When two pure tones (or primaries) of slightly different frequencies (f 1 and f 2; f 2 > f 1) are presented to the ear, new frequency components not present in the stimulus may be recorded in the ear canal. These new components are termed distortion product otoacoustic emissions (DPOAEs) and are generated by nonlinear interaction of the primaries within the cochlea. It has been conjectured that the level of the 2f 1 − f 2 DPOAE component is maximal when the primaries produce approximately equal excitation at the f 2 cochlear region because this is where and when the overlap between the traveling waves evoked by the two primaries is maximal. This region, however, almost certainly shifts as the level of the primaries increases following the well-known level-dependent shift of the cochlear traveling-wave peak. Furthermore, mutual suppression between the primaries may also affect the combination of primary levels that maximizes the DPOAE levels. This report summarizes our attempts to test these conjectures using psychophysical masking methods that are commonly applied to infer human cochlear responses. Test frequencies of 0.5, 1 and 4 kHz and a fixed frequency ratio of f 2/f 1 = 1.2 were considered. Results supported that maximal-level DPOAEs occur when the primaries produce comparable excitation at the cochlear site with CF ∼ f 2. They also suggest that the site of maximum interaction hardly shifts with increasing primary level and that mutual suppression between the primaries does not affect significantly the optimal DPOAE primary level rule.