Goshawk's diet and breeding success was studied in northern Finland in the vicinity of Oulu during years 1989–2004 in order to evaluate predation impact on four grouse species, the Black Grouse Tetrao tetrix, Capercaillie Tetrao urogallus, Hazel Grouse Bonasa bonasia and Willow Grouse Lagopus lagopus. Sporadic food data was also used from years 1965–1988. Number of studied territories raised from 12 to 37 during the study years. Food remains were collected from territories at least three times per year: in spring (around the nesting site, n = 1420), in summer (from the nest after chicks fledged, n = 1782) and late summer (around the nesting site, n = 826). Winter diet was assessed by telemetry and from museum samples (n = 88). Diet composed mainly of grouse species totalling highest in spring, around 50%, and lowest in winter, around 30%, by number. Black Grouse were the most numerous among grouse, but juvenile grouse outnumbered them during late summer. Preference of different grouse species in Goshawk's diet was measured by a simple catch/supply index. Willow Grouse was taken twice more among grouse than their abundance in the field suggested, while Black Grouse and Hazel Grouse were taken at the same ratio as their abundance in the field. Capercaillies (only females) were taken around half compared to their relative abundance. Goshawk's functional response (grouse found/nesting site as a response variable) against grouse density of the previous autumn was typically concave. Occupancy rate and productivity (chicks fledged/occupied territory) of the Goshawk territories declined as the grouse density declined but brood size remained at the same level. Combining functional and numerical responses for total response (kill rate) declined as well with
grouse density implying that Goshawk's predation impact on grouse has remained in a stable level. During years 1989–1998 year to year variation of total response tended to lag grouse density by two years, which implies destabilising effect of the Goshawk on grouse population. After 1999 this pattern, however, disappeared when grouse density fell to a very low level. Applied for the whole period, correlation with two year lag was observed but it was far from significant. Predation impact calculated for years 1989-1998 was 31% for the Willow Grouse, 15% for the Black Grouse, 2% for the Capercaillie and 16% for the Hazel Grouse.