Interaction strengths and spatial scale in community ecology: simulated quadrat-sampling and confinement experiments involving animals of different mobilities

Much recent effort has been directed at quantifying strengths of interactions between competitors, predators and prey, hosts and parasites, and mutualists. These measurements are used to quantify elements of ‘community matrices’, which are coefficients summarizing dynamics. In many cases, experimental manipulations are used to estimate certain interactions that are deemed to be likely to be of most dynamic significance (most ‘intense’), while other relationships within the community are estimated by using correlative means. These data are then linked by using statistical methods such as path analysis. The two forms of data are gathered by using very different methods: (1) experiments, usually involving confinement (or exclosure); and (2) quadrat- or transect-based information. A major issue is whether information from these two different sources are ‘compatible’, especially in relation to whether they have similar scalar responses. Can they be reliably mixed? Here, quadrat- and enclosure-based information are derived for the same model systems to evaluate whether scalar responses are consistent between the two forms of information. Interactions between pairs of three forms of forager are considered. For two of these forms, biases are generally consistent between enclosure and quadrat data indicating that they have similar scalar responses. However, for a third foraging form that has more ‘intelligent’ foraging and decision-making behaviour, major inconsistencies in deductions and biases arise. These latter results are discussed in terms of the implications for construction of community matrices.