OBJECTIVES--Examination of the locomotor system is frequently neglected. Therefore, the GALS locomotor screen (Gait, Arms, Legs, Spine) has been proposed by Doherty et al as a practical method of identifying functionally important problems. This study was designed to test whether this screen reflects functional impairment, as measured by accepted health status measures. METHODS--Two observers performed the GALS screen in a total of 83 patients with a variety of musculoskeletal conditions. The examination components of GALS were rated by a simple 0 to 3 scale. Physical ability was further assessed by Health Activity Questionnaire (HAQ), Barthel index and Steinbrocker's ARA classification. RESULTS--For the total patient group, Spearman correlations between GALS and the three functional indices were good (r = 0.62 to 0.71, p < 0.001). Correlations were equally good for rheumatoid arthritis patients alone (r = 0.65 to 0.70, p < 0.001), but less good although still significant for the other miscellaneous rheumatic conditions (r = 0.31 to 0.46, p < 0.05). Observed proportional agreement between the two observers for the individual scores was > 70%, with a kappa statistic k = 0.49 to 0.74. CONCLUSIONS--The GALS screen is a reliable and valid measure of functional ability, compared with standard accepted indices in a variety of musculoskeletal diseases. This supports the proposal for its use as a screening test by general practitioners and medical students.