The Analysis of Pattern in Vegetation: A Comment on a Paper by D. W. Goodall

Abstract
Goodall (Aust. J. Bot. 9, 162-196) has suggested that variance: block size graphs can be interpreted in terms of regression of variance on spacing. The apparent fit results from (a) failure to allow for occurrence of overall trends in representation of species along transects, (b) the rising curve to be expected for random distribution, (c) the application of logarithmic transformation to the variance data. Use of more appropriate tests shows that the occurrence of peaks in the graphs does not correspond to random expectation. Goodall''s criticisms of interpretation in terms of peaks in the graphs, corresponding to pattern at various scales, are thus invalid.