1 8 Willis . — The Evolution of Species in Ceylon , 
many wides VR and some endemics VC ? If we put it all down to age we 
get a simple and perfectly reasonable explanation, without making any 
appeal to ignorance. 
All these facts, taken together, go to show that whether the endemics 
be regarded as increasing in number, or dying out, their growth (or decay) 
is almost purely mechanical, and cannot therefore be explained on any 
other than a mechanical hypothesis. For this hypothesis I have adopted 
that of age. It is, however, absurd to imagine that the endemic species 
are older than the others, and consequently, if they are dying out, we 
have to explain the remarkable fact that the youngest species are dying 
out, and that mechanically, as rapidly in one family or genus as another. 
The whole goes to show that evolution, like so many other things, 
eludes our grasp just when we think that we have at last found a method 
which will give us a close insight into it. 
But if my interpretation be accepted — and it is very hard to see how 
any other can be found — great changes must follow, and in the present 
and succeeding papers I shall indicate a few of these. It may be noted 
here that we practically get rid of the idea of variation in a species until 
it has attained considerable age ; there is no room for it in the small 
number of individuals with which the species begins. 
Evolution is thrown back a stage, and we come once more, after many 
wanderings, back to the definition of Linnaeus, ‘species tot numeramus 
quot diversae formae in principio sunt creatae ’, but substituting evolutae 
for creatae. The species, in fact, in the Linnean sense, comes first, and 
varies into the innumerable varieties which may ultimately exist, later. 
One may further test the validity of using the figures for rarity as 
they were used in the preceding paper, by checking various other problems 
where one can with reasonable probability predict the answer. Thus, for 
example, it is obvious that on the whole parasites and Saprophytes can 
only come later than the other plants of the Angiosperm flora. With the 
figures for the Ceylon flora, we find that the widely distributed species 
of these groups show a rarity of 3-8, against a mean of 3-0, indicating their 
greater youth. 
Climbers, again, must on the whole be younger than the rest of the 
flora. Testing this, we find 179 species of widely distributed climbers, 
with 583 marks, numbers large enough to give a fairly accurate result, 
which shows rarity of 3-2, or again younger than the rest. 
In the case of water plants there is no reason to imagine them 
necessarily any later in arriving than the land plants, and in fact they 
show, 
Table XII. 
Dicotyledons of wide distribution 34 with 87 marks 2*5 
Monocotyledons 26 98 3*7 
