Pathogen infection influences the relationship between spring and autumn phenology at the seedling and leaf level.

Seasonal life history events are often interdependent, but we know relatively little about how the relationship between different events is influenced by the abiotic and biotic environment. Such knowledge is important for predicting the immediate and evolutionary phenological response of populations to changing conditions. We manipulated germination timing and shade in a multi-factorial experiment to investigate the relationship between spring and autumn phenology in seedlings of the pedunculate oak, Quercus robur, and whether this relationship was mediated by natural colonization of leaves by specialist fungal pathogens (i.e., the oak powdery mildew complex). Each week delay in germination corresponded to about 2 days delay in autumn leaf senescence, and heavily shaded seedlings senesced 5-8 days later than seedlings in light shade or full sun. Within seedlings, leaves on primary-growth shoots senesced later than those on secondary-growth shoots in some treatments. Path analyses demonstrated that germination timing and shade affected autumn phenology both directly and indirectly via pathogen load, though the specific pattern differed among and within seedlings. Pathogen load increased with later germination and greater shade. Greater pathogen load was in turn associated with later senescence for seedlings, but with earlier senescence for individual leaves. Our findings show that relationships between seasonal events can be partly mediated by the biotic environment and suggest that these relationships may differ between the plant and leaf level. The influence of biotic interactions on phenological correlations across scales has implications for understanding phenotypic variation in phenology and for predicting how populations will respond to climatic perturbation.

Integrating top-down and bottom-up effects of local density across scales and a complex life cycle.

Effects of group size (local conspecific density) on individual performance can be substantial, yet it is unclear how these translate to larger-scale and longer-term outcomes. Effects of group size can be mediated by both top-down and bottom-up interactions, can change in type or direction across the life cycle, and can depend on the spatial scale at which group size is assessed. Only by determining how these different processes combine can we make predictions about how selection operates on group size or link hierarchical patterns of density dependence with population dynamics. We manipulated the density of a leaf beetle, Leptinotarsa juncta, at three nested spatial scales (patch, plant within a patch, and leaf within plant) to investigate how conspecific density affects predator-mediated survival and resource-mediated growth during different life stages and across multiple spatial scales. We then used data from field predation experiments to assess how L. juncta densities at hierarchical scales affect different aspects of predation. Finally, we incorporated predator- and resource-mediated effects of density in a model to explore how changes in group size due to density-dependent predation might affect mass at pupation for survivors. The effects of L. juncta density on predation risk differed among scales. Per capita predation risk of both eggs and late instars was lowest at high patch-scale densities, but increased with plant-scale density. The final mass of late instars declined with increasing plant-scale larval density, potentially because of truncated development of high-density larvae. Predation incidence (i.e., group attack rate) increased with larval density at all spatial scales. A high coefficient of variation (i.e., greater aggregation) of L. juncta density was associated with lower predation incidence at some scales. Our model suggested that predator- and resource-mediated effects of density interact: lower survival at high larval density is mitigated by high final mass of larvae in the resulting smaller groups. Our results emphasize the importance of spatial scale and demonstrate that effects of top-down and bottom-up interactions are not necessarily independent. To understand how group size influences fitness, predator- and resource-mediated effects of density should be measured in their demographic and spatial context, and not in isolation.

Time since disturbance affects colonization dynamics in a metapopulation.

Disturbances are widespread in nature and can have substantial population-level consequences. Most empirical studies on the effects of disturbance track population recovery within habitat patches, but have an incomplete representation of the recolonization process. In addition, recent metapopulation models represent post-disturbance colonization with a recovery state or time-lag for disturbed ("focal") patches, thus assuming that recolonization rates are uniform. However, the availability of colonists in neighbouring "source" patches can vary, especially in frequently disturbed landscapes such as fire-managed forests that have a mosaic of patches that differ in successional state and undergo frequent local extinctions. To determine how time since disturbance in both focal and neighbouring source patches might affect metapopulations, we studied the effects of time since fire (TSF) on abundances of a specialist palmetto beetle within and between fire management units in Apalachicola National Forest, Florida. We measured beetle abundances at three distances from the shared edge of paired units, with units ranging from 0 to 64 months since fire and the difference in time since burning for a focal-source pair ranging from 3 to 58 months. Soon after fire, beetle abundances within management units were highest near the unit edge, but this pattern changed with increasing TSF. Between paired units, the more recently disturbed ("focal") unit's beetle abundance was positively related to source unit abundance, but the shape of this relationship differed based on focal unit TSF and the units' difference in time since burning. Results suggest that both focal and source habitat history can influence recolonization of recently disturbed patches and that these effects may persist over years. Thus, when predicting metapopulation dynamics, variation in habitat characteristics should be considered not only for patches receiving colonists but for patches supplying colonists as well.