A replicated cotton (Gossypium hirsutum) rotation experiment has been conducted for 22 years (1980-2001) on a Decatur silt loam (fine, kaolinitic, thermic, Rhodic Paleudults) in the Tennessee Valley of northern Alabama, USA. The highly productive soil with little disease and nematode problems resulted in cotton yield increases from rotations of generally less than 10% during the first 15 years of the study. A switch to no-tillage in all rotations except continuous cotton in 1995 greatly improved cotton yield response to rotations. From 1995 to 2001 cotton yield increases to rotation have averaged between 5% and 18%. In this study, yield increases due to rotations seem linked to increases in soil organic matter and consequent improvements in soil quality. From 1979 to 1994 using conventional tillage, the only rotation that produced a greater than 10% yield increase was cotton rotated with wheat ( Triticum aestivum) and double-cropped soyabean ( Glycine max). This rotation was also the only rotation that significantly increased organic matter levels under conventional tillage. From 1995 to 2001, all rotations were no-tilled and the greater yield increases to rotations can also be associated with higher soil organic matter levels. Wheat as a grain rotation or cover crop often produced the greatest yield increases to the following cotton crop. Under conventional tillage the wheat residue provided increased organic matter residue. With no-tillage the wheat cover crop reduced surface soil compaction. No-tillage and rotations that increased residue production were linked to increased cotton yields on this soil.
