1. Reliable predictions of population spread rates are essential to forecast biological invasions. Recent studies have shown that populations spreading through favourable habitat can rapidly evolve higher dispersal and reproductive rates at the expansion front, which accelerates spread velocity. However, spreading populations are likely to eventually encounter stressful conditions in the expanded range. How evolution during spread in favourable environments affects subsequent population growth in harsher environments is currently unknown. 2. We examined evolutionary change in performance under drought, interspecific competition and heat stress for populations of Arabidopsis thaliana that experienced six generations of spread through replicated experimental landscapes of favourable habitat. To quantify how population performance under stress differed between leading edge and founding populations, we combined individual tests of genotype performance under stress with knowledge of the genotype frequency changes that occurred over the replicate invasions. 3. After spreading through favourable environments, the average silique production of individuals exposed to drought or interspecific competition was lower in leading edge than founding populations. This change was driven by the evolution of lower intrinsic silique production, which was correlated with increased seed size, a trait that evolved as populations spread. The ability of plants to tolerate drought or interspecific competition, however, did not change markedly during spread. Heat tolerance did increase in leading edge populations, and this trait was associated with the evolution of taller plants during spread through favourable habitat. 4. Synthesis. We conclude that evolution during spread in favourable environments may affect the ability of populations to grow under stressful conditions as experienced in the expanded range, through changes in either intrinsic fecundity or stress tolerance. Thus, evolution during spread may constrain or extend the eventual range limit of non-native species invasions.