The effect of cobalt chloride oh zinc at different pre-treatment temperatures has been investigated by using a dual cavity platform, which makes differentiation between gas-phase and condensed-phase interferences possible. When the analyte and interferent are separated on the dual cavity platform, condensed-phase interferences disappear whereas gas-phase interferences continue to affect the sensitivity of the analyte. The effect of the addition of nitric acid on the atomic absorption signals of the interferent under the atomization conditions of the analyte, the atomic absorption and background signals for the analyte, the interferents and their mixed or separated solutions obtained in the atomization and pre-treatment steps, and thermal pre-treatment of the interferent at different temperatures before pipetting the analyte into the same cavity, served to elucidate the interference mechanism. The dominant interference mechanism is the formation of volatile zinc chloride upon reaction between the analyte and the interferent at low temperatures, which js expelled out of the tube by the gaseous hydrogen chloride generated in large amounts during the thermal hydrolysis of cobalt chloride hexahydrate. A condensed-phase-gas-phase reaction between the gaseous hydrogen chloride and the analyte may causes the formation of zinc chloride as well. Gas-phase reaction and/or expulsion mechanisms seem plausible but their effects are not very pronounced.