Several authors have suggested that estimating the levels of microalbuminuria will help in early screening for pre-eclampsia. The purpose of this work has been to look for the absence of microalbuminuria in normal pregnancies and to work out its predictive value for the risk of toxaemia of pregnancy when it does appear. The study was carried out on 257 women of whom 43 were controls and 214 women who were pregnant and had neither diabetes nor hypertension and had no kidney infections. The samples of urine were gathered in a 12 hour period of night and those that gave a positive reaction for albumin were rejected. RIA techniques were used to work out the levels of albuminuria and these were confirmed by immunoassay. We have compared microalbuminuria, the relationship between urine albumin and creatinine and the clearance of albumin in relationship to albuminuria (as defined by the relationship of albumin and creatinine clearance). We have calculated the sensitivity and the specificity and the prognostic value both positive and negative for these four parameters. Our results show that in a normal pregnancy there should not be any microalbuminuria, and on the other hand that if microalbuminuria does appear according to the four parameters studied, they are all equally sensitive for predicting pre-eclampsia. The relative clearance of albumin from the urine seems to be the most interesting parameter as far as we are concerned, and it could lead to early screening for toxaemia.