EVOLUTION 373 



Criticism demands now quantitative exactness in these 

 matters, and will henceforth not be content until it has 

 obtained it. Nor is this warning without meaning, even 

 in the case of more elaborate biological investigations. 

 For example, Weismann has propounded a theory of the 

 development of the mortal from the immortal by natural 

 selection. Now such a theory demands at least two, if 

 not more, preliminary investigations ; first, a statistical 

 inquiry into the actual duration of life in some simple 

 organism which reproduces by asexual division ; and 

 secondly, an inquiry as to whether duration of life is an in- 

 herited character. Both these inquiries were quite feasible, 

 but neither of these did Weismann attempt before pub- 

 lishing his theory. This is not an isolated instance,^ it is 

 only an illustration of what, according to this Grammar, is 

 a wrong scientific method, but one, alas ! too often applied. 

 It is imagination solving the universe, propounding a 

 formula before the facts which the formula is to describe 

 have been collected and classified. It does not fulfil 

 Faraday's notion of the scientific investigator crushing in 

 silence by his own criticism the many suggestive thoughts 

 which pass through his mind (p. 32). Every few months 

 we find in one journal or another some more or less 

 brilliant hypothesis as to a novel factor of evolution ; but 

 how few are the instances in which this factor is accurately 

 defined, or being defined, a quantitative measure of its 

 efficiency is obtained ! If the reader will only apply to 

 such hypotheses the tests of scientific method indicated by 

 my Summary on p. '">,'], he will at least keep the main 

 features of the Darwinian theory of evolution clear from 

 many of the overgrowths of recent years. What we need 

 in the theory of evolution is quantitative measurement 

 following upon precise definition of our fundamental con- 

 in the present author's " Socialism and Natural Selection," in The Chances 

 of Death, vol. i. 1897. 



1 Thus again variation has been attributed to sexual reproduction — a very 

 plausible hypothesis. But does variation not occur with parthenogenetic 

 reproduction, or even in the case where a single individual puts forth a number 

 of undifierentiated like organs ? It does. Hence had the quantitative test been 

 made, the hvpothesis would have been "crushed in silence." 



