142 THE JOURNAL OF INDIAN BOTANY.
from the same proximate stock as E. Elenorae, E. Margaretae, and E. minutum, through all show this character. The question could be decided only after examination of extra-Indian species.
These three types of sepal : the united, the boat-shaped and the crested, are quite distinct ; the only specific difference being that in the latter two the three sepals may not be equal in size, and one may be flat instead of boat-shaped, or without the crest. But in the species which show these exceptions individual variation does occur Thus in E, Elenorae the relative sizes of the 3 crested sepals varies, two may be of a size, one smaller, or the three all unequal, or one the smallest without a crest. These variations are shown in plants otherwise indistinguishable. They are therefore not of specific value. I am doubtful indeed whether E. Elenorae where the sepals are un- equal should really be separated from E. Margaretae (were they are all alike), but the species are slightly different in appearance, and I have found no variation in plants with the 3 sepals equal (E. Margaretae). So also in those with two sepals boat-shaped, one not. The odd sepal in E. Dianae may be lanceolate and as long as the other two, or short- er and bristle-like, or so slender that it is difficult to see.
We have in both these lines apparently a reduction in the size of one sepal till is nearly disappears, and if this be so a species with only two sepals may be derived from one with three. An instance where this has actually and unmistakably occurred even inside the species is afforded by E. Xeranthemum Mart. Hooker in F.B.I, gives the female sepals as 2. I find plants on the Himalayas have 2 sepals, but some at least on the Malabar coast have 3. Ruhland says that the sepals are 3, unequal. No one seeing the plants would wish to make several species of them. The same occurs in E. truncatum where the sepals may be 3, but are usually 2 only, and in the same plant I have found 3 equal, 2 and 1 smaller and 2 only ; and also in E. Thioaitcsii Hook. f.
We are thus faced by a set of conditions which must be unique among flowering plants. No other case is known to me of a reduction in the relative number of sepals and petals within the species of genus, though the number may be indefinite — e.g. in species of Banwncalus and Jasminum. The stamens certainly show a reduction in some genera (e.g. Cassia and Bauhinia), and occasionally between the genera of family (e.g. Caryophyllaceae, Acanthacae), but even this is not common. Characters in fact which in all other phanerogamic families are so constant as to be of the first importance in determining fami- lies and cohorts here vary even within the species, and so are of no use at all as guides to the phylogeny.
