W. Hatkson and Cakounk 1*ki,ij<;\\ 23 



11 lam il it's wen- from pjiivnts reckoned as class 2. They gave 

 G34 of classes 1 and 2, 

 17 of class 3, 



13 of classes 4 .■iml ">, 



or 634 types to 30 aberrants, viz. 211 : 1. 



The reniaiiiin^' 3 famili(\s contained a distinctly his/her proportion 

 of aberrants, though these were still a small minority. They gave 



162 of classes 1 and 2, 

 16 of class 3, 

 of classes 4 an<l 5, 



or 162 types to 25 aberrants, viz. 6"5 : 1. 



It is likely that these three form a distinct genetic group, and the 

 parents of all three were classified as belonging to class 3, which 

 combine the foliar structures of the types with the curved pods of 

 the rogues. 



The details of these three families were as follows : 



The details of the other families cannot as yet be regarded as 

 important, nor can any significance be attached to the particular com- 

 position of the first 57 families at all events. Since the frequency of 

 the aberrant plants is so small it is evident that many of the families 

 includeil among the 77 which contained no aberrants may have been 

 capable of producing them. All that can be po.sitively declared is that 

 individuals, thoroughly typical, are capable of throwing the lower forms 

 sporadically in small numbers, and that the genetic behaviour of classes 

 1 and 2 is approximately the same. 



Families of Group B. 



There are 16 families of this kind, namely containing a minority of 

 types and a large majority of aberrants. The proportions seem 

 altogether irregular, and they range from 1 type : 64 aberrants 

 (No. 118) to 1 type :2 aberrants (No. 112). Of the Iti parents 15 

 were of class 3 and 1 was of class 2. 



