Seeds of three species of annual grass weeds (Alopecurus myosuroides,
Avena fatua and Bromus sterilis) were placed in field margins around a
rable fields in a randomised block split-plot design experiment. The f
ield margin vegetation was either sown or naturally regenerated and ei
ther cut or uncut. The seeds were either placed in cages designed to e
xclude small mammals and birds or were uncaged. The proportion of seed
s removed was monitored on 10 occasions and mean seed loss was analyse
d. In general, a greater proportion of weed seeds was removed from unc
aged trays in uncut swards, suggesting predation by small mammals, whi
ch inhabit tall grass. This effect was mainly due to removal of seeds
of the two large-seeded species (A. fatua and B. sterilis), with A. fa
tua being especially favoured. It is therefore likely that small mamma
ls play a role in the population dynamics of major crop weeds by feedi
ng on their seeds in field margins, especially when these are dense an
d uncut.