Abstract
I examine here the epistemological and operational consequences of the use of ecological variability in comparative analyses. Ecological variability is used in comparative analyses to describe the possible states of ecological properties and to search for patterns in nature. Success in these goals requires (1) an explicit recognition of the structure and development of ecological variability and (2) appreciation of the important epistemological constraints inherent to the different methods used in the study of ecological variance. Ecological variance is often hierarchical in nature, containing a nested array of scales of variation, each scale contributing peculiar components to the global variability and presenting patterns that may differ across scales. The approach used to examine ecological variability needs, therefore, to be able to cope with the different types of patterns possible (e.g., continuous relationships, boundary conditions, threshold relationships, etc.) and to allow for changes in the existing patterns when sampling across scales. Thus, I contend that increased flexibility in the search for patterns in ecological variability is essential for successful comparative analysis.