It is currently a controversial issue whether epithelial ovarian cancers arise in the ovarian surface epithelium (OSE) or the fimbrial epithelium of the oviduct. The hypothesis presented here aims to reconcile these 2 views and provides a possible explanation for 2 questions arising: first, why tumors originating in the fimbriae and OSE, which are parts of different organs, express common features; second, why these epithelia are prone to neoplastic transformation whereas the remaining oviduct and the extraovarian mesothelium are not. We hypothesize that these questions relate to the common origin of the OSE and fimbriae in that region of the embryonic coelomic epithelium, which will eventually link the extraovarian mesothelium to the epithelium of the oviductal ampulla. OSE and fimbriae become separated during embryonic development but, like other transitional, interepithelial junctions in adults, this region might remain incompletely committed and thus prone to neoplastic progression. To define differentiation at the OSE-tubal junction, salpingo-oophorectomy specimens were stained immunohistochemically for mesenchymal differentiation markers of OSE and for epithelial markers and Pax8, characterizing oviductal fimbriae and ampullae. OSE and ampullae were distinctly different, but there was no sharp boundary between OSE and fimbriae. Rather, both mesenchymal and epithelial markers overlapped, and Pax8 and fimbrial epithelial markers diminished distally, near the OSE. The results support the hypothesis that the OSE and fimbriae are parts of a transitional epithelium of common origin rather than 2 independent sources of ovarian cancer, and suggest that their immature, incompletely determined phenotype contributes to their propensity to neoplastic transformation.