Seed and seedling predation by rodents along forest-field edges can in
fluence the dynamics of tree invasion into old fields. As predicted by
the Janzen-Connell escape hypothesis and the predator satiation hypot
hesis, the effects of rodents on tree propagules may be modified by va
riation in the distribution and abundance of tree propagules. We teste
d these hypotheses in old-fields using a series of U-shaped enclosures
in southeastern New York State. We maintained two densities (high ver
sus low) of meadow voles (Microtus pennsylvanicus), which are the prin
cipal seedling predators and dominant rodent competitors at our sites.
High density of voles significantly reduced density of white-footed m
ice (Peromyscus leucopus), which are the primary seed predators. Withi
n enclosures we manipulated the density and spacing of red maple (Acer
rubrum) seeds and black birch (Betula lenta) seedlings. We used three
densities of red maple seeds (1500, 750, and 150 seeds/2 m(2)) arrang
ed to mimic natural leptokurtic declines in seed density with distance
from a hypothetical seed source. The density of black birch seedlings
was also manipulated at three levels (50, 20, and 10 seedlings/2 m(2)
). Seed predation was uniformly high overall (approx. 99%) in three ad
jacent plots irrespective of seed density. However, predation was sign
ificantly lower in isolated, low-density plots. These results suggest
that mice were not satiated by the range of seed densities typical for
old fields and instead foraged in a distance-responsive manner consis
tent with predictions of the escape hypothesis. In contrast, seedlings
in high density plots experienced significantly lower predation than
those in medium- and low-density plots, suggesting that voles were sat
iated by plots of high seedling density. These patterns were dearest w
ithin enclosures having high vole density, highlighting the importance
of competitive interactions in modifying the foraging decisions of ro
dents in old fields.