During the growing session of 1995, the total soluble non-protein nitrogen (TSNN) composition and contents of mycorrhizal fine roots, xylem sap and phloem exudates of roots from a coniferous (Picea abies L.(Karst)) and a deciduous (Fagus sylvatica L.) tree species were analysed at a field site (‘Höglwald’, Germany) exposed to high loads of N. In April, TSNN in fine roots of spruce and beech trees amounted to 16 μmol N g−1 f. wt and 23·3 μmol N g−1 f. wt, respectively. It decreased to 9·2 μmol N g−1 f. wt and 18·1 μmol N g−1 f. wt, respectively, after bud break in June. The seasonal maximum of TSNN in fine roots of spruce was observed in July (32·7 μmol N g−1 f. wt) followed by a decline of c. 30% until the end of the growing season in September. TSNN in fine roots of beech trees showed a further decline between June and July, when its seasonal minimum was determined (15·6 μmol N g−1 f. wt), and increased to c. 29 μmol N g−1 f. wt until September. In spruce roots Gln and Arg were the most abundant TSNN compounds during the entire growing season. In roots of beech Asn played an important role alongside Gln and Arg, especially in April, when it was the most abundant TSNN compound. Other proteinogenic and non-proteinogenic N compounds comprised c. 20–30% of TSNN. Nitrate made up <1%, and ammonium <7% of TSNN in the fine roots of both species.

In April, TSNN in the xylem sap of roots of spruce and beech trees amounted to 3·4 and 8·6 μmol N ml−1, respectively. In roots of spruce trees xylem sap TSNN increased after bud break up to 12·7 μmol N ml−1 in July. At the end of the growing season TSNN had declined again to 3·9 μmol N ml−1. TSNN in the root xylem sap of beech trees decreased after bud break until July (2·4 μmol N ml−1 in July) followed by a slight increase until September (2·9 μmol N ml−1). Arg, Gln and Asp were the most abundant TSNN compounds in the xylem sap of spruce trees contributing together c. 90% to TSNN. The same TSNN compounds prevailed in the root xylem sap of beech trees in April and July, whereas in June and September Asp was replaced by Asn comprising 57% of TSNN in June. In addition to the N compounds mentioned above, a number of other proteinogenic and non-proteinogenic amino compounds were found in root xylem sap of both species. In either species, nitrate and ammonium were present in small amounts, contributing <1% and <4% to TSNN, respectively. Apparently, inorganic N taken up by the mycorrhizal roots is mainly assimilated in root tissues or by the mycorrhiza and N uptake by the roots is largely adapted to the assimilatory capacity of this organ.

In phloem exudates of spruce roots, TSNN amounted to 10·7 μmol N g−1 f. wt in April, increased in June to 23·4 μmol N g−1 f. wt and decreased again until September to a seasonal minimum of 4·8 μmol N g−1 f. wt. In contrast to spruce, TSNN content in phloem exudates of beech roots showed a seasonal maximum (c. 20 μmol N g−1 f. wt) in April with a subsequent decrease in June after bud break (c. 2 μmol N g−1 f. wt). A fourfold increase in July was followed by a decrease in September, when TSNN in phloem exudates of beech roots amounted to 4·3 μmol N g−1 f. wt. Arg was the most abundant N compound in the phloem of roots from spruce trees and made up c. 60–85% of TSNN during the entire growing season. In beech trees the seasonal course of TSNN correlated with the relative abundance of Arg. Arg comprised 69 and 57% of TSNN in April and July, respectively, but contributed <20% in June and September. Besides Arg, other proteinogenic and non-proteinogenic amino compounds could be detected in the phloem of both species. In addition, nitrate and ammonium were present in considerable amounts.

From these results and a previous report on TSNN in above-ground parts of spruce and beech at the same site, a whole-plant model for the cycling of TSNN in both species is proposed. Differences in the location of storage pools are assumed to be responsible for the differences in the seasonal course of TSNN composition and contents observed between the two tree species.