The usefulness of PCNA staining in paraffin-embedded specimens was assessed. Next, it was clinically assessed in 21 patients treated with radiotherapy alone for cervical cancer and 47 treated with chemotherapy for ovarian cancer whether there was a correlation between GF, DNA ploidy and the effect of radiotherapy or chemotherapy or not. Furthermore, it was assessed whether there was a correlation between GF and prognosis or not. The results were as follows. 1. A strong correlation (p < 0.01) was noted between the rate of labeling with Ki-67 and that with PCNA in fresh specimens from the same patients and between the rate of labeling with PCNA in fresh specimens and that in paraffin-embedded specimens from the same patients. 2. In patients with cervical cancer, GF was significantly higher (p < 0.01) in patients who were in the G4 classification of Oboshi and Shimosato, following radiotherapy at total doses of less than 40Gy than in those who required irradiation at total doses of 40Gy or more to become G4. No significant difference was noted in the incidence of DNA aneuploid between the two groups. 3. In patients with ovarian cancer, GF in patients who showed CR was significantly higher (p < 0.05) than in those who were PR, NC or PD. No significant difference was noted in the incidence of DNA aneuploid between the two groups. The prognosis in patients with cervical cancer and ovarian cancer who had a GF of 35% or more was better than that in those who had a GF of less than 35%.