The effects of scale on ecological dynamics are of increasing concern in the study of both natural and experimental ecosystem. Enclosed experimental ecosystems ('mesocosms') are by definition smaller than their natural counterparts, and are subjected to restricted material and energetic exchange. Experiments and a quantitative analysis of published research were undertaken to assess responses of experimental estuarine ecosystems to variations in physical dimensions and to determine how effects of energy are manifest at different levels of ecological organization.

An assessment of the design characteristics of 360 published experiments indicated that: key scaling variables (such as complete physical dimensions) are often unreported in the literature; volume and duration vary with habitat, treatment and trophic level; replication and degree of exchange vary with size. An evident bias towards scaling cylindrical mesocosms for a constant radius/depth ratio necessarily affects light, benthic-pelagic coupling, and wall artifacts. These patterns in the literature indicate the need for explicit examinations of scale.

Experiments conducted in different depth and radius mesocosms indicated that primary productivity expressed per unit area was independent of depth under light-limited conditions, but was proportional to depth under nutrient-limited conditions. Since depth is an inherent component of natural aquatic ecosystems these represent 'fundamental scaling' patterns. In contrast productivity was relatively unaffected by wall area per unit volume indicating that in this case 'artifacts of scale' were not a determinant of whole ecosystem productivity.

Experiments in mesocosms subjected to three mixing levels revealed that while zooplankton abundance was quantitatively related to turbulence, community level responses were less pronounced, and no effects were observed on nutrients or whole ecosystem metabolism. In contrast, mesocosms held at five light levels exhibited significant population, community and ecosystem level responses. Introduction of artificial feedback between ecosystem metabolism and light energy input did not dramatically alter patterns of ecosystem development, but did result in novel behavior which can be interpreted as evidence for the importance of ecosystem level cause.

These experimental results and analyses indicate that distinct scale dependent ecological dynamics result from different physical dimensions and energy exchange, and that unique integrative processes operate at the whole ecosystem level; scale matters.