The reversibility of acidification is being investigated in a full scale catchment manipulation experiment at Lake Gardsjon on the Swedish west coast using isotopes as environmental tracers. A 6300 m(2) roof over the catchment enables researchers to control depositional variables. Stable S isotope values were determined in bulk deposition, throughfall, runoff, groundwater and soil-extracted water during one year prior to and two years of experimental control. Data collected prior to experimental control suggest that the inorganic SO42- pool within the catchment has a homogeneous delta(34)S value of about +5.5 parts per thousand. Sprinkling of water spiked with small amounts of sea-water derived SO42- started in April 1991. The delta(34)S value of this SO42- is around +19.5 parts per thousand. Since April 1991, the SO42- concentration in runoff has decreased by some 30%, however, the delta(34)S value have increased by only 0.5 parts per thousand. This suggests mixing of sprinkling water S with a large reservoir of S in the catchment. Oxygen isotopes in SO42- suggest that less than one third of the SO42- in runoff is secondary SO42- formed within the soil profile. This is, however, no evidence for net mineralization of S. The SO(4)(2-)in runoff in the roofed catchment is a mixture of SO42- previously adsorbed in the soil, mineralized organic S and SO42- from the sprinkler water. Calculations based on isotope data indicate that the turnover time of S within the catchment is on the order of decades. Since SO42- facilitates base cation how, the acidification reversal will take a much longer time than concentration decreases of SO42- would suggest.