The effect of pattern adaptation on spatial frequency discrimination was examined in two experiments. In the first experiment Δƒ/ƒ was measured (with and without adaptation) as a function of stimulus contrast; in the second experiment the contrast required to discriminate a fixed Δƒ/ƒ was measured (with and without adaptation) as a function of the value of Δƒ/ƒ Maximum precision, as measured by the asymptotic value that Δƒ/ƒ approaches at medium and high contrasts, was not altered by adaptation. Rather, the effect of adaptation was to translate the functions relating Δƒ/ƒ and contrast along the log-contrast axis, i.e., to increase by a constant factor the contrast required to achieve a given precision of discrimination. This factor agreed closely with the factor by which adaptation raised the contrast threshold for detection of the test stimuli. The largest effects were observed when the reference and the adapting stimuli had the same spatial frequency (5 cycles/deg). Smaller or negligible effects were observed when reference and adapting frequencies differed by an octave.