THEORETICAL BASIS OF RUST RESISTANCE TESTING 301 



probably true for similar systems involving a space component (Pimentel, 

 Xagel, and Madden, 1963). 



The findings of some forest tree breeders of resistance reactions 

 apparently inherited in a simple Mendelian way (S0egaard, 1966, for 

 example) do not contradict our expectations about the genetic background 

 of resistance of tree species against fungal diseases in equilibrium 

 systems. But they might indicate the existence of different types of 

 host -parasite relations or of resistance barriers in one species which 

 might be lacking in a different but closely related species. We should 

 distinguish here between fatal diseases and others that rarely or never 

 are fatal to the infected host; we should further consider the various 

 ecological situations of host-parasite systems. 



Genes for resistance from a related host species might be particularly 

 useful for a breeder who is dealing with evolving host -parasite systems 

 like in the case of a newly introduced parasite. The genetical side of 

 host-parasite systems has been described by Pimentel (1968) as a genetic 

 feedback system. This is probably the most general model for dealing with 

 mutual effects of genetic changes in both populations. But genetic feed- 

 back means no more than a general description of the genetic interplay of 

 host and parasite. It seems, therefore, to be necessary to specify some 

 situations where this genetic feedback can work. The gene-for-gene system 

 represents only one of them. The general model of genetic feedback 

 described also all the other systems operating between hosts and parasites 

 having genetic equilibria, and where permanent survival of both populations 

 is possible as a consequence. 



We have to restrict ourselves here to discussion of 2 specific cases. 

 In one of them, resistance behaves like a typical quantitative character 

 in the sense of quantitative genetics, at least within a certain range. 

 In the other, resistance behaves like a threshold character with a more 

 or less narrow physiological or biochemical range, acting as a kind of 

 threshold above which resistance (or below which susceptibility) is 

 absolute for all genotypes. Selection might lead in both cases either to 

 absolutely resistant populations, to better resistance, or to less resis- 

 tance, the last two reactions being the results of changing the genetic 

 equilibrium points. 



MAIN .ASSUMPTIONS TO BE MADE WHEN THE CONCEPT OF HEPJTABILITY AND 

 GENETIC GAIN ARE TO BE APPLIED IN RESISTANCE BREEDING 



The outcome of mass selection or of related procedures over a few 

 generations of selection can often be predicted by using the well known 

 equation for "genetic gain" or "response to selection", E * h%S. P. stands 

 for response to selection and 5 for the difference between the mean of the 

 unselected population and the population of selected "superior" individuals 

 h e is the ratio of additive-genetic variance and total phenotypic variance 

 or its equivalent in an experiment. This ratio at the same time is a 

 measure of regression of progenies on parents which makes it an estimate 

 of response to selection for a selection experiment with a given S. 



But R can be estimated accurately in this way only if: 



1. There is no contraselection on the side of the parasite, 



