Objective: To study the pathophysiology of endometriosis. Design: Overview of observations on endometriosis. Setting: Not applicable. Patient(s): None. Interventions(s): None. Main Outcome Measure(s): The hypothesis is compatible with all observations. Result(s): Endometriosis, endometrium-like tissue outside the uterus, has a variable macroscopic appearance and a poorly understood natural history. It is a hereditary and heterogeneous disease with many biochemical changes in the lesions, which are clonal in origin. It is associated with pain, infertility, adenomyosis, and changes in the junctional zone, placentation, immunology, plasma, peritoneal fluid, and chronic inflammation of the peritoneal cavity. The Sampson hypothesis of implanted endometrial cells following retrograde menstruation, angiogenic spread, lymphogenic spread, or the metaplasia theory cannot explain all observations if metaplasia is defined as cells with reversible changes and an abnormal behavior/morphology due to the abnormal environment. We propose a polygenetic/polyepigenetic mechanism. The set of genetic and epigenetic incidents transmitted at birth could explain the hereditary aspects, the predisposition, and the endometriosis-associated changes in the endometrium, immunology, and placentation. To develop typical, cystic ovarian or deep endometriosis lesions, a variable series of additional transmissible genetic and epigenetic incidents are required to occur in a cell which may vary from endometrial to stem cells. Subtle lesions are viewed as endometrium in a different environment until additional incidents occur. Typical cystic ovarian or deep endometriosis lesions are heterogeneous and represent three different diseases. Conclusion(s): The genetic epigenetic theory is compatible with all observations on endometriosis. Implications for treatment and prevention are discussed. (Fertil Steril Ò 2019;111:327-40. Ó2018 by American Society for Reproductive Medicine.) El resumen está disponible en Español al final del artículo.