Studies were conducted at Adana, in the Cukurova region of southern Turkey, to evaluate the effects of the rate and timing of application of soil-applied potassium (K) on cotton (Gossypium hirsutum L.) in 1999 and 2000. Potassium rates of 0, 80, 160 and 240 kg K2O ha(-1) were soil-applied in single treatments (all at early boll development) or in split treatments (1/2 at first square and 1/2 at first white flower; 1/4 at first square, 1/4 at first white flower and 1/2 at early boll development). Data collected in the two years indicated that application of 160 kg K2O ha(-1) produced significant differences in seed-cotton yield, lint yield and boll weight compared with the untreated control. The best combination producing the greatest yield was application of 160 kg K2O ha(-1) with all of the K soil-applied at early boll development. Cotton yields did not respond to K fertilization above the rate of 160 kg K2O ha(-1) under the production practices typically found in the region. For application of K at a rate of 240 kg K2O ha(-1) there was a marked difference in fibre strength between years in this study, but micronaire and uniformity ratio were not different amongst K rates within each year. When the total amount of K was applied at early boll development, higher yields, boll weights and lint turnouts were obtained compared with split applications, but the single application did not have a large impact on fibre properties.