1. Often, closely related insect species feed on different host plant species, and the tremendous diversity of phytophagous insects is therefore attributed to host plant‐driven speciation. However, for most taxa, host use information comes from field observations of egg‐laying females or feeding caterpillars, which means that the underlying reason for a particular host‐affiliation is not easily determined.2. Therefore, it is often unclear whether an insect feeds on a certain host because it prefers that plant to alternative hosts, or because the host distribution overlaps with the habitat requirements of the insect.3. We ask to what extent a divergent host use in the field mirrors the host plant preferences of two closely related butterflies, Pieris napi and Pieris rapae (Pieridae). In nature, P. napi typically occurs in moister habitats than P. rapae.4. We scanned several microhabitats at a field site in Southern Sweden during multiple years, and collected Pieris eggs from three different plants, Cardamine pratensis (wet meadows), Barbarea vulgaris (drier micro‐habitats) and Alliaria petiolata (intermediate areas).5. As predicted, P. rapae eggs were more common than P. napi eggs on B. vulgaris, whereas all of the 358 individuals collected from C. pratensis were P. napi, indicating a divergence in host use between the Pieris species. However, under controlled laboratory conditions, both species had virtually identical oviposition preferences, laying eggs on all three plants, notably P. rapae also laying eggs on C. pratensis, indicating that habitat use, not plant preference, drives host plant use in nature.