195 
Willis.—Age and A rea. 
sequently one may infer with fair probability that Berkshire was the first 
original home of the name (2), for not even the most enthusiastic supporter 
of Natural Selection and adaptation would contend that a Goddard was 
especially adapted to Berkshire. The biologist fights shy of statistics, 
except when he wants to prove something in his own work, and at other 
times is apt to believe in the old gibe that statistics can be made to prove 
anything—a gibe which really refers to statistical fallacies, into which the 
untrained experimenter with statistics is very apt to fall. But in the case of 
the statistics upon which Age and Area is based, the fact that the work has 
been taken up con amove by a statistician of Mr. Udny Yule’s reputation 
should be sufficient to show that there is no fallacy involved, as one or two 
speakers at Hull seemed to imply. There are very many problems, and 
those often problems of wide range, which can only be solved by means of 
statistics, but from their very nature these statistics cannot be applied to 
individual cases. Statistics show that the average Scot is about 20 lb. 
heavier than the average Englishman, and it does not in the very slightest 
degree affect this result to say, ‘ Oh, but Smith weighs 13 st. and MacPherson 
only 10 \ And the bulk of the single-species arguments against Age and 
Area are almost exactly similar to, and equally as valid as this. 
It is probable enough that no two species , allied or not, travel at exactly 
the same speed in their diffusion over the surface of the earth. But each 
will, on the average of long periods, travel at more or less of an average 
rate, so that in twice the time it will cover twice the distance. A species 
A may travel at the rate of 50 miles per 10,000 years, say, and in 20,000 will 
cover a distance of 100 miles, yvhile B may only travel 5 miles in the 
10,000 years, and 10 in 20,000, requiring therefore 200,000 years to cover 
a distance traversed by A in 20,000. The area covered will be the same, 
yet B will be ten times as old as A. But both fall in with the Age and 
Area theory, and with equal exactness. 
If one were to take a group of closely allied species, e.g. species of 
similar habit belonging to the same genus, it would seem probable that as 
they will all have much the same mechanism for dispersal, and somewhat 
the same type of reactions to surrounding circumstances, they will spread at 
rates not very widely different. And taking at least ten will do away with 
the small differences that will occur between them, so that one group of ten 
will give much the same result as another group of ten allied to the first. 
One group, for example, might show rates of spread represented by 1, 3, 2, 
i, 4, 3, 2, 1, 2, 3, and another by 3, 2, 1, 4, 2, 4, 1, 3, 2, 1. On adding these 
up one finds the first to give an average of 22/10 = 2*2, and the second of 
23/10 = 2*3, a very small difference. If one deal with species always in 
groups of at least ten allied forms , and compare only with similar groups 
allied to the first , one will find that Age and Area is closely followed. 
That it is closely followed is shown by the extraordinary success of 
