The intra-observer reproducibility of three methods of measuring grip strength with a standard sphygmomanometer cuff was studied in 26 patients with rheumatoid arthritis (RA). Two techniques involved the measurement of peak pressure but differed in the period of time between consecutive attempts with each hand. The third method measured sustained grip. All three methods were highly reproducible. However, peak pressure recordings (either the best or the mean of three readings) were associated with least patient discomfort, were easier to read, and are suggested as the methods of choice for serial assessment of grip strength in RA.