Cytomegalovirus (CMV) is probably the most common agent of prenatal infection of the newborn, and one of 20 congenitally infected newborns shows serious symptoms. It was therefore considered important to be able to differentiate primary CMV from reactivation in pregnant females. A urea denaturation test was used to distinguish primary from secondary rubella infection in which the urea is included in the wash step of the standard IgG ELISA. This resulted in the removal of low-avidity antibodies, which are the antibodies produced early in infection. A group of CMV IgM-negative and -positive sera were tested, and all but one showed moderate to high avidity, with an avidity index reading of more than 30%. Among a group of babies 3-12 months of age, who were CMV IgM positive, 55% (16 of 29) showed low-avidity CMV antibodies. A small group of renal transplant patients and patients with clinically and laboratory-confirmed CMV gave more or less predicted avidity index results. It appears that, with the method used at this laboratory, the urea denaturation test can be applied to CMV to determine primary infection or reactivation in the majority of cases.