• PDF

    Final published version, 607 KB, PDF-document


Timing of the production of susceptible host stages can have a large impact on a plant's probability of infection by pathogens and on the extent of damage caused by herbivores. In the closely related plant species Silene alba and S. dioica, flowers are the site of infection by the host-sterilizing anther smut fungus Ustilago violacea, as well as the site of oviposition by the noctuid Hadena bicruris. In diseased populations of the two host species, the noctuid can act as pollinator, fruit predator, and vector of fungal spores. Hence, host traits that enhance or reduce visitation rates by the noctuid could affect fruit set, fruit predation and infection probability. A field study was conducted to investigate the magnitude and direction of effects of the timing of anthesis on reproductive success via these pollinator-mediated processes. More than 50% of the Silene fruits produced in a sympatric population of S. alba and S. dioica were predated by the noctuid. Predation was two times higher in S. alba than in S. dioica, annihilating the two times higher flower and fruit production by the former species. This finding emphasizes the potential impact of biotic factors on the relative fecundity of related plant species in natural populations. Onset of flowering was 4-8 weeks earlier in S. dioica than in S. alba. Tn S. dioica, early flowering individuals produced the largest fraction of undamaged fruits per flower because of their high fruit set, and their low rates of fruit abortion and predation. By contrast, in S. alba, late flowering individuals produced the highest fraction of undamaged Fruits per flower, because of a monotonic decrease in Fruit predation with time and the absence of a significant impact of phenology on fruit set. In principle, pollinator-mediated selection could thus favour earlier flowering plants in S. dioica and later flowering plants in S. alba, and an evolutionary response to such selection would increase the average difference in onset of flowering between the two species. However, differences in fitness among phenology classes were more strongly determined by differences in flower production per plant than by differences in the average reproductive success per flower. A response to pollinator-mediated selection is therefore only expected if effects of age, size, or other factors that determine differences in flower production per plant are averaged out over progeny. In both host species, ca 20% of the hosts only produced sterile flowers as a result of systemic infection by the fungal pathogen U. violacea. In S. alba, the impact of phenology on fungal infection was in the same direction as for fruit predation (decreasing with later onset of flowering). In S. dioica, phenology had no effect on the probability of infection. In both species, associations between infection and predation tended to be positive but were not statistically significant, suggesting that the presence of alternative vectors or variation in resistance unrelated to traits affecting vector contact map prevent a tight association between damage caused by a vector-herbivore and its vectored disease. [KEYWORDS: Anther-smut infection; ustilago-violacea;spore deposition experimental populations; disease transmission; genotypic variation; life-history; patterns; caryophyllaceae; epidemiology]
Original languageEnglish
Pages (from-to)467-480
Issue number3
StatePublished - 1996

ID: 188526