Endometriosis, defined histologically as the presence of endometrium-like glands and stroma outside the uterus, is a chronic, invasive and metastasising disease. It shares features with malignant tumours (invasion and metastasis) but is not neoplastic. Despite the fact that endometriosis is one of the most frequent gynaecological diseases, it is under researched, puzzling and highly debated. The aetiology and pathogenesis is little understood although it is agreed that implantation, at least in many cases, is responsible for endometriosis. This theory advocates retrograde menstruation as the underlying phenomenon, where cells of the menstrual efflux provide the cellular source for endometriotic lesion formation. Causative therapy and non-invasive diagnostics of endometriosis do not exist. Thus, there is a substantial but unmet need for molecular and cellular research to unravel the pathogenic mechanisms of endometriosis as a basis for developing novel diagnostic and therapeutic concepts. In this revi ew, we specifically focus on the cellular basis of lesion formation, the possible modulation of this by cytokines and other factors and the characteristics of endometriotic cells in terms of invasion and metastasis. Considering available experimental information, we concentrate on arguments and ideas in favour of an endometriotic founder cell population exhibiting substantial plasticity for differentiation and self-renewal. Perhaps present in the menstrual efflux or arising by metaplasia (a complementary theory to implantation), this cell type might respond to stimuli present in the ectopic host environment and establish the endometriotic phenotype.