The juniper Juniperus communis, a dominant plant in the high mountains of S
E Spain, produces a high proportion of empty seeds within well-developed co
nes. We tested the hypothesis that the production of empty seeds by jumper
reduces seed predation by the woodmouse Apodemus sylvaticus, thereby benefi
tting the plant. We performed laboratory and field experiments to determine
i) woodmouse discrimination ability between filled and empty seeds, and ii
woodmouse response to changes in the proportion of empty versus filled see
ds and in the seed density in seed clusters. In addition, we estimated, for
six juniper populations over three years, whether plants or populations sh
owing a higher proportion of empty seeds suffered reduced woodmouse predati
on. Experiments showed that woodmice can eventually discriminate externally
between filled and empty seeds, but in most cases had to bite the seeds to
identify and reject empty ones. The probability of predation for filled se
eds was independent of changes in the proportion of empty versus filled see
ds and in seed density per cluster. Seed predation suffered by plants in th
e field was unrelated to the proportion of empty seeds per plant for all po
pulations and years. The presence of empty seeds did not benefit juniper ag
ainst woodmouse predation, either in terms of individual reproductive outpu
t or in terms of offspring escape probability. Our study suggests that the
effect of empty seeds on seed predation should be rare in plant-seed predat
or interactions where predators are polyphagous and eventually able to disc
riminate against empty seeds, and therefore suffer a low cost when coping w
ith empty seeds.