To test the relation between the defibrillation threshold and the upper limit of vulnerability, the shock strength associated with 50% probability of successful defibrillation (DFT50) and that associated with 50% probability of reaching the upper limit of vulnerability (ULV50) were determined in 20 open chest dogs with use of the delayed up-down method, with pacing drive cycle lengths of 150 to 500 ms and either single 6-ms shocks (10 dogs) or 12-ms biphasic shocks (10 dogs) given at the mid-upslope, peak and mid-downslope of the T wave of electrocardiographic lead II. The shocks were given by means of a patch-patch configuration on the anterior and posterior surfaces of the heart, which was paced from a stimulating electrode attached to the left ventricular apex. Analysis of variance showed no statistically significant differences in ULV50 as determined with different pacing cycle lengths. For monophasic shocks, DFT50 (331 +/- 66 V or 5.8 +/- 2.7 J) was not significantly different from ULV50 determined at the mid-upslope of the T wave (318 +/- 64 V or 5 +/- 2 J). The correlation coefficients between the two values were 0.74 (p = 0.014) for voltage and 0.67 (p = 0.034) for energy. In contrast, DFT50 was significantly higher than ULV50 as determined at the peak of the T wave (219 +/- 43 V or 2.3 +/- 1 J) and mid-downslope of the T wave (200 +/- 38 V or 1.9 +/- 0.9 J). In three dogs, ventricular fibrillation could not be induced at the mid-downslope of the T wave with any baseline pacing (Si) cycle length.(ABSTRACT TRUNCATED AT 250 WORDS)