A method is described for the determination of bromides in biological samples using neutron activation analysis. The specificity of this method eliminates sources of interference from other halogens and contamination of analytical agents with extraneous bromide. In duplicate samples of blood, at varying concentrations of potassium bromide, the mean difference was 4.5% ±1.95 and the mean of the SD of twelve duplicate estimations was ±0.87. The 95% confidence limit of a single estimation was ±1.6%. Bromides are normal constituents of the human body, being present in ingested sodium chloride and food. The conventional way of measuring serum bromide is based on the colorimetric measurement of gold bromide formed by the reaction of this halide with gold chloride. Ucko (1936) and Conway and Flood (1936) have estimated the normal content of bromides in human blood and found that it was between 1 and 10 yug/ml. However, the gold chloride method has been reported to be insensitive and nonspecific (Wright, Smith and Black, 1958). Hunter (1953) proposed the use of a bromate-rosaniline reaction to estimate bromide in body fluids. This is sensitive and specific but requires an incineration step with alkali. Neutron activation analysis is a highly sensitive and specific method of detecting elements in biological material. Lyon (1966) gives the sensitivity for Br estimation as <0.002 ng. Samples of blood or tissue are placed in a source of neutrons to produce radioactive isotopes which can be separated chemically, positively identified and quantitated by plotting their characteristic gamma spectrum. Bowen (1959) and Belkas and Souliotis (1966) have applied this technique to quantitate bromides in biological samples. There is no exposure of the patient to radiation since the isotopes are produced after the samples are taken. It is known that debromination of halothane takes place since Rehder and colleagues (1967) have measured increased amounts of bromide excreted in urine after administration of halothane anaesthetics. Since one bromide ion is present in each molecule of halothane, the amount of bromide liberated can be used to quantitate metabolism of halothane. This paper describes a specific method of estimating bromide able to detect minor variations in small samples of blood. MATERIAL AND METHODS Irradiation of samples. Blood samples of 0.5 ml each were put in polyethylene snap-closure tubes of 1 ml capacity. These were stoppered and fused with a carbon-tipped soldering iron. The carbon tip was used in order to avoid tracer contamination from copper. The sealed containers were checked for leakage by immersion in boiling water. Standards of 200 /*g of bromide as potassium bromide in 0.5 ml of water were similarly sealed in polyethylene tubes. The tubes containing samples and standards were packed in aluminium cans, each containing four blood samples and one standard. These were sent for activation at the Universities Research Reactor at Risley, Lancashire, England. The cans were irradiated in a thermal neutron flux of 0.16...