SUMMARY. Physiotherapists commonly use motion testing techniques in the assessment of patients with spinal disorders. One of the techniques considered important in the assessment of low back pain is the application of posteroanterior (PA) forces to the lumbar spinous processes to evaluate spinal stiffness and symptom response. The purpose of this study was to investigate the effect of spinal position on the measured lumbar PA stiffness. The PA stiffness at the L3 and L5 vertebral levels was assessed in 12 painfree subjects in three prone lying positions (flexion, neutral and extension). PA forces were applied to the spinous processes, under quasi-static loading conditions, using a specifically built apparatus. The stiffness coefficient was calculated from the slope of the regression line fitted to the truncated (35N-80N) force-displacement curves. At both vertebral levels, the PA stiffness was position dependent, with the lowest values obtained in the neutral position. The mean stiffness increased by 12.4% in extension and 31.9% in flexion. In all positions PA stiffness was significantly greater at L5 than at L3. These results suggest that clinical assessments of lumbar PA stiffness may be influenced by the patient position in which the test is performed. Comparative assessments of PA stiffness over time should be performed with the spine supported in the same position. Copyright 1998 Harcourt Publishers Ltd.