A monaural loudness matching technique was used to study differential sensitivity to intensity as a function of tonal duration. The probable error (p.e.) of the loudness matches was used as the measure of differential sensitivity. With one technique, a standard tone of 500 milliseconds duration was followed by a tone of variable duration (10–500 milliseconds) after a silent interval of 50, 100, or 500 milliseconds. In another technique, both standard and comparison tones were of the same duration (10–500 milliseconds) with the same silent intervals between tones as before. (1) When the standard tone was always 500 milliseconds, the p.e. of the loudness matches increased with a decrease in the duration of the comparison tone from approximately 0.60 to 2.30 db, and the length of the silent interval had no effect on the function. (2) When both the standard and comparison tones had the same duration, the p.e. again increased with a decrease in duration; in this case, however, a silent interval of 500 milliseconds caused an increase from approximately 0.60 to 2.50 db while a silent interval of 50 milliseconds caused an increase to only 1.00 db. These differences are explained in terms of two different processes: adissimilarityeffect, and aninterferenceeffect. (3) When a standard tone of constant duration is used to obtain loudness matches, the mean of the matches becomes a measure of the loudness of tones as a function of duration. These measures showed a clear distinction between the six observers used. For three observers, the change in duration caused practically no change in loudness. For the other three, changes in loudness as great as 8.5 db were recorded. This order of loudness change agrees with that reported by Békésy, but is considerably less than that reported by Munson. Possible explanations for the differences are mentioned.