Accurate soil-test based fertilizer recommendations depend on the correlation between soil test nutrient concentrations and crop response to fertilization. The overall objective of these fertilization trials was to correlate Mehlich-3 P (M3P) with relative winter wheat (Triticum aestivum L.) grain yield to define the critical M3P threshold range and P-fertilizer rates needed to optimize wheat grain yield. Nineteen trials were conducted on silt loams between 2002-2006 with wheat following various summer crops. Phosphorus fertilizer (0-96 kg P/ha) was applied shortly before or after seeding. Grain yield was determined at maturity and adjusted to 130 g/kg moisture content. In the 0- to 10-cm soil depth, the mean M3P among sites ranged from 9-71 mg/kg. Eight site-years had M3P >28 mg/kg, produced relative grain yields >92% when no P was applied, and showed no significant (P>0.10) grain yield differences among P rates. When M3P was <28 mg/kg, significant, positive responses to P fertilization occurred at 8 of 11 sites. The unfertilized control at the eight responsive sites produced relative grain yields ranging from 73 to 91%. Relative wheat grain yield was significantly and linearly related to M3P (r² = 0.40). Omitting three outliers having M3P values from 23-26 mg/kg resulted in a significant non-linear relationship and improved the r² to 0.64. The predicted critical ranges of M3P to produce 90-95% relative winter wheat grain yields were 35-50 mg/kg for the linear relationship and 24-36 mg/kg for the nonlinear relationship. Wheat grain yields will respond positively to P fertilization in most soils having M3P <20 mg/kg. Consideration of previous crop and/or other soil properties that may influence wheat response to P fertilization may enhance the accuracy of soil-test based fertilizer recommendations. Phosphorus fertilization trials will be continued to improve our understanding of the soil properties that influence wheat response to P fertilization.