Recently, considerable effort was focused on unifying various aspects of cardiac pathophysiology in terms of nonlinear dynamics, particularly through application of chaos theory and the concept of fractal geometry.1-5 Goldberger et al4-7 have suggested that the specialized cardiac conduction network, the His-Purkinje system, has a fractal geometry. Fractals, first described by Mandelbrot8 and found ubiquitously in nature, are self-similar geometric structures composed of subunits that in turn are composed of smaller subunits in a cascade down to microscopic scales. Each subunit appears as a smaller but otherwise identical version of the super-structure. The normal QRS complex is a broad-band wave-form,9,10 and its power spectrum has been shown to fall with frequency according to the following relation: P(f) = kfβ (1) where k is a positive constant and β is a negative constant.6 This type of frequency dependence is generally referred to as an inverse power law, or more simply, a power law. Goldberger et al6 have provided a theoretical argument suggesting that the power-law nature of the QRS spectrum is a direct consequence of the proposed fractal geometry of the His-Purkinje system. We sought to test this hypothesis by studying the power spectra of QRS complexes from both normally and abnormally conducted beats. During a ventricular premature beat or an exogenously ventricular-paced beat, involvement of the His-Purkinje system in ventricular depolarization is reduced or absent.11 Thus, comparing power spectra for these beat types with spectra for normally conducted beats provides a means for establishing the importance of the conduction system in determining QRS spectral morphology. © 1993.