Trophic interactions were investigated on herbaceous communities in a
limestone dale in Northern England. Manipulative experiments involved
the application of molluscicides and foliar and soil insecticides alon
g natural productivity gradients. The results supported the theories o
f Fretwell and Oksanen in which trophic dynamics are predicted to be d
ependent upon primary productivity. Furthermore, the results extend th
e Fretwell-Oksanen model by the inclusion of invertebrates, and the ap
plicability of the model to the small, individual habitat scale. At ve
ry low productivity, the vegetation was dominated by slow-growing, unp
alatable species and did not experience a detectable amount of herbivo
ry. In circumstances of high productivity, 'top-down' control of herbi
vores by carnivores appeared to protect the resident fast-growing and
relatively palatable perennials from herbivory. Vegetation of intermed
iate productivity responded strongly to the removal of herbivores; her
e we conclude that herbivore pressure is high because productivity is
insufficient to sustain a high intensity of 'top-down' control from ca
rnivores.