Elevated atmospheric CO₂ can affect plant-insect herbivore interactions, primarily due to changes in leaf nutritional quality. However, current experimental evidence is predominantly based on greenhouse studies, and there is a paucity of data gained in natural plant communities growing under natural climatic conditions. We investigated the development of larvae of the satyrid butterfly Cocnonympha pamphilus in seminatural calcareous grasslands that had been exposed to elevated CO₂ for five growing seasons in the northwestern Swiss Jura mountains, Late second-instar larvae were grown until adult eclosion in cages containing a natural mixture of Grassland plant species. Leaf nitrogen concentration of the grass species present decreased at elevated CO₂ while nonstructural carbohydrates increased. Total polyphenolics were unaltered, but condensed tannins increased marginally significantly. Lipid concentration in freshly emerged butterflies of C. pamphilus increased, and females tended to have higher numbers of mature eggs in their ovaries at elevated CO₂ than under ambient conditions. Development time, and larval and adult mass of C. pamphilus were not affected by CO₂ enrichment. These findings contrast with previous greenhouse trials in which development time of C. pamphilus increased and larval masses decreased under elevated CO₂ despite similar changes in plant quality. A power analysis indicated that effects on herbivores were not missed due to a lack of statistical Power, but that effects were indeed absent in the field study. The absence of such effects could be explained by several mechanisms. In the field, insect growth was relieved from the N-limitation observed in the greenhouse because (1) animals were able to maintain optimal N:carbohydrate ratios in their diet by altering the proportions of different plant species they ingested, and (2) climatic conditions were harsher in the field. This may have limited animal Growth to such an extent that even the reduced N concentrations in plant tissue grown under elevated CO₂ did not limit growth any more. We therefore conclude that extrapolation of results on effects of elevated CO₂ on plant-insect interactions from,greenhouse experiments to natural situations should be interpreted with caution.