INRA, Thiverval Grignon, France,

Background and objectives
Host mixtures are considered to be an important disease control method, particularly for air-borne pathogens of cereals. A question has been raised, however, about the consequences of their use on a large scale on the evolution of pathogen populations. Different models have been proposed to describe the conditions in which complex races, able to develop on several components of the mixture, are selected and erode the resistance of mixtures. Most of these models are based on the existence of a selective disadvantage, or cost, associated with virulences that are unnecessary in a host-pathogen interaction. However, other mechanisms, such as differential adaptation of the pathogen to the host genetic background could also be involved in simple race-complex race competition. Differential adaptation has been described by Chin and Wolfe [1] for powdery mildew on barley. It can be described as a selection for increased reproductive ability on a given host genotype among isolates of the same virulence type. Because simple races always reproduce on the same host genotype but complex races can infect different host genotypes successively during an epidemic, differential adaptation could represent a selective disadvantage for complex races.

Results and discussion
Our objective was to evaluate the effect of differential adaptation in host mixtures, as a selective factor in favour of simple pathotypes. A field experiment was designed to measure differential adaptation in a powdery mildew population and it showed that the mean aggressiveness of the pathogen population increased in pure stands but not in mixtures of two cultivars. In this experiment, initial infection occurred naturally and initial populations were not differentiated for their relative aggressiveness to the two cultivars. After seven generations, isolates reproducing in pure stand had a better infection efficiency on their host of origin than on the other one: relative infection efficiency on the host of origin had increased, on average, by about 20%. For complex isolates reproducing in cultivar mixture, infection efficiency remained, on average, constant during the epidemic. In host mixtures, differential adaptation could then result in an increase in the reproduction rate for simple races but not for complex races. This hypothesis has been tested with a simulation model [2] and it has been shown, using Chin & Wolfe's data [1], that differential adaptation could have a greater effect than a cost of virulence of 10% in limiting the complex race development. An important point here is to note that the potential effect of differential selection in host mixtures depends on the genetic diversity within the pathogen population as well as within the host population. In multilines, in particular, such selection should not occur since all components have the same genetic background. Also, if the pathogen races develop as clonal populations, differential selection should not take place in host mixtures. When the different mechanisms involved in the race competition are considered together, it appears that the rate of complexification of the pathogen population is slow enough to allow a long-term use of host mixtures associated with a rotation of the mixed cultivars [2].

1. Chin KM, Wolfe MS, 1984. Plant Pathology 33, 535-546.
2. Lannou C, Mundt CC, 1997. Theoretical and Applied Genetics 94, 991-999.