The ultimate goal of defining cancer specific precursors is to facilitate early detection and intervention before the development of invasive malignancy. Unlike other malignancies involving the female genital tract such as cervical or endometrial carcinomas, precursor lesions of ovarian carcinomas have not been well characterised, resulting in a failure to develop effective screening programs. Recent clinicopathological and molecular studies have provided new insight into the origin and pathogenesis of ovarian carcinomas. It has been shown that ovarian cancer is comprised of different tumour types differing not only in morphology, but also in pathogenesis, molecular alterations and clinical progression. A dualistic model of ovarian carcinogenesis has been proposed. Type I tumours which include low grade serous, low grade endometrioid, clear cell, mucinous carcinomas and Brenner tumours, are generally indolent and tend to be genetically stable, although clear cell carcinoma would probably belong to an intermediate category. They demonstrate a step-wise progression from a benign precursor such as a benign to borderline tumour or endometriosis and are characterised by genetic aberrations targeting specific cell signalling pathways. Type II tumours comprise high grade serous, high grade endometrioid, and undifferentiated carcinomas as well as malignant mixed mesodermal tumours. They are clinically aggressive and exhibit high genetic instability with frequent p53 mutations. Mounting evidence suggests that many high grade serous carcinomas originate from the epithelium of the distal fallopian tube, and that serous tubal intraepithelial carcinoma (STIC) represents the putative precursor of these neoplasms. Low grade serous carcinomas arise via transformation of benign and borderline serous tumours, thought to be derived from inclusion cysts originating from the ovarian surface or tubal epithelium. Recently it has been suggested that papillary tubal hyperplasia may be a putative precursor lesion for serous borderline tumours. Both endometrioid and clear cell carcinomas develop from endometriosis, via alterations affecting different genetic pathways. The origin of mucinous and transitional cell neoplasms is not well characterised, although new data suggest a possible origin from transitional cell nests present at the tubal-mesothelial junction. Likewise, the pathogenesis of carcinosarcomas is also not well established because of their rarity but there is accumulating evidence that the carcinomatous component determines the course of the disease and gives rise to the malignant mesenchymal component. This review discusses recent developments in the pathogenesis of ovarian carcinoma, with particular emphasis on the putative precursor lesions that give rise to the major histological subtypes. Recognition of these lesions is not only important in improving the understanding of ovarian carcinogenesis, but it will also influence our approach to prevent, detect and treat these tumours.