Background: The upper limit of vulnerability (ULV) hypothesis for defibrillation predicts that maneuvers that alter the ULV will cause a similar alteration in the defibrillation threshold (DFT). The purpose of this study was to test this prediction by evaluating the effects of electrode polarity and waveform duration on the relationship between the DFT and the ULV. Methods and Results: Platinum spring electrodes were placed in the right ventricular (RV) apex and the superior vena cava in 12 pigs. Strength- duration curves were constructed for the DFT and ULV for each electrode polarity with monophasic waveforms (6 pigs) of different durations (2 to 14 ms) and biphasic truncated exponential waveforms (6 pigs) having phase. 1 equal to 4 ms and phase 2 of different durations (0 to 10 ms). ULV data were gathered by scanning of the T wave. The ventricular pacing threshold (VPT) and ventricular fibrillation threshold (VFT) were also determined with these same waveforms. For the RV electrode as a cathode for monophasic and the first phase of biphasic stimuli, VPTs for the same waveform duration were significantly lower than for the configuration with the RV electrode as an anode. VFTs were not significantly different for the two electrode polarities with either monophasic or biphasic waveforms. The DFT changed in a fashion similar to the ULV with changes in electrode polarity and phase duration for both monophasic and biphasic waveforms. The ULV and DFT for each waveform duration for each polarity were strongly correlated (r=.83 to .99). Conclusions: The almost identical changes in ULV and DFT with changes in electrode polarity and waveform duration provide new evidence to support the ULV hypothesis of defibrillation.