To sum up: several Dioscoreæ wild in Asia (especially in the Asiatic Archipelago), and others less numerous growing in America and in Africa, have been introduced into cultivation as alimentary plants, probably more recently than many other species. This last conjecture is based on the absence of a Sanskrit name, on the limited geographical range of cultivation, and on the date, which appears to be not very ancient, of the inhabitants of the Pacific Isles.
Arrowroot—Maranta arundinacea, Linnæus. A plant of the family of the Scitamineæ, allied to the genus Canna, of which the underground suckers[315] produce the excellent fecula called arrowroot. It is cultivated in the West India Islands and in several tropical countries of continental America. It has also been introduced into the old world—on the coast of Guinea, for instance.[316]
Maranta arundinacea is certainly American. According to Sloane,[317] it was brought from Dominica to Barbados, and thence to Jamaica, which leads us to suppose that it was not indigenous in the West Indies. Körnicke, the last author who studied the genus Maranta,[318] saw several specimens which were gathered in Guadaloupe, in St. Thomas, in Mexico, in Central America, in Guiana, and in Brazil; but he did not concern himself to discover whether they were taken from wild, cultivated, or naturalized plants. Collectors hardly ever indicate this; and for the study of the American continent (excepting the United States) we are unprovided with local floras, and especially with floras made by botanists residing in the country. In published works I find the species mentioned as cultivated[319] or growing in plantations,[320] or without any explanation. A locality in Brazil, in the thinly peopled province of Matto Grosso, mentioned by Körnicke, supposes an absence of cultivation. Seemann[321] mentions that the species is found in sunny spots near Panama.
A species is also cultivated in the West Indies, Maranta indica, which, Tussac says, was brought from the East Indies. Körnicke believes that M. ramosissima of Wallich found at Sillet, in India, is the same species, and thinks it is a variety of M. arundinacea. Out of thirty-six more or less known species of the genus Maranta, thirty at least are of American origin. It is therefore unlikely that two or three others should be Asiatic. Until Sir Joseph Hooker’s Flora of British India is completed, these questions on the species of the Scitamineæ and their origin will be very obscure.
Anglo-Indians obtain arrowroot from another plant of the same family, Curcuma angustifolia, Roxburgh, which grows in the forests of the Deccan and in Malabar.[322] I do not know whether it is cultivated.
CHAPTER II.
PLANTS CULTIVATED FOR THEIR STEMS OR LEAVES.
Article I.—Vegetables.