Author for correspondence: Arthur Geßler Tel: +49 761 203 8309 Fax: +49 761 203 8302 Email: email@example.com
• We assessed the effect of climatic and canopy density changes on the seasonal patterns of total soluble nonprotein N (TSNN) in naturally regenerated beech (Fagus sylvatica) seedlings grown on limestone.
• Leaves, roots, wood and phloem exudates from seedlings grown in control and thinned stands on a dry, warm SW-exposed site and a moist, cooler NE-exposed site were examined. The concentrations of amino compounds, ammonium and nitrate, comprising TSNN, were determined in May (new leaf expansion), July (mid-summer) and September (end of the growing season).
• In May, Asn was augmented in leaves and roots at the NE site, whereas Arg dominated in leaves and phloem at the SW site. In July, all TSNN compounds declined, independent of site and treatment. In September, TSNN, and particularly Arg, increased in roots, phloem and wood at the SW site, compared with the NE.
• TSNN indicates changes in the N status of beech seedlings, due to altered growth conditions. The drier and warmer climate at the SW site, relative to the NE, resulted in earlier N remobilization in spring and storage in autumn. Thinning improved the N status at the NE site, but impaired it at the SW site, by affecting differently the climatic conditions and soil nutrient balance of each site.
Thinning may affect both biotic and abiotic conditions of tree seedlings in different ways. On one hand, increasing soil temperature and reduced soil water, as well as microbial activity and decomposition of organic material, may facilitate nutrient availability (Thibodeau et al., 2000). Such a response, together with increasing irradiance may promote the growth of tree seedlings (Mizunaga, 2000; Thibodeau et al., 2000). However, thinning may also enhance interspecific competitive interactions between young tree seedlings and fast growing understory vegetation for nutrient and water resources (Fotelli et al., 2001; 2002).
As a consequence, this forest management practice may significantly affect the nutrient status of tree seedlings. In particular, the N balance of young trees seems to be decisive in their development (Woods et al., 1992), since N is a likely major growth limiting factor in forests not excessively exposed to reactive atmospheric N compounds from anthropogenic origins (Rennenberg et al., 1998).
Apart from N, water availability is a key factor affecting growth and competitive ability (Fotelli et al., 2000; Fotelli et al., 2001) and seedling nutrient balance (Fotelli et al., 2001; 2002). Water availability will be of major importance in the future, since actual climate models for Central Europe predict prolonged summer droughts (IPCC, 1997, 2001). Among the ecosystems most affected will be natural European beech stands growing on shallow, low water capacity – Renzina soils – derived from limestone, which are common in Central and Southern Europe (‘Schwäbische Alb’, ‘Frankische Alb’, ‘French Jura’, ‘Swiss Jura’).
Due to translocation and signaling functions performed by soluble nonprotein N, this pool reacts rapidly to changes in N availability, serving as a sensitive indicator of the plant’s internal N status and, hence, growth potential (Rennenberg & Geßler, 1999) in contrast to total N contents that do not greatly vary over a wide range of natural N availabilities (Rennenberg et al., 1998; Geßler & Rennenberg, 2000).
The aim of the present study was to characterize the seasonal changes in total soluble nonprotein N concentrations in different parts of beech seedlings grown in the field under different climatic and canopy density conditions. For this purpose, contents of free amino compounds, nitrate and ammonium in phloem exudates from stem, as well as in leaves, roots and stem wood of beech seedlings were determined. The study was performed with naturally regenerated beech seedlings grown at two sites, having different exposures (south-west vs north-east), and within each site, in stands of differing density. Our main hypothesis, based on previous records, was; that the differences in the meso- or local climatic conditions of the two sites would result in different seasonal patterns of N metabolism in beech seedlings; and that their N status would be favored when grown in thinned stands. An additional aim was to characterize soluble nonprotein N compounds as an integrative physiological mean for evaluating the N status of naturally regenerated beech seedlings on this soil type and under varying growth conditions.
Materials and Methods
The experimental sites are located in southern Germany, approx. 100 km south-south-west from Stuttgart (longitude: 8°40′ E; latitude: 48°00′ N). The site is in a low mountain range (Schwäbische Alb, 740–760 m asl) with mean annual air temperature of 6.6°C and mean growing season air temperature (May to October) 11.5°C. Average annual precipitation is 856 mm with monthly maxima between June and July. The sum of precipitation during the growing season (May to October) is 410 mm. In the year 2000 growing season the total precipitation reaching the forest floor was of about the same level, but mean air temperature was greater than long-time averages (Table 1).
Table 1. Climatic characteristics of the experimental sites examined in the present study during the growing season (May–September) 2000. The experiments were carried out on two differentially exposed sites, facing north-east (NE) and south-west (SW) and were either subjected to thinning (T) of the mature forest canopy or remained unthinned as controls (C). The climatic parameters were recorded at four climate stations c. 1.5 m above the forest floor. The soil water potential was measured with tensiometers in a soil depth of 20 cm (sources: Meteorological Institute and Institute for Soil Science, University of Freiburg)
Sum of precipitation throughfall (mm) refers to the amount of precipitation passing through the canopies of the trees and reaching the forest floor.
All means shown are averages over the entire growing season of 2000. The values of Tsoil, RH and soil water potential are averages of mean daily values. The values of Tair are averages of mean midday temperatures, measured from 12 : 00 h until 17 : 30 h. The values of PAR are averages of mean PAR, measured during the daylight phase.
Experimental sites are located on the two opposing sides (not more than 1000 m apart) of a single, narrow valley. One experimental site faces to the north-east (NE) and the other to the south-west (SW). Rainfall does not vary significantly across the valley (Geßler et al., 2001), but differences in the amount of throughfall were observed at the different experimental plots (Table 1). Soil profiles are characterized as Terra fusca–Rendzina derived from limestone (Weißjura beta and gamma series). On both sites (SW and NE) the soil profiles are shallow, averaging < 20 cm depth of topsoil before becoming dominated by parent rock interspersed with pockets of organic matter and mineral soil. Soil pH (H2O) is 5.7 in the surface organic layer and 7.5 at 0.6 m depth. More details on site characteristics are given in Table 2. On both sites European beech (Fagus sylvatica L.) is the dominant species making up > 90% of the total basal area of adult trees. The average age of the adult trees is 70–80 yr. The difference in exposure (NE or SW) produces a difference in radiation interception at the canopy level. The maximum daily radiation for the NE site amounts to 79% and 47% of the radiation available at the SW site in July and October, respectively (Geßler et al., 2001). On the forest floor (1.5 m above ground) PAR during the growing season did not vary greatly between the SW- and NE-facing site (Table 1; control plots). The differences in radiation at the canopy level, however, result in higher air temperatures and soil temperatures and lower water availability for the vegetation at the SW site (Table 1). According to past records of meteorological data, growth analysis and water status of adult beech trees (Geßler et al., 2001), the SW-facing site has lower water availability and higher air temperatures, due to greater solar radiation, than the NE-facing site.
Table 2. Soil and stand characteristics of the experimental sites. The experiments were carried out on two opposite-exposed sites, facing north-east (NE) and south-west (SW). The forest stands on both sites were either subjected to thinning (T) of the mature forest canopy, or remained unthinned as controls (C)
Total basal area of the adult beech trees (m2 ha−1)
Leaf area index of the adult beech trees
Coverage of vegetation in the understory layer (%)
Slope of the sites (%)
Content of rocks and stones in 20 cm depth (v : v percentage)
Content of rocks and stones in 50 cm depth (v : v percentage)
On each site (SW and NE exposure), beech seedlings of a stand where a thinning treatment was applied (designated T) and of a unthinned control stand (designated C), were used for the purposes of this study. The selective felling of trees was performed in March 1999 and it resulted in a substantial reduction of the mature canopy density and a simultaneous increase in the coverage of the understory vegetation (Table 2). Moreover, PAR was increased at the forest floor in the thinned stands on both NE and SW sites and was markedly higher at the SW site (Table 1). This increase in radiation caused daily mean temperatures of air and soil in the thinned stands to be consistently greater than in the control stands (Table 1). Thinning also resulted in higher concentration of nitrate in soil at the NE-exposed site and lower concentrations at the SW-exposed site (Fig. 1). In the thinned stands the throughfall of precipitation reaching the forest floor was increased (Table 1), but the amount of water leaching was probably dependent on the soil characteristics. As the content of stones and rocks found at the SW site was quite higher than that at the NE site (Table 2), it could be assumed that water, and furthermore, soil nitrate leaching occurs more rapidly and to a greater extent at the SW site leading thus, to lower nitrate content at this site, compared to NE. Neither site nor thinning resulted in differences in phenology of beech (initiation of leaf expansion, leaf yellowing, leaf fall; H. Mayer, pers. comm.).
Plant material was collected from young naturally regenerated beech trees in May (after bud break), July (in the middle of the growing season), and September (at the end of the growing season) during the 2000 growing season. During each campaign six beech seedlings (c. 2–3-yr-old) per site and treatment were harvested and samples of leaves, fine roots and phloem exudates of the stem (see below) were obtained. Wood was collected in July and September. Before collection the bark of the stem was removed, in order to obtain data for analysis of only woody and associated parenchymatic tissues. Samples of wood and phloem were obtained from a height c. 10 cm above soil surface. Fine roots (d < 2 mm), being almost always colonised by mycorrhizal hyphae (B. Metzler, pers. comm.), were collected from a c. 10 cm depth and were washed with double-demineralized water to remove adhering soil particles. Samples of all plant parts were collected between 08 : 00 and 12 : 00, immediately frozen in liquid N2 and stored at −80°C.
Collection of phloem exudate
Phloem exudates of beech seedlings were collected using the EDTA-technique described by Schneider et al. (1996). Small pieces of bark (c. 200 mg f. wt) were removed from the stem and incubated in 6-ml vials with 2 ml of an exudation solution containing 10 mM EDTA and 0.015 mM chloramphenicol at pH 7.0 for 5 h. Previous studies (Schneider et al., 1996) showed that contamination of phloem exudates of beech with cellular constituents are negligible under the experimental conditions applied.
Extraction of amino compounds, ammonium and nitrate
Samples of leaves, roots and wood frozen in liquid N2 were ground with a mortar and pestle. For extraction of amino compounds and ammonium aliquots of 0.1 g of the frozen powder were homogenised in 0.12 ml buffer containing 20 mM Hepes (pH 7.0), 5 mM EGTA and 10 mM NaF, in 1 ml of chloroform : methanol (1.5 : 3.5, v : v). The homogenate was incubated for 30 min at 4°C, and subsequently, water-soluble metabolites were extracted twice with 0.670 ml double-demineralised water, after centrifugation at 16 000 g and 4°C for 5 min. The aqueous phases were combined, centrifuged at 16 000 g and 4°C for 20 min, and freeze-dried (Alpha 2–4, Christ, Osterode, Germany). The dried material was dissolved in 1 ml lithium citrate buffer (0.2 mM, pH 2.2).
For the extraction of nitrate aliquots of 0.25 g of the frozen powder of leaves, roots and wood were incubated for 2 h with 1.5 ml double-demineralised water and 70 mg PVPP at 5°C (Sigma Chemie, Deisenhofen, Germany).
Determination of total soluble nonprotein nitrogen (TSNN)
TSNN was determined as the sum of amino compounds, ammonium and nitrate in leaves, roots, wood and phloem exudates of young naturally regenerated beech seedlings. In order to express TSNN in the context of plant N, total N was also determined in samples of leaves, roots and wood using an elemental analyzer (NC 2500, Carlo Erba, Milan, Italy).
For ammonium and amino acid analysis, the pH of the extracts of all plant parts was adjusted with HCl to 2.2. An aliquot of 70 µl of each sample was injected into an automated amino-acid analyser (Biochrom, Pharmacia LKB, Freiburg, Germany). Amino compounds were separated on a PEEK column (Ultrapac 8 Resin, Lithium 250 × 4.6 mm, Biochrom, Pharmacia, Freiburg, Germany) using a system of five lithium citrate buffers that produced a pH gradient from 2.8 to 3.55. The separated amino compounds and ammonium were derivatised with ninhydrin and were measured spectrophotometrically at 440 and 570 nm. Peaks were identified and quantified using of a standard solution containing 39 amino compounds and ammonium (Sigma Chemie, Deisenhofen, Germany).
Before the analysis of nitrate in phloem exudates aliquots of 1 ml of were shaken for 2 h with 70 mg PVPP at 5°C to remove phenolic compounds. Subsequently, all extracts (leaves, roots, wood and phloem exudates) were heated to 95°C for 5 min, shortly incubated in ice for recovery to room temperature and centrifuged for 10 min at 16 000 g and 4°C. 0.5 ml aliquots of the clear supernatants were injected into an ion exchange chromatography system (DX 100; Dionex, Idstein, Germany). Anions were separated on a IonPac® column (AS9-Sc 250 × 4 mm; Dionex, Idstein, Germany) eluted with a solution containing 1.8 mM Na2CO3 plus 1.7 mM NaHCO3 at a flow rate of 1.0 ml min−1. Nitrate in leaf, root and wood extracts was detected with a conductivity detector module (CDM, Dionex, Idstein, Germany) with a detection limit of < 0.3 nmol ml−1. In phloem exudates determination of nitrate contents was performed with an UV-VIS detector (SDP-6AV; Shimadzu, Duisburg, Germany).
For validation of the extraction methods applied the recovery rates of ammonium, aspartic acid (Asp), asparagine (Asn), glutamic acid (Glu), glutamine (Gln) and arginine (Arg) applied as internal standards have been previously determined in the phloem exudates, in roots and in leaves of beech seedlings. Recovery rates amounted to c. 82 ± 16% and 108 ± 12% for the different TSNN compounds in the different tissues (Geßler, 1999).
Determination of nitrate contents of the soil
Soil samples were taken from the A-horizon between 0 and 10 cm depths and were homogenised. 5 g of the dried sample were extracted according to Breuer (2000) with 50 ml of a 0.01-N KAl(SO4)2 solution for 20 min on an incubator shaker. After centrifuging twice (15 min at 7000 g and 15 min at 12 000 g) 15 ml of the supernatant were filtered in a two-step procedure with a 5-µm and a 0.2-µm microfilter (Satorius, Göttingen, Germany). Nitrate concentration in the filtrate was determined by ion exchange chromatography combined with conductivity detection (DX 500, Dionex, Idstein, Germany) and referred to as a soil dry weight basis.
All statistical analysis was carried out using SPSS 10.05 (SPSS, Inc., USA). The effect of exposure and thinning treatment on TSNN and on the concentrations of the most abundant amino compounds was assessed using a two-way factorial ANOVA procedure. Significant differences of amino acids and TSNN concentrations between sites and treatments during each trial were detected with one-way ANOVA.
TSNN contents and composition in leaves
Figure 2 shows the contents of total soluble nonprotein N (TSNN) and its four most abundant constituents in leaves of seedlings grown at the two sites (SW vs NE) in the controlled and thinned stands. TSNN varied between 4.3 and 7.3 µmol N g−1 f. wt. in May comprising 0.30–0.35% and 0.65–0.75% of total N content in leaves, at the NE and the SW site, respectively. Neither a significant site, nor a thinning effect on TSNN could be observed. However, a tendency (P < 0.13) of thinning to result in increased TSNN is supported by the higher mean square effect (MS) of thinning (12.48), compared to site (5.80). By contrast to TSNN, a significant site effect was found for Asn contents of leaves in May. In plants of the NE site, Asn amounted to 1.4 and 2.3 µmol g−1 f. wt in the control and the thinned stand, respectively, and comprised c. 30% of TSNN. At the SW site, Asn contents were significantly lower compared to the NE and amounted to 0.8 and 0.5 µmol g−1 f. wt in the control and the thinned stand, respectively. The opposite effect of site was observed for Arg that contributed c. 13% to TSNN at the SW site and only c. 6% at the NE site. The Glu content in leaves increased significantly with thinning. In addition, its contribution to TSNN in May was slightly higher at the NE site (10–12%) than at the SW site (c. 5%). When the NE site was tested alone, there was a significant difference (P < 0.05) in the TSNN contents of leaves between the control and the thinned site with thinning resulting in a 1.6-fold higher TSNN. This increase could be mainly attributed to the amino compounds Asn, Glu and Gln.
In July, TSNN was not significantly affected, neither by site nor by treatment, and it was drastically reduced in all treatments, compared to May, amounting to c. 2.0 µmol N g−1 f. wt and to c. 0.1–0.2% of total N in leaves. This decline could be mainly attributed to the strong decrease in Asn and Arg contents. Only a slight reduction of the amount of Glu was observed in July, and leaves from the SW site exhibited significantly higher concentrations, than those from the NE site. Due to the low TSNN contents in July, the relative contribution of Glu to TSNN increased at this time, comprising > 15% at the NE and > 20% at the SW site.
In September TSNN contents increased slightly to 2.6–4.0 µmol N g−1 f. wt. comprising c. 0.15–0.30% of total N, but no significant effect of site and/or thinning was found, in agreement to May and July. However, when each site was analysed separately, thinning resulted in significantly higher TSNN at the NE site, whereas the opposite was found at the SW site. None of the four most abundant amino compounds was significantly affected by site or treatment.
Asn, Glu, Gln and Arg comprised together 38–64% of TSNN in leaves throughout the growing season. In addition to these amino compounds, other proteinogenic and nonproteinogenic N compounds were found. The most abundant proteinogenic N compounds were alanine (Ala), serine (Ser) and valine (Val) and comprised together c. 9–18% of TSNN. The most abundant nonproteinogenic amino compounds were γ-aminobutyric acid (GABA) and ethanolamine. Together they contributed to c. 2–7% of TSNN. Ammonium contributed to c. 2.5–10% of TSNN and it remained constant during the growing season. Nitrate was almost absent in leaves (< 1–2% of TSNN).
TSNN contents and composition in fine roots
In May, lower concentrations of TSNN in the roots of seedlings grown at the SW site (2.6–3.0 µmol N g−1 f. wt, being c. 0.35% of total N), compared to the NE site (c. 5.5 µmol N g−1 f. wt, being c. 0.20–0.25% of total N) resulted in a significant site effect, whereas thinning had no significant effect (Fig. 3). Comparable to TSNN, Asn contents were significantly higher in the roots at the NE site, than the SW site. Moreover, Asn was also significantly increased due to thinning on the NE-site. No significant influence of site and/or treatment on Glu, Gln and Arg was observed in May.
TSNN decreased in July at both sites and treatments, relative to May, to c. 2.0 µmol N g−1 f. wt, being c. 0.10% of total N. The contents of Asn and Arg were significantly higher at the SW site, as compared to the NE site but, due to the somewhat lower contents of the other amino compounds, this did not result in a site-dependent effect on TSNN.
TSNN content increased again in September and was significantly higher at the SW site (c. 4.5 µmol N g−1 f. wt, being c. 0.1% of total N), compared to the NE (c. 2.5 – to 3.5 µmol N g−1 f. wt, being c. 0.1–0.3% of total N). This effect could be mainly attributed to the higher Arg and Asn contents in seedlings grown at the SW site. In September, no treatment effect was found. However, when the NE site was tested separately, a significant increase in TSNN contents due to thinning was found, caused mainly by an increase in Arg, Asn and Glu contents.
Asn, Glu, Gln and Arg comprised together 50–73% of TSNN in roots throughout the growing season. Alanine showed also a high abundance and comprised 12–23% of TSNN in May and 4–14% of TSNN in July and September. Among the other proteinogenic N compounds, Asp, Ser and Val were also of considerable abundance and amounted together to 9–12% of TSNN throughout the growing season. The nonproteinogenic amino compounds GABA and ethanolamine made up c. 2–5% of TSNN during the growing season. Ammonium was constantly found in roots and fluctuated between 4 and 16% of TSNN. Nitrate contributed < 3% to TSNN.
TSNN contents and composition in phloem
TSNN contents of phloem exudates in May differed significantly between sites (Fig. 4), with higher values at the SW site. This difference was mainly due to the significantly higher Arg contents in the phloem of seedlings from the SW site. No general treatment effect on TSNN was found. However, as observed for leaves in May, when the NE site was tested separately, TSNN contents in the phloem were significantly higher in the thinned stand.
Similar to leaves and roots, TSNN contents of all sites and treatments decreased in July, compared to May. The substantial decrease of c. 65% observed for the seedlings at the SW site was mainly caused by the decline in Arg contents. TSNN contents in July were not significantly affected, either by treatment, or by site. However, testing the NE site separately revealed a significant increase in TSNN contents as a consequence of thinning. Of the four most abundant amino compounds only Glu was significantly affected by site showing significantly higher contents (9–13% of TSNN) at the NE, compared to the SW site (1–7% of TSNN).
In September, TSNN contents in the phloem increased significantly in all treatments, compared with July, and amounted to c. 4.7–6.0 µmol N g−1 f. wt. No significant effects of treatment or site were found. However, the Arg contents were significantly higher at the SW, compared to the NE site.
Beside the four most abundant amino compounds shown in Fig. 4, other proteinogenic (Asp, Ala and lysine) and nonproteinogenic (GABA, α-aminoadipine acid and ethanolamine) amino compounds contributed c. 4–27% and 5–22% to TSNN, respectively. Both, ammonium and nitrate in the phloem were below the limit of detection.
TSNN contents and composition in wood
In July, TSNN contents in the wood of the stem of beech seedlings were not affected by site or thinning and amounted to c. 13–17 µmol N g−1 f. wt and to c. 3–6% of total N (Fig. 5). From the four most abundant amino compounds, only Gln showed differences due to treatment and was significantly increased by thinning.
From July to September, TSNN contents decreased in the wood of seedlings at the NE-control site, remained constant on the NE-thinned stand and increased under both densities at the SW site. Their contribution to total N content amounted c. 1.0% at the NE site and c. 2.5–5.5% at the SW site. In September, a highly significant site effect was observed for TSNN with higher contents at the SW, compared to the NE site. When the NE site was tested separately, thinning resulted in a significant increase in TSNN from 8.5 to 16.8 µmol g−1 f. wt. The opposite was found for the SW site, where TSNN was significantly decreased (from 38.0 to 27.9 µmol g−1 f. wt) due to thinning. The contents of Asn increased in September, particularly in seedlings grown in the control stand of the SW site (16% of TSNN). The effect of site appeared to be greater (MS = 32.00), than the effect of thinning (MS = 9.86), but interaction between the two factors prevents significant influences from being determined. A similar pattern was observed for Glu. Gln contents were significantly affected by site and were enhanced at the SW site, compared to the NE site. The most abundant amino compound in September was Arg which comprised c. 33–70% of TSNN. There was a significant effect of both site and treatment on the Arg contents with highest values at the SW-thinned stand.
Asn, Glu, Gln and Arg contributed together to 45–86% of TSNN during the growing season. Other proteinogenic amino compounds (mainly Asp and Ser) comprised c. 24–32% of TSNN in July and c. 6–16% in September. The concentration of the main nonproteinogenic amino compounds (GABA and ethanolamine) remained constant during the growing season and comprised together c. 1–5% of TSNN. Ammonium made up 2–11% of TSNN in the wood whereas nitrate was below the detection limit.
In the present study we assessed the effects of thinning and different meso- or local climatic conditions (temperature, radiation, water availability), as induced by variations in exposure (SW vs NE), on the contents and composition of TSNN in leaves, roots, phloem and wood of beech seedlings. From previous studies with young and adult beech trees, this parameter is known to react rapidly to changes in N availability (Geßler et al., 1998c; Geßler & Rennenberg, 2000) and N demand (Rennenberg & Geßler, 1999) and is therefore a sensitive indicator of plant internal N status, contrary to total N which remains stable over a wide range of nutritional and climatic conditions (Rennenberg et al., 1998; Geßler & Rennenberg, 2000).
Previous studies revealed that the SW-exposed site not only received more radiation, leading to higher air and soil temperatures (Table 1), but also had lower soil water storage capacity, compared to the NE site (Geßler et al., 2001). Hence, adult beech trees were found to experience water shortage during periods of low rainfall in summer only at the SW site, whereas water supply was sufficient during the entire growing season at the NE site (Geßler et al., 2001). Since prolonged drought periods in summer are expected to increase in frequency in Central and Southern Europe in the future (IPCC, 2001), the site design is regarded suitable for assessing climate change effects on natural regeneration of beech.
In May, during development of new leaves, significant site effects (SW vs NE) were observed for TSNN in roots and phloem (Figs 3 and 4). However, TSNN of the two tissues responded differently; in the roots it was higher at the NE site (Fig. 3) and in phloem at the SW site (Fig. 4). Although the acropetal transport of N takes place mainly via the xylem stream, transport of amino compounds from storage tissues via phloem is known to contribute substantially to the N demand of newly developing leaves (Da Silva & Shelp, 1990; Geßler et al., 1998b) during N remobilisation in spring. Furthermore, Arg which was mainly responsible for the TSNN patterns found in the phloem in May, is the form in which stored N is remobilised and transported (Flaig & Mohr, 1992; Gezelius & Näsholm, 1993; Geßler et al., 1998a). Therefore, a possible explanation for the patterns found is the transport of remobilised N from storage tissues, such as wood (Millard & Proe, 1992, 1993) to the newly developing leaves via the phloem. This process seemed to be of greater importance in seedlings from the SW site, as indicated by higher Arg and TSNN contents in the phloem (Fig. 4).
On the other hand, the higher TSNN content in the roots of beech seedlings from the NE site was mainly due to higher Asn and Glu concentrations (Fig. 3). A correlation analysis between the nitrate contents in the soil (Fig. 1) and the contents of Glu in the fine roots from all sites and treatments during the growing season (Fig. 3) produced a highly significant correlation (R2 = 0.78, P < 0.0001; Fig. 6a).
Preferential uptake of ammonium by roots of woody plants (Plassard et al., 1991; Kronzucker et al., 1996) and specifically also of beech (Geßler et al., 1998b) is reported. In the studied field site, however, ammonium availability in the soil is very low (E. Hildebrand pers. comm.), due to a soil pH > 5.5 which favors high nitrification rates. As a consequence nitrate is the main pedospheric N source for beech. The small amounts of nitrate present in the fine roots of beech independent of site and treatment indicate that the uptake of inorganic N is adapted to the N assimilation capacity of the mycorrhizal fine roots of beech, a finding consistent with other studies on beech (Martin & Amraoui, 1989; Geßler et al., 1998b). Finlay et al. (1989) and Kreuzwieser et al. (2000) showed that the several types of Fagus sylvatica ectomycorrhizas were able to absorb and reduce nitrate rapidly. Consequently, since Glu is the first product of N assimilation in mycorrhizal beech roots (Martin & Lorillou, 1997) it can be assumed that its concentration increased in May in the roots of seedlings grown at the NE site due to higher N uptake and assimilation rates as a consequence of higher soil nitrate availability.
The differences in TSNN contents and composition between the two sites in May are also reflected by differences in TSNN composition in leaves (Fig. 2). The dominant amino compound was Asn at the NE site and Arg at the SW site. As in roots, Asn contents in the leaves may also be regarded as an indicator of nitrate availability in the soil since it is one of the major compounds translocated from the roots to the leaves in beech (Geßler et al., 1998a). This is further supported by the highly significant correlation of the Asn concentration in leaves with the nitrate concentration in soil, during the entire growing season (R2 = 0.71, P < 0.0001; Fig. 6b), indicating that Asn was produced due to higher nitrate uptake at the NE site and was augmented in the leaves, where it was allocated via the xylem stream.
Thinning resulted in a significant TSNN increase in the leaves of beech grown at the NE site in May (Fig. 2). This could be attributed to the higher soil nitrate availability in the thinned NE-stand (Fig. 1) and is, further, supported by the significant correlation between TSNN contents in leaves and nitrate contents in soil (R2 = 0.76, P < 0.0001; Fig. 6d). Moreover, significant correlations between soil nitrate concentrations and the contents of dominant TSNN compounds like Glu (R2 = 0.53, P < 0.0001; data not shown), Gln (R2 = 0.61, P < 0.0001; Fig. 6c) are consistent to this hypothesis.
In leaves, roots and phloem, TSNN contents decreased in July independent of site or treatment (Figs 2, 3 and 4), relative to values in May. As no significant growth increment of beech seedlings occurred between July and September (data not shown), the decline of TSNN contents in July could not be due to dilution by tissue growth. On the contrary, it can be attributed to; reduced soil nitrate availability (Fig. 1) in agreement to the above mentioned correlations (Fig. 6); and reduced N remobilisation from storage tissues, indicated by the generally decreased amounts of Arg in the phloem, compared to May (Fig. 4). However, the concentration of Arg in roots of the seedlings from the SW site was significantly higher than that at the NE site (Fig. 3) and it also tended to be higher in the thinned stand of the SW site, than in the control one. This may indicate earlier reallocation for storage of N from leaves to roots, because of less favorable conditions at the SW site (high temperatures, reduced water availability) or particularly in the thinned stand (lower soil nitrate supply). Such responses of Arg, indicating reallocation to storage tissues, were reported by Millard & Proe (1992) and Gezelius & Näsholm (1993) in Scots pine and Sitka spruce grown under similar suboptimal conditions.
In September, at the end of the growing season, N is reallocated from senescing leaves via phloem transport to the storage tissues (Millard, 1989; Coleman et al., 1991; Pate & Jeschke, 1995; Millard, 1996). The increase in TSNN, particularly in Arg contents in the phloem of beech in September (Fig. 4) could result from increased transport of leaf-derived N to storage tissues, as already observed for adult beech trees (Geßler et al., 1998b). This could also explain the increased Arg content in wood (Fig. 5), the main location of N storage in deciduous trees (Millard & Proe, 1991), and in roots (Fig. 3) which also contribute to N storage in young broadleaf trees (Tromp, 1983; Suzuki & Kohno, 1983; Millard & Proe, 1991). Changes in air and soil temperature, resulting from global change, are known to affect the phenology of different tree species. However, leaf senescence of beech is not influenced by higher temperatures (Kramer, 1995). Consistent with this is that neither exposition (SW vs NE), nor thinning resulted in differences in phenology of beech (leaf yellowing, leaf fall) in autumn in the field site studied (H. Mayer, pers. comm.). Despite that, it cannot be excluded that physiological responses – such as the onset of N storage in beech – to increases in temperature and/or limitations in N and water availability did occur, even when signs of senescence were still not visible.
In September, a thinning effect on the TSNN contents and composition in the leaves was evident. The increase of TSNN concentration in the thinned stand of the NE site and, at the same time, the decrease in the thinned stand of the SW site agrees with the pattern of soil nitrate (Fig. 1). Marmann et al. (1997) reported differences in the amount of N stored in autumn and remobilised in spring in ash grown at two sites with different soil water availability and concluded that soil N supply was the main factor affecting the intensity of N storage. Similarly in the present study, the effect of thinning on climatic conditions and nitrate availability seemed to result in earlier and/or more intensive autumn reallocation in the thinned stand of the SW site, as indicated by the decrease of TSNN in leaves and the concurrent increase of Arg, the main soluble N storage compound, in phloem, wood and roots (Figs 2, 3 and 4).
The present study shows that TSNN can indicate seasonal changes in the N status of young beech seedlings, according to changes in growth conditions. Moreover, the patterns of certain amino compounds, like Glu, Gln and Asn in leaves, and Glu in roots are well correlated to soil nitrate availability, as modified by local climatic and canopy density conditions. The combined effect of climate and thinning during the entire growing season produced noticeable differences between treatments at the end of it, in September. TSNN was higher in all plant parts in the NE – thinned stand, and lower in leaves and wood in the SW – thinned stand, compared to the controls. These responses indicate that in the long term the application of thinning in beech forests grown on limestone may improve the growth conditions on cold and humid sites (by increasing radiation and temperature), as it was our initial hypothesis, but it may result in less favourable conditions in already warmer and drier sites (by further increasing temperature and reducing soil nitrate availability). Since climate models prognosticate higher temperatures and prolonged drought periods in Central Europe, that is, conditions comparable to those at the SW site used within the present study, relative to the NE, selective felling as a mean to promote natural regeneration of beech may result in adverse effects concerning the N status and, hence, growth potential of beech seedlings in the future.
MN Fotelli thanks the DAAD (Deutscher Akademischer Austauschdienst) for the financial support during part of the present study. This research was part of the SFB 433 funded by the DFG (Deutsche Forschungsgemeinschaft). The authors thank T. Holst and H. Mayer (Meteorologisches Institut, University of Freiburg) for the meteorological data, and S. Augustin and E. Hildebrand (Institut für Bodenkunde und Waldernährungslehre, University of Freiburg) for the soil water potential data. We are also grateful to Gail Jackson (University of Edinburgh) for helpful criticisms on the manuscript.