TJie "Amilbeds " and the Pummelos. 39 



appears to be an elephant form of the warty khatta, 

 orange group, pis. 32 and 35. . . . The at Anni, also 

 tomentosc, appears to be the " elephant " variety of the 

 suntara group, pi. 1 1 2. The Kumaon lemon, or 

 gulgul, may be taken as the " elephant " variety of the 

 lemon group, pis. 187 and 188, and the mawling of 

 Mangalore, and the madhkakree of Almora, pis. 148 and 

 171, would, perhaps, be the " elephant " representatives 

 of the Citron group. For all I know, there may be 

 others, which I may not have come across. It would, 

 therefore, appear that each type or race of Citrus has, 

 or may have, its variety decumana. If so, it is no 

 wonder, that the wild original Q{ the pummelo has never 

 yet been discovered. It probably only required the 

 evolution of a yellow orange, like that of Kandy, for 

 this " decumana " citrus to have come into being, first 

 as a product of luxuriant growth, and afterwards as the 

 result of selection. The only parts of this Citrus which 

 have not been elephantised are the spines. These, 

 however, are usually well developed under conditions 

 just the opposite of luxuriance, some of the branch buds, 

 instead of developing normally, remaining dwarfed as 

 spines. If grown from seed, however, the spines 

 of the pummelo might also be large, like that of most 

 other seedlings. 



Alphonse de Candolle, in searching for the cradle of 

 the pummelo, seems to lay great stress on the observa- 

 tions of Sleeman, that in the Fiji Isles, it is extremely 

 common, and covering the banks of rivers. De Can- 

 dolle appears to infer from this that, being so common, 

 there must be in that vicinity some centre containing 

 the wild parent of this supposed distinct species, the 

 Citrus decumana. 



From experiments I have made, I found that most 

 of the citrus fruit will float in water ; so will that of the 



