Does variation in census area confound density comparisons?

1. Estimates of density are not always independent of the area over which populations are censused, Instead, lower densities tend to be recorded for species when they are censused over larger areas. This may have serious implications for both intra- and interspecific comparisons of density, with relevance to such issues as conservation prioritization and management strategies, because differences in density map simply result from differences in census areas. 2. Here, we use long-term population census data for British birds in the period 1968-91, drawn from the Common Birds Census (CBC), to examine relationships between density and census area within species, and the consequences of these relationships for intraspecific and interspecific density comparisons. 3. Most British bird species exhibit statistically significant negative relationships between density and census area. We used these relationships to standardize mean density estimates for all species to a common census area. These area-adjusted estimates were usually very similar to geometric mean density estimates calculated without reference to census area. 4. For a subset of species recorded from a large number (> 30) of census sites in each year in the period 1968-91, we used intraspecific density-area relationships for each year to standardize mean density estimates to a common census area in all years. Again, the area-adjusted estimates for each year were usually very similar to the simple geometric mean density estimates calculated for the species in the same year. 5. These results are encouraging, but are certainly a consequence of the relatively limited range of census areas used here, and the fact that the mean census area varies little across species, or across pears within species, Moreover, those species occupying few sites are the most likely to have area-biased densities, but are the species for which area-correction will be most difficult. 6. Overall, the results suggest that past analyses conducted using the CBC data are unlikely to have been seriously confounded by variation in census area. Nevertheless, they do highlight that the effects of census area will require consideration by anyone planning to measure or use densities for comparative purposes either within or among species.