We have previously observed a light-induced suppression of circadian rhythms of body temperature and locomotor activity in 50/55 male Sprague Dawley rats (Am. J. Physiol., 1995). This study presents a description of 4 records of body temperature obtained by telemetry and compares three methods of period analysis (Fourier analysis, autocorrelation and cosinor). In case of persistant circadian rhythms, Fourier analysis and cosinor method agree and allow description of rhythm parameters. However, in case of an alteration of the circadian system in continuous light, these 2 methods can disagree on the dominant period. Autocorrelation analysis permits to conclude on the absence of circadian and ultradian rhythmicity. These results demonstrate the necessity to associate several methods for period analysis and plaids for the development of a strategy for data processing of longitudinal time series.