Assessing the lifetime consequences of plant-animal interactions for the perennial herb Lathyrus vernus (Fabaceae)

Abstract

Interactions with animals are potentially important in shaping the life histories of plants. However, we still largely lack the quantitative information that is necessary to compare plant-animal interactions and plant life histories. This is because most studies have focused on only a single interaction and a single component of fitness. This review suggests a protocol for how plant-animal interactions, as well as other, interactions can be quantitatively investigated. I review results from the forest herb Lathyrus vernus to examine how plant performance is influenced by mollusc damage of shoot buds, vertebrate grazing, leaf damage, pre-dispersal seed predation, flower herbivory and pollinator visitation. More specifically I ask: (1) What are the effects of these interactions on components of plant fitness? (2) How do such effects translate into total fitness? (3) Do trade-offs among components of plant fitness influence the net effect of animals? (4) Are antagonistic and mutualistic interactions influenced by the same plant traits? The results illustrate why it is often necessary to monitor plants over several seasons and over all life cycle phases to assess the total effects of environmental factors. In L. vernus many components of fitness were affected by multiple interactions with animals. A large proportion of these effects was carried over to subsequent seasons and could only be detected after one or two seasons. Assessing the effects on overall performance in perennial plants is particularly difficult because fitness components have very different influence on total fitness. The matrix model approach used to estimate the effect of herbivores, seed predators and pollinators on L. vernus provides one possible way to make such assessments for organisms with complex life cycles. The results show that the cryptic damage of emerging shoots has the largest negative effects on total fitness. Correlations between different components are an essential feature of many life histories and must be incorporated into demographic models if we want to assess the net effect of environmental factors. In L. vernus costs of reproduction were present and demographic modelling showed that the positive effects of an increased seed production after additional hand-pollination on total fitness was completely offset by poorer performance in the following year. Lastly, this study provides an example of how optimal trait values can be assessed when traits influence several fitness components and several interactions. In L. vernus a higher number of flowers was correlated to both an increased fruit set and an increased risk of grazing damage and these effects were, in turn, carried over to the next season. The results clearly illustrate the limitations of inferences about phenotypic selection pressures that are based on single fitness components or single interactions. Taken together the results for L. vernus show that the selective pressure constituted by a given amount of animal interference is highly dependent on the context, in terms of both plant life history and other interactions. A life cycle approach like the one outlined in this study may avoid several problems and provide the basis for assessments of the total effects of animals on plants. Such an approach must be based on in-depth studies of plants.