Thirteen long-term soil fertility experiments (circa 1928 and 1954) at five locations in Alabama have been in a residual P mode since 1982, producing crops on variable soil P levels from P rates applied prior to 1982. Soil test calibration for cotton using Mehlich-1 extract during 1992-1998 indicate that critical values in use by state soil testing laboratories are well within reason for new cultivars and modern cultural practices. There is no need to adjust the critical soil-test values currently in use in Alabama, 25 mg P/kg for soils with CEC < 9 cmol/kg (Coastal Plain soils) and 15 mg P/kg for soils with CEC>9 cmol/kg (Limestone Valley soils). Based on a comparison between ‘Deltapine NuCotn35B' (DP35B) and ‘Deltapine DP5690' (DP5690) cultivars at one location, there is no difference in responsiveness to residual soil P. However, the new Bollgard® variety, DP35B, averaged 80 pounds lint per acre per year more than the conventional variety. Soil-test P does not appear to change dramatically over a 15-yr period when no fertilizer P is applied. Decreases in Mehlich-1 extractable P at the higher residual P levels is probably associated with shifts in the soil P fractions in the soil rather than with crop removal, runoff, or leaching.