Research on heart rate variability (HRV) received increasing attention. This study analysed the reliability of the most common HRV parameters for baseline measurements. 103 healthy students (83 women, M = 21.72 +/- 3.31 years) participated in five short-term HRV sessions, each including supine, sitting, and standing positions, respectively, spanning a time interval of eleven months. Relative reliability was evaluated by intraclass correlation coefficients, and absolute reliability by standard errors of measurement, smallest real differences, and 95 % limits of random variation. No systematic mean differences between measurements emerged. Intraclass correlation coefficients were quite low (supine: .49-.64, sitting: .40-.57, standing: .35-.56). Absolute reliability indicators revealed pronounced variations between test and retest. Influences of posture and time between measurements on reliability were small and unsystematic. We conclude that such high levels of within-subjects variability in HRV measurements (a) hamper the detection of changes over time, and (b) should be considered carefully in future analyses.