Specimens of SCS-6/Timetal®21S in 4 and [0/90]S orientations were subjected to three types of tests covering temperatures of 650, 760, and 815‡C. In the first, low frequency cycles with no superimposed hold times were applied. In the second, low frequency fatigue cycles were applied with superimposed hold times of various durations at maximum loads. In the third, hold times were superimposed at minimum load. In all tests, the number of cycles to failure was recorded. The behavior of the composite in these tests was modeled by considering damage to consist of a linear sum of a cycle-dependent component and a time-dependent component. It was found that the behavior in all tests was basically time dependent Using time at temperature and stress as a failure criterion, it is shown that data from creep, low frequency fatigue with and without hold times, and in-phase thermomechanical fatigue tests can be consolidated for all 4 tests into a single curve and for all [0/90]S tests into a different curve. Both curves can be explained in terms of a single fiber bundle strength value by considering the time-dependent behavior of the matrix material. Additionally, hold times at minimum load are found to be more detrimental to fatigue life than identical tests where the hold time is applied at the maximum load.