Abstract : OBJECTIVE: To address the advantages and drawbacks of quantitative polarized light microscopy for the study of myocardial cell orientation and to identify its contribution in the field. METHODS: Quantitative polarized light microscopy allows to measure the orientation of myocardial fibers into the ventricular mass. For each pixel of a horizontal section, this orientation is the mean value of the directions of all myosin filaments contained in the thickness of the section for each pixel of the section and is accounted for by two angles, the azimuth angle, which is the angle of the fiber in the plane of the section, and the elevation angle, which measures the way the fiber escapes from the section. The azimuth is accurately measured, and its range of definition is complete from 0 degrees to 180 degrees . The elevation angle can be defined only in the range 0 degrees to 90 degrees . It is accurately measured between 20 degrees and 70 degrees . From 0 degrees to 20 degrees , there is a systematic bias raising the measured values, and from 70 degrees to 90 degrees , the angle is not accurately measured. RESULTS: With this method, we validated Streeter's conjecture concerning the architecture of the left ventricle. We formulated a pretzel conjecture about the fiber architecture of the whole ventricular mass during fetal period. In our model, elaborated by visual analysis of registered maps of orientation, the fibers run like geodesics on a nested set of 'pretzels'. Next, the validity of the helical ventricular myocardial band model of Torrent-Guasp has been examined. It appears that the band model does not account for the patterns observed in our data during the fetal period. However, after the major events of postnatal cardiovascular adaptation, our data can neither discard nor confirm Torrent-Guasp's model. CONCLUSIONS: Present limitations of quantitative polarized light analysis can neither confirm nor discard the existing models of fiber orientation in the whole ventricular mass after the neonatal period. However, the problems of mathematical and experimental validation of these two models have been posed in a rigorous manner. Non-ambiguous fiber tracking and demonstration of these models will require significant improvement of the definition range of the elevation angle that should be extended to 180 degrees .