A Chemical Aqueous Phase Radical Mechanism (CAPRAM) for modelling tropospheric multiphase chemistry is described. CAPRAMcontains (1) a detailed treatment of the oxidation of organic compounds with one and two carbon atoms, (2) an explicit description ofS(IV)-oxidation by radicals and iron(III), as well as by peroxides and ozone, (3) the reactions of OH, NO3, Cl-2(-), Br-2(-), and CO3-radicals, as well as reactions of the transition metal ions (TMI) iron, manganese and copper. A modelling study using a simple box modelwas performed for three different tropospheric conditions (marine, rural and urban) using CAPRAM coupled to the RADM2-mechanism(Stockwell et al., 1990) for liquid and gas phase chemistry, respectively. In the main calculations the droplets are assumed asmonodispersed with a radius of 1 mu m and a liquid water content of 0.3 g m(-3). In the coupled mechanism the phase transfer of 34substances is treated by the resistance model of Schwartz (1989). Results are presented for the concentration levels of the radicals in bothphases under variation of cloud duration and droplet radius.The effects of the multiphase processes are shown in the loss fluxes of the radicals OH, NO3 and HO2 into the cloud droplets. Fromcalculations under urban conditions considering gas phase chemistry only the OH maximum concentration level is found to be 5.5 . 10(6)cm(-3). In the presence of the aqueous phase (r = 1 mu m, LWC = 0.3 g m(-3)) the phase transfer constitutes the most important sink(58%) reducing the OH level to 1.0 . 10(6) cm(-3). The significance of the phase transfer during night time is more important for the NO3radical (90%). Its concentration level in the gas phase (1.9 . 10(9) cm(-3)) is reduced to 1.4 . 10(6) cm(-3) with liquid water present. Inthe case of the HO2 radical the phase transfer from the gas phase is nearly the only sink (99.8%). The concentration levels calculated in theabsence and presence of the liquid phase again differ by three orders of magnitude, 6 . 10(8) cm(-3) and 4.9 . 10(5) cm(-3), respectively.Effects of smaller duration of cloud occurrence and of droplet size variation are assessed.Furthermore, in the present study a detailed description of a radical oxidation chain for sulfur is presented. The most important reactionchain is the oxidation of (hydrogen) sulphite by OH and the subsequent conversion of SO3- to SO5- followed by the interaction with TMI(notably Fe2+) and chloride to produce sulphate. After 36 h of simulation ([H2O2](0) = 1 ppb; [SO2](0) = 10 ppb) the direct oxidationpathway from sulfur(IV) by H2O2 and ozone contributes only to 8% (2.9 . 10(-10) M s(-1)) of the total loss flux of S(IV) (3.7 . 10(-9) Ms(-1)).