A Quantitative Method for Analysing Competition where Individuals Cannot be Identified

Abstract

Methods of analysing the genetic diallel cross, for which different genotypes are mated in all possible combinations, have been modified for the study of competition or other interactions between indi-viduals or between subgroups, even if it is impossible to identify each participant. Two alternative analyses of variance provide a test for any overall effect on performance through combining in a group individuals differing in some respect, as well as distinguishing between more complex forms of interaction. In addition, analysis of the variances and covariances among the different groups can show (1) if there is any consistent interaction, (2) if any divergence from such consistency is due to a single type of individual and (3) if any correlation exists between the magnitude of the interaction and the performance of the individuals tested separately, either on the trait under consideration or on any other behavioural or physical variable.