Blood prolactin (PRL) variations have been linked to temperature and osmotic changes in several species. The latter factors are here explored to better understand blood PRL responses frequently induced during physical exercise. Since body heat generated by exercise can lead to marked body fluid shifts, it was postulated that PRL changes observed during exercise could be associated with variations in body temperature and/or blood osmolality (OSM). A wide range (38.5–40.5°C) of rectal temperatures (Tr; used here to appreciate core temperatures) were theoretically selected and randomly assigned as targets to male runners. Measured by thermistor probe, target Tr were obtained by a combination of factors: (a) ↑ heat production by treadmill running, and (b) ⇓ heat losses by appropriate clothing (⇓ evaporation) in warmed (⇓ radiation) and hypo ventilated (⇓ convection) laboratory conditions. For each subject, target Tr was attained not prior to 30 min after initiation of running, and had to be maintained for at least 10 min, for a mean ( ± SD) running time of 52.6 ± 10.0 min. In a first protocol, hypohydration was provoked in 26 runners (23.9 ± 4.7 years) by total restriction of water intake. In a second protocol (10 different runners: 22.3 ± 3.3 years), euhydration was maintained by water intake (20 ml/kg body weight). Venous blood was sampled at rest before and immediately after the run. PRL was assayed by RIA; OSM was measured by freezing point depression; sodium was analyzed by flame photometry. At rest, before the heat-producing exercise, mean PRL values were 9.4 ± 3.4 ng/ml for both eu/hypohydrated groups. In the hypohydrated runners, exercise-induced hyperthermia was significantly (r = 0.82; p < 0.0005) associated with blood PRL responses. Moreover, these changes in Tr were also significantly (r = 0.54; p < 0.0025) related to changes in OSM, the latter variations being mostly explained (78 %) by the accompanying hypernatremia. In the euhydrated group of runners, the hyperthermic exercise failed to induce significant changes in OSM (r = 0.22; p > 0.15) and, as expected, variations in blood sodium levels were also not significant under these conditions. However, hyperthermic running in these iso-osmolar conditions did not prevent blood PRL levels from rising (r = 0.77; p < 0.0005). It was thus concluded that, in male trained runners, exercise-induced blood PRL responses could be derived more from thermic than from osmolar stresses.