Introduction

Several methods have been proposed to detect the scale of pattern in vegetation; most of them analyze density data in strings or rectangular arrays of contiguous quadrats by examining how variance depends on the size of blocks of quadrats which are lumped together in the analysis (e.g., Greig-Smith 1952; Hill 1973; Usher 1975). In this chapter, we will review and illustrate the basic methods for studying the spatial pattern of a single species in one dimension along which there is no environmental gradient. The kind of data under consideration are therefore density or presence/absence data collected in a string of contiguous quadrats (see Chapter 2).

Data

We will begin by considering a standard pattern consisting of a regular square wave and let the scale of the pattern be B quadrat units. Throughout the transect, gaps of B quadrats, each of density 0, alternate regularly with patches of B quadrats, each with density d. There are several ways in which this basic pattern can be modified to be made less regular:

1. The pattern is ‘unbalanced’ with the patch:gap ratio different from 1:1 but the patch size (p quadrats) and the gap size (g quadrats) are both constant for the entire length of the transect. For a given value of d, unbalanced patterns have a lower intensity than balanced patterns. In forest communities, the spatial pattern of the canopy will often be unbalanced in this way with the patches of canopy being considerably larger than the gaps between them.