Agriculture Reference
In-Depth Information
major factors with any confidence. This is partly because of the numbers of inter-
actions involved and partly because of their dynamic nature. For these reasons, it
may be more useful in practice to analyse the gross effects of different levels of
diversification in relation to a range of major biotic and abiotic stresses, rather than
to concentrate on a particular form of diversification and a single effect.
10.6 RESPONSES OF PEST AND PATHOGEN POPULATIONS TO
DIVERSIFICATION STRATEGIES
For crop diseases, host selection is the most important of the evolutionary forces
affecting the structure of pathogen populations so that mixtures may be expected to
have a strong within-crop influence on the evolution of pathogen populations
relative to pure stands.
A pure cultivar selects strongly for a pathogen race able to overcome it (and thus
for pathogen uniformity within the crop). However, there is relatively weak
selection across different monocultures for races able to overcome more than one
resistance (Hovmøller
et al.,
1993). Thus, in a crop monoculture system, simple
races of the pathogen, able to attack cultivars carrying only one resistance (or none
at all), are able to survive in large numbers (Hovmøller
et al.,
1993; Wolfe and
McDermott, 1994; Caffier
et al.
, 1996). Under these conditions, gametic
disequilibrium (i.e. non-random association of gene combinations) may be common
in the pathogen population even if sexual recombination occurs (Wolfe and Knott,
1982; Wolfe and McDermott, 1994). Within a cultivar mixture, diversity may also
decline but more slowly because of selection for complex (able to attack more than
one component) or super-races (able to attack all components), caused by intimacy
of the host components. If the mixture is common, then gametic disequilibrium
generally may decline. It is clearly important to determine whether this form of
selection is important and, if so, how it might be avoided or reduced (see section
10.7).
10.6.1 Theoretical analyses and models
The development of models dealing with the effects of selection by host populations
on pathogen populations has helped to reveal some of the important factors. Simple
models have been based on many assumptions among which is a fitness cost for the
pathogen if it possesses unnecessary virulence. The reason for a strong adherence to
this assumption (and a corresponding cost of resistance) was mainly that such a cost
was needed to prevent selection of complex or super-races in mixture models (e.g.
Leonard, 1977; Leonard and Czochor, 1980; Lannou and Mundt, 1995). However,
such fitness differences are not observed in practice and, in asexual or partly asexual
organisms, they would be small relative to selection for associated characters in
common genotypes (Brown, 1995).
Nevertheless, even in the absence of a cost of virulence, a number of factors tend
to favour polymorphism and to delay or prevent unidirectional selection for complex
races. For example, simple gene-for-gene relationships as assumed in most models
