The clinically established proarrhythmic effect of bradycardia and antiarrhythmic effect of lidocaine (10 μM) were reproduced in hypokalaemic (3.0 mM K + ) Langendorffperfused murine hearts paced over a range (80-180 ms) of baseline cycle lengths (BCLs). Action potential durations (at 90% repolarization, APD 90 s), transmural conduction times and ventricular effective refractory periods (VERPs) were then determined from monophasic action potential records obtained during a programmed electrical stimulation procedure in which extrasystolic stimuli were interposed following regular stimuli at successively decreasing coupling intervals. A novel graphical analysis of epicardial and endocardial, local and transmural relationships between APD 90 , corrected for transmural conduction time where appropriate, and VERP yielded predictions in precise agreement with the arrhythmogenic findings obtained over the entire range of BCLs studied. Thus, in normokalaemic (5.2 mM K + ) hearts a statistical analysis confirmed that all four relationships were described by straight lines of gradients not significantly (P>0.05) different from unity that passed through the origin and thus subtended constant critical angles, θ with the abscissa (45.8°±0.9°, 46.6°±0.5°, 47.6°±0.5°and 44.9°± 0.8°, respectively). Hypokalaemia shifted all points to the left of these reference lines, significantly (P<0.05) increasing θ at BCLs of 80-120 ms where arrhythmic activity was not observed (∼63°, ∼54°, ∼55°and ∼58°, respectively) and further significantly (P<0.05) increasing θ at BCLs of 140-180 ms where arrhythmic activity was observed (∼68°, ∼60°, ∼61°and ∼65°, respectively). In contrast, the antiarrhythmic effect of lidocaine treatment was accompanied by a significant (P<0.05) disruption of this linear relationship and decreases in θ in both normokalaemic (∼40°, ∼33°, ∼39°and ∼41°, respectively) and hypokalaemic (∼40°, ∼44°, ∼50°and ∼48°, respectively) hearts. This extended a previous approach that had correlated alterations in transmural repolarization gradients with arrhythmogenicity in murine models of the congenital long QT syndrome type 3 and hypokalaemia at a single BCL. Thus, the analysis in terms of APD 90 and VERP provided a more sensitive indication of the effect of lidocaine than one only considering transmural repolarization gradients and may be particularly applicable in physiological and pharmacological situations in which these parameters diverge.