There are conflicting data about chemoresistance and prognosis in ovarian clear cell carcinoma (CCC). This could be due to significant interobserver variation in the diagnosis of CCC and other ovarian surface epithelial tumors containing clear cells. Thirty-two cases previously diagnosed as CCC, high-grade ovarian serous carcinoma (SC), and mixed surface epithelial carcinoma (SEC) with clear cell and serous components were reviewed by 4 gynecologic pathologists blinded to the original diagnoses. Interobserver reproducibility was evaluated. Each case was also assessed using immunohistochemical markers Wilm tumor 1, estrogen receptor, and p53. The interobserver reproducibility was greatest for pure CCC (kappa of 0.82), and lowest for the mixed SEC (kappa of 0.32). Moderate agreement was seen in the pure SC category (kappa of 0.59). All pure SC and most mixed SEC presented as stage III or IV diseases. Most pure CCC presented as stage I or II diseases. Immunoreactivities of the mixed SECs were similar to those of pure SC, but significantly different from those of pure CCC for Wilm tumor 1 (P=0.0011 for both components), estrogen receptor (P=0.0003 for clear cell component, P=0.0001 for serous component), and p53 (P=0.0062 for both components). The serous and clear cell components of mixed SEC showed higher mitotic rates than pure CCC (P=0.004 and P=0.023, respectively), but the mitotic rate of pure SC was similar to the mixed SEC. We conclude that (1) pure CCC is reproducibly diagnosed. (2) The diagnosis of mixed ovarian SEC with clear cell component is not reproducible. (3) Mixed SEC with clear cell and serous components show similar stage, mitotic activities, and immunoreactivities to those of pure SC, and likely represent SC with clear cell changes.