The Detection of Small-Scale Patterns in Vegetation

SUMMARY (1) Pattern at the scale of a few centimetres may be important for understanding the determinants of species-distributions, but may be difficult to demonstrate using quadrats. (2) A transect sampling system is described, together with two methods of analysis, for detecting small-scale positive and negative associations between species. (3) The first method considers the frequency of juxtaposition of species, and tests whether this frequency is significantly greater or less than a random expectation. (4) The second method divides the transect into regions, and tests whether the number of occurrences of each species within the regions of occurrence of each other species is significantly greater or less than outside those regions. (5) Indices of association may be generated by either method. (6) The second method may also be used to examine the effect on association patterns of changing the size of regions, and to determine whether some species may show stronger patterns when both of two other species are present than when either one alone is present. (7) Both methods were used to analyse transect data collected in an old-field in Illinois, and the results were compared. These methods made it relatively simple to detect patterns of association for the stems and tillers of the seven most abundant species at a scale of 4 cm.