Field vole abundance was assessed on a range of upland habitats in spring and autumn, using live—trapping, snap—trapping and vole sign index methods. Three estimates of vole numbers were derived from live—trapping, five from snap—trapping and one from vole signs. Estimates were compared using a correlation matrix in order to establish the most accurate and consistent method. Snap—trapping for five nights provided enough information to calculate vole density, though the method was time consuming. Vole abundance indices derived from two nights' snaptrapping were significantly correlated with indices from five nights' trapping. Live—trapping and vole sign index were discounted as methods for this study area, as both proved unreliable.