562 HANS H. HATTEMER 



variation has to be the result of the differential frequency of certain 

 rare preadapted genes, since the introduction of the pathogen meant a 

 sudden and fundamental change of the host environment. This variation 

 may or may not be related to provenance, depending on the adaptive 

 significance of correlated gene effects. 



The percentages of healthy trees are estimates of the probability 

 that seedlings are not infected (or show no disease symptoms) under given 

 conditions. But, strictly speaking, no information exists on whether 

 these given conditions represent the environments where the select material 

 will be planted in the future. 



Should testing be put on a much broader base since differential 

 response of host genotypes has to be envisaged as does variability in 

 response to the spectrum of pathogen genotypes? I will not talk about 

 horizontal vs. vertical resistance (van der Plank, 1968) at this point 

 since I am unable to see the idea of this terminology. Evidence exists 

 in forest trees (SchUtt, 1964; Heybroek, 1969) that both general and 

 specific resistance may occur in the varieties entering one and the same 

 experiment. In fact, a certain individual may carry genes showing 

 different types of variation. 



THE PROBLEM OF PERSISTENCE 



Many investigations have shown that resistance longevity depends on 

 the type of genetic variation. Polygenically inherited resistance 

 generally fails to provide immunity; consequently, the pathogen is allowed 

 to complete its life cycle. But a certain level of the resistance attained 

 lasts longer since a newly emerging pathogen genotype is less likely to 

 overcome the host resistance. The risk in many instances has turned out 

 to be greater if resistance was based only on a few major genes. 





If the genetic basis of host resistance is broad enough, general 

 resistance may prevail and one need not breed for resistance against 

 specific fungal genotypes. To ascertain this condition, one has to 

 utilize a large gene pool. It is also understood that one need not 

 bother about loss of part of a resistant stock that we mass-grow for 

 timber production since the spacing of a long-lived crop at harvest will 

 be wider than at the time of establishment (Heimburger, 1962). 



The process leading to the breakdown of resistant varieties is 

 usually described in terms of the emergence of new virulent or adapted 

 pathogen genotypes. In other words, failure is attributed to an unpre- 

 dictable event. If we keep human behavior and facts about resistance 

 in mind, we must suppose that many such "breakdowns" can be attributed 

 to inadequate testing. For instance, nobody would talk about a breakdown 

 of spring-frost resistance if a spruce variety that was never exposed to 

 late frost failed the first time it was planted in a frost pocket. 

 Unfortunately, the overall fraction of infected or killed individuals in 

 a resistance test does not indicate the value of the testing method since 

 this fraction depends on the genetic make-up of the host material, the 

 environment, and the genotype of the inoculum. Different host varieties 

 and individuals may become infected in different tests. 



Selection for resistance solely in the field sometimes is insufficient 

 because of the existence of more chance variation of escapes. The trees 

 growing in the field could gain resistance because of their age and the 



