144 THE JOUENAL OF INDIAN BOTANY.
not allow me to say very much ; but certain general conclusions may be drawn from an examination of the very full and careful descriptions of the flowers given by Kuhland in his monograph. It might be supposed at the outset that since the plants grow in water and marshy ground their seeds would be carried on the feet of migrating birds, and that this, coupled with the universally accepted similarity of the conditions of water and marsh tbe world over, would result in a very wide distribution of most of the species. Some certainly are scattered widely, but the majority seem to be confined to comparatively small areas. This question is of course bound up with that of the limits of the species : thus E. Sieboldianum Sieb ot Zucc, aj understood by Hooker in the F.B.I., occurs all over S. E. Asia from Bombay to Japan and N. Australia ; but Euhland separating from it several smaller species gives to them a much narrower distribution, though he retains almost as wide a one for E. Sieboldianum itself. Of the groups which I have proposed in this account the SETACE^j group has one representative, E. bifistulosum Van Huerck, in West Africa and probably others elsewhere. The jglMPLICES being all those with no special modification of the floral parts are no doubt primitive and world-wide. The HIRSUTE and ANISOPETAL^I are spread over S. Eastern Asia from Cochin to China, probably on the mountains of the warmer parts, and the latter seem to have a second centre of distribution in British Guiana. The CRISTATO- SEPAL^ also seem to have a centre in tropical South America, reaching from Mexico to Brazil. But the CONNATO-SEPALAE, which have in India only one representative on the Himalayas, belong almost entirely to China and Japan. Of the LEUCANTHERAE one species, E. Sieboldianum Sieb. et Zucc, is widely spread over tropical S. E. Asia, Malaya and Australia, but the others seem confined to India. E. Sie- boldianum is probably ."the most widely distributed of all the species, and E. Brownianum Mart, with its varieties (or related species of Euh- land) covers almost as wide an area.
Inside India there appear to be on the plains and lower hills no species at all north of a line from Mt. Aboo to Dacca, and not many northwards on the Himalayas, though there are one or two in Kash- mir. They occur all over South India. The hirsutae belong almost entirely to the mountains above 3,000 ft. of Burma, Bengal, S. India and Ceylon, but extend far southwards to Singapore. The ANI- SOPETALAE are developed chiefly in Ceylon, with one species in Ben- gal, one in the Central Provinces and the Deccan and another on the Niligiris ; but not curiously enough collected hitherto on the Palnis which are nearer Ceylon and floristically show closer affinities. Of the CRISTATO-SEPAL2E the smaller species belong to the Western
