In this study nine elite athletes each participated in three different 24- h trials, as follows: (1) complete bed rest (REST), (2) one bout of exercise at 1515 hours (ONE-EX), (3) two exercise bouts, one at 1100 hours and one at 1515 hours (TWO-EX-3 h), and (4) two exercise bouts, one at 0800 hours and one at 1515 hours (TWO-EX-6 h). Exercise was performed on a cycle ergometer with 10 min of warm-up and then 65 min at an exercise intensity of 75% of maximum oxygen uptake (VO2max). The polymorphonuclear neutrophil (PMN) counts increased consistently in response to exercise, and more in trial TWO-EX-3 h than in the two other exercise trials (P < 0.01). The respiratory. burst of PMN was measured as chemiluminescence (CL), obtained with phorbol myristate (PMA) and serum-opsonised zymosan (SOZ) as stimulators. Exercise triggered the CL response for a defined number of PMN, significantly above baseline (REST) values (P < 0.05) for ONE-EX and TWO-EX-3 h, but not for TWO-EX-6 h. The strongest response was observed for TWO-EX-3 h, but the difference between exercise procedures was not significant. However, as a novel approach, a comparison was made using total oxidative potentials per litre of blood, as obtained by combining CL values and PMN numbers. TWO-EX-3 h yielded significantly higher values than the other experimental treatments. Thus, by this measure the total oxidative potential of PMN(.)1(-1) blood remains at a higher level with short intervals between exercise bouts (i.e. 3 h instead of 6 h), possibly due to a combined effect of cell number increase and the priming state of PMN. This may suggest that for intensive training twice a day, a recovery phase of 5-6 h is preferable. The elevation in cell number is best explained by a combined effect of catecholamines and cortisol. Growth hormone is one probable candidate as a stimulator of CL, but other molecular participants that respond to exercise may exert roles as either stimulators or inhibitors of CL.