70 WILLIS : GEOORAPHICAL DISTRIBUTION 



species, which most commonly occur also beside, or near to, 

 the local endemics. 



If we take the genera of this family in order, from Hooker's 

 Flora, the first is Delima, and in this there is one very widely 

 ranging species, D. sarmentosa, occurring in Ceylon, and from 

 Assam to Singapore, the Malay Arcliipelago, the Philippines, 

 and South China, while the other species, D. Icevis, has only 

 been once collected, at Malacca. Now, if we accept the theory 

 of infinitesimal variations, we must either admit that D. sar- 

 mentosa was evolved near Malacca, and afterwards spread 

 enormously, while D. Icevis has not spread, or else that there 

 has been a vast amount of destruction, reducing D. Icevis to 

 one locality, or destroying the other species that were evolved 

 with it, or make some other bold assumption. It is, however, 

 very nearly allied indeed to D. sarmentosa, differing mainly in 

 the size and number of veins of the leaf, and in the size of the 

 flower, and if anything, one would expect that it would be a 

 successful species, and spread as widely as D. sarmentosa. But 

 it has not spread, while D. sarmentosa is very widespread and 

 very common. But if we accept the theory of mutation, the 

 explanation is comparatively simple ; we have only to suppose 

 that D. sarmentosa spread freely through the country, and 

 mutated off D. Icevis at Malacca. This is a far simpler expla- 

 nation than that which can be given upon the theory of infini- 

 tesimal variations, besides being a good deal more probable ; 

 we have absolutely no warrant to assume that there has been 

 much, or indeed any, destruction, nor to assume that D. sar- 

 mentosa was evolved at or near Malacca. The small spread of 

 D. Icevis is easily accounted for on the mutation theory, for 

 it may have been quite recently evolved, and not having a very 

 efficient distribution mechanism would not travel very far 

 except in a great length of time.* 



The other species of the genus Delima is South American, 

 D. aspera, in French and Dutch Guiana. So far as I am aware, 

 there are no species yet recorded from Africa, and should it 

 transpire that there are none, we shall be met by a very diffi- 

 cult problem to the infinitesimal variationist , but one which is 



* Ridley, on Dispersal of Seeds by Wind. Ann. Bot., XIX, 190.5, 

 p. 351,. 



