219 



has gone to great pains in destroying his own theory. 



[In what follows tlie pages quoted refer to Prof. Pear- 

 sons paper in parts I and II of tire présent volume of 

 this Journal]. 



On page 210 (and sevoral other places) the équation : 



1 dy — je 



(2) y dx ~ „^2 y (x\ 



is stated by Prof. Pearson to represent the generalised 

 „probability curve for an. infinité numberof cause groups". ') 

 On page 211 again he asserts that ail discussion of asym- 

 metrical frequency must turn on this équation. Only, in 

 accordance with p. 178, it is hère written, with a slightly 

 différent notation : 



(3) 1 dy _ — X 



y dx Oo* F{x) 



According to Prof. Pearson this équation leads at once 

 to his (Prof. Pearson 's) generalised probability curves 



by expanding fi — ) in a séries of ascending porwers of 



l~\ (p. 210, 211) „A very few terms of the expansion, 



„however, suffice for describing practical frequency distri- 

 „bution" (p. 211). According to p. 204 and 212 Prof. Pear- 

 son 's curves stop at three terms, in fact he puts (p. 204 

 and 212) 



(4) F {x) = a^ + «1 ic -f- «2 x^. 

 Now this équation (2) or (3), which thus is stated to be 



1) The express condition oi' very numerous causes of déviation 

 has been adliered to throughout in my //skew curves". Considérations 

 based on a supposed very restricted number of causes can be easily 

 shown to be illusory in nearly every case oi' asymmetric frequency. 



