Previous evidence has shown that detection threshold in humans and olfactory neural discharge rate in animal preparations both depend on flow rate of odorous vapor. But no data have been reported that show the effects of flow rate in humans on perceived odor strength at suprathreshold intensities. Subjects learned to inspire at two flow rates, one twice as great as the other, by adjusting (on a cathode ray tube) the transduced trace of a sniff-produced pressure change to match either of two target contours. They then made magnitude estimations of odor strength, while producing either weak or strong sniffs, for odorants presented over a wide range of concentrations via a specially designed sniff-bottle system. The odorant, diluted in diethyl phthalate, was n-butanol in two experiments and n-amyl acetate in two others. Subject-controlled flow rate had no effect on odor strength for either odorant. There was an apparent contradiction between these data and those on neural discharge rate that may, however, be resolved by adopting an odor constancy model: When sniff intensity varies during the olfactory exploration of an odor source, information about the rate at which odorant molecules are established at the receptor site is combined with information about sniff vigor so that the resulting percept is of invariant odor strength.