A new regression analysis was proposed to evaluate the degree of spatial he
terogeneity for individual species comprising a plant grassland community.
The weighted average of the heterogeneity value of all the species comprisi
ng the community provides a measure of community-level heterogeneity. A fie
ld survey was carried out, as an example, in order to analyze the spatial h
eterogeneity of a pasture with grazing cows, using 100 quadrats 50 cm x 50
cm, each of which was divided into four smaller quadrats 25 cm x 25 cm, on
a 50 m long line-transect. The frequency of occurrence for all the species
in each small quadrat was recorded. The regression associated with the rati
o of the theoretical and observed variances of occurrence counts was used t
o analyse the frequency distribution of species in a pasture community. A g
ood fit to the regression for the whole community was obtained. These resul
ts indicate that (i) each species in the example was distributed more heter
ogeneously than a random pattern; and (ii) the regression could well descri
be the spatial heterogeneity of the grassland plant community. In most of t
he observed species, spatial heterogeneity is often characterized by specie
s-specific propagation traits and the architecture of plant bodies. Thus, t
he spatial patterns of a grassland community can be evaluated in detail by
this power-law approach. This measure is suitable for field surveys and com
parative studies of grassland communities, and for other plant communities
that are generally short in height.