23 



like proportion. It results from this, tliat tlie parents do not 

 precisely reproduce themselves in their offspring, but that one-third 

 of their influence evaporates, so to speak ; by whom is the missing 

 one-third exercised? Mr. Galton answers, by the grandparents, 

 and more remote ancestors. 



The question arises, since nature seems always to seek to 

 revert to the average, do we never improve? The answer is 

 probably, that while nature resists the perpetuating extreme 

 variations, she is always seeking to raise the average. I am not 

 aware that evidence exists upon which we can rely to prove this, 

 but one cannot help thinking that it ought to be true. 



In the course of his researches, Mr. Galton desired to investigate 

 the frequency of error resulting from the data supplied by his 

 observations, and he submitted a (question to a mathematical friend, 

 Mr. Dickson, of Peterhouse, for his solution. He says, " T never 

 felt such a glow of loyalty and respei;t towards tlie sovereignty and 

 magnificent sway of mathematical analysis as when his answer 

 reached me, confirming, by purely mathematical reasoning, my 

 various and laborious statistical conclusions with far more minute- 

 ness than I had dared to hope. His calculations corrected my 

 observed value of mid-parental regression from i to -^, that is. 

 it increased it from ^|ths to ^|ths, a difference too small to be 

 taken into consideration." In other respects, the differences between 

 the mathematically-calculated law and the actual results observed 

 were only 3 per cent., or two degrees in an angle : showing that 

 while on the one hand the mathematical doctrines apply to the 

 case in question, the observed statistics range over a sufficient 

 number of cases to make them trustworthy. 



The next important measurement to be referred to is that of 

 chest girth. It is of course essential that that should be taken in 

 a uniform manner, as you may make a difference of several inches 

 by the way you hold the tape. Compared with the other dimen- 

 sions, it appears that chest girth increases more slowly than weight : 

 the boy of 10 having 21bs. weight for each inch of chest girth, 

 while the man of all ages above 17 has 4^1bs. weight for eacli inch : 

 while with regard to height, the increase is slightly the other way, 

 though not so much as to disturb the general relation that the chest 



