Published soil test recommendations do not call for K fertilization on cotton when surface soil 1M NH₄OAc-extractable k levels are >100 ppm, or if <100 ppm K, when H₂SO₄-extractable K is >500 ppm. In field experiments, however, significant lint yield responses are obtained when surface soil (0-20 cm) K levels test well above these critical limits. A survey study was conducted at 25 sites in 1985 to evaluate the influence of soil K status and distribution in the 0-20, 20-40, and 40-80 cm depths on plant k status and lint yield using multiple regression techniques. Results indicated that lint yield was better correlated with 1 M NH₄-extractable K (K(E)) in the 20-40 cm soil zone (r= 0.78, P, 0.01) than with K(E) levels in the 0-20 or 40-80 cm zones (r=0.71 and 0.66, respectively). Lint yield was also highly correlated with the potential percent gravimetric water at field capacity (0.03 MPa) minus percent gravimetric water at the permanent wilting point (1.5 MPa). PAW in the surface 0-20 cm was better correlated with lint yield (r+0.76, P<0.01) than PAW in the 20-40 of 40-80 cm soil depths (r=0.72 and 0.55, respectively). A new soil K parameter named the potassium availability index (K(AI)) was defined as:

K(E)/percent gravimetric water at field capacity

Because soil extraction with 1 M NH₄ salts removes both exchangeable and soil solution K, correction of K(E) for differences in cation exchange capacity (as reflected by water retention at field capacity in soils with similar mineralogy) may improve assessment of actual plant available soil K. Stepwise multiple regression with lint yield as the dependent variable and K(E), K(AI), and PAW for each depth and related squared terms as independent variables gave the following regression equation: