(CI-, NO•-, SO42-, oxalate, NH•-, Na +, K +, Ca 2+, Mg2+), trace metals and crustal elements (Fe, Mn, Ca, Mg, Pb, Cu, Ag, Cd), "black carbon" (BC), total carbon (TC); and in selected samples for aerosol humidity, for mono-and dicarboxylic acids and other polar organic components, and for cellulose as a tracer for plant debris. Aerosol mass balances for the individual samples were constructed from the sum of following groups: Humidity (determined as weight loss at 100øC with a microthermobalance), ammonium sulfate, C1-and NOj, organic material, BC, and soil dust. On the average, the fit between the sum of the determined groups of components and the gravimetrically determined mass was within 6%, whereas deviations ranged for individual samples from -12 to +27% relative to the gravimetrically determined mass. The main component in aerosol was soil dust (36%), followed by organic material (28%), and ammonium sulfate (27%). Humidity was 6%, BC was 2%, and C1-and NOj were i% of aerosol mass. The low BC/TC ratio of 0.09 indicated little influence from combustion sources. The diurnal trend of the BC/TC ratio of 0.06 during day and 0.14 during night indicated a daytime source for organic components with no BC associated. As there were no industrial or other anthropogenic sources evident, the daytime source of organic components is assumed to be gas-particle conversion from biogenic emissions. Plant debris and organic acids were the major analytically accessible groups; however, together they form only 7.2% of the carbonaceous material. With 0.4% other polar organics and 5.7% BC in this group, 86.7% of the carbonaceous material remains unidentified. In this work we present a detailed mass balance of the atmospheric aerosol of a biogenically dominated southern African site for "nonburning" conditions. The balance includes organic and elemental carbon as well as aerosol humidity. In addition to ion and elemental analysis, we apply a combination of thermal methods originally proposed by Puxbaum [1979a, 1979b]