116 NATURAL SCIENCE [August 



yet a quite unproven principle." The " contention " of the paper 

 is that there is no proof forthcoming of man's greater variability. 

 Whether either sex is the more variable is left for the future to 

 settle in the following words : — 



" When more material is available, and finer methods are 

 applied, then perhaps it will be possible to detect a more note- 

 worthy preponderance of variability in the one or other sex." 

 And, again, in referring to the slight preponderance of variability 

 observed in woman : — " I strongly suspect that this preponderating 

 variability of women is mainly due to a relatively less severe 

 struggle for existence." These are not the words of one whose 

 " object is to support the contention that women are, on the whole, 

 more variable than men." They seem to me the words of one who 

 wishes to reach a scientific conclusion without any party or sex 

 bias. 



In the next place, Professor Weldon objects to my use of the 

 co-efficient of variation. He apparently wishes to assert that 

 absolute variation is the real test of most things. I am some- 

 what surprised to see him advocating this test. It is not so 

 many months since an American critic pointed out how fatally 

 this measurement of variation affected the conclusions of a certain 

 paper of Professor Weldon's on selective mortality in crabs. I 

 have not seen any answer to that criticism, and I very much doubt 

 if one can be found. Some years ago I pointed out to him that 

 the same measurement of variability led to absurd results in the 

 case of the selective mortality of men. 



But even here Professor Weldon puts in my mouth opinions I 

 have never expressed. He writes : — " The violent assertion that 

 there is only one ' scientific ' measure of variability is therefore to 

 be regretted." Now, so far from asserting the validity of only one 

 measure of variability, I carefully state in paragraph (c) of my 

 conclusions : — 



" There is more than one method of quantitatively measuring 

 variability, but the measure which is really significant for pro- 

 gressive evolution has not hitherto been determined." 



On p. 343 I write, " We may stay to ask whether the statistics 

 of skull capacity do not in themselves give us any information with 

 regard to the superiority of either standard deviation or the co- 

 efficient of variation as a test of that variability which is valuable 

 for progressive evolution," and on p. 345 I conclude that the 

 results do not enable us to say offhand that absolute or percentage 

 variation is a better measure of the variability which is a source 

 of progressive evolution. These are hardly the words of one who 

 lias made a " violent assertion that there is only one measure of 

 variability." 



