Topsoil samples (0--75mm) collected at the beginning ofeach growing season over a 35-year period from grazed pastures receiving annual superphosphate applications at rates of 0, 188, and 376 kg/ha per year were analysed for changes in soil organic nitrogen (N), total carbon (C), organic sulphur (So), hydriodic acid-reducible S (HI-reducible S), carbon-bonded S (C-bonded S), and readily-soluble plus adsorbed S. In addition, soils from subsoil depths (75-300 mm) were also sampled at the beginning of each growing season for 3 years to investigate the effects of superphosphateapplications on So, readily-soluble plus adsorbed S, microbial S, and arylsulphatase activity. Results obtained showed that over 95% of soil S was in So and the remainder was in readily-soluble plus adsorbed S. Applications of superphosphate to increase S fertiliser inputs from 0 toeither21 or42 kg S/ha per year, substantially enhanced the accumulation of So, HI-reducible S, and C-bonded S fractions. However, there was no further increase in these So fractions by increasing annual S inputs from 21 t042kg S/ha. Approximately 50% of this additional S was considered to be leached from the top 300 mm of the soil, which has a low capacity to adsorb S. The accumulation of So reached an apparent steady state after 25-27 years of superphosphate applications, whereas N and C attained a steady state at 3-4 and 9-12 years earlier respectively, thus suggesting that these elements were stabilised independently in soil organic matter. Both C-bonded S and HI-reducible S were the major So fractions constituting 50--60% and 40--50%, respectively. The proportion of So as C-bonded S decreased with depth, whereas the reverse occurred with HI-reducible S. The arylsulphatase activity reflected the previous topdressing history and decreased with depth. In contrast, microbial S was not affected by superphosphate applications and showed inconsistent trends with depth. It accounted for less than 3% of total soil S in all treatments.