The aim of this study was to investigate physical impairment in patients with chronic low back pain, to develop a method of clinical evaluation suitable for routine use, and to consider the relationship between pain, disability, and physical impairment. Twenty-seven physical tests were investigated. Permanent anatomic and structural impairments of spinal deformities, spinal fractures, surgical scarring, and neurologic deficits were excluded as not relevant to the patient with low back pain in the absence of nerve root involvement or previous surgery. Three consecutive 20-patient reproducibility studies were used to develop reliable methods of examination for 23 of the tests. Only four tests were excluded as unreliable: sacral angle, pelvic tilt, and separate lumbar and pelvic extension, none of which are part of routine clinical examination or have any proven relationship to disability. The remaining 23 physical tests were evaluated in 70 asymptomatic subjects and 120 patients with chronic low back pain. Passive knee flexion, passive hip flexion, hip flexion strength, hip abduction strength, pain reproduction on each of these tests, and the prone extension strength test were excluded because they were too closely related to nonorganic and behavioral responses to examination. Eight tests successfully discriminated patients with low back pain from normal subjects and were significantly related to self-report disability in activities of daily living: pelvic flexion, total flexion, total extension, lateral flexion, straight leg raising, spinal tenderness, bilateral active straight leg raising, and sit-up. Factor analysis failed to demonstrate an underlying statistical dimension of physical impairment. However, an empirical combination of total flexion, total extension, average lateral flexion, average straight leg raising, spinal tenderness, bilateral active straight leg raising, and sit-up provided an equally satisfactory alternative. Simple cut-offs from normal subjects made the scale simple and quick to use. This final scale successfully discriminated 78% of patients and normal subjects and explained 25% of the variance of disability, with a specificity of 86% and sensitivity of 76%. This scale provides an objective clinical evaluation that meets the criteria for evaluating physical impairment, yet is simple, reliable, and suitable for routine clinical use. It should, however, be emphasized that all the tests included in the final scale are measures of current functional limitation rather than of permanent anatomic or structural impairment. This raises questions about the physical basis of permanent disability due to chronic low back pain.