PLANTS CULTIVATED FOR THEIR STEMS OR LEAVES. 157 



Ganges, which they regard as a proof that it is indigenous. 

 Macfadyen says so without giving any proof. It was an 

 assertion made to him in Jamaica by some traveller ; but 

 Sir W. Hooker adds in a note, " Dr. Roxburgh, in spite 

 of his long residence on the banks of the Ganges, has 

 never seen the seeds of the sugar-cane." It rarely flowers, 

 and still more rarely bears fruit, as is commonly the case 

 with plants propagated by buds or suckers, and if any 

 variety of sugar-cane were disposed to seed, it would 

 probably be less productive of sugar and would soon be 

 abandoned. Rumphius, a better observer than many 

 modern botanists, has given a good description of the 

 cultivated cane in the Dutch colonies, and makes an 

 interesting remark.^ " It never produces flowers or fruit 

 unless it has remained several years in a stony place." 

 Neither he, nor any one else to my knowledge, has de- 

 scribed or drawn the seed. The flower, on the contrary, 

 has often been figured, and I have a fine specimen from 

 Martinique.* Schacht is the only person who has given 

 a good analysis of the flower, including the pistil ; he 

 had not seen the seed ripe.* De Tussac,* who gives a 

 poor analysis, speaks of the seed, but he only saw it 

 young in the ovary. 



In default of precise information as to the native 

 country of the species, accessory means, linguistic and 

 historical, of proving an Asiatic origin, are of some 

 interest. Ritter gives them carefully; I will content 

 myself with an epitome. The Sanskrit name of the sugar- 

 cane was ikshu, ikshura, or iJcshava, but the sugar was 

 called sarkara, or sakkara, and all its names in our Euro- 

 pean languages of Aryan origin, beginning with the 

 ancient ones — Greek, for example — are clearly derived 

 from this. This is an indication of Asiatic origin, and that 

 the produce of the cane-was of ancient use in the southern 

 regions of Asia with which the ancient Sanskrit-speak- 

 ing nation may have had commercial dealings. The 

 two Sanskrit words have remained in Bengali under the 



• Ramphius, Amboin, vol. v. p. 186. ' Hehn, No. 480. 



• Schacht, Madeira vmd Teneriffe, tab. i. 



• Tussao, Flore des Antilles, i. p. 153, pi. 23. 



