SummaryThe objective of this study was to determine the effects of 2 different gas collection systems on locomotor and ventilatory responses to exercise in a group of 8 Standardbred pacers. Each horse completed the same incremental treadmill exercise test (8, 9, 10 and 11 m/s on a 3° incline) using an open (flow‐through) system (OS) and a closed system (CS) incorporating one way valves. Only 2 horses completed the final (11 m/s) step using the CS. Peak values for V̇O2, O2pulse, plasma lactate and respiratory exchange ratio (RER) were significantly higher for the OS compared with the CS. There was no change in breathing frequency (Vf) or stride frequency (Sf) (mean ± s.e. 128 ± 2 strides/min) throughout either exercise test, however Vfwas significantly lower when horses exercised in the CS (68 ± 6 breaths/min) when compared to the OS (100 ± 3 breaths/min). There was a linear increase in stride length (SL) with speed (3.8 ± 0.1 m/stride at 8 m/s to 5.0 ± 0.2 m at 11 m/s) but no difference in SLbetween systems. There was also a linear increase in VTmeasured in the CS. Incremental exercise did not alter the Sf:Vfratio, however Sf:Vfwas significantly lower in the OS (1.3:1) when compared to the CS (2.1:1). There was a significant difference between systems for PaO2and PaCO2values at all increments of exercise. The lower Vfmeasured in the CS was associated with marked hypercapnia (59 ± 7 mmHg at 10 m/s) and hypoxaemia (45 ± 3 mmHg at 10 m/s). The higher Vfmeasured in the OS was associated with hypocapnia (35 ± 2 mmHg) at 8 m/s. It was concluded that although Vfvaried between OS and CS, Vfand Sfdid not change during incremental exercise with either system. Therefore the increase in speed was achieved solely by an increase in SL. Although the CS did not alter locomotor parameters or V̇O2at lower increments of exercise, the altered Vfappeared to limit