A hundred buckets and some lengths of plastic tubing may not sound like the most high-tech array of equipment, but it’s pretty much all Kristin France needs to build ecosystems.

Based at the College of William and Mary in Gloucester Point, Virginia, France created the ecosystems to find out what happened to them when organisms migrated from one area to another. The loss of biodiversity is thought to damage the health and stability of an ecosystem, but few studies have considered the role of migration and immigration in the process, France says.

To find out more, France created a basic marine habitat consisting of sea water and seagrass in each bucket. The water was filtered so that only algae and other microorganisms made it through. The buckets were put in groups of five, giving France 20 ‘metacommunities’. Within these metacommunities, some of the buckets were connected together with plastic tubing and some were left isolated, so that in some cases organisms would be free to disperse between the buckets. “We wanted to incorporate the processes that create diversity in the real world,” says France.

Each bucket was then given 30 tiny crustaceans, which graze on the algae that grow on the seagrass. The actual number of grazer species put in each bucket varied, but over all each metacommunity received either low biodiversity in the shape of three species, or high biodiversity with eight species. France and her labmates spent about two months in the heat, setting up and monitoring the system — one of the reasons “people have not done this kind of experiment”, she smiles.

France spent the next six weeks sampling each bucket to see how the ecosystems were changing. “It took me about eight months to count and characterize everything under the microscope with the help of a fantastic technical assistant,” she says. But if the data collection was labourintensive, working out what the data meant was an even bigger challenge. The fruits of that labour appear on page 1139 of this issue.

Earlier studies had indicated that greater biodiversity leads to a more stable ecosystem, but France found that in some cases the oppo site was true. Increasing biodiversity often pro duced greater variability in organisms from bucket to bucket. The only time France saw a stabilizing effect from increased diver sity was when she looked at the ‘sum’ of all the buckets within a metacommunity. “So, in one way, we got the effect that we were expecting, that increasing biodiversity increases stability through time,” she says. “But only when we looked at the metacommunity scale.” How ever, this was true only for the isolated buckets — when grazers were allowed to move from bucket to bucket, the stabilizing effect vanished.

The results should help conservationists who are dealing with increasingly fragmented habitats. "It may be that to maintain ecosystem function you have to think about both biodiversity and the level of connectivity in different habitats," France says. "You may not be able to maximize both and also stabilize ecosystem function." France now wants to follow up her work by looking at how well-connected and unconnected communities respond to a disturbance, such as a sudden rise in temperature. "