The magnitude of fractionation during photorespiration and the effect on net photosynthetic (13)C discrimination (Delta) were investigated for three Senecio species, S. squalidus, S. cineraria, and S. greyii. We determined the contributions of different processes during photosynthesis to Delta by comparing observations (Delta(obs)) with discrimination predicted from gas-exchange measurements (Delta(pred)). Photorespiration rates were manipulated by altering the O(2) partial pressure (pO(2)) in the air surrounding the leaves. Contributions from (13)C-depleted photorespiratory CO(2) were largest at high pO(2). The parameters for photorespiratory fractionation (f), net fractionation during carboxylation by Rubisco and phosphoenolpyruvate carboxylase (b), and mesophyll conductance (g(i)) were determined simultaneously for all measurements. Instead of using Delta(obs) data to obtain g(i) and f successively, which requires that b is known, we treated b, f, and g(i) as unknowns. We propose this as an alternative approach to analyze measurements under field conditions when b and g(i) are not known or cannot be determined in separate experiments. Good agreement between modeled and observed Delta was achieved with f = 11.6 per thousand +/- 1.5 per thousand, b = 26.0 per thousand +/- 0.3 per thousand, and g(i) of 0.27 +/- 0.01, 0.25 +/- 0.01, and 0.22 +/- 0.01 mol m(-2) s(-1) for S. squalidus, S. cineraria, and S. greyii, respectively. We estimate that photorespiratory fractionation decreases Delta by about 1.2 per thousand on average under field conditions. In addition, diurnal changes in Delta are likely to reflect variations in photorespiration even at the canopy level. Our results emphasize that the effects of photorespiration must be taken into account when partitioning net CO(2) exchange of ecosystems into gross fluxes of photosynthesis and respiration.