Considerations in experiments for the control of codling moth (Cydia pomonella (L.)) in apples

Abstract

A study was made of some aspects to be considered in experiments for the control of codling moth in apples. Examination of the number of fruit damaged by either deep entries or stings in samples of 100 fruit showed that the counts of damaged fruit could be considered as random samples from a Poisson (or binomial) distribution both within trees and also in plots up to a size of four test trees. An optimum plot size and sampling procedure was determined in experiments for the control of codling moth in apples. A single-tree plot with four samples per tree is optimum for unguarded plots, and plots of either one, two, three, or four trees with four samples per tree are satisfactory for guarded plots where the trees are hand-sprayed. Guarding seems an unnecessary safeguard and expense for hand-sprayed plots. Where "Air-Blast" spray machines are used, large plots with a double guard row are necessary, and it is sufficient to examine two to four samples per tree. It is suggested that Smith's heterogeneity law can be used to obtain the optimum plot size in experiments for the control of insect pests, and possibly of diseases on fruit trees. No reduction in experimental error was obtained by using the crop size of individual trees as a covariate in an analysis of covariance.