The statistical problems of intercropping experiments (which involve the growing of two or more crops together) are investigated in this study. Measures of combined yield are discussed; the Land Equivalent Ratio (LER) is shown to be the 'best' index for intercropping. Problems that arise in the standardization of LER are investigated, and use of a single pair of divisors is recommended. The use of systematic designs are advocated for yield-density studies, to reduce the number of guard rows. A 3-way systematic design is proposed and methods of analysis are suggested. A regression model is employed for the combined yield data (LER), from which estimates of the optimum densities can be calculated. The study also deals with varietal trials in intercropping. Methods for reducing the large number of possible varietal combinations to be tested in the field and ways of reducing the block size are given. The field layout is discussed, and illustrated by examples. Stability measures that can be used in intercropping are derived and it is shown how they can be used in evaluating stable varietal combinations. It is also shown how information about the contribution to stability of each crop can be obtained. The best proportions of the component crops in the intercropping mixture is also investigated. Design and analysis for an experiment on proportions in conjunction with varying densities is given. / Ph. D. / incomplete_metadata