SUMMARY1. The maximum rate of rise of left ventricular pressure (DP) and action potential duration (a.p.d.) were measured in closed-chest anaesthetized dogs with atrioventricular dissociation and fl-adrenergic blockade. Right ventricular stimulation was carried out with protocols consisting of a conditioning 'priming' period and a test period.2. When a single test stimulus was introduced at varying intervals after the priming period, DP was found to be maximal at 800-1000 ms. With this single test stimulus fixed at the optimum, DP was found to be a variable inverse function of the a.p.d. of the same beat; no positive correlation could be found between DP and a.p.d.3. When a second test stimulus at the optimum interval was introduced after the first, the DP (DP2) was found to be strongly dependent on that elicited by the first test stimulus (DP1); the relationship was positive, linear, independent of the method used to vary DP, and independent of whether DP1 was depressed or potentiated. The slope of the relationship was less than 1I0 and the line passed through the point where DP2 = DP1; this is the point of continuous stimulation at the optimum interval in a steady state. 4. This result is consistent with the hypothesis that the coefficient relating DP1 to DP2, at constant a.p.d. of the first test pulse (AP1), is an index of the proportion of the activator of contraction stored during relaxation of test beat 1 which is released again on beat 2.5. In order to test the hypothesis that the remaining contractility depended on the action potential of test beat 1, AP1 was varied by changing the intervals between the priming stimuli.