Scale-dependence of pollinator responses to floral resource density

Background/Question/Methods

Many studies have found decreased pollinator visitation rates with reductions in plant or flower density, raising concern that reduced reproductive success could exacerbate population declines in threatened plant species.  A few studies have also found effects of floral resource density on the species composition of visitors, which could have additional, significant impacts on plant reproductive success because of variation amongst flower visiting taxa in their quality as pollinators.  However, the scales that are most relevant for conservation applications are much larger than the local scales at which most of these data have been collected.  Little is known about how pollinators respond to flower or plant density on large scales, nor about the degree to which large-scale responses are similar to local-scale responses.  Here I present data from an observational field study in which I measured the effects of floral resource density on per-flower visitation rates and species composition of visitors to Holocarpha virgata, at both local (4m²) and large (12.5 ha) spatial scales.  At the large scale, I analyzed effects of variation in density both across space and over the course of the flowering season.

Results/Conclusions

Total per-flower visitation rate was positively correlated with large-scale floral resource density as it changed through the season, but did not respond to resource density on a local scale.  The species composition of visitors was influenced by resource densities at both local and large scales. The most abundant flower visitors were the anthophorid bee Melissodes lupina, honeybees, and two species of halictid bees. The proportion of visits that were by M. lupina decreased with local flowerhead density but was positively correlated with large-scale floral resource density as it changed during the season. The proportion of visits by honeybees, in contrast, increased with local density and did not respond significantly to density on a large scale. Visitation by halictid bees did not respond to floral resource density on either scale.  These data show that the effects of flower density are scale-dependent, and therefore that the large body of data collected at small scales cannot safely be applied to larger-scale contexts.