The reliability of road transects as a census method for wintering and breeding Red Kites was studied in Spain. Road transect data were compared with censuses of wintering populations made by roost counts and of breeding populations made by a combination of nest searching and detection of territorial pairs. The variation in population density explained a high percentage of the variance of abundance indices provided by road transects during winter (> 90%) and breeding seasons (85%). The estimation of densities from distances to the transect did not provide more accurate results than unadjusted counts. On the contrary, strip counts may consistently bias population estimates in species that are attracted by roads, or may provide unpredictably unrealistic results when the number of observations is low. Winter roost counts made in areas of high breeding density apparently underestimated the real populations, because resident Red Kites usually do not use roosts, and this was detected through the results obtained in road transects. No clear effect was found of forest cover on the detectability of Red Kites, probably due to their selection of open areas for hunting. Road transects seem to be a reliable method to estimate the numbers of Red Kites, and probably of other species of similar habits and size.