Vascular plant N natural abundance in heath and forest tundra ecosystems is closely correlated with presence and type of mycorrhizal fungi in roots.

In this study we show that the natural abundance of the nitrogen isotope 15, δN, of plants in heath tundra and at the tundra-forest ecocline is closely correlated with the presence and type of mycorrhizal association in the plant roots. A total of 56 vascular plant species, 7 moss species, 2 lichens and 6 species of fungi from four heath and forest tundra sites in Greenland, Siberia and Sweden were analysed for δN and N concentration. Roots of vascular plants were examined for mycorrhizal colonization, and the soil organic matter was analysed for δN, N concentration and soil inorganic, dissolved organic and microbial N. No arbuscular mycorrhizal (AM) colonizations were found although potential host plants were present in all sites. The dominant species were either ectomycorrhizal (ECM) or ericoid mycorrhizal (ERI). The δN of ECM or ERI plants was 3.5-7.7‰ lower than that of non-mycorrhizal (NON) species in three of the four sites. This corresponds to the results in our earlier study of mycorrhiza and plant δN which was limited to one heath and one fellfield in N Sweden. Hence, our data suggest that the δN pattern: NON/AM plants > ECM plants ≥ ERI plants is a general phenomenon in ecosystems with nutrient-deficient organogenic soils. In the fourth site, a␣birch forest with a lush herb/shrub understorey, the differences between functional groups were considerably smaller, and only the ERI species differed (by 1.1‰) from the NON species. Plants of all functional groups from this site had nearly twice the leaf N concentration as that found in the same species at the other three sites. It is likely that low inorganic N availability is a prerequisite for strong δN separation among functional groups. Both ECM roots and fruitbodies were N enriched compared to leaves which suggests that the difference in δN between plants with different kinds of mycorrhiza could be due to isotopic fractionation at the␣fungal-plant interface. However, differences in δN between soil N forms absorbed by the plants could also contribute to the wide differences in plant δN found in most heath and forest tundra ecosystems. We hypothesize that during microbial immobilization of soil ammonium the microbial N pool could become N-depleted and the remaining, plant-available soil ammonium N-enriched. The latter could be a main source of N for NON/AM plants which usually have high δN. In contrast, amino acids and other soil organic N compounds presumably are N-depleted, similar to plant litter, and ECM and ERI plants with high uptake of these N forms hence have low leaf δN. Further indications come from the δN of mosses and lichens which was similar to that of ECM plants. Tundra cryptogams (and ECM and ERI plants) have previously been shown to have higher uptake of amino acid than ammonium N; their low δN might therefore reflect the δN of free amino acids in the soil. The concentration of dissolved organic N was 3-16 times higher than that of inorganic N in the sites. Organic nitrogen could be an important N source for ECM and, in particular, ERI plants in heath and forest tundra ecosystems with low release rate of inorganic N from the soil organic matter.