K. Pearson 
321 
be shown to have no basis whatever beyond that of confused and undefined 
notions. It will seem almost incredible to those readers of Biometrika who have 
been working for years statistically that some of these notions can still be accepted 
and propounded. They will say that variation, correlation, and heredity are 
concepts of which they have quite clear and quantitatively definite ideas ; yet 
they will be startled to find how little the great body of English biologists have yet 
studied, or at any rate digested the biometric work of the last eight years. But 
the fact has to be recognised ; biometricians have not only to collect material, 
analyse it, and see its bearing on vital phenomena, but they have still to convince 
the great body of biological workers that their methods are the only logical 
methods for solving, not necessarily every problem, but certainly many problems 
in the evolution of life. 
It is therefore with considerable sense of the gravity of the contest that I take 
up the gauntlet thrown down by Mr Bateson, but it seems necessary to do so for 
the sake of our infant science. I should have been content for the present to 
continue my own work, leaving the old school of biologists rigidly alone. It is 
Mr Bateson who has forced the controversy by a brilliant but logomachic attack. 
He does not attempt to meet biometric conclusions by new measurements, he 
appeals to the significance of words, and to what he holds to be fundamental 
biological conceptions. A reply to Mr Bateson must therefore in the first place 
be an analysis of terms, and only in the second place a personal defence. The 
discussion accordingly tends to become dialectical, rather than ontological ; we 
have to discuss the definition and use of words, rather than put observation 
against observation, fact against fact. Partly on this account, — because the 
controversy may be long and disputatious, and so, even were it free*, hardly 
fitting to the proceedings of a learned society, — partly because it is of fundamental 
interest to all biometricians, I have changed the venue to this journal. 
In the paper of Mr Bateson's to which I have referred there is a very free use 
of the terms Variation, Discontinuity, Differentiation, etc., but he does not provide 
a definition of any one of these terms. He must therefore either be using them 
(i) in the sense of the memoir which he is criticising, or (ii) in the sense accepted 
by biological writers, or (iii) in some sense of his own which he has elsewhere 
defined. Now I will at once put aside (ii) for I can find no common denominator 
in the use of these terms by biological writers. If it exists at all, I must presume 
that Mr Bateson has not neglected it, when he formed his own conceptions on 
these points. Mr Bateson is therefore either using his terms in my sense, which 
I believe is in the main in accordance with current biometric practice, or he is 
using them in some other sense, somewhere or other defined by himself. 
(i) Is Mr Bateson using these terms in their current biometric sense? 
We might possibly expect such use from him when he is criticising a biometric 
memoir. But unfortunately Mr Bateson and I speak in totally different tongues. 
* I am officially informed that I have a right to a rejoinder, but only to such a one as will not confer 
on my opponent a right to a further reply ! 
Biometrika i 33 
