W. JJateson and Caholink 1'km,k\v 29 



figs. () .111(1 7). In sdiuo of the tiiinilios the course of di'velopiiKMit w;is not 

 noted, but of 41 families which were more carefully watched most of 

 the members of 34 families are recorded as beginning like types and 

 three families only (10 plants) as beginning in the rogue-like condition. 



It might be supposed that the more luxuriant growth of the F^ 

 seedlings was, as it so often is, an incidental result of crossing ; but this 

 suggestion is negatived by the experience of 10 crosses made between 

 rogues, giving (i4 plants which were characteristic rogues from the first, 

 without any symptoms of extra luxuriance. 



Besides the 50 families raised from rogue x type or reciprocal, which 

 behaved in the way described, there were two exceptional cases'. 



\st Exception. In this (f|) an E. G. rogue fertilized by a true type 

 gave 2 seeds of which one grew to be adult. This t\ plant was recorded 

 as having leaves like type, but curved pods. Judging from the offspring 

 it produced, this plant was probably of the kind which we came 

 subsequently to recognize as class 2, namely type-plants with some 

 degree of pod-curvature, behaving genetically as types. In F^ it gave 

 a family of 38 plants. They were not very well grown, but their 

 characters were in all respects those of types, several having pods 

 slightly curved, but none in any way approaching rogues. Two of 

 these gave moreover offspring in F^, 21 and 57 plants respectively, all 

 well grown plants in all respects typical, occasional pods being slightly 

 curved, as may happen in any family of types. 



There is in this case complete certainty that the original mother 

 was a rogue, for the normal offspring of that plant have been grown for 

 three generations. All are thorough rogues, and the strain has been 

 used repeatedly for crossing, giving the results which we have come to 

 look on as normal. 



Seed of the original father was unfortunately not kept. 



' In JH F\ from E.G. rogue x E.G. type was a rogue as usual, but among 19 F-, plants 

 derived from it were 4 steriles. The flowers did not open properly and the pollen was 

 deficient in amount. One artificial pollination failed. Probably this sterility was 

 a recessive condition, but whether it affected one sex or both was not clearly made out. 

 The seed of the parents used in the original cross was not preserved and it cannot be said 

 if sterility existed as a recessive on either side of the parentage, llecessive sterility is of 

 course not uncommon among plants, especially in connexion with the male side. As 

 already stated, we had on one occasion sterile rogues in D.A. It is interesting to notice 

 also that from Andrew Knight's description of his experiments in crossing peas he clearly 

 had a strain affected with male sterility, which phenomenon, in accordance with the 

 scientific ideas of those days, he attributes to prolonged cultivation in one locality (Vhil. 

 Trans. 1799, p. 196). 



