For the carcinoma of the cervix the vaginal wall can be an object of cancer invasion, and it is closely associated with the postoperative recurrence at the stump of the vagina and also with the prognosis. In the present study the author carried out histological investigations of the vaginal wall of the patients who had Okabayashi's radical extensive hysterectomy in our clinic; and in Section Ⅰ are presented the results of investigations on the manner of cancer invasion into the vaginal wall in 160 recent unselected cases, and in Section Ⅱ the relationship between the recurrence at the vaginal stump and the vaginal-wall invasion as well as prophylactic effect of concomitant use of radium on the recurrence at the vaginal stump are described concerning 102 older specimens. Section 1. As for the invasion of the cervical carcinoma 62 cases (38.8%) out of 160 cases have been verified to have the cancer invasion, and there are some differences between the macroscopic and histological findings; and even in the 39 cases whose findings coincide with one another, in about one third of them the cancer invasion has been found 5 mm deeper in the histological examination than in the macroscopic examination. Next, the incidence of the cancer invasion into the vaginal wall is greater in more advanced age; more in the cervical canal cancer than cancer in the vaginal portion; more frequeut in Stage Ⅱ than in Stage Ⅰ; in a retrocessive type of growth; in the case where the outward appearance of the lesion is ulcerous; in the C.P.L. classification more in L-type; and more in the case where the cancer propagated to the endometrium than in the case without such apropagation; likewise in the case where the cancer propagated to the corpus uterus; and more frequent in the case with lymph node metastasis. As for the incidence of the cancer infiltration to various layers of the vaginal wall, the epidermal infiltration, the epidermal + submucous tissue, and the epidermal + submucous + muscular layers occupy two thirds of the total, and as for the invasion of each layer it is slightly more frequent in the epidermal layer. No difference can be observed in the frequency of the cancer invasion as regrads the sites of the vaginal wall (anterior, posterior, right, and left), and the cancer infiltration into the vaginal wall is always contiguous. Section Ⅱ. As regards the relationship between the cancer invasion into the vaginal wall and the recurrence at the stump of the vagina, out of 47 cases with recurrence 35 cases (74.4%) proved to have infiltration into the vaginal wall, and 8 cases had the cancer invasion penetrated as far as the vaginal stump. In those that had the recurrence at the vaginal stump the majority of them consisted of the vaginal cancer, cancer of Stage Ⅱ and Ⅲ, ulcerous exterior appearance, retrocessive type of growth, L Type, and moderately matured type. In comparing the 35 recurrent cases having the cancer invasion into the vaginal wall with those cases without recurrence at the vaginal stump, it has been clarified that the recurrence at the vaginal stump is more liable to occur in Type L with the vaginal-wall invasion and in the case whose distance between the tip of the infiltration and the stump end is less than 5 mm. Moreover, for the prevention of the recurrence at the vaginal stump taking the data in Section I into consideration, for the cases without vaginal-wall infiltration the conventional 3 cm-resection of the vaginal wall is indicated, while for the cases with vaginal-wall infiltration it is necessary to resect the vaginal wall over 1.8 cm. below the tip of macroscopic infiltration in view of the shrinkage of the specimens examined and the fact that histologically the infiltration goes deeper than macroscopic finding. In addition, when the cancer infiltration is marked and there is a danger of recurrence it is possible to prevent such a recurrence by radium therapy. In our laboratory the proximal dosage of radium used is 1,000 mg elh.