Correlation of Soybean Yield Response to Phosphorus Fertilization with the Mehlich-3 Soil Test.

Soybean is grown on more than 1.2 million ha annually in Arkansas on soils that usually have medium or lower Mehlich-3 P (M3P) soil-test levels. Our research objective was to correlate relative soybean yield with M3P. The goal was to verify current guidelines or develop more accurate soil-test based P-fertilizer recommendations. Phosphorus fertilization trials were conducted at 35 site-years between 2004 and 2009 using three to five P rates ranging from 0 to 98 kg P ha^-1. Sites were classified as P responsive (P<0.10) or unresponsive using analysis of variance. For each site, the mean percent relative yield of soybean receiving no P was calculated and regressed against mean M3P using quadratic and linear-plateau models. Mehlich-3 extractable P ranged from 3 to 107 mg kg^-1 among sites with 32 of the 35 sites having <50 mg M3P kg^-1. Ten of 35 sites responded positively to P fertilization. Nine of the 10 responsive sites had <23 mg M3P kg^-1. Soils having <11 mg M3P kg^-1 responded positively and frequently (5 of 6 sites) to P fertilization. In comparison, only 3 of 11, 1 of 7 and 1 of 5 sites having 11-20, 21-30, and 31-40 mg M3P kg^-1, respectively, responded positively to P fertilization. The quadratic and linear-plateau models were both significant, had r² values of 0.33, and predicted a critical M3P of 22 mg kg^-1 for soil samples collected from the 0-10 cm depth. Based on results from these 35 site-years, M3P predicts with reasonable accuracy soils that do not usually require P fertilization (>22 mg kg^-1) for near maximum soybean yields and those that require P when M3P is <11 mg kg^-1. Research continues to evaluate methods of improving recommendations for soils with 11-22 mg M3P kg^-1.