342 Willis.—The Sources and Distribution of the 
In my ‘ Catalogue of Ceylon Plants 5 ( 15 ) is a list of 387 species known 
to have been introduced and more or less naturalized there, and the rapid 
spread of some of which, e. g. L ant ana, has been used by the supporters 
of Natural Selection as an argument for the greater adaptability of foreign 
species to local conditions. In a recent paper ( 20 , p. 197), I pointed out 
that this assumption ignores three important facts, (1) that foreign conditions 
have also been introduced, (2) that such weeds are also common in con¬ 
tinental areas, and (3) that they spread just as much at the expense of the 
wides already in the country as of the endemics. We may supplement 
this by a brief analysis of the list. Of the 387, 204 are cultivated only, 
47 are semi-wild cultivated plants found only near houses and gardens, 
and 125 are weeds of open ground, a feature which was almost unknown 
in Ceylon before the advent of man. This leaves only 11, of which two 
have only been once recorded, and have not been seen in recent times, and 
five exist only as clumps of two or three planted trees. There are thus 
only four left, of which Passijlora ednlis might almost be added to the 
semi-wilds, as it is only found fairly near to places where it is or has been 
cultivated. Bocconia cordata and Sapium sebiferuni have spread a few 
hundred yards down-stream from Hakgala Botanic Garden, and there 
remains only Aloe vera var. littoralis , which is common on the northern 
coast, and which is quite possibly a real native of Ceylon. The introduced 
plants of Ceylon are thus plants which are better adapted to the new 
conditions created by man, but that is all. 
Even in the cases recorded of rapid spread of introduced plants, those 
without special mechanisms have often been dispersed just as rapidly as 
those with such. None of the Ceylon introductions has spread more 
quickly than Tithonia diversifolia (Compositae), which has no pappus, and 
has spread largely by vegetative methods. Elodea in western Europe was 
a similar case, and there are several others. It would seem a priori 
probable that the possession of means for dispersal should improve the 
chances of rapid spread, but there are few facts to support this view. An 
examination of the ‘ adaptations ’ for dispersal shows at once that they are 
usually confined to very small groups, and are therefore not very old, 
regarded from the general evolutionary point of view. In Compositae, 
practically the whole family has the same mechanism, and as it is 
usually regarded as young, it may be that its wide distribution in the 
comparatively short time is due to its possession of such a mechanism. 
Dr. Sinnott, in bringing up objections to my hypothesis, to which he 
gives a much wider application than I have yet claimed for it, passes over 
without mention such extraordinary results as those given in the Tables IV, 
V, and VI of my New Zealand paper ( 19 ). It is perfectly obvious that 
such results must be explained, and it is almost equally obvious that no 
explanation other than that they are the result of a mechanical cause will 
