Get reliable rangeland science

Three experiments were conducted to measure variability in sampling for dry matter and crude protein determination in small?plot cutting trials. At one harvest of a trial comparing two strains of Panicum maximum, triplicate samples were drawn from the massed harvested material from each plot. The crude protein contents of these samples were determined in duplicate. There was little variation between the crude protein determinations for a single sample, but marked variation in dry matter and crude protein percentages between samples from the same plot. This suggested the need for examining sampling procedures. Two sampling methods were compared in a trial comparing four grasses, a single sample of herbage being taken from every plot by each method. Sampling the standing herbage before harvest gave a lower estimate of dry matter content and an error mean square 42% less than taking the sample from the harvested material after massing. Three sampling methods were later compared in the same trial, triplicate samples being taken by each method. Sampling the standing herbage again gave the lowest estimate of error mean square when the results from the four grasses were pooled, but the pattern varied when the grasses were analysed individually. Pre?harvest sampling again produced a lower dry matter percentage than post?harvest sampling. It is concluded that further testing is required before pre?harvest sampling can be generally recommended.