Duplicate endocervical swabs were collected for detection of Chlamydia trachomatis by PCR (Roche Diagnostics). One swab was swirled in Specimen Transport Medium (Roche) for PCR testing and discarded. A saline aliquot from the other specimen, sent as a dry swab to the laboratory, was Papanicolaou stained to determine specimen adequacy, and the remainder was PCR tested. Significantly more (24%) PCR-positive results (118 versus 95; P < 0.001) were obtained,vith the dry specimens than with the swirled specimens when first tested. In addition, PCR-positive results were obtained with 107 (10.6%) of 1,007 microscopically adequate specimens but with only 3 (0.9%) of 341 inadequate specimens (P < 0.001).