Six different methods for assessing amniotic fluid bilirubin were compared in 111 samples from 37 patients. When the Liley methods were compared, the chloroform extraction and 5% correction methods resulted in 20% and 4% reductions, respectively, in the mean change in optical density at 450 nm when compared with the uncorrected mean (0.086 +/- 0.06 [p less than 0.05] and 0.104 +/- 0.07 [p, not significant] vs 0.108 +/- 0.08). This reduction was observed in spite of significant visual blood contamination being rarely noted. The predictions based on amniotic fluid analysis were compared with the first fetal or neonatal hematocrit. The chloroform-extracted change in optical density at 450 nm accurately predicted fetal status in all patients, whereas lesser degrees of accuracy were observed with other methods. Hydrops fetalis did not occur during the observation period, and fetuses needing transfusion (n = 5) or early delivery (n = 10) were indicated correctly. The chloroform extraction change in optical density at 450 nm accurately predicts fetal status, and its use should continue.