A nonisotopic probe (Gen-Probe PACE; Gen-Probe, Inc., San Diego, Calif.) for detection of Chlamydia trachomatis in endocervical specimens was evaluated in 344 women attending a dysplasia clinic or an obstetrics clinic and 158 women who visited an emergency room. For each patient, the probe, a tissue cell culture, and a direct immunofluorescent-antibody test (DFA; MicroTrak; Syva Co., Palo Alto, Calif.) were used. C. trachomatis was detected in 54 specimens by at least one method. Forty-four, 44, and 37 specimens were positive by culture, probe, and DFA, respectively, and 31 were positive by all three methods. Considering culture-positive plus both probe- and DFA-positive results as the "gold standard," we determined the overall sensitivity, specificity, and positive and negative predictive values of the probe to be 80, 98, 82, and 98%, respectively. These values were 94, 98, 84, and 99%, respectively, in emergency room patients and 71, 98, 80, and 97%, respectively, in clinic patients. The sensitivities, specificities, and negative predictive values of the DFA and probe were comparable. The positive predictive values of the DFA in all patients and in emergency room and clinic patients were 97, 100, and 95%, respectively. Given the number of probe-positive results that were not confirmed by culture, we do not recommend using the Gen-Probe PACE to screen for C. trachomatis in women with a low to moderate risk for infection.