In order to understand, more fully, some of the factors affecting the S requirements of pastures, S³⁵-labelled gypsum fertilizer was applied to several field trials located on soils within the recent, yellow brown earth and high country yellow brown earth soil groups in Canterbury, New Zealand. Among the major processes studied were the incorporation of the S³⁵-labelled gypsum fertilizer into the soil organic matter, its movement in the soil and its uptake by pasture plants. The results obtained showed that on the same soil type, plant uptake of fertilizer S and its incorporation into the organic matter were greater under improved pastures. However, the movement of fertilizer S was relatively unaffected by the pastoral improvement. Rainfall conditions affected all three processes studied, particularly the downward movement of fertilizer S. Over the growing seasons of the first year almost all of the added fertilizer S remained within the upper 15 cm of the soil at a low rainfall trial. On a similar textured soil, receiving almost twice as much rainfall over a similar time period, fertilizer S was distributed evenly to a depth of 60 cm. Where the rainfall was similar, but soils differed in their water holding capacity, large differences in the downward movement of fertilizer S occurred. On a steepland yellow brown earth soil almost 70% of the applied fertilizer S was lost from the 0-45 cm soil depth two months after application. While on a recent soil, with a much higher water holding capacity, greater amounts of fertilizer S remained within the upper 45 cm over the same period.. Large differences in the downward movement of fertilizer S were shown by soils with contrasting retention capacities, but under similar rainfall conditions. On a yellow brown earth soil, with a medium S retention in the subsoil, fertilizer S was found not to have moved beyond 60 cm in the winter following the spring application. In contrast, fertilizer S penetrated beyond the 60 cm on a recent low S retentive soil. Under improved pastoral conditions, applying fertilizer S in late spring rather than early spring, to two different soils gave larger (almost twofold) residual effects the following year. Total recovery of the added fertilizer S in the soil-plant systems studied were generally greater than 80%. The pattern and rate of the downward movement of fertilizer S was little affected by fertilizer rates. Pasture recovery (%) of fertilizer S at the lower rate (22.5 kgS/ha) was slightly higher than that at the higher rate (45 kgS/ha). Fertilizer rates did not produce a consistent effect on the extent of fertilizer S incorporation into soil organic matter. The magnitude of the sulphur cycling pool was determined at each trial location by using the equilibrium specific activity of pasture. This pool was found to vary between trials and showed no relationship to their respective S response patterns. The rate of S cycling appeared to be more important. At some trials a strong correlation was found between the specific activity of the S in pasture and that for phosphate-extractable sulphate in particular soil depths. This relationship was used to assess indirectly, the likely soil zones from which plants extract S to meet their nutritional requirements. A new technique was developed to study, directly, the soil zone from which plants extract S to meet their-nutritional requirements. Carrier-free S³⁵, injected horizontally into various soil depths in a recent and a yellow brown earth soil, showed that both grasses and clovers recovered S³⁵ to a depth of 52 cm, in the spring. However, most of the S³⁵ uptake by plants was found to occur in the upper 30 cm (A horizon) in both soils. In the autumn, S³⁵ uptake experiments on these soils showed uptake to occur at 100 cm on the recent soil, and cease at 52 cm on the yellow brown earth soil. Growth cabinet experiments, using S³⁵-labelled pasture residue, recovered from the field plots, and added to various soils, showed that ryegrass recovery of labelled fertilizer S ranged from 6-16% and it varied according to soil type, degree of pastoral development, type and form of plant residue added. This extent of recovery represents only a small proportion (1%) of the fertilizer S recovered by pasture in the first year. The overall experimental results were considered in relation to both the S response patterns obtained from the present field trials and also those from other similar trials conducted in recent years in the Canterbury area. Some possible reasons for the variable S responses obtained were discussed and methods for improving the efficiency of fertilizer use, at farm level, were proposed.