Introduction

The most important criteria that determine how sampling will proceed in the study of spatial pattern are the question being asked and the scale at which we wish to answer it. Secondly, the kind of analysis that is required to answer the question must be considered because particular methods of analysis require certain kinds of data. Then, the sampling method will be determined by the interaction of a number of factors including the morphology, size and density of the plants of interest; the topography, accessibility, and area of the study site; the availability of time, money, technological and field assistance. It will be influenced fundamentally by whether the spatial pattern is to be treated as the arrangement of points in continuous space or as a mosaic of domains. We must also consider how much disturbance the sampling technique will cause, because we will want to minimize disturbance in long-term studies or in ecologically sensitive areas.

The methods used will also depend very much on whether the focus is on the spatial pattern of plants relative to a fixed frame of reference, on the elucidation of a community's response to an environmental gradient, or on the spatial arrangement of plants relative to other plants (species association). Kenkel et al. (1989) make the important point that the considerations for sampling design that are traditionally emphasized in statistics textbooks may not apply in studies of spatial pattern, because they are designed to provide efficient and unbiassed estimates of parameters such as mean cover or diversity.