Forty-eight licensed pilots flew three cross-country flights in which certain aspects of perceptual workload were varied by altering the rate and number of instrument-displayed incipient danger conditions. A moving-base simulation of a single-engine general aviation aircraft was used. The sensitivity of eight mental workload estimation techniques investigated to changes in perceptual workload was within a univariate factorial design. Concurrently, the differential intrusion of the eight techniques on four primary task measures was investigated using multivariate analysis. Of the eight techniques, six displayed statistically-significant sensitivity to load level. These included two opinion rating scales, secondary task measures of time estimation standard deviation and tapping regularity, respiration rate, and a primary task measure of danger condition detection/identification time. No intrusion effect was found. Recommendations for applying the various techniques, based on the relative sensitivity of those showing significance, are discussed. / Ph. D.