Four triticale (xTriticosecale Wittmack) genotypes were grown under rainfed conditions with limited irrigation support in Lleida in northeast Spain. For each variety, samples consisting of 10 tillers with half-sterilized spikes were taken three times from anthesis to maturity. Carbon isotope ratios (deltaC-13) were then determined in water extracts from ear bracts (glumes, paleas, and lemmas), awns and flag leaves, and in powdered kernels. For the half-sterilized spikes, carbon isotope analysis was carried out separately in bracts and awns from fertile and nonfertile spikelets. The deltaC-13 in the water-soluble fraction of awns, glumes, and glumells from fruitless spikelets was significantly higher than that from fertile spikelets sampled at mid-grain filling. Differences in deltaC-13 among sterile and fertile spikelets were not significant in samples taken a few days after anthesis or at maturity. These results are in accordance with some degree of refixation by awns and ear bracts of the CO2 respired by grains during grain filling. There was progressively higher deltaC-13 from flag leaf blades to awns, glumes, and glumells. This variation in deltaC-13 along plant parts may be caused by differences in the ratio of assimilation rate to CO2-diffusive conductance. Values of deltaC-13 of mature kernels were between the values at anthesis and mid-grain filling for the water-soluble fraction of flag leaves and inner bracts and were fairly similar to those of glumes and awns.