70 Mendclian hilierltance aiul Yelloiv Jiusf in Whral 



(C) The lioniozj'jious "iuiiiimic" cultures {'!.) wtue characterized by: 

 (a) Kemarkable resistance to attack under the most adverse external 



conditions. 



(6) Extreme lateness of infection where it occurred. 



(c) The comparatively mild nature of llie attack (and in some cases 

 its complete absence). 



(D) The cultur(!s in u liich segregation occurred (3.) took generally an 

 ijitermediate position as regards the period of infection and rate of rust 

 spread; though liiuilly a proportion of the plants were as severely rusted 

 as those in the pure susceptible cultures. 



The 5() cultures in which segregation was clear, and for which full 

 statistics are available, contained 3045 plants, of which 2385 were either 

 moderately or badly rusted, while 660 had only a slight attack or none. 

 The numbers 2284 : 761 would exactly represent the 3 : 1 Mendehan 

 ratio, and if it exists here it follows that 101 plants out of a probable 

 761 recessives were rusted beyond the slight extent indicated by grade 1 , 

 i.e. 13-2 per cent. But this degree of "disturbance in rust resistance" 

 is quite comparable with that which occurred in the 15 homozygous 

 "immune"' cultures given in Table 111 where, out of a total of 594 plants, 

 114, or ID per cent., were rusted beyond grade 1. Furtlier, since the F^ 

 results have shown that these 15 cultures wi^re the progeny of genetically 

 immune parents, there is sufficient evidence to show that such genetically 

 immune plants may under very adverse conditions be subject to a mild 

 attack. This being the case, it appears safe to conclude that in the 

 above-mentioned segregating cultures one-quarter of the plants were 

 genetically immune, and that these cultures were the product of F.^ 

 heterozygotes for rust resistance. 



5. Application of thk F^ results to the i^.^ .statlstics. 



Table VIII gives a summary of the results obtained from all the F^ 

 cultures. It will be seen that it is comparatively easy to pick out homo- 

 zygous susceptible and immune types in the F^ generation by merely 

 selecting the extreme cases of attack or non-attack. On the other hand, 

 one cannot be so sure about the constitution of the moderately rusted 

 individuals, for in this particular instance nearly one-quarter of these 

 proved to be pure susceptibles. IMost of the slightly rusted plants of 

 1918, however, turned out to be heterozygotes. Finally, a ])lant that 

 bears traces of attack in one season may be shown by its offspring to be 

 a geneticalh' iinnninc iii(li\idu:il, w liile a rust-free ])lant niny occasionally 

 be proved a heterozygote which has eseajietl infection. 



