The critical point hypothesis for the upper limit of vulnerability (ULV) states that the site of S1 pacing should not affect the ULV S2 shock strength for a single S2 shock electrode configuration but may affect the S1-S2 interval at which sub-ULV shocks induce ventricular fibrillation (VF). Furthermore, early post-S2 activations leading to VF should arise in areas with low potential gradients of similar magnitude, regardless of the S1 site. This hypothesis was tested in 10 pigs by determining ULVs for three S1 sites [left ventricular apex (LVA), LV base (LVB), and right ventricular outflow tract (RVOT)] with one S2 configuration (LVA patch to superior vena cava catheter). T-wave scanning was performed with biphasic S2 shocks incremented from 60 V in 40-V steps and stepped up or down in 20- and 10-V steps. Activations and S2 potential gradients were recorded at 528 epicardial sites. Although shocks just below the ULV induced VF significantly earlier in the T wave when the S1 site was the RVOT than when it was the LVA or LVB, ULVs were not significantly different for the three S1 pacing sites. Early post-S2 activations arose closer to the S2 electrode for weak S2s but moved to distant low potential gradient areas as the S2 strengthened. Just below the ULV, early post-S2 activations arose in the RVOT when the S1 site was the LVA or LVB but arose along the RV base when the S1 site was the RVOT. Early site potential gradients were not significantly different just below the ULV (LVA: 8.2 ± 4.1 V/cm; LVB: 8.6 ± 4.9 V/cm; RVOT: 8.7 ± 4.4 V/cm). At the ULV, early post-S2 activations arose from the same areas but did not induce VF. The results support the critical point hypothesis for the ULV. For this S2 configuration, no single point in the T wave could be used to determine the ULV for all S1 sites.