Few scientists can say that their research project grew up as they did. But when ecologist Christopher Clark joined David Tilman at the University of Minnesota, St Paul, as a graduate student in 2001, Tilman's prairie grassland project to understand the effects of nitrogen deposition had already been running for almost 20 years. “I was 7 years old when Dr Tilman started it,” says Clark.

Nitrogen is an essential element for plants, but too much of it in the soil results in decreased plant biodiversity. Nitrogen also affects the global carbon cycle. Normally, soil holds twice as much carbon as the atmosphere. The presence of extra nitrogen may cause less carbon to be stored in soil, contributing to global warming.

During the past 50 years, fossil-fuel combustion and fertilizer use have greatly increased the amount of nitrogen deposition in land ecosystems. “It has been estimated that human activities rival all natural processes combined in terms of how much nitrogen is being put into ecosystems,” says Clark. “We're basically doubling the nitrogen supply globally.”

Tilman's project to study the effects that nitrogen is having on plant biodiversity began in 1982. He deposited varying amounts of nitrogen-containing fertilizer onto a patchwork of grassland plots in three fields of Cedar Creek Natural History Area in Minnesota every summer until 2004. The fields were chosen because they had fallen out of agricultural use, and had become dominated by a species-rich mixture of native grasses and forbs. During the spring and summer months, Tilman — and later Clark — and various teams of summer interns harvested and recorded the plants that had emerged in the various plots.

In their analysis, described on page 712, Clark and Tilman found that even the lowest amounts of nitrogen in their treatments, which mimicked rates of nitrogen deposition over much of the developed world, resulted in the loss of 1 in 6 plant species. This is a large attrition, says Clark, because there are hundreds of plant species at their sites. Contrary to expectation, there was little difference in the loss of biodiversity between sites that received high and low nitrogen input. Clark suggests that because most species in this region are adapted to grow in nutrient-poor conditions, nitrogen inputs at any rate above the low historical levels may have an effect on the plant community.

But the study's results do not bear only bad news. From 1992, as part of a second experiment, Tilman stopped treating half of the plots in one of the three fields with fertilizer. Thirteen years later, he and Clark found that changes in biodiversity in these plots occurred at the same rate as in plots that had never received any fertilizer. “It may take a while, but our ecosystem seems to be able to recover,” says Clark, now a newly minted postdoctoral fellow in Jianguo Wu's lab at Arizona State University in Tempe. He is about to embark on a biodiversity study of the Eurasian Steppe in China.

“Our findings suggest that if we start some sort of coordinated national–international effort we can either prevent or reverse some of these losses, which are probably occurring across much of the globe,” Clark says.