Over the last decade, chlorine dioxide has been increasingly used for disinfecting drinking water in many countries. A guarantee for the protection of the consumer is the presence of a sufficient residual concentration of the bactericidal reagent in drinking water. Thus it is important to determine exactly and accurately the levels of chlorine dioxide at the tap. During water treatment and subsequent distribution, chlorine dioxide can undergo a variety of reduction and disproportionation reactions producing primarily chloride but also chlorite and chlorate, which have been shown to cause haemolytic anemia. Reliable analytical methods are needed to identify and determine levels of chlorine dioxide, chlorite and chlorate in drinking water. A procedure based on the use of indigo carmine for the determination of each species in natural waters is suggested in this paper. In phosphate buffer (pH 6.8), two moles of chlorine dioxide oxidize one mole of indigo carmine. The concentration of the bactericidal reagent can be determined by measuring the difference in absorbance of the dye at 610 nm before and after reaction with chlorine dioxide. This method is selective as chlorite and chlorate do not react with indigo carmine in phosphate buffer at pH 6.8. Although the spectrophotometric method can be used successfully used at levels of chlorine dioxide down to 30 μg/l, the determination of lower levels in tap water requires a more sensitive method such as an electrochemical stripping procedure. This analysis is based on the measurement of the decrease in the indigo carmine signal after addition of chlorine dioxide. The detection limit is around 1 μg/l. At pH = 2, one mole of indigo carmine reduces one mole of chlorite. Thus the chlorite concentration can be determined by measuring the indigo carmine absorbance at pH = 2. At pH = 0, indigo carmine reacts with both chlorite and chlorate. A measurement at pH = 0 allows chlorate concentrations to be determined since the decrease in absorbance due to the presence of chlorite can be calculated. The stability of indigo carmine absorbance has been studied. An indigo carmine solution prepared in phosphate buffer is stable over several days if kept in light-proof bottles. It is not surprising that the presence of chlorite and chlorate does not lead to a change in absorbance as they do not react with the dye at pH = 6.8. A slight decrease in absorbance of an indigo carmine solution containing chlorine dioxide is observed after about twenty hours. This means that the chlorine dioxide concentration has to be determined in the first hours, which follow the addition of the dye to the sample in order to avoid errors. Interferences can arise from other residual oxidants, which may also be used in water treatment, or from substances present in the sample, which may react with indigo carmine, chlorite and chlorate. Accordingly, we have considered the influence of humic substances, ozone and hypochlorite. The absorbance of indigo carmine at pH = 2 and at pH = 0 does not change in presence of natural organic matter (1 mg/l). Chlorite and chlorate react with humic substances but the kinetics are much slower than those of the reactions with indigo carmine. Errors arising from humic substances in chlorite and chlorate measurements are thus very weak. Ozone may interfere in analyses as it reacts with indigo carmine. However its existence in the distribution network is unlikely as it also reacts with chlorine dioxide, which is in excess, and chlorite to give chlorate. Hypochlorite causes errors in chlorine dioxide, chlorite and chlorate determinations as a result of a reaction with indigo carmine. In the case of chlorine dioxide determinations, errors can be eliminated by adding ammonia to the sample before indigo carmine. Once the validity of the procedures had been proven in synthetic media, the methods were applied to a natural water, that of the water distribution network of the city of Brest, France. The results have been compared with those of other analytical techniques.A procedure based on the use of indigo carmine for the determination of each species in natural waters is presented. The method is selective as chlorite and chlorate do not react with indigo carmine in phosphate buffer at pH 6.8. Results obtained with this procedure are compared with those of other analytical techniques.