Spotlight transect surveys with distance sampling were used to estimate spring (pre-breeding) and autumn (post-production) fox Vulpes vulpes densities in three contrasting rural areas of Britain during 1995–97. This was the first attempt in Britain to measure and compare fox densities over large geographical areas (630–1460 km2). Mean post-production fox abundance was estimated to be 0.90/km2, 2.62/km2, and 0.59/km2 in mid-Wales, the east Midlands and East Anglia, falling to pre-breeding levels of 0.41/km2, 1.17/km2, and 0.16//km2 in spring. As relative measures of regional density, these estimates are strongly supported by independent indices of fox abundance, and by the simultaneous survey of two sympatric species, the badger Meles meles and brown hare Lepus europaeus, which demonstrate the absence of any terrain-related bias. Absolute abundance is less easy to verify, but estimates of spring density based on breeding earth censuses support the transect surveys. For two of the three regions, fox density was close to levels predicted by extrapolation on the basis of landscape, but in the third region (East Anglia), fox density was substantially below prediction. Thus, results failed to support a hypothesis that fox abundance can be predicted solely on the basis of landscape and its close correlates. Rather, they favoured a competing hypothesis that an independent factor determines fox abundance in some regions. A likely factor is culling by man.