Levels of neck and shoulder myoelectric activity (EMG) were recorded in eight female cashiers during work with different cash register systems. Static, median and peak levels of load were determined, and comparisons were made between conventional keyboard operation and operation with horizontal scanner, vertical scanner and pen reader respectively, and between sitting and standing work postures. Each complete transaction (33 articles) lasted 74–166 s (mean 104 s). Cash register operation, regardless of register system, caused static load (mean) in the thoracic erector spinae at 4·1–6·2 TAMP% (Time-Averaged Myoelectrical Potential), in the infraspinatus at 2·9–5·8 TAMP%, and in the trapezius at 2·8–3·8 TAMP%. The cervical erector spinae and levator scapulae were very low in activity (0·6–1·3 TAMP% at the static load). Except for the left infraspinatus during scanner and pen reader operation, the median loads were low. Mean peak loads never exceeded 27·4 TAMP%. When the cashier was standing, the muscle activity at all levels of load was lower than when sitting, e.g., the left infraspinatus mean static load was 3·7 TAMP% (standing) and 5·4 TAMP% (sitting) (p≤0·001). Differences were found between the cash register systems at median and peak loads. Work with scanners and pen reader caused, in ranking order, higher loads than conventional keyboard operation in the trapezius, infraspinatus and thoracic erector spinae. However, the main problem seems to be the repetitive handling of merchandise and the constant static load regardless of cash register system used.