FORMS OF FLOWERS IN VALERIANA DIOICA. 93 
flower ages, the style undergoes marked elongation and the ovary increases 
rapidly in size. 
The stamens of these flowers are the merest rudiments, in by far the largest 
number of cases, and are scarcely visible from above (figs. 3 a, 4). There is, 
however, considerable variation in this respect, not only as regards the 
flowers of different female plants, but also between different flowers of the 
same individual plant. ОЁ course, the degree of visibility of the stamens 
depends to some extent upon characters, such, for instance, as the openness 
of the throat of the corolla, whieh have nothing to do with the stamens 
themselves. It depends, too, on the age of the flower, the stamens being 
somewhat more conspicuous in the young flower. But there is also a real 
variation not dependent on such factors as these, which is sufficiently great 
to render the stamens, in some cases, easily visible from above. The variation 
may be either in the size of the rudimentary anthers, or in the length of the 
filaments, or in both (see figs. 5, 6). 
In the more extreme forms of this variation, the staminal rudiments are 
sufficiently obvious to give rise to some doubt as to whether the plants might 
not correctly be placed under Müller's third type—‘ Female, with distinct 
rudiments of anthers.” Especially is this the case when this variation is 
accompanied by another, in the form of an unusually large corolla. The 
combined effect of masses of such flowers upon the general appearance of the 
flower-head is such that the plants could be picked out from the other female 
plants of more usual type, among which they were growing. The plants were 
therefore kept under observation for two or three years, when it became 
quite clear that no real line of separation could be drawn between them and 
the more usual types into which they grade. 
Although such plants as these just mentioned cannot be separated from the 
females, on the other hand, in their more extreme forms (see figures 10, 11, 
lla, 116) they closely approach another type of flower which is herm- 
aphrodite both in structure and function. 
Plants with Hermaphrodite Flowers. 
In making the records of the plants examined, I have included under the 
title * Hermaphrodite,” all plants which produced pollen and indicated, by 
the continued growth of the ovary after the corolla had fallen and by the 
development of a pappus, the power to produce good seed. — The possibilities 
of testing the seed produced by these plants were limited; the wild plants 
had already begun to set seed from the earlier flowers when they were 
transferred to the garden, but they suffered so much from the transplanting 
during their flowering period that only one plant ripened a few seeds, from 
which one seedling was obtained. Exceedingly poor harvests were obtained 
in all cases from plants which had been raised from seed in the garden, and 
