The existence of a circadian rhythm in plasma prolactin of the ram is controversial. Differences among authors can be related to both data sampling (e.g. interindividual changes, time of day, time of year, sampling interval among others) and statistical analyses. To test this hypothesis six adult "Préalpes du Sud" rams were studied individually during 72 hr in January (8 hr of light-16 hr of darkness), April (13L-11D), June (16L-8D) and September (13L-11D). Blood was sampled (vacutainer) from a jugular vein every hour, centrifuged and plasma samples stored at -20 degrees C until prolactin determinations (radioimmunoassay) were made. Individual time series were analysed according to three complementary methods: display of raw data (chronogram), best fitting cosine functions with different period tau (iterative cosinors) and power spectra. Seasonal changes in the 24 hr mean (peak time in June) were confirmed. A circahemidian rhythm (tau = 12 hr) and a circadian rhythm (tau = 24 hr) were validated, respectively in January and April while time series documented in June and September exhibited no rhythmic organization. It seems, therefore, that animals adjusted their rhythmic patterns of prolactin secretion to the increasing (January, April) rather than decreasing (June, September) photofraction (duration of the light span/24 hr).