Stable isotopes provide a powerful means of elucidating the trophic ecology of organisms. Analyses of variation in the ratio of nitrogen isotopes (δ15N) can provide insights into the trophic position of species with broad diets and the ability to occupy multiple positions in food webs, such as ants. The most powerful studies compare subjects across various spatial scales, but to do so, local variation in δ15N baselines must be taken into account. To date, a wide variety of baseline calibration methods have been employed, leading some authors to suggest that a standard approach is needed, and that the reality of environmental variation necessitates that this should be at fine scales. In this study, we examine the fine-scale variation in δ15N value of colonies of the ant Formica kozlovi Dlussky (Hymenoptera: Formicidae: Formicini) along a sloped transect in Mongolia, and compare these with values for associated soils in an effort to shed further light on this issue. We find variation in ant δ15N to the order of one trophic level (ca. 3‰), over a distance of only 1 km. Ant δ15N was highly correlated with soil δ15N, and variation in mineral soil δ15N explained ca. 81% of the variation in ant δ15N. This study underlines the importance of local-scale baseline corrections for isotopic studies, particularly in environments where baseline variation might be expected. It further suggests that δ15N of mineral soils may provide a suitable baseline for ecological studies of terrestrial arthropods.