BACKGROUND Gap junction resistivity, R(j), has been proposed as a key determinant of conduction velocity (CV). However, studies in connexin-gene knockout mice demonstrated significant CV slowing only with near-complete connexin deletion, and these findings led to the concept of a significant redundancy of myocardial gap junctions. We challenged this prevailing concept and addressed the hypothesis that there is a continuous relationship between R(j) and CV, each independently measured in human and guinea-pig myocardium. METHODS AND RESULTS R(j) and CV were directly measured by oil-gap impedance and microelectrode techniques in human left ventricular myocardium from patients with hypertrophic cardiomyopathy and in guinea-pig atrial and ventricular myocardium before and during pharmacological uncoupling with 20-µmol/L carbenoxolone. There was a continuous relationship between R(j) and CV in human and guinea-pig myocardium, pre- and post-carbenoxolone (r(2)=0.946; P<0.01). In guinea-pig left ventricle, left atrium, and right atrium, carbenoxolone increased R(j) by 28±9%, 26±16%, and 25±14% and slowed CV by 17±3%, 23±8%, and 11±4% respectively (all P<0.05 versus control). As a clinically accessible measure of local microscopic myocardial conduction slowing in vivo in the intact human heart, carbenoxolone prolonged electrogram duration in the right atrium (39.7±4.2 to 42.3±4.3 ms; P=0.01) and right ventricle (48.1±2.5 to 53.3±5.3 ms; P<0.01). CONCLUSIONS There is a continuous relationship between R(j) and CV that is consistent between cardiac chambers and across species, indicating that naturally occurring variations in cellular coupling can account for variations in CV, and that the concept that there is massive redundancy of coupling is not tenable.