Human alteration of the global nitrogen (N) cycle has had major impacts on naturally N-limited ecosystems worldwide. Peatlands, dominated by peat mosses, Sphagnum species, represent one such sensitive ecosystem. I have studied how this ecosystem is affected by increased N availability, using a small-scale N fertilization experiment in combination with a gradient study of three peatlands with varying N deposition. I found both in the experiment and in the gradient a similar pattern of Sphagnum decline accompanied by an increase of vascular plants associated with enhanced N supply. For one common Sphagnum species - both in the experiment and in the gradient study - I also found an identical, linear increase in soluble amino acid N (NAA) accumulation. As soluble amino acids function as N storage compounds among Sphagna, NAA is a suitable measure for Sphagnum N status, and indicates accumulation of excess N not used for growth. My results show that NAA can be used as a sensitive indicator to signal N pollution before the slow, and gradual, regime shift from Sphagnum to vascular plant dominance is visible. In an N-uptake experiment using Sphagnum specimens from the three peatlands varying in N deposition, I found a reduced N-uptake by both investigated Sphagnum species from a high N deposition site, in south-western Sweden. This potential of Sphagna to adjust to high N loads through N uptake regulation will, however, not prevent tissue N accumulation, and as a result a shift from Sphagnum to vascular plant dominance. In general I found similar patterns of N induced changes both in Sphagnum tissue chemistry and vegetation structure in the experiment and along the gradient study. Thus, I conclude that long-term, small-scale field experiments seem to offer reliable estimates of both the direction and strength of key vegetation responses in Sphagnum dominated peatlands. This is likely related to the key role of Sphagna as ecosystem engineers. In the experiment I found a marked time lag in vegetation response to N application treatments. The closed Sphagnum carpet did not collapse until after eight years of continuous treatments. Another result was that dwarf shrubs, e.g. cranberry Vaccinium oxycoccos, first increased, but later declined due to severe attacks by fungal diseases. One important conclusion is that long-term, manipulative field experiments are necessary for our ability to understand how ecosystems will respond to environmental change.