Our goal was to design a stimulation-contraction paradigm using an isolated in situ dog gastrocnemius muscle preparation that would provide an experimental model for brief intense intermittent (IC) exercise in humans. Second, acid-base and ion exchanges across the muscle were investigated using four 30-s bouts of isotonic tetanic contractions (2 s(-1), 100-ms train, 50 impulses/s) with 4 min of rest between bouts. During the bouts, peak power output (W) was 18.2 mW/g in the first bout; it declined by 4.4% by the fourth bout and by 12-16% in each bout. Compared with repetitive continuous contractions (CC) at maximal O-2 uptake (Vo(2)), W was greater and Vo(2) (similar to 3.5 mu mol.g(-1).min(-1)) and CO2 production (similar to 4.5 mu mol.g(-1).min(-1)) were less with IC. Venous-arterial (v-a) differences and lactate output peaked immediately after each bout and were not different from the values reported for CC at maximal Vo(2). Thus, with IC, Vo(2)/W as lower and the CO2 production/Vo(2) and lactate output/Vo(2) ratios were greater than those seen with CC. These differences suggest that this stimulation-contraction paradigm may be an appropriate model for brief intense exercise. The v-a [H+] difference was a direct result of the v-a Pco(2), difference. The venous strong ion difference was always greater than or equal to the arterial strong ion difference because the v-a [Cl-] difference was opposite and greater than the v-a lactate concentration difference, whereas the v-a [Na+] and [K+] differences were small. Quantitative estimates for the contribution of active muscle to arterial [H+] and lactate concentration during in vivo exercise from the in situ experiments suggest that active muscle does not account completely for the change in either.