The purpose of this study was to compare test-retest reliability when measuring peak physiological capacities of children and adolescents (age = 13.6 ± 2.9 yr) with mental retardation (MR) and their peers (12.0 ± 2.9 yr) without mental retardation (NMR) using a discontinuous treadmill (TM) protocol. Forty-six participants (23 MR = 12 male and 11 female; 23 NMR = 12 male and 11 female) completed two peak performance treadmill tests with 3 to 7 days of rest between tests. Physiological values measured included V̇O2peak (1 $$ min-1 and ml $$ kg-1 $$ min-1), V̇Epeak (1 $$ mhr-1), HRpeak (bpm), and RER (V̇O2 $$ V̇O2-1). Test-retest reliability coefficients ranged from .85 to .99 for participants with MR and from .55 to .99 for participants without MR. Test reliability and accuracy in the present study does not appear to differ between the NMR and MR participants.