A system for monitoring multiple scatter during a clinical Compton scatter densitometry measurement of bone density is described. Multiple scatter from the measurement site was measured using a supplementary collimated detector positioned so that only multiply scattered photons could enter the detector. The data from the detector were used to form a multiple-scatter correction factor (mcf) to correct for the bias attributed to multiple scatter. The results of experimental and computer simulations are presented which demonstrate the relationship between the multiple-scatter reading and calculated mcf values. The influence of bone size on the values of mcf in large measurement sites, such as the femoral neck, was shown to be negligible. A simulation was used to produce a visualization of the multiple scatter in order to ascertain the optimum position of the supplementary detector. This technique was shown to be a rapid and accurate method of measuring the multiple-scatter bias and suitable for use during clinical CSD measurements.