Get reliable science

A natural pasture in which there was contagious distribution of the species was sampled using a single wheel-point apparatus on which the interpoint distance exceeded the size of individual plants but was less than that of the plant clusters. The standard error of mean basal cover calculated from repeated independent samples was lower than that e\pected from a binomial distribution. The standard errors for five levels of sampling, with and without stratification of the plot are presented and these can be used to predict the sampling intensity needed to achieve an acceptable standard error for each mean basal cover. Either an increase in the number of points sampled over the whole-plot or shatification of points within the plot reduced the standard error of the mean estimate of basal cover. At all levels of sampling, stratification of point samples gave a substantially lower standard error and was more efficient in terms of field sampling time, than an increase in the number of points sampled. -