The ghost of granivory past.

We report a 36-month echo of vole (Microtus pennsylvanicus) seed-selection in the composition of synthetic forb communities in Wisconsin. An initial study showed direct suppression of an exceptionally large-seeded (seed > 15 mg) species (Silphium integrifolium, Asteraceae) by seed-size predation by rodents during the winter, resulting in indirect release of small-seeded (seed ≤ 1.5 mg) species. Twelve months after planting, plant diversities (Simpson’s D) were 33% higher in plots exposed to winter seed selection by rodents. Thirty-six months after planting, Silphium integrifolium had almost equalized densities in rodent access (20.8 ± 4.1 m–2) and exclusion (29.0 ± 5.2 m–2) plots, but still suppressed both large-seeded (≥ 3.5 mg) and small-seeded (≤ 1.5 mg) species (F1,16=11.84 and F1,16=10.42, P ≤ 0.005, respectively). A multivariate analysis of covariance (MANCOVA) distinguished effects of Silphium integrifolium (Wilk’s lambda P=0.029) from echoes of earlier winter granivory (Wilk’s lambda P=0.014). Thirty months after rodent exclusion, diversity (D) remained 27% higher in plots once opened to winter granivory (adjusted mean 4.70 ± 0.37 SE) as compared with closed plots (3.70 ± 0.26; F1,16=5.12, P < 0.05). Echoes of earlier granivory remained after rodent-induced imbalances in the abundances of the competitive driver of this system, Silphium integrifolium, all but disappeared. [ABSTRACT FROM AUTHOR]