Clinicians often fabricate foot orthotic devices at the subtalar joint neutral position (STNP) to mimic the position of the rear foot during midstance. However, rear foot motion during gait, relative to the resting standing foot position, not the STNP, is often reported in the literature. The motion of the rear foot relative to a valid estimate of the STNP is unknown. In this study, six experienced foot care specialists manually placed the rear part of the feet of nine subjects at the STNP seven or eight times to obtain a valid estimate of each subject's STNP. The worst-case mean and 95% confidence interval of the STNP estimate for any one subject was 0.0° ± 0.7°. These nine subjects then walked on a motor-driven treadmill, set at 0.89 meters/sec, and three-dimensional estimates of each subject's rear foot inversion/eversion motion were obtained, then averaged over 6 to 26 strides. For most subjects, the rear foot was always everted during stance with mean and standard deviation maximal eversion (7.2° ± 1.2°) occurring at 44% of the total gait cycle. The inversion/eversion orientation during swing was characterized by 1 ° to 2° of eversion, with a small amount of inversion in early swing. These findings have implications for the fabrication of foot orthoses, since the rear foot is rarely near the STNP during stance.