Regulations have been adopted in several countries of the European Union which prescribe that dental treatment waste water must be discharged via an amalgam separator device. Since the effectivity of the device strongly depends on the size of the amalgam waste particles, this size was evaluated in waste water samples from eight dental offices. Per sample, all solid particles were separated from the waste water by pressure filtration, then dried and divided into six fractions by being sieved over five sieves with decreasing mesh width. Of the particles in each fraction, the density was determined by picnometry, the mass by weighing, and the area and width by image analysis. For this analysis, width was defined as the dimension perpendicular to the length of the particles. By combining the density, area, and width determinations of all fractions, we obtained mass distributions per waste particle width of the samples. The proportional amalgam mass of the distributions was estimated with the measured density of the particle fractions and with a number of assumptions for the density of amalgam particles only and of other waste particles only. Each waste sample has its own characteristics with respect to the mass and density of the particle fractions. The size distribution of waste particles has a bimodal shape and consists of a distribution of small (width, 2 to 90 microns) and large particles (width, 160 to 5500 microns). For small particles with a width up to 60 microns, the influence of the assumptions on the estimation of the proportion amalgam of the waste mass distribution is minor when compared with the sampling error. By averaging the estimations over the samples, one can estimate the weight of amalgam particles with a width < 10 microns and < 50 microns, respectively, between 4 and 15% and between 15 and 30%. The smallest particles comprising 5% of the amalgam mass have an estimated width of up to 15 microns.