The most popular periodicities in biology and medicine-the circadians and circannuals-stem undoubtedly from the Earth's rotation and its revolution around the sun. The problem is how to explain the existence of circaseptan, i.e. 5-9-day, and other infradian rhythms. They may correspond to the lunar cycles and their 2nd to 6th harmonics. To test such hypothesis, the calendar dates of 127 attacks of atrial fibrillation in one male subject (M.M.) between 1980 and 1994 were transformed into the days numbered 0-29 for the synodic, and 0-26 for tropic lunar cycle. The daily frequencies obtained in this way were smoothed by moving averages of three successive days each. Considerable fluctuations of frequencies of attacks during both cycles were visible by inspection of the corresponding graphs, called lunar plexograms. Thus, a conspicuous nadir is found under the full moon in the synodic cycle, and a marked peak shortly after the extreme southern position of the moon in the tropic cycle. Halberg's cosinor analysis testing the presence of the 1st to 6th harmonic of either lunar cycle rejected the null hypothesis at the alpha = 0.05 level for all harmonics. Accordingly, the occurrence of attacks was cycling with the period lengths of synodic and tropic lunar cycles, and with those of their 1/2-1/6 period lengths, i.e. with a cluster of approximately circa(di)-septan rhythms. This conclusion is supported by similar findings obtained earlier for various medical and biological events.