Person and situational determinants of cognitive ability test performance and subjective reactions were examined in the context of tests with different time-on-task requirements. Two hundred thirty-nine first-year university students participated in a within-participant experiment, with completely counterbalanced treatment conditions and test forms. Participants completed three test sessions of different length: (a) a standard-length SAT test battery (total time 4 1/2 hr), (b) a shorter SAT test battery (total time 3 1/2 hr), and (c) a longer SAT test battery (total time 5 1/2 hr). Consistent with expectations, subjective fatigue increased with increasing time-on-task. However, mean performance "increased" in the longer test length conditions, compared with the shorter test length condition. Individual differences in personality/interest/motivation trait complexes were found to have "greater" power than the test-length situations for predicting subjective cognitive fatigue before, during, and at the end of each test session. The relative contributions of traits and time-on-task for cognitive fatigue are discussed, along with implications for research and practice. (Contains 3 figures, 7 tables, and 2 footnotes.)