34 Genetical and Cytological Studies on Pisum 



Crossing-over between B and B (A-a) and between G and G (Gp-gp) 

 would occur, and if crossing-over in the two regions is independent, the 

 result would be the same as if the factors A and Gp were in different 

 chromosomes. It is, however, conceivable that in the chromosome 



B (a) (Gp) G 



the frequency of simultaneous crossing-over might be reduced. The efiect 

 of such "interference" might be seen in the families raised from self- 

 fertilised seed, but if the effect is slight it could perhaps only be detected 

 by back-crossing with the original double-recessive variety; the smallest 

 classes would then be, among the fertile plants, A Gp, and among the 

 half -sterile plants, a gp. 



Although in Hammarlund's experiments the regularity of the 

 "coupHng" ratios is remarkable, a point open to criticism is the irre- 

 gularity of the ratios for the same two factors in the "control" crosses 

 in which there is no sterility. In two groups of back-crosses the irre- 

 gularity is in the same direction for the two pairs of factors, e.g. coloured, 

 82 : white 111, green pod 81 : yellow pod 112 (Table IV, p. 214, 1929). 

 It is unfortunate that in these back-crosses the heterozygote was always 

 used as the male parent, and consequently divergences may be the result 

 of differential pollen growth. It would be satisfactory to know the female 

 gametic ratio, for although there is no sign of linkage, it is conceivable 

 that some other relation may be found between the two factors. 



Of Hammarlund's and Hakansson's second case of ring formation, no 

 genetical account has yet been published. 



Other factors, especially those for colour-white (Aa), yellow-green 

 cotyledons (Yy) and stumpy pointed pods (Btbt), have been investigated 

 both in the fertile and the partially sterile Thibetan hybrids. For all these 

 factors the zygotic ratios in the F^ and F^ self-fertilised progenies are ex- 

 tremely irregular, and their inter-relations appear to be very complicated. 

 Some of the dif&culties met with are doubtless a consequence of differential 

 pollen growth, for it has been observed that green seeds may occur more 

 often at the distal than at the proximal end of the pod. At the proximal 

 end on the other hand there is sometimes an excess of round seeds. 

 Another factor of importance is differential viability; in the upper pods 

 the end seeds often fail to set, and the proximal end suffers most loss. 

 It is evident that in order to imderstand the variable ratios obtained 

 by self-fertilisation, it is necessary to examine the properties of the 

 ovules and pollen separately by back-crossing. From a few back-crosses 



