Is the observed recurrence rate after curative surgery for rectal cancer always a good measure of therapeutic improvements? In an attempt to answer this question, the rates of local and distant recurrences were studied in two complete series of patients operated on for cure for rectal carcinoma. One hundred one consecutive patients were followed for five years in one series and 231 were followed for at least 18 years in the other series. The recurrence rate in the first series was 39 percent and in the second, with the longer observation time, 54 percent. The local recurrence rates were 24 and 38 percent, respectively. Both local and total recurrence rates increased with the length of the follow-up period. This was especially true for patients with combined local and distant disease. Autopsy sometimes demonstrated recurrences, clinically undiscovered. It is concluded that completeness, long follow-up, and intensive search for recurrence, including a high autopsy rate, are factors that raise both total and local recurrence rates. All these factors are important to consider when comparing results of different treatment modalities.