Abstract

No-tillage (NT) is becoming increasingly attractive to farmers because it clearly reduces production costs relative to conventional tillage (CT). However, many producers in southern Italy are reluctant to adopt this practice because NT can have contrasting consequences on grain yield depending on weather conditions. The effect of NT and CT on continuous durum wheat (Triticum durum Desf.) under rainfed Mediterranean conditions was studied, over a 3-year period (2000–2002) at two locations (Foggia and Vasto) in southern Italy. Yield, grain quality [thousand kernel weight (TKW), test weight (TW) and protein content (PC)] and soil water content were assessed. Higher yield was obtained with NT than CT in the first 2 years at Foggia. In contrast, mean yield and quality parameters at Vasto were similar for the two treatments, except in the third year in which CT produced more than NT (4.6 Mg ha−1 versus 2.9 Mg ha−1). At Foggia, TW and TKW were higher in NT than CT in all years. Highest PC was obtained under CT (19.6% and 15.5% for CT versus 14.7% and 11.4% for NT, respectively, in the growing season 2000–2001 and 2001–2002) indicating that grain was shriveled with low starch accumulation. At Foggia, where this study was part of a long-term experiment started in 1995, a strong correlation was observed between yield and rainfall during the wheat growing season. The coefficient of determination (R2) values for CT and NT were 0.69* and 0.31 ns, respectively, and the regression straight line crossed around 300 mm of rainfall. These results indicate that NT was superior below this rainfall value, whereas more rainfall enhanced yield in CT. We conclude that NT performed better at Foggia with limited rainfall during the durum wheat growing season. The superior effect of NT in comparison to CT, was due to lower water evaporation from soil combined with enhanced soil water availability.