To evaluate whether fetal cardiac measurements can be made in the second trimester, we examined a cohort of normal pregnancies between 12 and 18 weeks' gestation using state-of-the-art ultrasound equipment. We examined this population longitudinally at intervals of 2 weeks, as well as at 32 weeks' gestation. From the 4-chamber view we measured the ventricular and atrial cavity dimensions, the thickness of the ventricular walls and septum at end-diastole, and the annulus dimensions of the mitral and tricuspid valves. Using a variety of views we also measured the long and cross-sectional diameters of the atria, the aorta, the pulmonary artery and its main left and right branches, the ductus arteriosus, and the superior and inferior vena cavae. To test the frequency with which measurements could be made, we divided them into measurements that were clear and easy to define (statistically good), to those that were unclear (statistically bad), or those that were not measured at all (none). Data were then analyzed by regression analysis, analysis of variance, and covariance. The frequency of reliable measurements varied inversely with gestational age. The inflection point for measurements was approximately at 16 weeks. Data from this longitudinal study were evaluated against those obtained from our previous study. Because no statistical differences were found in measurements between these studies where they overlapped, the data were pooled into 1 large group and the mean and SEEs calculated for all variables. Our study demonstrates that with current transabdominal imaging, fetal cardiac measurements can be made reliably in normal fetuses from 16 weeks' gestation onward. The frequency of obtaining data in younger normal fetuses suggests it is unlikely that reliable observations can be made routinely in abnormal fetuses < 16 weeks old, although this might be possible in individual fetuses.