This study examined the effect of mild hypothermia (a 0.5 degrees C decrease in rectal temperature) on heart rate variability (HRV), with the identical hypothermia protocol performed twice and compared using intraclass correlation coefficient (r) analysis to study the repeatability. Twelve healthy males each completed 1 neutral (23 degrees C) and 2 cold (0 degrees C) trials. In the neutral trial, participants sat quietly for 30 min. In the cold trials, baseline data were obtained from a 5-min sample following 30 min of quiet sitting at 23 degrees C, followed by passive exposure to 0 degrees C; hypothermic measures were taken from a 5-min period immediately prior to rectal temperature decreasing by 0.5 degrees C. HRV was obtained from a 3-lead electrocardiogram. There were no differences (all p > 0.05) in baseline measures between the neutral and the 2 cold trials, suggesting no precooling anxiety related to the cold trials. Heart rate, together with HRV measures (i.e., root mean square difference of successive normal RR intervals, triangular interpolation of NN interval histogram, low-frequency oscillations (LF), and high-frequency oscillations (HF)), increased (all p < 0.05) with mild hypothermia and showed excellent reliability between the 2 cold trials (all r >/= 0.81). In contrast, the LF/HF ratio decreased (p < 0.05) and had only fair reliability between the 2 cold trials (r = 0.551). In general, hypothermia led to increases in heart rate, together with most measures of HRV. Although it was counterintuitive that both sympathetic and vagal influences would increase simultaneously, these changes likely reflected increased stress from whole-body cooling, together with marked cardiovascular strain and sympathetic nervous system activity from shivering to defend core body temperature. An important methodological consideration for future studies is the consistent and repeatable HRV responses to hypothermia.