Bitumen has been used as a matrix for inorganic radioactive waste salts. If this waste would appear to be unacceptable for deep geological or shallow land disposal, a stabilizing chemical re-treatment could be considered. In this framework, we present some lab-scale experiments on the removal of the maltenes and water-soluble salts from a waste simulate mixture of blown bitumen (Shell type R85/25) plus water-soluble and -insoluble salts. Two secondary waste streams are produced, that is, an organic waste containing the maltenes and an aqueous waste mainly containing soluble (nitrate) salts. The radionuclides are concentrated in the remaining mixture of asphaltenes and water-insoluble salts. If one waste salt composition is used, the relative elemental composition of the two secondary waste streams is constant over the experimental range presented in this paper. The quantitative distribution of the elements among the obtained fractions depends on the bitumen-salt concentration: at low {bitumen + salt} to {solvent + nonsolvent} ratios, the purest organic secondary waste stream is produced. A seemingly inverse relationship between the amount of maltene ash on one hand and the settling velocity of asphaltenes and salts on the other hand is observed. If this relationship could be quantified, long and laborious separation experiments could be replaced at least partially by turbidimetric analysis of the settling velocity.