The problem of scaling in ecology
It is clear from the vegetation sampling literature (e.g. Curtis and Mcintosh, 1950) that ecologists have long been aware of the importance of scale. Scaling effects were seen as an obstacle to achieving properly representative data. The early efforts to identify the scale of aggregation of plants were cast in more positive terms ,(Greig-Smith, 1952, 1983). Unraveling the different scales at which underlying processes occurred was seen as a way of uncovering critical causes of ecological phenomena. More recently there has been work on the mechanical influences of size in allometric studies of animals (Schmidt-Nielsen, 1972, 1984) and plants (McMahon, 1975). The relationship between size, physiology and temperature of the habitat has been particularly well studied for animals (Porter and Gates, 1969; Porter et al., 1973). Scale also plays a very large role in the study of biomechanics (Alexander, 1975; Alexander and Goldspink, 1977). Detailed and extensive studies of fluid mechanics have kept this line of research on mechanical limits actively growing (Vogel, 1981, 1988). A main source of modem theory of scaling in biology comes from hierarchy theory, a branch of general system theory which has its origins in the works of Simon (1962), Pattee (1973) and Whyte et al. (1969). Hierarchy theory works on the tension between the contributions of the observer and the observed to phenomena. Rather than focusing on a methodology of observation, as did the vegetation scientists midcentury, or upon the scaledependency of the material system, as do the modem allometric biologists, hierarchy theory is a theory of observation, really a meta-theory which puts observation protocol in a general theoretical setting. Hierarchy theory emphasizes that scientists only deal with the mechanics of the material system through observation in phenomenological terms (Allen et al., 1987). Modem technology allows moving across scales of observation and analysis in systems with great detail over enormous extent (King et al., 1991; Turner and Gardner, 1991). Thus a rekindled interest in landscape ecology in North America is one of the implementations of a hierarchy theoretical approach. Issues of scaling are well known in the physical sciences, particularly in applied science where there is a scaling up from model designs to the final full-scale product. Despite considerable experience with small scale models, designers are still much relieved when the full scale prototype works. For example, research on the design of the hulls ofboats in model form is complicated by the fact that the smaller model must be tested in water, with its particular scaling effects. Substitution of another fluid may be able to compensate for the viscosity of water as it influences the small model system. In this way the model is tested in a fluid that might in some respects behave as would water with the full size hull, but there will be other changes like surface tension that will not allow the substituted medium to pass for water.