Continuous-wave (CW) Doppler recordings of ascending aortic maximal blood flow acceleration (PkA), maximal velocity (PkV) and systolic velocity integral (SVI) were taken at each stage of a graded exercise treadmill test on two separate days with 30 physically active adult males. Signals were measured (Quinton Exerdop) for all cardiac cycles in the 3rd minute of each stage using a hand-held probe positioned at the suprasternal notch. A dedicated microcomputer, programmed to select "valid" beats on the basis of value consistency in the sample set, determined the acceptability of signals. No significant differences were found between the three trial means within each stage on either day for PkA, PkV or SVI. Significant (p <.01) intraclass reliability estimates ranged from r = 0.89 to 0.97 (Pk.A), r = 0.90 to 0.98 (PkV) and r = 0.85 to 0.95 (SVI) . Coefficients of variation were calculated at each stage to estimate the relative consistency of each measure. A gradual reduction of the coefficient of variation was observed for each blood flow measure between stages one and four. The test-retest (between days) reliability coefficients for PkA, PkV and SVI for stages one to four ranged between r = 0.51 to 0.78 (P <.004), but correlations for the pre-exercise baseline and stages five and six were lower. These results indicate that (1) PkA, PkV and SVI demonstrate greater measurement stability within each stage of a graded exercise test than is the case between separate days of measurement at the same stage; and (2) there is modest day-to-day response stability for clinical testing with the Doppler parameter of PkV. Reliability/stability was best in exercise stages which encompass the speed and grade range of 45.0 m•min⁻¹/10%. - 111.7 m•min⁻¹/14%, i.e., those in which all subjects can walk. / Master of Science