Rain-fed dairy pastures on sandy soils common in the high rainfall (>800 mm annual average) Mediterranean-type climate of south-western Australia comprise the annual species subterranean clover (Trifolium subterraneum L.) and annual and Italian ryegrass (Lolium rigidum Gaud. and L. multiflorum Lam.). In wet years, clover becomes potassium (K) deficient and shows large dry matter (DM) responses to applied fertiliser K due to leaching of K in soil by rainfall. In contrast, ryegrass rarely shows DM responses to applied K. Many dairy pastures in the region are now intensively grazed to maximise pasture use for milk production, and nitrogen (N) fertiliser is applied after each grazing. It is not known if frequent applications of fertiliser N to these pastures changes pasture DM responses to applied K. Therefore, a long-term (2002–07) field experiment was undertaken on an intensively grazed dairy pasture in the region to quantify pasture DM responses to applied fertiliser K with or without applications of adequate fertiliser N (141–200 kg N/ha per year). Soil samples (top 10 cm of soil) were collected from each plot of the experiment each February to measure soil test K by the standard Colwell sodium bicarbonate procedure used for both K and phosphorus soil testing in the region.

When no N was applied, pasture comprised ~70% (dry weight basis) clover and 25% ryegrass, compared with ~70% ryegrass and 25% clover when adequate N was applied. Significant linear responses of pasture DM to applied K occurred in 3 of the 6 years of the experiment only when no N was applied and clover dominated the pasture. The largest response varied from ~1.7 to 2.0 t/ha DM consumed by dairy cows at all grazings in each year, giving a K response efficiency of between 8 and 10 kg DM/ha per kg K/ha applied. Significant pasture DM responses to applied N occurred at all grazings in each year, with ~2–3 t/ha extra DM consumed by dairy cows at all grazings in each year being produced when a total of 141–200 kg N/ha was applied per year, giving an N response efficiency of ~7–19 kg DM/ha per kg N/ha applied. Soil test K values were very variable, attributed to varying proportions of soil samples per plot collected between and within cow urine patches, containing much K, arbitrarily deposited on experimental plots during grazing. Soil test K values were not significantly affected by the rates of K applied per year. A re-evaluation of results from the major soil K test study conducted for pastures in the region confirm that ryegrass rarely showed DM responses to applied K, and that for clover, soil K testing poorly predicted the likelihood of K deficiency in the next growing season.