Forests within agricultural landscapes can act as safe harbourages for species that conflict with neighbouring landowners' interests, including mammalian predators. The agency responsible for the management of forests in upland Wales, UK, has permitted the killing of foxes (Vulpes vulpes) on their land as a "good neighbour policy" with the aim of reducing fox numbers. The principal method used was the use of dogs to drive foxes to a line of waiting shooters; a small number of foxes were also killed by shooting at night with a rifle. However, it has been postulated that recent restrictions on the use of dogs to kill foxes in Britain could lead to an increase in fox numbers in plantations. The aim of this study was to determine whether over-winter culling (i.e. driving foxes to guns using dogs and rifle shooting) in these forests acted to reduce fox density. Fox faecal density counts were conducted in commercial forests in Wales in autumn 2003 and spring 2004. Data were analysed from 29 sites (21 individual forests and 8 forest blocks, the latter consisting of pooled data from 20 individual forests). The over-winter change in faecal density was negatively related to the proportion of felled land and the proportion of land more than 400 m altitude, these associations probably reflecting reduced food availability. Over-winter change was positively associated with culling pressure (i.e. more foxes were killed where more foxes were present, or vice versa), but this was not significant. The number of foxes killed was large relative to the estimated resident population, but losses appeared to be negated (most likely) by immigration. Pre-breeding (spring) faecal density counts were significantly positively related to culling pressure, i.e. more foxes were killed with increased fox density or vice versa. Overall, there was no evidence to suggest that culling reduced fox numbers. Consequently, restrictions on the use of dogs to control foxes are unlikely to result in an increase in fox numbers in commercial forests.