Confocal scanning laser microscopy (CSLM) is increasingly being used to image antibody-labeled structures visualized with a fluorescent secondary antibody. Such digital images are routinely stored on computer and are well suited to quantitative analysis. Although theoretical aspects of CSLM imaging and resolution are well defined, information is lacking on the relationship observed between measurements of fluorescent antibody-labeled structures and the size of the same structures as determined by electron microscopy (EM). In the present study we examined this relationship for the cardiac gap junction. Data on the size of immunofluorescent-labeled gap junctions were acquired by two methods of analysis from CSLM images and compared statistically with measurements of gap junction size obtained by freeze-fracture EM. The freeze-fracture data were compared before and after exclusion of small junctions, corresponding to those that theoretically would not have been detected in CSLM analysis. The data obtained by the different methods were similar but not identical, reflecting the advantages and limitations of each technique. However, the comparison did indicate that with appropriate sample preparation and orientation, accurate and rapid analysis can be achieved by CSLM, particularly when digital semi-automated techniques are employed.