Dispersion statistics and sample size estimates for monitoring mite populations in commercial apple orchards.
Sample size estimates for monitoring population densities of Tetranychus urticae and Panonychus ulmi in apple orchards in South Africa were derived from dispersion statistics in a 2-stage sampling system. Taylor's power law proved to be a more suitable model than the regression method of Iwao & Kuno. Using dummy variables in Taylor's power law, it was found that in some cases spray applications affected the dispersion statistics and, therefore, sample size estimates. The background variances were unstable from farm to farm and season to season, contributing to unstable sample size estimates. As the unstable background variances were thought to be partly due to differences in the size of the orchards on the different farms, it was suggested that in future orchards be sub-divided into blocks not exceeding 2 ha for monitoring mite population densities. There was a trend suggesting that the sampling error decreased more rapidly when the number of trees sampled was increased than when the number of leaves per tree was increased.