The aim of this study was to examine whether flux through the pathways of carbohydrate oxidation is accurately reflected in the pattern of (14)CO(2) release from positionally labelled [(14)C]substrates in conventional radiolabel feeding studies. Heterotrophic cell suspension cultures of Arabidopsis thaliana were used for this work. The presence of an alkaline trap to capture metabolically generated (14)CO(2) had no significant effect on the ratio of (14)CO(2) release from specifically labelled [(14)C]substrates, or on the metabolism of [U-(14)C]glucose by the cells. Although the amount of (14)CO(2) captured in a conventional time-course study was only about half of that released from a sample acidified at an equivalent time point, the ratios of (14)CO(2) released from different positionally labelled [(14)C]glucose and [1-(14)C]gluconate were the same in untreated and acidified samples. Less than 5% of radioactivity supplied to the growth medium as [(14)C]bicarbonate was incorporated into acid-stable compounds, and there was no evidence for appreciable reassimilation of (14)CO(2) generated intracellularly during oxidation of [1-(14)C]gluconate by the cells. It is concluded that the ratio of label captured from specifically labelled [(14)C]glucose is a valid and convenient measure of the relative rates of oxidation of the different positional carbon atoms within the supplied respiratory substrate. However, it is argued that failure to compensate for the incomplete absorption of (14)CO(2) by an alkaline trap may distort estimates of respiration that rely on an absolute measure of the amount of (14)CO(2) generated by metabolism.