AbstractOne method of correcting K deficiencies in cotton (Gossypium hirsutum L.) is by in‐row deep placement of K fertilizer. At present, the mechanisms involved in cotton yield response to deep placement of K have not been elucidated. A field study was conducted in 1990 and 1991 to evaluate root development and dry matter yield of cotton as affected by in‐row subsoiling and placement of K fertilizer. The experiment was located in central Alabama on a Norfolk fine sandy loam (fine‐loamy, siliceous, thermic Typic Kandiudult). Five treatments were compared: (i) check, without in‐row subsoiling; (ii) check, with in‐row subsoiling; (iii) 84 kg K ha−1 surface‐applied, without in‐row subsoiling; (iv) 84 kg K ha−1 surface‐applied, with in‐row subsoiling; and (v) 84 kg K ha−1 deep‐placed, in‐row. Penetrometer readings taken in 1991 demonstrated that the soil has a well‐developed traffic pan at a depth of approximately 15 to 38 cm. In‐row subsoiling disrupted the pan up to 25 cm away from the in‐row position. Root density measurements taken in‐row showed that root growth at depths > 20 cm was improved by in‐row subsoiling and K fertilization. Cotton root growth at depths > 20 cm was generally better for the treatment receiving the deep applied K. However, broadcast K in combination with in‐row subsoiling resulted in the highest productivity and K accumulation per plant. Results of this study suggest that, for cotton production in Alabama, deep placement of K is not superior to broadcast applications of K.