Application of diallel analysis to experiments in plant competition

Abstract

Graphical and statistical analyses based on the methods of Durrant (1965) are applied to data on plant competition experiments previously presented by Williams (1962) and McGilchrist (1965) and to new data from a comparable experiment in barley.

The primary objectives of the analysis are to determine: 1. whether the differences between species grown together are a function of the differences between them when grown separately (beta competition), (2) whether the individual species are uniformly increased or decreased when grown with the other species (alpha competition), (3) which species influence the other species most and which species are influenced most by other species, (4) to what extent the change of one species results in a compensating change in the species grown with it, (5) whether the mixtures give more or less yield than the unmixed species.

Data from both experiments are similar in showing large competitive effects in their reciprocal differences which are compensating, the larger species increasing and the smaller decreasing, but whereas the one could be interpreted in terms of beta competition and the mixture means tended to the smaller species, the other could only be interpreted in terms of alpha competition, and the mixture means tended towards the larger species. It is possible that the differences may be a reflection of the environmental differences imposed by the pot culture of the first, and the field plot conditions of the second.