To examine the effect of elastic loading on variational activity of breathing, we studied 11 healthy subjects breathing at rest and with inspiratory elastic loads of 9 and 18 cm H2O/L, applied randomly for 1 h each. Compared with rest, a load of 18 cm H2O/L decreased gross variability, quantitated as standard deviation, of tidal volume (VT) and expiratory time (TE) (p < 0.01 in both instances) but increased that of inspiratory time (TI) (p < 0.03). The autocorrelation coefficients at a lag of 1 breath for each breath component were not altered by elastic loading, although the number of breath lags with significant serial correlations for TE tended to increase with a load of 18 cm H2O/L (p = 0.08). A load of 18 cm H2O/L decreased only the fraction of variational activity of VT and TE due to uncorrelated, random behavior (white noise), while it increased that fraction for TI (p < 0.05 in each instance); the correlated and oscillatory fractions did not change. Uncorrelated random behavior constituted > 87% of the variance of each breath component, correlated behavior represented 3 to 11%, and oscillatory behavior represented < 1.5% during both rest and loaded breathing. Elastic loading changed the gross variability of each primary breath component by altering the random fraction of variational activity; it had no significant effect on the structured, correlated fraction. We speculate that the observed changes in variational activity may reflect an attempt by the controller to compensate for the increased load while simultaneously minimizing load-induced dyspnea.