Live trout spermatozoa initiate flagellar motility for only a short period (30 sec at 18 degrees C) during which their mean beat frequency decreases steadily from 60 to 20 Hz. Motility then stops abruptly. Investigations of the activation of movement in demembranated sperm points to cyclic-AMP being necessary for reactivation (half effect at 0.5 microM) in some conditions. cAMP acts mainly by increasing the percentage of motile cells and not the beat frequency (BF) of the flagellar axoneme. Dibutyryl cAMP does not initiate movement or prolong motility of live sperm. The initiation of movement of demembranated trout sperm was investigated in various incubation conditions relative to previous phases of in vivo movement and to ATP concentration. In the absence of cAMP and in the presence of ATP lower than 25 microM, all sperm cell models were active with BF up to 15-20 Hz whatever their previous physiological conditions. In contrast, at ATP concentrations above 100 microM, the fraction of active spermatozoa decreased proportionally but the BF of the active ones increased so that, at 1 mM ATP, only 5% were active but with a BF of 65 Hz: the addition of cAMP up to 20 microM restored activity to 100% sperm models with a similar BF of 65 Hz. At ATP concentrations higher than 25 microM, cAMP was necessary in a concentration dependent manner in the reactivation, but not in the demembranation medium. This dependence was found to be unrelated to a previous in vivo phase of movement. The antagonistic effects of ATP vs. cAMP were tested at various concentrations of both nucleotides: the apparent affinity for cAMP, measured as the concentration restoring movement of 50% cell models, was decreased from 15 nM at 0.1 mM ATP to 0.5 microM at 1 mM ATP; conversely, the affinity for ATP, measured as the concentration giving rise to the half maximal beat frequency, was not significantly affected when the concentration of cAMP was raised to 0.5 mM. Preincubation with phosphodiesterase (PDE) resulted in motility of 100% of sperm models even at low ATP concentration. This tends to show that cAMP must be constantly present to sustain motility.