SYNOPSIS An automated assay suitable for estimating urinary oestrogens in pregnant women has been investigated. Fluorimetry was found to have considerable advantages over colorimetry. The fluorimetric assay was simpler, more precise, more sensitive, and eliminated the need for correction for non-specific chromogens; in the assay of oestriol in pregnant women there was no need for correction for non-specific fluorescence. Spectrofluorimetric and photometric analyses, recoveries, and reproducibility show that the method offers a robust means of providing values for urinary oestrogen in pregnant women on a scale of up to 100 tests a day, the time of the assay being one and a half hours. Muir (1967 and developed an automated assay for total oestrogens in pregnancy in which the Allen correction for non-specific chromogens was replaced by that of Fournier, Shields, Neil, Hayes, and Papineau-Couture (1966). Spectral studies of the effluent obtained during steady-state analysis suggested that this correction was justified. Brown, Mcnaughtan, Smith, and Smyth (1968) published a study of the estimation of oestrogens by the Kober colour reaction (Kober, 1931) coupled with the Ittrich extraction (Ittrich, 1958) using both colorimetry and fluorimetry. These workers found that at 120°C the Kober colour complex could be developed in five minutes and that a simple factor could be used to correct for non-specific fluorescence. We have repeated our original spectral studies and carried out spectrofluorimetric studies based on Brown's original observations. It was hoped that the use of fluorimetry would eliminate the need for two colorimeters and simplify the automated system. Methods Figure 1 shows the manifold used to compare Received for publication 15 October 1969. colorimetry and fluorimetry; for either the fluorimetric or colorimetric methods the other instruments were bypassed. Where necessary, urine samples were made up to a 24-hour volume of 1,500 ml. Five ml of diluted urine was hydrolysed with 5 ml of 50% hydrochloric acid at 100°C for one hour. After hydrolysis, the samples were placedin8 ml AutoAnalyzercups and sampledat20 samples per hour, with a sample:wash ratio of 1:3. The samples were segmented with ether, passed through two double mixing coils, after which the phases were separated. The ether phase was then segmented with saturated quinol in 50% sulphuric acid and entered the digestor; during its passage down the digestor the ether was evaporated off and the Kober colour was developed. The first heater was set at 1000 and the second heater at 1800; the motor gears were set at 6 rpm.