In an attempt to determine the natural history of early invasive adenocarcinoma of the cervix, defined as depth of invasion of 5 mm or less, an extensive review of the literature was undertaken, together with recent unpublished material of the author. Many of the cases had to be extracted from series dealing with microinvasive squamous cell carcinoma. The pons asinorum for the pathologist is the differentiation between adenocarcinoma in situ and early invasion. The criteria for microinvasion are: 1.) obvious invasion to 5 mm or less; 2.) usually complete obliteration of the normal endocervical crypts; 3.) extension beyond the normal glandular field; and 4.) a stromal response characteristic of invasive carcinoma. Not all of these criteria are present in every case. In all 436 cases were collected. Allowing for vagueness of reports, 126 patients were treated by radical hysterectomy, and none had parametrial involvement. No cases of adnexal tumors were found in the 155 patients in whom one or both ovaries were removed. Of the 219 patients with pelvic lymph node dissection, five (2%) had metastasis. There were 15 recurrences and six tumor-related deaths in the 436 patients. Only 21 patients had conization as the only treatment, and none has suffered a recurrence. It appears that early invasive adenocarcinoma behaves in the same way as its squamous counterpart. Cold knife conization is acceptable treatment only when the cone biopsy has been adequately sampled and the margins are free, especially when preservation of fertility is an issue. Loop excision procedures obscure depth of invasion and margins and are not acceptable either for diagnosis or therapy. Multicentricity does not appear to require cylindrical cones. If hysterectomy is contemplated, removal of the adnexa, per se, is unnecessary.