Predictions of global warming require predictions of how much CO₂ will be taken up by the oceans, how much by land plants, and how much will stay in the atmosphere. Measurements of atmospheric oxygen (O₂) help with these predictions if we also know the ratio of O₂ release to CO₂ uptake in land plants. We have measured this ratio in a midlatitude forest and find a lower value than the one in wide use. If truly applicable, our results call for a modest adjustment in the global carbon budget.