SELECTION OF SEED 109 



has so far been made seems to me to be that undertaken 

 in Madagascar by Prudhomme. The Colonial Govern- 

 ment collected, by visiting different lands with well- 

 developed coco-nut industry, the best-known varieties 

 of Ceylon, Indo- China, and the islands of the Indian 

 Ocean. These were grown together, but no publications 

 have so far been made which show how the different 

 varieties have behaved when grown side by side. 



As to the number of varieties which may be 

 recognized, Prudhomme presents a plate showing 

 photographs of twenty apparently very distinct forms. 

 These represent the most conspicuously distinct varieties 

 from Ceylon, Noumea, the Seychelles, and continental 

 India. The list of names for varieties can be carried 

 very much higher. Simmonds lists thirty varieties, 

 and Shortt names thirty for Travancore alone. On 

 the other hand, Watt reduces all the Indian forms to 

 five varieties. Jumelle states that at least twenty-five 

 varieties are known in Java. But both he and Hubert 

 also state that there are more than forty known in 

 the Philippines ; and this figure, if it has any basis 

 whatever, can only rest on the number of dialect 

 names which it is possible to distinguish. 



The more conspicuously different varieties are for 

 the most part unquestionably duplicated in the different 

 ]ands where coco-nuts are grown. It is possible to 

 distinguish varieties in various ways. For instance, 

 the colour of coco-nuts varies from green to yellow 

 and brown. The brown nuts and green nuts are very 

 different in appearance ; but a careful analytical study 

 made by Walker at San Eamon failed to show any 

 constant differences whatever in size, in yield of copra, 

 or in richness in oil. Different, then, as these nuts 

 appear, there is no practical object whatever in dis- 

 tinguishing between them. There is no reason to 

 suppose that this statement does not apply equally 

 in every part of the world. 



So far as any careful observations show, differences 

 in form are likewise unimportant. It is easy in any 



