The emergence or the persistence of a high microbial diversity in natural ecosystems results in the balance between the probability of having the best adapted species of the local niche, and the colonization of new species through immigration from neighboring communities. Designing microbial microcosms seemed well adapted to delineate the respective role of the competitive exclusion at the local scale and the immigration events at the regional scale according to the theory of island biogeography. We started from an activated sludge of a domestic wastewater treatment plant as inoculum. It was repeatedly grown in either commercially available 96 wells microplates with 31 substrates (EcoPlate Biolog) or custom designed microplates. The productivity of each community was monitored on line with a microplate absorbance reader. The community structure and genetic diversity were assessed after each growth by CE-SSCP fingerprinting profiles. Successive inoculations allowed us to follow the dynamics during 100 generations. In order to gradually increase the immigration pressure, we added increasing amounts of cells from neighboring wells (with bacteria having different history) for the successive inoculations. The inoculum diversity first decreased then stayed constant at equilibrium, but differed according to the substrate used (as expected by the niche theory). The immigration from the regional pool induced a larger diversity as compared with the undisturbed communities, but a slightly lower performance. These results were in accordance with the island biogeography theory; the dispersal of non adapted species increased the apparent local diversity but encountered unfavorable conditions. We were able to follow the dynamics of many communities in parallel with a perfect control of environmental conditions. This experimental set up allowed us manipulating artificial ecosystems to test whether some fundamental ecological theories applies to microbes.