We studied the effects of prolonged physical activities on resting heart rate variability (HRV) during a training session attended by 23 cadets of the French military academy. This course lasts 1 month and is concluded by a 5-day field exercise simulation with physical and psychological stress. Data collection took place before (B) and immediately at the end (E) of the course. It included HRV recordings during a stand test (5 minutes lying down and 5 minutes standing), with a Polar R-R monitor, followed by blood sampling to assay plasma testosterone. The results (B and E) showed that the testosterone level fell by approximately 28.6 +/- 7%, indicating a high level of fatigue. During the stand test, the total power (TP) of the HRV spectrum increased in a supine position. The TP of B was 5,515.7 ms2 (SE, 718.4) and of E was 13018.9 ms2 (SE, 2,539.2; p < 0.001). High-frequency (HF) normalized values increased and low-frequency (LF) normalized values fell, regardless of position (HF normalized values and LF normalized values: supine, p < 0.01, p < 0.05; standing, p < 0.05, p < 0.01, respectively). LF:HF ratio fell 66.2 (SE, 12.9%; p < 0.01) in a lying position. During the time-domain analysis of HRV, differences between adjacent normal R-R intervals more than 50 milliseconds, expressed as a percentage, and differences between the coupling intervals of adjacent normal RR intervals increased in the lying position (p < 0.001). These results as a whole suggest that parasympathetic nervous system activity increases with fatigue.