Heredity, Variation and Genius 57 



and think is worth more than a thousand 

 observations by persons most of whom are 

 incapable of looking accurately and have never 

 learnt to think — the more prone they commonly 

 to see what they wish to see and the more sure 

 of what they think they see — and none per- 

 haps adequately equipped with the requisite 

 precedent knowledge of the subject rigidly to 

 adopt and exactly pursue a right method of 

 enquiry. Multiplication of errors cannot add to 

 the value of the product. To one not competent 

 to rate mathematical values, entitled therefore 

 to speak only with much diffidence, it would 

 seem that the statistical data on which to base 

 profound calculations are too uncertain to 

 warrant sound inferences. Not that the calcula- 

 tions are therefore wrong, for given the arbitrary 

 data those may be accurate even when the data 

 are fallacious or of quite doubtful value. But 

 what is the conclusion worth when there is no 

 security that the data are sound ? 



Anyhow, whatever the value of statistical in- 

 formation, it does not help in the particular case 

 nor dispense with the necessity of pursuing exact 

 positive research into matters of individual in- 

 heritance. That which statistics seem to do is 

 to formulate a general statement intimating or 

 perhaps indicating the direction in which to 

 prosecute useful research. So far, though it is 

 not far, they do something. Take for example 

 I he conclusion formulated and lauded as Galton's 



