This study assessed whether more potassium (K) was required for optimal growth and grain yield of cereal crops under drought and salinity than under non-stressed conditions. In 2011, three experiments on wheat (Triticum aestivum L.) with four K rates (0, 20, 40, 80 kg K/ha), four application times (0, 5, 10, 15 weeks after sowing, WAS) and two sources (KCl, K₂SO₄) were conducted in the central and southern grainbelts of Western Australia. The lack of plant response to K supply at the sites of Bolgart (36 mg K/kg at 0–30 cm) and Borden (25 mg K/kg at 0–30 cm), compared with significant gain in K uptake, dry matter and grain yield at Dowerin (29 mg K/kg at 0–30 cm), was not explained by differences in soil K levels. However, rain fell regularly through the growing season at Bolgart and Borden, whereas a dry spell occurred from stem elongation to grain development at Dowerin. The effectiveness of K application time followed the trend of 0, 5 > 10 > 15 WAS. In 2012, barley (Hordeum vulgare L.) was grown on a moderately saline (saturation extract electrical conductivity ∼4 dS/m) and low K (20 mg K/kg) farm in the central grainbelt and treated with 0, 20, 40 and 120 kg K/ha. Applying K increased K uptake but decreased Na uptake, especially at 120 kg K/ha. Plant growth and grain yield increased with K supply, but the difference between the K rates was relatively small, indicating possible partial K substitution by Na. Higher than normal fertiliser K supply on low K soils would enhance the adaptation by cereals to water-limited environments, but K-fertiliser management on moderately saline soils may need to account for both K and Na uptake and use by the crops.