Ventricular late potentials are obtained by signal averaging of surface electrocardiograms. Late potentials have been reported to be affected by body size or left ventricular mass. However, the effect of subadipose tissue, which is known to influence QRS amplitudes of surface ECG, on the variables of late potentials has not been evaluated. The relationships between the variables of late potentials and various obesity indices were assessed in 45 men, aged 24-38 years, without structural heart disease and bundle branch blocks. QRS duration, root mean square voltage in the last 40 ms and low-amplitude signals < 40 µV were obtained by signal-averaged ECG. Left ventricular mass was determined by echocardiogra-phy. QRS duration and root mean square voltage had no correlation with body height, weight, body mass index, sum of skin folds (triceps and subscapular) or left ventricular mass. Positive linear correlations were found between low-amplitude signals and weight (r = 0.48, p < 0.002) body mass index (r = 0.54, p < 0.002), sum of skin folds ( r = 0.57, p < 0.002), percent body mass index (r = 0.54, p < 0.002). Subadipose tissue may shift the onset of the 40-µV point of low-amplitude signals to the left with a consequent prolongation of low-amplitude signals by attenuation of the QRS complex. These data suggest that the use of low-amplitude signals alone or as a combination in an obese population for the definition of positive late potentials is inappropriate. The low-amplitude signal has to be used with caution in obese patients. In an obese population, the definition for normal for low-amplitude signals may need revision.