Natural microcosms in ecology: fulfilling the promise of model systems?

Natural microcosms (NMs) have been proposed as model systems for ecology based on their ubiquity, ease of study and natural context. We assess whether this potential has been met by examining 824 studies of six NMs (bromeliads, moss patches, nectar microbiomes, pitcher plants, rockpools, treeholes). Of these, we judged 314 studies to use NMs as model systems, testing a broad range of ecological theories. Although these theories spanned large biological, spatial and temporal scales, most studies concentrated on community-level niche processes occurring at local scales and within a generation-especially trophic control theories. However, metacommunity theory, which integrates over spatial and temporal scales, was also commonly studied. NMs were particularly effective in studying environmental stressors, often in combination with multitrophic effects or responses. While NMs have many advantages as model systems, there are limits to which theories can be tested due to specific life-histories of their inhabitants, uniqueness of certain ecosystem processes and frequency of disturbance. Particularly under-represented were tests of behavioural and ecosystem theory, long-term processes and impacts of invasive species. Greater use of molecular methods, community science and collaborative research networks could enable NMs to reach their full potential as model systems. This article is part of the theme issue 'Life in natural microcosms'.