Electrical mapping from exposed rabbit right ventricular endocardium provides an attractive experimental model to study Purkinje-myocardial interactions. Such mapping is complicated by three main factors: (i) the nonuniform tissue structure in which the marked trabeculation limits surface contact with standard plaque electrode arrays; (ii) the need for sufficiently high sampling rates to resolve the rapid Purkinje (P) deflections in the electrograms; and (iii) the advantage of eliminating far-field contributions in those electrograms by using a reference electrode retracted from the surface recording electrode by 0.5-2.0 mm. The present study analyzes the impact of these factors on the P deflection yield. Yield was defined as the percentage of recording sites at which unique P deflections were identified during sinus beats or direct pacing of proximal conduction system tissue. During unipolar mapping with simultaneous recordings from 528 electrodes embedded in a flat plaque and archived at 2000 samples/s, P deflection yields measured 22±9% (n=5 experiments). During sequential mapping from fewer sites at 8000 samples/s with a filament electrode that included a retracted reference, P deflection yields increased to 60±8% (n=3 experiments). A method for simultaneous mapping with multiple filament electrodes whose positions vary with the contour of the endocardium is therefore proposed.