230 MENDELISM 



number of determinations of parental correlation 

 have, however, since been made in the case of all 

 kinds of characters. The values show considerable 

 variation, but the average which they indicate is much 

 nearer to O'5 than to 0-33. Pearson therefore con- 

 cluded that in none of these cases could anything 

 resembling Mendelian inheritance be taking place, 

 and that the latter is, in fact, the exception rather than 

 the rule. 



Mendelians, aware of the certainty of their own 

 results, and being convinced that these facts must 

 have a very wide application, were thereupon driven 

 reluctantly to the conclusion that something was 

 seriously wrong with the methods adopted by biome- 

 tricians for determining the coefficients of correlation. 

 It seems, however, that this conclusion may have been 

 arrived at with undue haste. 



In August, 1906, Mr. Yule read before the Inter- 

 national Congress of Hybridization assembled in 

 London a very interesting paper on ' The Theory of 

 Inheritance of Quantitative Compound Characters on 

 the basis of Mendel's Laws/ Though some difficulty 

 was then experienced in following his argument by 

 an audience unaccustomed to statistical methods, 

 Yule's conclusion is really very simple. 



Yule points out that the only character dealt with 

 in Pearson's memoir is the number of protogenic or 

 allogenic couplets present in the individual, and it is the 

 proportionate number of these couplets present in the 

 parent and in the offspring respectively which is taken 

 as determining the value of the correlation coefficient. 



