Abstract
Examination of pattern within superficially homogeneous vegetation is more revealing of relationship between species performances and controlling factors than if obviously different communities are compared. Both vegetation and environmental factors are recorded in grids or transects of contiguous samples and the variance analysed into parts appropriate to different'' block sizes. Peaks in the mean square/block size graph indicate scales of pattern present. Length of transect is determined by maximum scale of pattern to be detected, size of basic unit by the smallest scale, and number of units by intensity of pattern and abundance of species involved. Sig-nificance of peaks is assessed by (a) consistency between observations, (b) use of calculated significance bands. Terms for co-variance with position may be deducted to allow for trends along the transects. The relation between patterns of different species may be examined by the use of correlation coefficients for different block sizes.

This publication has 1 reference indexed in Scilit: