438 Willis. — -The Distribution of Species in New Zealand . 
that area occupied goes on the average with age. I propose, in other words, 
to substitute age for natural selection as the chief agent in determining the 
area occupied by any given species, so long as no important barrier inter- 
feres. It must be clearly understood that in this work, as in dealing with 
figures of breeding under Mendel’s law, one must always deal with twenty 
or so at once. In individual cases results may vary greatly, owing to the 
operation of chance, of natural selection, of local adaptation, and other 
factors which appear to have little or no aggregate effect in the long run or 
on large numbers. 
The law which appears to me to govern the distribution of species may 
be thus tentatively expressed : ‘ thfe area occupied by any given species 
(taken in groups of twenty or so) at any given time in any given country in 
which there occur no well-marked barriers depends upon the age of that 
species in that country.’ 
Had it not been that observations on the rich flora of Rio de Janeiro 
bore out this hypothesis, and that I found that sample genera, taken 
at random in the floras of the Himalaya and the Malay States, conformed 
to the law which I have suggested, I should have hesitated to put it forward, 
though the evidence in its favour from the Ceylon flora (which there is 
no reason to consider as unique) was very strong. 
The adoption of my hypothesis will be found to render simple in very 
many cases the puzzles of geographical distribution, such for instance 
as that Ceylon contains 222 widely distributed species which are each con- 
fined to a very small area, usually not over eight or ten miles in diameter, 
while in New Zealand (with one-fifth the flora of wides) the number is only 
twenty-one, several of which (see below) are fairly evident recent intro- 
ductions or doubtful determinations. 
By showing- — as has been done for Ceylon, and will in this paper 
be done for New Zealand — that the distribution of plants follows simple 
arithmetical rules, grave doubt has been thrown upon the idea that natural 
selection is the main agent in determining it. 1 This position has been taken 
in flank, and new defences must be prepared. One of these is based upon 
the hackneyed argument that introduced species spread rapidly over islands 
at the expense of the native flora, and I have endeavoured elsewhere to show 
that this position is equally unsound. 
This law of ‘ age and area’, if accepted, has such important bearings 
upon the whole subject of geographical distribution as well as upon 
evolution itself, that it is highly desirable to bring up as early as possible 
confirmation of the results which have been obtained by a study of the 
Ceylon flora. A cursory investigation of many groups of plants — and 
1 i. e. so far as area occupied is concerned, which has always been supposed to be determined 
by natural .selection. It must be clearly understood that in this work we are dealing always with 
area occupied, not commonness ivithin the area. 
