Endometrial serous carcinomas constitute no more than 10% of endometrial adenocarcinomas, but frequently present at an advanced stage and have a significantly worse prognosis than the more common low-grade and intermediate-grade endometrioid adenocarcinomas. The neoplasm's potential for rapid tumor progression and the high mortality that is associated with advanced-stage disease underscore the importance of understanding endometrial serous carcinogenesis so that its precancers can be diagnosed and an effective therapeutic intervention can be administered. In this study, the authors summarize the current state of knowledge on endometrial serous carcinogenesis and propose a model for its development based on recent work from our group and published data from other researchers. In this model, endometrial serous carcinoma arises predominantly in the resting endometrium, manifesting first as p53 immunoreactive, morphologically normal endometrial cells (p53 signatures), evolving to endometrial glandular dysplasia (which is the first morphologically identifiable precursor lesion), then to serous endometrial intraepithelial carcinoma (a carcinoma with a noninvasive growth pattern in the uterus but which is not infrequently associated with extrauterine disease), and finally into fully developed serous carcinoma. Endometrial glandular dysplasia is a lesion, which can be diagnosed by routine microscopic evaluation, whose ablation or removal may potentially offer the opportunity to prevent the development of the associated malignancy. The diagnostic criteria, practical applicability, and evidentiary basis for the delineation of this lesion are studied.