!3 
with reference to the Dying Out of Species . 
of which extends into the valley of Maskeliya 2,000 feet below. There 
seems to be something about a mountain top which causes it to be an 
especially favourable place for the commencement of new species. Nearly 
every isolated mountain in Ceylon has its own species. On the other hand, 
the conditions are the same to all intents on these summits, except if 
averages be taken over many years. And in the large forests, where the 
conditions must be equally uniform, one does not find the VR endemics 
coinciding in localities ; each has its own. 
The R and the RR species show distribution areas forming a kind of 
chain-mail pattern over the country. No two have the same area of distri- 
bution. If we were to find a number of species jointly occupying a certain 
area, and a number of others occupying a second, it might be possible to 
say that they were adapted to some special local conditions there existing. 
But how can such an explanation be brought forward for a chain-mail 
pattern ? What conditions can one find to distinguish four or five areas 
which overlap one another ? At the parts where they overlap, all species 
must be growing in the same conditions. What then prevents A from 
spreading into the area occupied by B ? 1 The mere demonstration of the 
actual distribution of the endemics of Ceylon is almost sufficient to show 
that no explanation involving response to local conditions can possibly 
serve to explain it, but that there must be at work some cause which is 
much more purely ‘ mechanical \ 
In the second place, the number and proportion of endemics are far 
greater in the wet south-western zone, i. e. in the broken hilly country of 
Ceylon, than in the flat and uniform dry country which surrounds it to 
north, east, and south-east. If all the endemics were VR, this would fit in 
admirably with Natural Selection, for each species could be looked upon 
as suited specially to its own mountain top or other locality. But 
unfortunately for this hypothesis, while there are 233 of these VR species, 
there are also 192 R, 136 RR, 139 RC, and so on in diminishing numbers 
upwards. Even the R species cover so much ground that they come into 
a great variety of conditions, and this becomes more marked with each 
step upwards in the scale. Many RC species, for instance, cover the bulk 
of the upper montane zone ; this is composed of hills, valleys, and plains, 
of regions of richer and of poorer flora, of granitic and gneissose soils, of 
higher and of lower rainfall, temperature, &c. How can a species become 
adapted to such a region in which no two places show the same 
conditions ? 
Even if one takes the most local VR species, one does not in reality 
encounter uniformity of conditions. In one year the climate is wet, in the 
next dry, and if a form settled, or arose by mutation, in a wet season, how 
1 Of course on my hypothesis there is nothing. Given time enough, A may spread into the 
whole area now occupied by B. 
