It is generally accepted that bundle branch block (BBB) may be corrected simply by selective His bundle pacing (S-HBP) distal to site of block. However, clinical HBP is often non-selective (NS), the site of block unknown and moving the HB lead is not always an option. Thus, while correction of right (R) BBB in NS-HBP has been reported, its mechanism is unknown. Threshold testing was done during His bundle pacing in 39 patients with RBBB. We compared NS-HBP and S-HBP and analyzed the role of pacing voltage and capture threshold of the parallel pathway, in partially or completely correcting RBBB. During NS-HBP, higher pacing voltage completely corrected RBBB in 22/35 patients as compared to 5/18 patients at lower voltage. Remaining patients showed incomplete correction of RBBB. NS-HBP group with lower capture threshold of 1.3 ± 0.5 V completely resolved RBBB in 9/14 vs 3/11 patients in NS-S HBP group with higher capture threshold of 2.4 ± 0.8 V. QRS activation time in NSHBP was 91 ± 8 ms (baseline 97 ± 7 ms p = ns). S-HBP resolved RBBB in only 1/21 patients. 1.) Lack of correction with S-HBP suggests that RBBB was distal to site of HBP and yet was corrected with NS-HBP. 2.) Voltage dependent correction of RBBB in NS-HBP suggests that conduction via a specialized parallel pathway maintains normal ventricular activation time. 3.) Correction of RBBB in all patients with NS-HBP, suggests that conduction block was either bypassed or right ventricular free wall pre-excited by conduction via parallel pathway.