Ecology and host interactions of the angiosperm parasite <italic>Cuscuta gronovii</italic> (Convolvulaceae) Willd. in southeastern Michigan wetlands.

Abstract

Ecologists have long recognized that generalist parasites are capable of impacting the nature of their hosts' interactions with other species and thus have a potentially significant role in shaping community structure. I use a plant host-parasite system to study the impacts of a generalist parasite the structure of its hosts' communities. The parasitic vine, <italic>Cuscuta gronovii</italic> (Convolvulaceae), grows in wetlands where it often infects the annual <italic>Impatiens capensis</italic> (Balsaminaceae), and perennials <italic> Aster puniceus</italic> (Asteraceae) and several <italic>Solidago species </italic> (Asteraceae). Field observations and experiments were carried out to test the hypotheses that <italic>C. gronovii</italic> infectiveness and virulence varied with life-history stage of the parasite and among the species of host. Data show that, although <italic>C. grononvii</italic> infects a number of host species by the end of the growing season, all initial infections occurred on one species, I. capensis. Surveys also demonstrate that the virulence of <italic>C. gronovii</italic> varies among host species. Field grown <italic> Solidago gigantea</italic> plants show no significant deleterious effect of <italic>C. gronovii</italic> infection; however, <italic>Aster puniceus </italic> and <italic>I. capensis</italic> are impacted negatively by <italic> C. gronovii</italic> infection. At the plot level, increased intensity of infection was associated with reduced per capita perennial rhizome production, but was not significantly correlated with per capita fruit production of <italic> I. capensis</italic>. This suggests an indirect mutualism between <italic> I. capensis</italic> and <italic>C. gronovii</italic>, in which <italic>I. capensis</italic> enabled <italic>C. gronovii</italic> to infect perennials hosts that compete with <italic>I. capensis</italic> for space. Mathematical models based upon the natural history of this system were constructed to predict the conditions that result in indirect mutualism between <italic>C. gronovii </italic> and <italic>I. capensis</italic>. These models suggest that indirect mutualism is possible under a wide variety of conditions and depends primarily upon the species composition of the neighborhood in which the pair occurs. Lattice simulation models were used to predict the effect of selection on <italic> I. capensis</italic> susceptibility to <italic>C. gronovii</italic>. These models predict that evolution of either highly resistant, or susceptible populations of <italic>I. capensis</italic> populations are possible, depending on the dispersal distance of <italic>I. capensis</italic> and the virulence of <italic> C. gronovii</italic>. My research demonstrates that fitness outcomes of multi-species interactions can vary widely in space and time and that knowledge of the range of natural variability within which the interaction occurs and how this variability effects both the direct and indirect interactions between species will be necessary to predict where along a continuum from antagonistic to mutualistic lies the interaction between a pair of species.