ON THE STATISTICAL VIEW OF NATURE, 



G23 



phenomena, such as correlation, heredity, regression and 

 panmixia, and he has shown how to analyse tliese graphi- 

 cal tracings so as to indicate the several possible elements 

 out of which they are compounded, representing separate 

 agencies which are at work in nature. The mathemati- 

 cal inventions of Fourier had similarly enabled physicists 

 to analyse the complicated periodicity of tidal curves 

 into their elements, and, under the hands of Ohm and 

 Helmholtz, to resolve the harmonies of music. 



We have here arrived at the last stage of the devel- 

 opment of tlie statistical view of nature. It has been 

 variously judged l)y biologists according to the special 

 views they take of their problems, and also according 



ity of variation " (p. 568). Mr Bate- 

 son expects great assistance from 

 the statistical methods. '"There 

 is," lie says, "no common shell or 

 butterfly of whose variations some- 

 thing would not be learnt, were 

 some hundreds of the same species 

 collected from a few places and 

 statistically examined in respect of 

 some varying character. Any one 

 can take part in this class of work, 

 though few do" (p. 574). Not- 

 withstanding the general resem- 

 blance noted above between the 

 ideas of ^Ir Bateson and of Prof. 

 Pearson, they differ so much in 

 detail as to be led to confess 

 that they do not understand one 

 another's languages. Cf. W. Bate- 

 son, " Heredity, Differentiation, and 

 other Concei)tions of Biology," 

 'Roy. Soc. Proc.,' vol. Ixix. pp. 

 193-205; K. Pearson, "On the 

 Fundamental Conceptions of Biol- 

 ogy," ' Biometrika,' vol. i. pp. 320- 

 344. Prof. Pearson's view is that, 

 for the working out of the theoiy 

 of evolution, " biological conceptions 

 can be accurately dethied, and so 

 defined measured with quantita- 



tive exactness" {loc. cit., p. 324). 

 Mr Bateson, on the other hand, 

 regards them as to some extent 

 out of the reach of mathematical 

 definition and measurement. "Dis- 

 continuous variation" in Mr Bate- 

 son's special sense — by which we 

 niaj- perhaps understand great as 

 distinguished from small but num- 

 erous deviations from the average 

 — Prof. Pearson regards as "statis- 

 tically negligible for the purpose 

 of vital statistics" (pp. 33-3, 334). 

 He, in fact, holds closer to Dar- 

 winism as understood by Darwin, 

 who never looked with much 

 favour on Huxle\'"s view, for ex- 

 ample, that "sports," as distin- 

 guished from the sum of small 

 differences in individuals, might 

 furnish an ap[)reciable part of the 

 materials for natural selection. 

 Mr Bateson's view found favour 

 with Huxley, as may be observett 

 in the ' Life and Letters.' On the 

 novelty and value of Prof. Pearson's 

 methods, see ealso the Address by 

 Prof. Weldon to the Zoological 

 Section of tiie British A.s.sociation 

 in 1898. 



