The quantification of maximum oxygen uptake (V(O2 max)), a parameter characterizing the effective integration of the neural, cardiopulmonary, and metabolic systems, requires oxygen uptake (VO2) to attain a plateau. We were interested in whether a VO2 plateau was consistently manifest during maximal incremental ramp cycle ergometry and also in ascertaining the relationship between this peak VO2 (V(O2 peak)) and that determined from one, or several, maximal constant-load tests. Ventilatory and pulmonary gas-exchange variables were measured breath by breath with a turbine and mass spectrometer. On average, V(O2 peak) [3.51 +/- 0.8 (SD) l/min] for the ramp test did not differ from that extrapolated from the linear phase of the response in 71 subjects. In 12 of these subjects, the V(O2 peak) was less than the extrapolated value by 0.1-0.4 l/min (i.e., a "plateau"), and in 19 subjects, V(O2 peak) was higher by 0.05-0.4 l/min. In the remaining 40 subjects, we could not discriminate a difference. The V(O2 peak) from the incremental test also did not differ from that of a single maximum constant-load test in 38 subjects or from the V(O2 max) in 6 subjects who undertook a range of progressively greater discontinuous constant-load tests. A plateau in the actual VO2 response is therefore not an obligatory consequence of incremental exercise. Because the peak value attained was not different from the plateau in the plot of VO2 vs. work rate (for the constant-load tests), the V(O2 peak) attained on a maximum-effort incremental test is likely to be a valid index of V(O2 max), despite no evidence of a plateau in the data themselves. However, without additional tests, one cannot be certain.