Some Statistical Analyses for a Maize and Beans Intercropping Experiment

Several univariate and multivariate analyses were applied to observations from a single experiment involving sole crop and intercropped combinations of two maize ( Zea mays L.) and four bean ( Phaseolus vulgaris L.) cultivars. Some of the strengths and weaknesses of the analyses are indicated. For joint analyses on maize and bean yields, analyses for crop value or income, land equivalent ratios, and a multivariate analysis with maize yields as one variable and bean yields as a second variable, were performed on the data. The last analysis necessarily ignores sole crop yields. These different types of analyses provided insight into different aspects of an intercropping experiment. It was found that an intercropping system which maximized the yield of any one crop did not necessarily maximize total crop value or income or land equivalent ratio. This indicated that the type of analysis required clearly depended on the goals of the experiment, and that all these analyses may usefully be performed on an intercropping experiment. The conclusions from the multivariate analyses for this experiment were found to agree with those from the univariate analyses. All eight mixtures had a land equivalent ratio greater than one. Combined crop values computed for bean prices three times that of maize were larger for three of the eight mixtures than for the highest sole crop value for a tall maize cultivar. Considering total yield of maize plus beans, none of the mixtures yielded as high as the higher yielding maize cultivar. In considering combined yields in a multivariate analysis, yields of beans were relatively twice as important as maize yields in determining differences in total yield.