Use of paired plots and multivariate analysis for the determination of goat grazing preference

Abstract
Goat grazing is examined as a possible tool for grassland management. In this paper, goat grazing preferences are described. The proportion of biomass of particular species eaten by a goat was estimated on the basis of comparison of paired quadrats. The control quadrats were clipped before grazing and their biomass was compared with the biomass in grazed quadrats. Estimates of the proportion consumed were calculated for frequent species; they were significantly different among species; the most preferred species was Polygonum bistorta, a species with broad, soft, nutritious leaves. The least grazed species were Agrostis tenuis and Galium harcynicum. Various indices of food selectivity (e.g. forage ratio, electivity index) were calculated and their behaviour was compared. Canonical Correspondence Analysis and Redundancy Analysis with appropriate covariates reflecting the block design of the experiment were used to evaluate overall changes in species composition and yielded results similar to those from the direct estimation of the proportion consumed.