Cardiac fibrillation is a leading cause of death in the modern world. Its treatment has so far been limited to painful and damaging defibrillation shocks. In this thesis, we experimentally investigate the possibility of using low-energy electric pulses to terminate fibrillation. We built two high-resolution fluorescence optical mapping systems to investigate the phenomenon of wave emission from heterogeneities in in vitro canine cardiac tissue preparations and whole rabbit hearts. The activation sequence from single pulses of varying amplitude was captured and quantified. The global activation time is found to obey a power-law with exponent -0.5 in rabbits and -0.15 in canines. We also present a novel model of tissue activation based on the density of wave sources within the tissue, valid for both 2d and 3d. Using data from CT scans of our tissue preparations, we found that the size distribution of the cardiovasculature also follow a power-law distribution over an order of magnitude, with scaling exponents -1.8 in canines and -2.13 in rabbits. Our model of activation time links the activation time with the size distribution via their scaling exponents. Finally, we tested the hypothesis that anti-fibrillation pacing (AFP) can terminate ventricular fibrillation with lower pulse energy than conventional defibrillation. We found an 81 % reduction in pulse energy in canine experiments. We saw no such reduc- tion in rabbit hearts, suggesting additional interactions between the fibrillatory activity and the far-field induced wave emission. Our results are presented as a contribution to our understanding of excitable systems and to the development of new treatment approaches in clinical cardiology.