An analysis of factors affecting wheat yield on farmers' field in the Buberuka highlands of Rwanda
1990
Burleigh, J.R. (California State Univ., Chico, CA (USA). College of Agriculture) | Yamoah, C.F. (Arkansas Univ., Fayetteville, AR (USA). Dept. of Agronomy) | Eylands, V.J. (Illinois Univ., Urbana, IL (USA). Dept. of Agronomy) | Barasebwa, L. (ISAR, Ruhengeri (Rwanda))
From 30 farmers' wheat fields during each of two seasons, ten 0-15 cm depth soil samples were taken, bulked and analyzed by standard methods for pH (KCl), Al, H, C, total N, organic matter, P, Ca, Mg, Na, and K. Each field was also characterized by altitude slope and orientation (in degrees) and severity estimates were made three times on all foliar diseases. Yield and yield components (dependent variables) were regressed on soil and site characteristics and on disease estimates made at the soft dough growth stage (independent variables) singly and in combination for each season and for combined seasons to identify variables which were consistently and closely associated with yield and yield components. Principal components regression was used to identify collinearity among independent variables. Percent total nitrogen and exchangeable aluminum explained 35 % of the variability in tiller and ear number for 1988B, whereas % field slope and stripe rust severity comprised the best model for 1989A. No other variable or combination of variables gave higher R2 values or had larger standard partial regression coefficients. When seasons were combined, N and Al along with severities for powdery mildew and stripe rust and % field slope explained 55 % of the variability in tiller and ear numbers. Variables associated with grains/ear and yield showed more seasonal variability than variables for tiller and ear numbers, but Al and N emerged as important explanatory variables for grains/ear (season 1989A) and for yield (combined seasons). Regression coefficients for N, Al, slope, orientation and stripe rust severity usually were not significantly different among seasons and among yield components. Al and N emerged as significant and stable independent variables for all yield components and their standard partial regression coefficients usually were greater than for other independent variables, confirming their importance as selection criteria for wheat improvement in the Buberuka, whereas stripe rust and powdery mildew were implicated only in tiller and ear formation
Afficher plus [+] Moins [-]Mots clés AGROVOC
Informations bibliographiques
Cette notice bibliographique a été fournie par International Maize and Wheat Improvement Centre
Découvrez la collection de ce fournisseur de données dans AGRIS