Lechea) in tropical American pentandrous Malpighiaceae, and several other plants—but these have exposed stamens to their perfect flowers whilst in Viola and Papilionaceae they are concealed.⁶

Your’s very sincerely | George Bentham