Application and evaluation of Dale's non-parametric method for detecting community structure through zonation

Dale's non-parametric method for analysing zonation was tested on three coa stal ecosystems in New Zealand. The method, intended to investigate communi ty structure, is based on a prediction that the upslope (i.e. top) boundary of one species will be followed upwards, more often than expected at rando m, by the downslope (i.e. bottom) boundary of another species. In the prese nt study, the predicted pattern was not found: only one transect out of 12 had significantly more top/bottom boundary (TB) contiguities than would be expected on a random basis. Rather, the opposite trend was seen in two tran sects; of a deficit of TB contiguities. One of these deficits was significa nt even after Bonferroni correction for the number of tests made. Such TB d eficits might be expected if there were community boundaries? and indeed TB deficits are predicted by several theoretical models of community structur e. The results contrast with Dale's own results, in which there were consid erably more TB contiguities than would be expected on a random basis. This confirms suggestions that the pattern envisaged by Dale is likely to be unc ommon.