IV] STATISTICAL STUDY 47 



there is as yet no certainty that this account of 

 continuous variation is sufficient to cover all cases ; 

 it is a suggestion of possibility rather than a state- 

 ment of fact. 



We have seen that there is reason to believe that 

 the Law of Ancestral Inheritance is true only when 

 applied to a large number of individuals considered 

 in mass, or, as it has been put, that it is a statistical 

 rather than a physiological law. In individual cases 

 it is not true that the offspring need be influenced 

 by ancestors beyond the parents, but in other cases, 

 as will be seen in dealing with Mendelian heredity, 

 these ancestors have important effects, so that 

 statistically it is possible to say what is the average 

 influence of the ancestors of any generation upon 

 the offspring. Now in cases where it is possible to 

 define rigidly single characters, much more is learned 

 from the physiological than from the statistical 

 method, but where no such rigid separation of 

 characters is possible the statistical law is the only 

 one that can be applied. This is particularly the 

 case in characters which vary continuously, or where 

 the categories into which the character falls overlap 

 one another, as for example in Johannsen's beans. 

 Further, the statistical method is frequently the only 

 one which is available when experiment is impossible 

 and when our knowledge of the facts is based solely 

 on numerical data from observed cases, and this of 



