Problems in labeling of pharmaceutical tablets are detected in the marketplace several times each year. When the offending tablet has an entirely different therapeutic application from the labeled drug, the problem often initiates a full recall of the product by the FDA. This report examines two varieties of tablets that had a labeling problem, 17-alpha-methyltestosterone and meclizine, using near-IR spectrometry and a new technique, acoustic-resonance spectrometry (ARS). The two types of spectrometry were combined (NIR/ARS) and the results were compared with near-IR spectrometry alone to determine which gave the most accurate results. A bootstrap algorithm for estimating the errors in each principal component of the spectra was used in interpreting the column vectors of the principal-component transformation matrix. Both near-IR and ARS techniques could identify the tablets individually, and the combined NIR/ARS data produced even more reliable identifications.