SUMMARY1. The effect of electrically induced 'exercise' on the respiratory oscillation of arterial pH was studied in chloralose-anaesthetized dogs with spinal cord transaction at T8/9 (dermatome level T6/7).2. Respiratory oscillations of arterial pH (presumed to be due to oscillations of arterial Pco2) were sensed with a fast-responding electrode in one carotid artery.Breath-by-breath estimates of the maximum rate of change of pH of the downstroke of the pH oscillation (dpH/dtjmax) were obtained by differentiating the pH signal.3. Consistent with the findings of the previous paper (Cross et al. 1982), the ventilatory response to exercise could not be explained on the basis of sensitivity to C02; the A T4/APaco, was significantly greater for 'exercise' than for CO2 inhalation.4. On average, the amplitude of the pH oscillations decreased during 'exercise'.The change in the phase relationship (0) between respiratory and pH cycles, although significant from the second breath onwards, was not thought to be responsible for the increased ventilation V,; the direction of the change was opposite to that previously found to increase VI. 14 and dpH/dtjmax were also linearly related during the on-transient, although the same relationship did not hold true throughout 'exercise'.6. The dpH/dtjmax was related to CO2 production (1rco,) lending support to the prediction that the slope of the downstroke of the pH oscillation is a function of TCo2.7. It was concluded that the dpH/dtimax (dpCO2/dttmax) is a potential humoral signal in 'exercise' and could account totally for the shortening of te. Since there was a late rise in VI (due to an increase in tidal volume VT) in the absence of a change in dpH/dtjmax, it was considered unlikely that the dpH/dtjmax was the only humoral signal present during 'exercise'.