This article describes investigations of dynamic biomechanical stresses associated with lifting in stooping and kneeling postures. Twelve subjects volunteered to participate in two lifting experiments each having two levels of posture (stooped or kneeling), two levels of lifting height (350 or 700 mm), and three levels of weight (15, 20, or 25 kg). One study examined sagitally symmetric lifting, the other examined an asymmetric task. In each study, subjects lifted and lowered a box every 10 s for a period of 2 min in each treatment combination. Electromyography (EMG) of eight trunk muscles was collected during a specified lift. The EMG data, normalized to maximum extension and flexion exertions in each posture, was used to predict compression and shear forces at the L3 level of the lumbar spine. A comparison of symmetric and asymmetric lifting indicated that the average lumbar compression was greater in sagittal plane tasks; however, both anterior-posterior and lateral shear forces acting on the lumbar spine were increased with asymmetric lifts. Analysis of muscle recruitment indicated that the demands of lifting asymmetrically are shifted to ancillary muscles possessing smaller cross-sectional areas, which may be at greater risk of injury during manual materials handling (MMH) tasks. Model estimates indicated increased compression when kneeling, but increased shear forces when stooping. Increasing box weight and lifting height both significantly increased compressive and shear loading on the lumbar spine. A multivariate analysis of variance (MANOVA) indicated complex muscle recruitment schemes--each treatment combination elicited a unique pattern of muscle recruitment. The results of this investigation will help to evaluate safe loads for lifting in these restricted postures.