This article addresses the problem of flight-skill decay. The complexity of the problem is outlined with regard to identifying the nature and extent of decayed skills. It is suggested that cognitive/procedural skills are more prone than control-oriented skills to decay over periods of disuse. Effective recurrent training methods offer the most promise in forestalling loss of proficiency. Several such methods are described, and the importance of cognitive training is emphasized. Finally, criteria are noted for evaluating the utility of recurrent training media. The theories and practical measures discussed apply to many skills other than those of flying.

Twenty-five minutes after ingesting one of three possible dosages (placebo, 200 mg, or 400 mg) of caffeine, 48 subjects individually participated in two short-term visual target scanning tasks (subject-paced and experimenter-paced). Assignment of subjects to treatment groups was made on the basis of a priori coffee consumption rates. Subjects who reported that they normally consumed less than three cups of coffee per week (average) were assigned to the low usage rate group (LR), while those who reported average consumption rates of three cups or more per week were assigned to the high usage rate group (HR). Latencies were obtained for the subject-paced measure by allowing subjects to scan ten 10 × 10 matrices until the target was found. Correct detection percentages were taken for the experimenter-paced task by allowing subjects to scan ten 10 × 10 matrices for 5 s and report whether or not a pre-announced target number was contained in the matrix. Significant differences in latencies were obtained between LR and HR only with 400 mg caffeine dosages. LR exhibited significantly higher latencies as a function of these dosages than did HR. No reliable differences occurred between LR and HR for correct detection percentages.

Visual vigilance performance was investigated with respect to environmental stimulation (90 db, 1000 cps noise) and intra-organismic stimulation (simple vs complex response). 140 Ss monitoring a display for 1 hr. under 1 of 2 noise types were instructed either to press a button upon detection of aperiodic signals (simple response) or to perform also a checklist operation subsequent to the signal detection (complex response). Significant correct detection differences occurred between response groups with complex groups showing higher performance ( p < .05). Differences in correct detection were obtained for noise conditions (for continuous, M = 83%; for intermittent, M = 78.5%). Results were evaluated in terms of the activation hypothesis.

Human performance data and theory are introduced relative to Bradley's (1975) optimalpessimal paradox. An attempt is made to demonstrate that the paradox, although theoretically sound, is grounded upon assumptions that are empirically untenable. Alternatives to Bradley's recommendations when optimal-pessimal conditions obtain are presented.

Sixty-four subjects served in a 50-min auditory monitoring task. Task complexity was examined with regard to both signal and response demands in an effort to determine variance contributing to each of these variables. Signals were presented at a mean rate of one per min, and no intersignal interval was greater than 2 min. Results showed that signal demands were of greater importance in affecting performance than were post-detection response contingencies. Statistically significant differences were obtained between groups monitoring only one signal and those monitoring any ofseven signals, with the former condition exhibiting better performance Over time. Groups in which signals were left unspecified exhibited lower detection percentages and higher false alarm rates than conditions in which signals were specified. NO statistical differences between simple and complex response conditions were ohserved. Application of the present findings to applied environments is discussed.