The purpose of this study was to examine the concurrent validity of the Personal Activity Monitor (PAM) accelerometer relative to the Actigraph accelerometer using oxygen consumption as a reference, and to assess the test-retest reliability of the PAM. Thirty-two fit, normal weight adults (aged 21-54) performed two activities, treadmill walking and stair walking, while wearing the PAM, the Actigraph and the Cosmed K4b(2). Correlation coefficients and agreement in absolute energy expenditure (EE) levels between PAM, Actigraph and Cosmed were calculated. The test-retest reliability was examined among 296 PAM's using a laboratory shaker. Intraclass correlation coefficients (ICC) and coefficient of variation (CV) were determined. Correlations for treadmill walking and stair walking, respectively, were r(2)=0.95 and r(2)=0.65 for PAM with Actigraph, r(2)=0.82 and r(2)=0.93 for PAM with VO(2) and r(2)=0.64 and 0.74 for Actigraph with VO(2). Both the PAM and Actigraph underestimated EE during treadmill and stair walking by a substantial amount. The test-retest reliability of the PAM was high [ICC=0.80; 95% confidence interval (CI) (0.28;0.92) and intra-CV=1.5%]. The PAM and Actigraph accelerometer are comparable in assessing bodily movement during treadmill and stair walking. The PAM is a valid device to rank subjects in EE and can be useful in collecting objective data to monitor habitual physical activity.