Willis.—Age and Area. 203 
New Zealand, contains five species, the second (20) three, while the whole 
100 miles from 101 to 200, or any of the higher hundreds, never contains 
more than four. The areas are obviously concentrated towards the bottom. 
If one made the unit one mile instead of ten, it is clear that there would 
still be five species in the first ten classes, while in the one hundred classes 
from 601 to 700 there would only be one species. I think that one may 
say that the curve is not modal, but hollow, but to get a curve at all one 
must work with units of appreciable size. If one take all the genera in 
the New Zealand flora that have over 20 species, and add up their 
endemics, one gets the result that while the endemic species with a range 
of 100 miles or less are no less than 105, those with a range from 101 to 
200 miles or any of the other hundreds never exceed 38. Splitting up 
the 105, one finds that 43 have a range of ten miles or less ; more, that is, 
than in any hundred above the first. Of these 43 I feel confident that 
more will be near to the unit of one than to ten miles of range, but much 
more detailed study is yet needful before a definite decision can be given 
on this point. The flora of New Zealand is too small fora proper decision, 
and one requires that of a continent. 
The importance of the question lies in the fact that if the curve be 
shown to have a mode at, say, five miles, with diminishing numbers below 
that (when large numbers are taken), it will mean that that is the probable 
size of area upon which species commenced, and that those with smaller 
areas are perhaps dying out. Even so, however, the area is too small to 
have allowed of Natural Selection of infinitesimal variations. Judging from 
the great numbers of species that are confined to very small areas indeed, 
I should imagine, however, that a smaller unit than five miles of range will 
be found to show the maximum. 
Mr. Tate Regan did not think that the hollow curves had any particular 
significance. Such a statement can only be attributed to an insufficient 
consideration of the matter. Can it for a moment be supposed that all the 
first fifteen families of flowering plants should give such a series of curves 
for sizes of genera as those in Fig. 1, without a very definite reason? 
In no single instance does the curve for one family even approach closely 
the curve for another, though the points of origin are only five squares 
apart upon the tracing-paper. Why should every family, animals and 
plants alike, show such a large proportion of monotypes ? Why should 
they all turn the corner between the threes and the fives ? These curves 
would not show such an exact similarity, and, for animals and plants alike, 
always turning the corner between the figure for three species and the 
figure for five, unless there were some overmastering reason for it. 
Not only are these curves hollow curves, but they are all hollow curves 
of the same mathematical type , and when plotted logarithmically they give 
straight lines or close approximations thereto, which is a thing that does 
