A prospective study was performed to determine the incidence of antinuclear antibodies in 214 normal pregnant women and in 50 age-matched controls. Serum samples of 23 pregnant women (10.7%) yielded positive results (1 + or more at a dilution of 1:20) in contrast with only one sample (2%) in the control group (p less than 0.05). Of the pregnant women found to be positive, five (9.2%) were in the second trimester, and 18 (13.4%) were in the last trimester. Only two also had positive anti-DNA antibodies. A review of their hospital records disclosed no reason to suspect systemic lupus erythematosus or other rheumatologic diseases nor any significant increase in neonatal morbidity of their infants. We conclude that the incidence of antinuclear antibodies in pregnant women is significantly higher than in nonpregnant women and that the finding probably does not correlate with any adverse clinical effect.