Amplitudes of inaudible “subjective” signals are inferred from tone‐on‐tone masking measurements. Several methodological issues are involved, e.g., the problems of how many and which measurements to include. To explore such sampling questions, sets of masked thresholds using a 250‐Hz masker and a 500‐Hz maskee are determined at various phase angles. These measurements are distributed into subsamples of (n =) 13, 9, 7, 5, or 4 each, and repeated estimates of the second aural harmonic inferred. For n > 4, these estimates averaged to within ± 1 dB of one another, although individual estimates vary considerably with the smaller n's (to 13 dB with n = 4]. Sets of five consecutive measurements also were analyzed when the masker : maskee phase angle was incremented in 90° steps through two complete periods (of the maskee). Again, the reestimates for individuals can vary by at least 6 dB, but the averages over seven subjects remain within 2 dB of each other. Apparently, for n's of six or more, it matters little which masker: maskee phase angles are used to estimate the aural harmonic amplitudes.