The utility or futility of flexion-extension radiographs in the diagnosis of lumbar spine segmental instability is a controversial issue. Previous investigations have reported a large range of normal motion and a significant overlap of symptomatic and asymptomatic motion patterns. The authors' goal was to define normal lumbosacral motion in vivo using ordinary weight-bearing lateral flexion-extension radiographs from 40 volunteers without the use of computers or special X-ray equipment. Calculation of dynamic vertebral translation, defined as the change in relative position from flexion to extension, provided a more accurate assessment of vertebral motion than measurement of static displacement on a flexion or extension view alone. Normal lumbar vertebral levels should have less than 3.0 mm of dynamic anteroposterior (AP) translation (less than 8% of vertebral body width). Although 42% of the normal subjects had at least one level with a static olisthesis greater than 3.0 mm in either flexion or extension, only 5% had a dynamic AP translation greater than 3.0 mm. These data have suggested that the previously reported large range of motion and frequency of overlap between symptomatic and asymptomatic patients may be significantly decreased (eightfold) by calculating dynamic motion, rather than static vertebral positions. The authors believe these data will provide the basis for reassessment of flexion-extension radiography in the diagnosis of lumbar spine instability.