Diallel analysis of competition between grass species

Abstract

SUMMARY Timothy [S. 50], meadow fescue [S. 53], perennial ryegrass [S. 23], Italian ryegrass [S. 22] and cocksfoot [S. 143] were grown together in all possible pair combinations at two levels of fertility, in a pot experiment. Dry weight of herbage and tiller counts were determined at 136 and 198 days from sowing. The data were analysed for competitive effects by the application of the diallel analysis technique of Durrant (1965). The competitive effects were largely compensatory in type with dominant species such as Italian ryegrass and perennial ryegrass increasing in mixtures, relative to their pure stand values, more than species such as timothy and meadow fescue declined. This resulted in a tendency for the mixture means to exceed their mid-constituent values. The competitive effects on the reciprocal differences were seen to be predominantly of the alpha type; that is, there was a constant increase or decrease in one or more species when grown with others. The alpha values were found to be correlated with the unmixed species values so that ƀ, the mean effect, wasappropriately calculated from alpha. The over-all value of ƀ was –1.216, which means that for each gram difference in weight between the species when grown by themselves there was 2.452 g increase in difference between the species when grown in mixtures, the larger species increasing and/or the smaller species decreasing. There were significant differences in ƀ values between harvests, the ƀ competitive effects being approximately twice as large in the more mature plants. Although the alpha competitive effects increased in magnitude with time the species order remained relatively constant, and in terms of competitive ability the species could be placed in the ascending order, meadow fescue, timothy, cocksfoot, perennial ryegrass and Italian ryegrass. Tiller number showed a similar pattern, but the ƀ values were smaller, and not significant. Nevertheless competition had differential effects upon the plant weight/ tillering relationship and it is possible that the resultant morphological changes would influence the re-growth quality of the species. The conclusions derived from the analysis of reciprocal differences in this work is compared with earlier examples.