Several recent, high-impact ecological studies feature natural microcosms as tools for testing effects of fragmentation, metacommunity theory or links between biodiversity and ecosystem processes. These studies combine the microcosm advantages of small size, short generation times, contained structure and hierarchical spatial arrangement with advantages of field studies: natural environmental variance, 'openness' and realistic species combinations with shared evolutionary histories. This enables tests of theory pertaining to spatial and temporal dynamics, for example, the effects of neighboring communities on local diversity, or the effects of biodiversity on ecosystem function. Using examples, we comment on the position of natural microcosms in the roster of ecological research strategies and tools. We conclude that natural microcosms are as versatile as artificial microcosms, but as complex and biologically realistic as other natural systems. Research to date combined with inherent attributes of natural microcosms make them strong candidate model systems for ecology.