PLANTS CULTIVATED FOR THEIR SEEDS. 435 



Christ, but the most unmistakable descriptions are in 

 works later than the ninth century of our era. It is true 

 that the ancient writers scarcely knew the south of 

 China, the only part of the empire where the cocoa-nut 

 palm can live. 



In spite of the Sanskrit names, the existence of the 

 cocoa-nut in Ceylon, where it is well established on the 

 coast, dates from an almost historical epoch. Near Point 

 de Galle, Seemann tells us may be seen carved upon a 

 rock the figure of a native prince, Kotah Kaya, to whom 

 is attributed the discovery of the uses of the cocoa-nut, 

 unknown before him ; and the earliest chronicle of Ceylon, 

 the Marawansa, does not mention this tree, although it 

 carefully reports the fruits imported by different princes. 

 It is also noteworthy that the ancient Greeks and Egyp- 

 tians only knew the cocoa-nut at a late epoch as an Indian 

 curiosity. Apollonius of Tyana saw this palm in Hin- 

 dustan, at the beginning of the Christian era.^ 



From these facts the most ancient habitation in Asia 

 would be in the archipelago, rather than on the continent 

 or in Ceylon ; and in America in the islands west of 

 Panama. What are we to think of this varied and 

 contradictory evidence ? I formerly thought that the 

 arguments in favour of Western America were the 

 strongest. Now, with more information and greater 

 experience in similar questions, I incline to the idea of an 

 origin in the Indian Archipelago. The extension towards 

 China, Ceylon, and India dates from not more than three 

 thousand or four thousand years ago, but the transport 

 by sea to the coasts of America and Africa took place 

 perhaps in a more remote epoch, although posterior to 

 those epochs when the geographical and physical 

 conditions were different to those of our day. 



* Seemann, Fl. Vitiensis, p. 276; Pickering, Chronol. Arrangement, 

 p. 428. 



