The reproducibility of spirometric measurements made at 6-mo. intervals for 18 to 24 mo. in a group of 38 Be workers is analyzed. The vital capacity and 1-, 2-, and 3-sec. forced expiratory volume had standard deviations from 2 to 3% of the mean value for each subject; the values for the maximal mid-expiratory flow and maximal voluntary ventilation were 8 to 10%. The per cent of standard deviation was slightly greater in smokers than in nonsmokers. These values are similar to those reported in the literature for observations made over a shorter period of time. In nonsmokers there was a significant correlation between repeated observations of the change produced by an aerosol bronchodilator on the forced expiratory volume and maximal mldexpiratory flow.