SUMMARY Fungal pathogens can alter host fitness by affecting a plant's ability to survive, reproduce, compete, grow or defend itself against herbivores and other parasites. However, infections need not have uniformly negative effects, and the outcome of a fungus‐plant interaction can range from strongly parasitic, through commensalistic, to mutualistic. The outcome of the interaction is determined by characteristics of the fungus and plant, as well as ecological conditions. Damping off and root rot diseases severely reduce plant survivorship. These pathogens appear to be highly aggressive, a trait that may be evolutionarily constrained because of the necrotrophic infection habit. Incidence is locally high with patch size being variable. Canker and wilt pathogens reduce plant survivorship and growth, and tend to be highly aggressive. Two diseases within this group, chestnut blight and Dutch elm disease, have caused sustained highly destructive pandemics on their hosts. Neither pathogen appears to be evolving towards genetically based reduced aggressiveness despite the fact that disease incidence is near 100 % over a wide area. Less aggressive strains may not be selectively favoured because plants can be multiply infected, which would allow more aggressive strains to invade trees previously infected by less aggressive strains. Obligate foliar diseases can affect plant fitness by reducing survivorship, reproduction, growth or competitive ability of the host. However, overall effects are often slight and the relationship between plant and pathogen range from strongly parasitic to nearly commensalistic. Disease incidence is highly variable across space and time, and much of the variability appears to be influenced by climatic and environmental variables. Non‐systemic foliar diseases reduce plant reproduction, but some interactions may be nearly commensalistic because of low seed set in healthy individuals. Thus, disease effects are minimized in environments where reproduction by seed is not important. Systemically infecting pathogens can alter plant survivorship, reproduction, growth, competitive ability and susceptibility to herbivores. The effects can be positive as well as negative, and the net outcome of the interaction can range from strongly parasitic to unconditionally mutualistic. Trends within this group support Clay's New Function Hypothesis that pathogens can reduce their aggressiveness by acquiring new functions which increase plant fitness, and not by the amelioration of the original disease symptoms.

Isolates of Colletotrichum lindemuthianum (138 total) from Argentina, Brazil, the Dominican Republic, Honduras, Mexico, and the United States were characterized into 41 races based on virulence to 12 differential cultivars of Phaseolus vulgaris. These 41 races were categorized into two groups: those found over a wide geographic area and those restricted to a single country. Races 7, 65, and 73 were widespread. Race 73 was the most common (28%). Race 7 was found once in Argentina and Mexico but at a higher frequency in the United States. Race 65 was found repeatedly in Brazil and the United States. Although 39% of the races were detected repeatedly and three races were widespread, no race was isolated from both P. vulgaris gene pools. Phenetic analyses showed no obvious patterns correlated with virulence clusters. No geographic pattern was evident. Molecular polymorphism generated by random amplified polymorphic DNA confirmed the extensive variability in virulence of C. lindemuthianum. Virulence phenotypes were grouped into 15 clusters. The two largest clusters contained isolates from all the geographic regions sampled. Molecular polymorphism was observed among isolates from races 65 and 73 within and among countries, except among Bra-zilian isolates of race 65. The genetic diversity of C. lindemuthianum was greatest in Mexico and Honduras. Our data suggest that C. lindemuthianum may not be highly structured to specific Phaseolus gene pools.

No abstract

Spatial variation in the resistance structure of Linum marginale (wild flax) populations to the rust fungus Melampsora lini, and in the racial structure of this pathogen, was investigated by sampling 10 populations distributed throughout the Kosciusko National Park, New South Wales, Australia. Considerable differences were found among populations in the structure of both host and pathogen. Host populations were divided into three broad categories: (1) populations susceptible to all testing races; (2) populations containing a strictly limited number ofresistant phenotypes; and (3) populations with a considerable diversity ofresistant phenotypes. The pathogen populations also showed a range ofdiversity. The major differences between these populations were determined by the occurrence and frequency of four common races of pathogen (races A, E, K., and N). These differences were apparent both at a regional spatial scale (over the 100 km separation of the most distant populations) and at a local scale where major differences were detected between two populations only 300 m apart. The distribution of the four common races of M. lint was consistent with the hypothesis that a fitness cost was associated with unnecessary virulence. In general, however, differences in the structure of pathogen populations from genetically very similar host populations implied that, in addition to host resistance genes, other evolutionary forces are also important in determining the genetic structure of individual pathogen populations.

No abstract

Abstract. A model by Levin and Udovic (1977) emphasizes the need for integration of studies of the numerical abundances and genetic composition of host and pathogen species. We use their conceptual framework to summarize our collaborative research on the ecological genetics of the anther-smut disease of Silene alba caused by the fungus Ustilago violacea. Our theoretical investigations have revealed the importance of the rate and mode of disease transmission on the likelihood of coexistence between host and pathogen. Our empirical studies have quantified patterns of disease spread on local and regional spatial scales and have shown that host genotypes differ greatly in resistance. Comparable genetic variation in pathogen virulence has not yet been demonstrated. The genetic composition of host populations alters numerical dynamics in experimental populations: disease declines in resistant populations, while host and pathogen appear to coexist in susceptible populations. Ecological outcomes also may be affected by the negative relationship between disease incidence and host flowering time, which may constitute a "cost" to resistance. We are currently expanding our work on the ecology and genetics of metapopulation dynamics of host and pathogen.

Summary1 Matrix models were used to evaluate the effect of chestnut blight infection on transition probabilities and population growth rates for American chestnuts. Diseasefree, epidemic and recovering (i.e. pathogen infected with a double-stranded (ds) RNA hypovirus) populations were compared. 2 Population growth rates ( λ ) did not differ significantly over time or with disease status. However, predicted stable stage distributions differed between population types, with disease-free and recovering populations more similar to each other than either was to epidemic populations. 3 Survival had the highest proportional contribution to population growth rates as revealed by elasticity analyses. However, reductions in stasis of the largest trees contributed most to reductions in population growth rate when comparing diseased with disease-free populations using LTRE. 4 The presence of hypovirus reduces pathogen virulence, allowing individual American chestnut trees to increase in size. Where dsRNA has spread, chestnut populations in Michigan have attained population dynamics similar to those found in disease-free populations. 5 Matrix models and life table response experiments can be used to detect important pathogen-mediated changes in the dynamics of host populations.

No abstract