The pathogenesis of complete transposition of the great arteries (TGA) is still controversial because useful animal models have not been established. We previously reported that all-trans retinoic acid induced complete TGA at a high proportion in mice. The aim of the present study was to clarify the morphogenesis of the cardiac outflow tract in the retinoic acid-treated embryos destined to develop TGA. We first examined the morphology of TGA in mouse fetuses treated with retinoic acid to establish an animal model of TGA (experiment 1) and then examined the retinoic acid-treated embryonic hearts by means of ink injection and histology (experiment 2). All mouse fetuses and embryos showed visceroatrial situs solitus and d-ventricular loop. In experiment 1, among 45 embryos treated with retinoic acid 70 mg/kg at day 8.5 of gestation, 35 (78%) had TGA and 3 (6.7%) had a double-outlet right ventricle with a subpulmonary ventricular septal defect. In experiment 2, all hearts already exhibited d-loop at gestation day 8.5. At gestation day 9.5, conus swellings, composed of acellular cardiac jelly, where hypoplastic, and the conotruncal cavity was nonspiral or tubular. At gestation day 11.0, aberrant conus swellings located anteroposteriorly to give a straight orientation to the conotruncal cavity. At gestation day 12.0, side-by-side great arteries were transposed in that the aorta arose from the right ventricle and the pulmonary artery arose above the interventricular foramen. These results suggest that a reproducible animal model of TGA can be produced in mice by treatment with retinoic acid; that there was no loop anomaly, such as an A-loop or L-loop, in our model; and that hypoplasia of the conus swellings appears to be the primary event leading to TGA.