The effect of euglycaemic hyperinsulinaemia on the recovery of 13C in expired CO2 has been assessed in six normal subjects. Each was studied on three occasions: once with a 6 h primed constant infusion of NaH13CO3 combined with a euglycaemic hyperinsulinaemic clamp for the last 3 h (study 1), once with a 6 h primed constant infusion of NaH13CO3 alone (study 2) and once with a 6 h infusion of normal saline combined with a hyperinsulinaemic clamp for the last 3 h (study 3). Measurements of 13C enrichment of expired CO2 were made in the third and sixth hour of each infusion. 2. There was no significant increase in enrichment during study 3 (3 h 0.00047 +/- 0.00016 versus 6 h 0.00069 +/- 0.00028 atom per cent excess) with potato-starch-derived D-glucose used to maintain euglycaemia. 13C recovery increased in the sixth hour of both study 1 and 2 (study 1: 3 h 74.4 +/- 2.0 versus 6 h 85.5 +/- 2.6%, P < 0.01; study 2: 3 h 72.1 +/- 2.4 versus 6 h 81.7 +/- 1.4%, P < 0.01). There was no significant difference in recovery between studies 1 and 2. 3. These results suggest that increased recovery during a sequential euglycaemic clamp is predominantly time-dependent. Studies which use this technique to examine the effect of insulin on substrate oxidation should take this into account.