A cotton field study established in 1986 was used to further evaluate the responses of cotton (Gossypium hirsutum L.) and soybeans (Glycine max (L) Merr.) to potassium (K) fertilizer rates and residual soil K levels on a soil with "low" subsoil K (Cope et al., 1981). The site on a Dewey silt loam (Typic Paleudults) had been in alfalfa (Medicage sativa L.) for 5 years prior to 1986. Cotton yield response to K fertilizer rates and timing from 1986 to 1989 were reported by Mullins and Burmester, (1990). During 1986 to 1989 two cotton varieties, Stoneville 825 a longer season wilt susceptible variety, and DPL 50, an early fruiting, shorter season variety with some wilt resistance, were grown. Cotton yields and K deficiency symptoms were used to determine if these varieties differed in susceptibility to K deficiency. In 1990, soybean was grown without K fertilization in the study area to evaluate soybean response to soil K fertility levels. In 1991, the cotton variety DPL 50 was planted in the study area. Half the area remained at residual K levels and the other half received the K fertilizer rates as applied in 1986 to 1989.

Cotton K deficiency ratings and yields in 1989 indicated that the faster fruiting cotton variety (DPL 50) and the longer maturity variety (Stoneville 825) were equally affected by K deficiency. In fact, Stoneville 825 had more severe K deficiency symptoms in the check (O lbs ac^-1) K treatment than DPL 50. In both varieties, fiber micronaire was increased significantly by K fertilizer applications. No other cotton fiber quality factors were found to be affected by K fertilization rates.

Soil samples were taken to a depth of 25 inches following the 1989 cotton harvest. High surface application of K fertilizer or plow-down applications had little effect on subsoil K levels indicating little downward K movement. All plots had subsoils that were "low" or "very low" in K.

Soybeans in 1990 showed no visual K deficiency symptoms and no yield increase with higher residual soil K levels. This supports the idea that cotton is more susceptible to K deficiency than soybeans.

In 1991 cotton yields were reduced by a drought from mid-July until September. Cotton yields, however, were still increased with all K fertilizer treatments compared to the check (0 lbs K₂O ac^-1). Plots fertilized in 1991 consistently produced higher seed cotton yields than cotton in the residual K plots, but differences were not significant.

Cotton response to rates and timing of K fertilizer has been inconsistent between years, with greatest response in years of higher cotton yields. Multiple annual application of K seems necessary to correct the K deficiency on this soil. In this test, spring applied K fertilizer was as effective as fall plow-down K fertilizer in correcting K deficiency in cotton.