Prematurity is the leading cause of infant morbidity and mortality in the United States, so research into the causes and prevention of preterm birth occupies the highest priority in the field of obstetrics. Before a prevention program can be tested, a reliable means of identifying pregnancies at risk must be available. Scoring systems, based mainly on demographic characteristics, were developed a decade or more ago. Initially, they seemed promising, but experience has shown them to be inadequate as clinical predictors. The most promising candidate as a predictor of preterm labor is fibronectin, a cellular glycoprotein normally present as a specific isoform in various fetal tissues (as well as in certain malignant neoplasms). Fetal fibronectin can be measured with an immunoassay using a monoclonal antibody, and in cervical secretions the critical level seems to be about 50 ng/mL. When levels exceed this cutoff, presumably reflecting disruption or other alteration of the