THE INDIAN SPECIES OF EKIOCAULON. 137
The Arrangement of the Species.
As far as the Indian species are concerned, if we except the local floras founded on the F.B.I, there are, as stated ahove, only two modern works giving descriptions, the F.B.I, and Euhland's mono- graph of the whole family. Previous to this excellent descriptions were given hy Steudel (Syn. PI. Cyperacearum 1858) and Koerniche (Linnsea XXVII, 1854, pp. 577-692).
The species are arranged in these two works on entirely different plans. Hooker after separating the purely aquatic and submerged forms, divided the remainder according to the external appearance of the heads and the presence or absence of hairs on the receptacle. Ruhland on the other hand arranged the species according to the num- ber of parts in the flower, placing in his first section, which though he does not so identify it I take to be Naysmithia Huds., those with 2 parts to each whorl, in his second section those with 3 parts, and in his third those with 3 parts in the staminal and carpellary whorls but with fewer sepals or petals ; these sections being further divided for convenience into the eastern or old-world species and the western or new-world. He then took the nature of the stem whether disciform or elongated, with such characters of the flower as white or black anthers, crested or plain sepals, for the lesser divisions. The difference in the two systems is very great. Two plants classed by Hooker as dimerous and trimerous varieties of the same species, E. sexangulare, appear in Euhlands monograph, as also in Steudel's Syn. PI. Cyp., in different main divisions of the old world species, and in the former s list as numbers 25 and 186 respectively. The plant named by Trimen E. atratum Koerniche var. major was raised by Hooker to the rank of a species, E. caulescens, and placed next to E. robustum Steud. of the Nilgiris, from which it hardly differs except in having a tall and branched stem ; whereas Ruhland separated the two by no fewer than twelve Indian species, and placed E. robustum next to E. quinqucangulare Linn., which in Hooker's arrangement is separated from it by almost the whole of the Indian forms, one being No. 4 and the other No. 35 out of 43.
It would probably be correct to state that except in his main divisions Ruhland in fact did not attempt to arrange the species in phylogenetic groups, but only to provide a general clavis for aid in their identification. Hooker made tentative groupings, but apart from the separation of aquatic from terrestrial species made no definite sections.
Before attempting to classify the species of a genus it is clearly necessary to determine what characters if any are liable to vary with age or with the conditions of the environment, and further
