Summary 1. From July 1994 to September 1995, at six censuses, the herbivore community associated with understorey (< 2.5 m height) and canopy (15–20 m) leaves of Quercus alba and Q. velutina was sampled in south‐eastern Missouri, U.S.A. 2. Across all censuses, herbivore densities were not significantly different between canopy and understorey for Q. alba and Q. velutina, except in August 1994 when herbivore densities were 60% higher in the canopy on Q. alba. Little significant spatial variation in herbivore densities or community composition was found during the study years. 3. The herbivore community was diverse, consisting of 138 species of leaf‐chewing insects. Species richness was significantly greater (by 5–20%) in the understorey than in the canopy for both tree species, and the relative abundance of the main families, different feeding guilds, and most common species differed significantly between understorey and canopy. 4. To determine the extent to which leaf quality explained the observed patterns, percentage nitrogen and protein binding capacity were measured in canopy and understorey leaves of Q. alba and Q. velutina. Per cent nitrogen was higher in canopy leaves for Q. velutina while protein binding capacity was higher in canopy leaves for Q. alba. 5. These results suggest that the herbivore community associated with these two species of Quercus comprises species that appear to respond individually to environmental and biological conditions encountered in the understorey and the canopy.

Specialization on ephemeral resources (e.g. new leaves) should produce large annual variation in herbivore population size when the timing of availability of those resources is unpredictable. Despite considerable evidence for impacts of synchrony with budburst on survival of larval Lepidoptera, previous studies of adult Geometridae and Noctuidae found no correlations between insect phenology and population variability.2.# We surveyed larval Lepidoptera feeding on Quercus alba and Q. velutina in Missouri from 1993 to 2003 and examined population variability, measured as the coefficient of variation of population density (CV), in a subset of abundant species. We compared CV values among species whose larvae feed only in spring, early summer, mid-summer, late summer, or all season. We predicted that univoltine species whose larvae eclose and complete development in spring during leaf expansion would have higher variability than species feeding later in the season, having multiple generations, or having longer development times.3.#As predicted and consistent with hypotheses, spring-feeding species had CV values 32% higher than species feeding in summer months. Coefficients of variation were also 34% higher in leaf-rolling and mining guilds compared with free-feeders, suggesting that mobile species may compensate for asynchrony with budburst by dispersing to higher quality plants or plant parts. Multivoltine species, however, did not differ from univoltine species in population variability.4.#Our results suggest that asynchrony with plant phenology and factors that might exacerbate it, such as climate change, will have the largest impacts on the dynamics of spring-feeding Lepidoptera, particularly species with limited mobility

No abstract