Background/Question/Methods Understanding plant physiological constraints to heat stress may help to predict the responses of plants to climate change. Using growth chamber experiments and a common garden experiment we asked two fundamental questions about within-species variation in constraints to heat stress in tall goldenrod, Solidago atlissima: (1), Is it possible to predict carbon gain at different levels of biological organization (from the cell to the whole plant)? under warming scenarios and (2) Does variation in cell-level carbon gain scale whole ecosystem carbon gain across populations? We collected Solidago genotypes from northern and southern latitude populations and exposed individual genotypes to a gradient in temperatures ranging from 14 °C to 42 °C. We measured cell-level production of reactive oxygen species, and total antioxidant capacity), leaf-level (carbon gain), and plant-level responses (net ecosystem carbon exchange). Results/Conclusions We found intraspecific variation in temperature response at the cell level where both reactive oxygen species and total antioxidant capacity were 35% and 30% greater in southern than northern genotypes. In addition, leaf-level carbon gain in southern genotypes was 25% lower than in northern genotypes at higher temperatures. This is due, in part, to reduced maximal Rubisco activity, a photosynthetic enzyme that catalyzes CO2. These cell – and leaf-level patterns do not scale to the field -- perhaps because of increased total leaf area in the southern genotypes. These data indicate that intraspecific variation in physiological constraints to heat stress at lower-levels of biological organization are not predictive at the ecosystem-scale. They also suggest that when investigating plant responses to climate change one needs to measure across scales to better predict future constraints.