Harmonic variations in heart period between 0.15 and 0.40 Hz (high frequency spectral power, HF) measure cardiac vagal modulation. Data obtained under autonomic blockade show either a linear (Hayano et al.1991) or curvilinear (Goldberger et al. 1994) relationship between HF and beat to beat (RR) interval. Kiviniemi et al. (2004, 2006) demonstrated a curvilinear relationship between 5-min epochs of HF and RR data taken from 24-h recordings in some subjects. The aim of this study was to determine the magnitude and nature of the HF/RR relationship during wakeful, resting ECG recordings. Two hundred and fifty volunteers took part in the study (140 men; mean age 35.9 ± 12 years and 110 women; mean age 34 ± 10 years). ECG recordings were made using a two lead ECG with a sampling frequency of 1 ms (Polar Electro Ltd, Oy, Kempele, Finland) while subjects lay supine for 12 min with either free or paced (0.25 Hz) breathing. RR intervals were filtered and a fast Fourier transformation used to calculate HF power. HF was log transformed, HF(ln), and plotted as a function of corresponding mean RR interval. Linear and quadratic regression models were used to evaluate the relationship between RR and HF(ln). The point at which HF(ln) failed to increase for an increase in RR was derived from the differentiated quadratic equation and defined as the deflection point (DP). The criterion used to determine saturation of HF(ln) was that DP lay within the range of measured RR intervals. With free breathing (n=179 samples), the adjusted R² was similar between linear (25.2%) and quadratic (25.0%) fits and DP occurred outside the recorded RR interval range at 2042.0 ms. With paced breathing (n =234), adjusted R² values were smaller, but again similar between linear (19.7%) and quadratic (20%) fits. DP was outside the range of recorded RR intervals at 1422.9 ms. The HF(ln)/RR relationship does not show saturation when the measures are derived from single epochs of ECG data. The HF(ln)/RR relationship is weaker when derived from recordings made between subjects as opposed to 5 min epochs of a single tachogram (Kiviniemi et al.2004). Previous data have shown DP within the recorded RR interval range and perhaps, more importantly within the range that is likely to be observed. During free breathing, DP was found at a heart rate of 29 BPM. This is outside the normal range, indicating that HF(ln) shows no potential saturation characteristics. However, during paced breathing, a plateau in HF(ln) was found at 42 BPM. Such low heart rates were not recorded in the present study. However, they do occur and the HF(ln)/RR relationship may, therefore, saturate in elite athletes and beta blocked subjects, particularly if assessed during sleep. Such low heart rates are unlikely during wakeful, short-term ECG recordings.