The diet of tawny owls Strix aluco was determined from pellets and prey items in owl nests in Kielder Forest, a planted spruce forest in northern England. Field voles Microtus agrestis were their most important food, and formed the highest proportion of tawny owl diet in winter and early spring. Common shrews Sorex araneus, common frogs Rana temporaria and birds were taken more frequently in late spring and summer. Clear cuts, areas from which timber had been felled at the end of the rotation, provided the main ®eld vole habitat in the forest and remained suitable for voles for 10±15 years after re-planting. Field vole abundance was measured three times a year on numerous clear cuts throughout the study area using a vole sign index based on fresh grass clippings in runways. Tawny owls responded functionally to the 3 to 4-year cycles of ®eld vole abundance. In years when voles were scarce, adult owls took more common shrews and common frogs, as determined from pellet analysis. In contrast, more bird prey was fed to nestlings when ®eld voles were scarce, as determined from prey items in nests. The proportions of the main prey in nests changed over a 19-year period. More bank voles Clethrionomys glareolus and wood mice Apodemus sylvaticus occurred in every year after 1992 than in any year before this. Numbers of wood mice in owl nests increased signi®cantly throughout the study period, whereas bank vole numbers exhibited noncyclic, multi-annual¯uctuations that were unrelated to ®eld vole cycles. It is argued that¯uctuations in rodent prey re¯ected changes in rodent guilds in the study area; reasons for this are discussed. This is the ®rst study of tawny owl diet in spruce forests in Britain and highlights the value of such large-scale dynamic habitats for rodent populations and their predators.