We studied 32 transverse left ventricular slices of myocardium from 16 pigs after 45 to 100 minutes of coronary artery occlusion followed by 180 minutes of reperfusion. Infarct area for each slice was determined as follows: (1) grossly, by triphenyl tetrazolium chloride staining of each slice, and (2) microscopically, by complete histologic sectioning of the triphenyl tetrazolium chloride-stained surface of each slice. Planimetry of necrotic and nonnecrotic areas was performed from tracings and photographs of triphenyl tetrazolium chloride-stained slices and from actual histologic sections. When triphenyl tetrazolium chloride and histologic measurements were compared, necrotic tissue area had decreased 11.4% +/- 15.0% (2.59 +/- 1.04 vs 2.09 +/- 0.86 cm2). Nonnecrotic tissue area decreased 20.6% +/- 24.0% (8.31 +/- 3.79 vs 5.16 +/- 2.73 cm2). In this model of ischemia followed by reperfusion, with fixation and processing, viable tissue shrank almost twice as much as necrotic tissue. This differential shrinkage introduces an error resulting in overestimation of infarct size by histologic quantitation.