In continuous darkness for about 5.5 days in eleven replications, the pineal shows neither only a decreasing trend nor only a damped circadian rhythm; there is an infradian modulation: after an initial decline, melatonin production increases again, with a pattern compatible with the assumption, for at least one cycle, that a built-in pineal "week" has a larger extent of predictable change than that of a pineal "day". Melatonin release from the isolated pike pineal reveals a circadian rhythm which, on the average, is damped, but demonstrable in the second half of a time series covering 5.5 days (P < 0.05). At fixed environmental temperatures of 10, 15, 19 or 20 degrees C (the latter two pooled), single and population mean cosinor analyses reject the assumptions of zero circaseptan and circadian amplitudes in the chronome of the pineal from the pike (Esox lucius L.). The circaseptan pattern is confounded by possible trends. The time series analyzed are brief, yet validate, by a one-way analysis of variance in plexograms as well as by cosinor, an about-weekly pattern, if not (yet) rhythm, on the basis of 745 determinations. The circaseptan- and circasemiseptan-to-circadian amplitude ratios are numerically larger than unity, on the average (2.014; 95% confidence interval, CI: 1.120 to 3.641 and 1.107; CI: 0.724 to 1.693, respectively). A positive correlation of the urinary excretion rate of a major breakdown product of melatonin of a cancer patient with a circulating cancer marker serving as a gauge of disease progression lends an oncologic perspective to the amplitude ratios of melatonin release in vitro here reported.(ABSTRACT TRUNCATED AT 250 WORDS)