To evaluate the difference of ventilatory and gas exchange response differences between arm and leg exercise, six healthy young men underwent ramp exercise testing at a rate of 15 W·min⁻¹ on a cycle ergometer separately under either spontaneous (SPNT) or fixed (FIX) breathing modes, respectively. Controlled breathing was defined as a breathing frequency ( f _b; 30 breaths·min⁻¹) which was neither equal to, nor a multiple of, cranking frequency (50 rev·min⁻¹) to prevent coupling of locomotion and respiratory movement, i.e., so-called locomotor-respiratory coupling (LRC). Breath-by-breath oxygen uptake (V O ₂), ventilation (V E ), CO₂ output (V CO ₂), tidal volume (V T ), f _b and end-tidal P CO ₂ (P ET CO₂) were determined using a computerized metabolic cart. Arm exercise engendered a higher level of V O ₂ at each work rate than leg exercise under both FIX and SPNT conditions. However, FIX did not notably affect the V O ₂ response during either arm or leg exercise at each work rate compared to SPNT. During SPNT a significantly higher f _b and lower P ET CO₂ during arm exercise was found compared with leg exercise up to a f _b of 30 breaths·min⁻¹ while V E and V T were nearly the same. During fixed breathing when f _b was fixed at a higher rate than during SPNT, a significantly lower P ET CO₂ was observed during both exercise modes. These results suggest that: 1) FIX breathing does not affect the V O ₂ response during either arm or leg exercise even when non-synchronization between limb locomotion movement and breathing rate was adopted; 2) at a f _b of 30 breaths·min⁻¹ FIX breathing induced a hyperventilation resulting in a lower P ET CO₂ which was not associated with the metabolic rate during either arm or leg exercise, showing that V E during only leg exercise under the FIX condition was significantly higher than under the SPNT condition.