Comparisons of cropping systems often have more of an illustrative than an explanatory value, since there are so many factors differing. There is a risk that the conclusions that can be drawn from comparisons between very different systems are either self-evident or obscure.

In this case, two carbon-exporting systems are compared with a carbon-recycling one (manure). Soils in all treatments received equal amounts of carbon input. But manure input to soil is a more stable carbon source than fresh plant residues. Manure is produced by feeding, for example cattle, and the cattle gain energy from the more accessible carbohydrates, leaving more recalcitrant compounds in the manure. Therefore, it should be no surprise that soil carbon levels over a 15-year period increase more from adding recalcitrant (manure) carbon than from adding carbon in fresh crop residues.

The increase in soil carbon owing to legumes in the crop rotation may partly be due to differences in the quality of crop residues, but may also be due to differences in longevity of the crop, higher water uptake (drier soil gives lower carbon losses) or differences in soil cultivation. Usually, legume or grass leys increase soil carbon compared with annual crops, owing to these factors.

When looking at cumulative nitrate leaching, there was a 50% higher rate from the fertilizer-based system, although not statistically significant (P = 0.06). But the whole difference for the 15-year period is based on the observation that only one year out of the five measured had a higher leaching rate. This is interesting, but there is no attempt to discuss this difference.

Much valuable information can be obtained by comparing different cropping systems in long-term experiments. However, as long as the interpretation is biased and the information provided to the reader is incomplete, conclusions will remain more ideological than scientific.

Olof Andrén
Department of Soil Sciences, Swedish University of Agricultural
Sciences, Box 7014, S-75007 Uppsala, Sweden

Holger Kirchmann
Department of Soil Sciences, Swedish University of Agricultural
Sciences, Box 7014, S-75007 Uppsala, Sweden

Olle Pettersson
Department of Soil Sciences, Swedish University of Agricultural
Sciences, Box 7014, S-75007 Uppsala, Sweden

Tilman replies - Andrén and co-authors express scepticism about the possibility that 'ecological' methods may help solve agricultural pollution(2), but do not dispute the equality of yields from manured versus fertilized crops. I highlighted (1) experimental work on 'ecological' farming (2) not because it offered a definitive solution to agricultural pollution, which it did not do, but because it demonstrated the plausibility of approaches that have been ignored in most conventional research. Any practices that seemingly reduce pollution while maintaining yields and profitability are worthy of study.

Agricultural intensification during the past 35 years has led to a doubling of world grain production, but this required 6.9- and 3.5-fold increases in annual global rates of nitrogen and phosphorus fertilization, respectively, and a doubling of irrigated land (3). Use of pesticides also increased markedly, with many of these accumulating far from points of application (4). If these trends presage the future, the next doubling of global food production, expected within 35 years, will require a tripling of nitrogen and phosphorus fertilization and a doubling of irrigation (3).

Nitrogen and phosphorus losses from agricultural fields are already the major source of nutrient loading into freshwater and nearshore marine ecosystems (5), and are a major source of terrestrial nutrient loading (6). This pollution is having serious impacts on non-agricultural ecosystems (5,6).

Research is needed that pursues all reasonable approaches to this problem. The apparent distrust between conventional and 'ecological' schools of agricultural thought must not blind either side to novel insights, nor slow the development of solutions to a global problem.

David Tilman
Department of Soil Sciences, Swedish University of Agricultural
Sciences, Box 7014, S-75007 Uppsala, Sweden

References

1. Tilman, D. Nature 396, 211-212 (1998).
2. Drinkwater, L.E. et al. Nature 396, 262-265 (1998).
3. Tilman, D. Proc. Natl. Acad. Sci. USA 96 (in the press, 1999).
4. Blais, J.M. et al. Nature 395, 585-588 (1998).
5. Carpenter, S. et al. Ecol. Applications 8, 559-568 (1998).
6. Vitousek, P. et al. Ecol. Applications 7, 737-750 (1997).