In mice circadian body temperature curves are masked due to the effect of motor activity. However, body temperature will not immediately reflect activity, but rather the integrated activity over IT minutes (integration time) and after a certain delay (lag), and the sensitivity to such masking may change throughout the circadian cycle. The aims of the present investigation were to estimate IT and lag, to quantify the effect of motor activity on body temperature at different times of the day, and, using these results, to draw temperature curves that are closer to the endogenous one. Activity and body temperature of adult male laboratory mice were recorded telemetrically at 10-min intervals. Animals were housed in air-conditioned rooms (T = 22+/-2 degrees C; relative humidity: 55-65%) with a light-dark cycle of 12 h:12 h (light from 0700 to 1900 hours) and food and water available ad lib. The diurnal activity and body temperature rhythms were similar with a main maximum during the dark time and a secondary maximum immediately following lights-on. Nearly all changes of activity were reflected in body temperature. IT and lag were established on the basis of the best correlation between body temperature and activity (overlapping 4-h sections of 12 days) for all combinations of IT from 10 to 90 min and lag from 0 to 50 min (10-min steps each). The overall means of IT and lag were 40 and 0 min, respectively. During the dark time the values were somewhat larger, but not significantly so. The correlation between activity and body temperature was significantly better in the light time compared to the dark time. The sensitivity of the body temperature to changes in activity was investigated by linear regression analysis for every hour over 12 days (IT = 40 min, lag = 0 min). The gradients assessed by regression analysis showed a diurnal pattern with maximal values during the light time (p < 0.01). Thus, body temperature was raised by activity more during the light time (minimum of body temperature and activity) than during the dark time. The intercepts showed a nearly sinusoidal diurnal pattern with maximal values in the middle of the dark time. Accepting that the intercepts correspond to zero activity at a certain time of day, one might use them to get a curve that is closer to the endogenous body temperature rhythm. Mechanisms (circadian and thermoregulatory) that might cause the diurnally changing sensitivity of body temperature to activity are discussed.