Malt, wort and beer samples representing a single pilot brew were extracted from Celite columns with Freon 114. Total extracts were analyzed by gas chromatography/mass spectrometry. Compounds were identified that were unique or common to beer, wort or malt and quantities present were determined for those common to the three sample types. Concentrated extracts were separated by liquid chromatography on partially deactivated silica gel packed in Teflon tubes, using Freon 11 as the solvent. Fractions from liquid chromatography were analyzed by descriptive sensory evaluation. Those fractions selected as bearing cereal-like, grassy, nutty or malty aroma characteristics were analyzed by gas chromatography. Freon extracts and fractions were found suitable for descriptive sensory evaluation and this was used to advantage in locating fractions with the desired aroma characteristics. Cereal-like, grassy, nutty or malty aroma characters were associated with liquid chromatographic fractions containing primarily alcohols, aldehydes and ketones. The Ascending Method of Limits for determination of the threshold of added substances in beer was critically reviewed and a new threshold test was developed. The new test design reduces the sample requirement from eighteen to twelve for each session. In contrast to the Ascending Method of Limits, the new test design was amenable to statistical evaluation and statement of a level of confidence for the threshold intervals determined. Both a nine-point intensity scale of difference from the labeled reference and a category scale were applied to the new test design. Intensity scale data were normalized and subjected to analysis of variance and a series of one-sided t-tests to determine individual and combined panel threshold intervals. Category scale data were analyzed by a rank sums test to determine individual thresholds and the R-index values thus obtained were subjected to analysis of variance and orthogonal contrasts of means to establish the combined panel threshold. The intensity scale method applied to the new test design provided the lowest and narrowest estimates of the threshold interval in comparative testing with the category scale method and the Ascending Method of Limits and, in addition, required fewer replications than the category scale method to establish the threshold interval. / Graduation date: 1989