The palliative role of nonhormonal cytotoxic chemotherapy in the treatment of endocrine-resistant prostatic carcinoma has not been established. Conventional means of quantifying tumor response are most frequently not applicable in this disease because of the lack of measurable objective parameters to allow for a reliable estimation of antitumor effects. While this problem is not unique to prostatic carcinoma, this review illustrates its magnitude in this disease. Only approximately 5% of patients studied fulfill the various criteria for complete response (CR), partial response (PR), or both, while the vast majority of patients reported as responders are actually in the stable disease category. Stable disease is highly questionable as an indicator of antitumor response and should not be used as a criterion for response in conventional phase II studies unless it is convincingly demonstrated that it occurs as a result of treatment. A study design that may allow a more reliable assessment of the value of the stable disease category is described in the text. More effective means for assessing tumor responses and better instruments to measure aspects of quality of life are needed. Review of several prospective randomized clinical trials showed that no treatment program tested during the last decade resulted in a survival advantage when compared with a concurrently treated control group. Furthermore, in two such trials, four different single chemotherapeutic agents widely used in the treatment of this disease (cyclophosphamide, 5-fluorouracil, estramustine phosphate, and streptozocin) either alone or in combination, did not produce any prolongation of survival when compared to a no chemotherapy (standard treatment) control arm. Survival curves for endocrine-resistant patients fall within a relatively narrow and possibly predictable range that may be used as an additional endpoint in conjunction with response (CRs and PRs only) in phase II trials. More definitive evidence of therapeutic efficacy in this disease should derive from phase III trials using survival as one of the major endpoints. Because of the poor results observed with chemotherapy thus far, we suggest that the appropriate control arm for phase III testing in endocrine-resistant patients continues to be a no chemotherapy control arm consisting of a best symptomatic care or a uniformly applied second-line endocrine manipulation.