Introduction

Growth disturbances on trees characterized by a loss of apical dominance and shoot dieback and, consequently, the development of multiple-shoots resulting in bushy topped trees, have been reported widely throughout the world (Stone 1990). These symptoms have often been connected to boron deficiency (Silfverberg 1980, Aronsson 1983; Braekke 1983; Raitio 1983; Tamminen and Saarsalmi 2004), as well as to magnesium and calcium deficiencies (Rehfuess et al. 1983) and to diseases like Pucciniastrum areolatum (Fr.) Otth. (Lilja 1967) and Sirococcus blight (Sirococcus conigenus (DC.) P. Cannon & Minter) alone (Halmschlager et al. 2000), or together with nutrient (Mg, Ca, P, Mn) deficiencies (Anglberger et al. 2003). Spring and autumn frost (Silfverberg 1980), and in some cases insects (Saalas 1949), may also kill the leader shoot and lead to a bushy appearance of the trees.

The occurrence of bushy topped trees in half of the young spruce stands on fertile sites in the province of North Savo (Tamminen and Saarsalmi 2004), have raised increasing concern about forest health and productivity in eastern Finland. Healthy-looking and bushy topped trees grow side-by-side. Over the years, the tops of the trees may have recovered and only forked stems and ramicorns are a reminder of earlier disturbances. The boron-related visible symptoms, the fact that boron concentrations in the bedrock of much of Fennoscandia are relatively low, and the low annual deposition of boron due to the long distances from the ocean (Wikner 1983), indicate that boron deficiency is the cause of the damage. Furthermore, slash-and-burn agriculture were widely practiced in eastern Finland in the 19th century (Heikinheimo 1915) and forest grazing up to the middle of the 20th century (Jäntti 1947), and this may have disturbed the nutrient balance in the soil (Tamminen and Saarsalmi 2004).

The effects of boron deficiency have been studied at the visible, biochemical and microscopic levels on both broad-leaved and conifer trees (e.g., Stone 1990; Cakmak and Römheld 1997; Matoh 1997). The main role of B in plants is currently thought to be in the structural development and maintenance of cell walls (Loomis and Durst 1992; Matoh 1997; Power and Woods 1997). Furthermore, the assumptions concerning the mechanism behind the disturbances of boron deficiency include the effects on the plasmalemma and phenol metabolism (Cakmak and Römheld 1997; Liakopoulos and Karabourniotis 2005). Changes in phenol metabolism have been suggested to be secondary or tertiary effects of B deficiency (Dordas and Brown 2005). The structural abnormalities described at the light and electron microscopic level vary widely depending on the species studied, and on the severity and timing of the boron deficiency (Dell and Huang 1997). In conifers, abnormal structures such as underdeveloped thin-walled sclerenchyma cells and cavities in the central cylinder in the needles have been described in connection with boron deficiency, and also with low K, P and a high N/B ratio in the needles of Scots pine (Raitio 1983; Jokela et al. 1995). However, the time of onset of the disturbance at the cellular and tissue levels in the needles and buds preceding visible disturbance are not known.

The aim of this study was to determine (i) the time of onset and appearance of the damage in the emerging buds and needles in growth-disturbed Norway spruce trees growing on mineral soils, and (ii) possible connections between the damage and imbalance in nutrient concentrations, especially of boron. To achieve the aims we followed at the light microscopic level bud emergence, development and structure of the current and previous year’s needles, as well as needle nutrient concentrations, from early spring (April) to early winter (November) in three ca. 35-year-old spruce stands showing both bushy and healthy growth.

Materials and methods

Study areas

Three spruce stands with growth disturbances of different severity were selected for the study. All the sites represented the relatively fertile mineral soil site type (Oxalis-Maianthemum type (OMT) according to the classification of Cajander (1949)). The density of the stands was about 2000 stems/ha, which meant that the canopy was closed and the ground vegetation therefore relatively undeveloped.

The stand with bushy tops and multiple leaders (later called leader candidates), and with dead shoots in more than 90% of the trees, called the damaged stand (D stand), was about 3 ha in size and located in Central Finland (62°36′N, 26°43′E, 95 m a.s.l.). During the study year (2000), the trees were 34 years old (plus seedling age). Before afforestation the area had been a hay field, which was then left to regenerate naturally with pubescent birch (Betula pubescens). In 1966, the site was planted with bare-root spruce transplants (provenience unknown) at a density of 2500 plants/ha. After planting, the area was used for grazing for 5 years. Naturally regenerated alder trees (Alnus incana) growing among the spruces were felled a couple of years before the study started. The mean height and breast height diameter of the sample trees were 9.9 m and 18 cm, respectively. According to the dense clusters of branches observed on the spruce stems, the first disorders were traced back to the time when the trees were 2–3 m high and about 15 years old. Some of the trees of about this size were heavily bushed and dead.

The second stand (2 ha) had mainly healthy looking trees, i.e. less than 10% of the trees were bushy, and is called the partly damaged stand (PD stand). The PD stand is located next to the D stand, but the altitude is about 3–5 m higher. Spruce transplants of unknown origin were planted 5 years later than those in the D area. Otherwise the history of the PD stand is the same as that of the D stand. The mean height and breast height diameter of the sample trees were 12.5 m and 15 cm, respectively.

The third stand (acreage about 3 ha) was characterized by normal, healthy (H stand) crown tops apart from a few trees attacked by Pucciniastrum areolatum, which was not found in the other stands. The H stand is located 20 km northeast (62°42′N, 27°03′E, 170 m a.s.l.) from the D and PD stands. Four-year-old spruce transplants (the seed, M28-65-30, from a selected stand in Lammi, 61°N 25°E) were planted in 1975 on the forest-ploughed site. The mean height of the sample trees was 11.5 m and breast height diameter 15 cm.

Sampling

Needle and bud samples were taken on seven occasions in all three stands between April and November 2000. Trees with severe growth disturbances from the D stand and trees with normal crowns from the PD and H stands were randomly selected and felled, and a 1.5 m section was cut from the top of each tree. After immediately taking bud and needle samples for microscopic analysis, the tops were packed in plastic bags and transported to Suonenjoki Research Station (19 km), and stored at +5°C for one night prior to needle nutrient analysis. The sampling dates in 2000, the number of sampled trees, the 1999- and 2000-needle generations sampled, temperature and precipitation data for both years (the weather station of Suonenjoki Research station, 62°39′N, 27°03′E, 140 m a.s.l.), and the relative length of the current shoot at each sampling date compared with the final height estimated as a function of the temperature sum according to Raulo and Leikola (1974), are given in Table 1. As the monthly means of the meteorological variables do not necessarily fit with meristemic cell production (Deslauriers and Morin 2005), the precipitation and mean temperature for 6 days before the sampling date are also given in Table 1.

Table 1 Sampling dates and number of sampled spruce trees, from which current (2000; Cx) and second (1999; C+1x) year needles were sampled for nutrient (n) and microscopic (m) analysis during 2000. “x” refers to the age of the C and C+1 needles in months
Full size table
Table 2 \({\rm pH}_{{\rm H}_2{\rm O}}\), and total concentrations of N, P. K, Ca, Mg, Mn, Fe, Zn, Cu and hot water extracted B in the humus
Full size table

Nutrient analysis

Three bulked mineral soil samples (0–10 cm under the humus layer with a core sampler, d=22 mm) and three humus samples (core sampler, d=55 mm) were taken in each study stand on 5 September 2000. Each bulked mineral soil sample and humus sample consisted of 20 sub samples. The samples were dried (at +35°C for 1 week), after which stones and larger roots were removed from the samples. Samples of humus were ground in a mill fitted with 2 mm bottom sieve. pH was determined by adding 25 ml of distilled water to 10 ml of sample, stirring, leaving to stand for 2 h, followed by measurement on a calibrated pH meter. Total element (P, K, Ca, Mg, Mn, Fe, Zn, and Cu) concentrations in the humus samples were determined by dry digestion (550°C/3 h) and, following extraction of the ash with HCl (Halonen et al. 1983), by inductively coupled plasma atomic emission spectrometry (ICP/AES, ARL 3580). Acid ammonium acetate (pH 4.65) extractable P, K, Ca, Mg, Mn, Fe, Zn, and Cu (Halonen et al. 1983), and hot water soluble B were determined from mineral soil samples and humus samples (Saarsalmi and Tamminen 2005). As the main trends in extractable and total soil nutrients were similar in the humus and mineral soil, only the total element concentrations and hot water extracted B concentration in the humus are given in this paper (Table 2). In general, the chemical soil properties of the D stand, and in some cases also of the PD stand, differed from those of the H stand. The total K, P, Ca, and Mg concentrations were higher in the D and PD stands than in the H stand. The B concentrations in all the stands were low, and there were no significant differences among the stands. However, the concentrations were close to those reported earlier for eastern Finland (Tamminen and Saarsalmi 2004). The pH of the humus was relatively low, whereas in the mineral soil it was close to the average values reported for Finnish forests on mineral soil sites (Tamminen 1991).

Microscopy

Buds

The emerging new terminal buds of the three longest leader candidates from each D tree were sampled from May to August. In contrast, the true terminal emerging buds were sampled from the dominating leader shoots of the PD and H trees. The emerging terminal buds of one shoot from the second whorl on each tree were also sampled. The shoot tips with the emerging terminal buds were transferred to individual tubes containing the fixative (1.5% glutaraldehyde, 1.5% paraformaldehyde, 0.15 M sucrose, 2 mM CaCl2 in 0.05 M cacodylate buffer, pH 7.0), and transported to the laboratory. The next day the emerging buds (for the D trees the most healthy looking bud out of the three leader buds sampled) were cut longitudinally under a stereomicroscope into two parts, one of which was used for preparing longitudinal sections and the other for cross-sections. The bud sections were kept in the fixative overnight (22 h), and fixed further according to Soikkeli (1980). The cross-sections, about 2 μm thick, were cut for light microscopy on a LKB Ultratome III and double-stained with 1% toluidine blue and 1% p-phenyldiamine (Ryyppö et al. 1997).

The bud sections were digitally photographed (Leica CD Camera, Switzerland) with a light microscope (Leitz Laborlux12, Germany) at 100× (10× lens) and 250× (25× lens) magnifications. The condition of the meristematic regions of the emerging terminal buds (one from the dominating shoot and one from the side shoot per tree) were qualitatively analysed. On 28 August, the proportion (as percentage) of the area of tanniferous cells, i.e. living cells accumulating tannin, out of the total primordial shoot area was calculated. From 28 August onwards, the buds from healthy spruces were monitored by means of fresh cuttings up until 17 September when the needle primordia had developed up to the base of the apex, i.e. bud development was complete.

On 27 November, 10 cm-long shoot pieces were cut from the 1.5 m tree tops as follows. From the D trees the eight longest leader shoot candidates were cut. Of these, the sixth and seventh shoot were used for the other studies. From the PD and H trees the true leaders were cut. In addition, five 10 cm-long C+1 side shoots per tree top were cut. The shoot pieces were put into plastic bags and stored at −15°C for analysis of the buds.

The terminal bud and lateral buds along the section running 5 cm from the base of the terminal bud of each shoot were classified on fresh-cut longitudinal sections under a stereomicroscope into three categories: (1) healthy buds (the pith, scales and the primordial shoot and needles were well developed) (Fig. 1a), (2) slightly damaged buds, where the scales showed disorganization such as curling or slight accumulation of resin, and primordial shoot/needles were deformed, but still green (Fig. 1b), and (3) fatally damaged buds, where the primordial shoot/needles were not developed at all or were dead, although the bud scales may have been partly developed (Fig. 1c). The third category also included dwarf buds that had only a few bud scales (Fig. 1d) and small pith, but no green primordial shoot (Fig. 1e).

Fig. 1
figure 1

Longitudinal fresh sections of terminal buds from leader shoots of Norway spruce in November. A Well-developed bud from a H tree, B slightly damaged bud, and C fatally damaged as well as D and E dwarf buds from D trees. White stars show bud scales, and black stars the pith above which a normal primordial shoot in (A), underdeveloped shoot in (B) and no primordial shoots in (C) and (E). All bars = 2 mm

Full size image

Needles

In May, the cross-sections included both the apex of the emerging bud and differentiating needles around the apex. At that time, when the bud and primordial needles were elongating inside the bud scales, the stage of tissue differentiation of the needles was determined. Furthermore, a rough estimate of the accumulation of tannins in the differentiating epidermial, mesophyll and parenchymatic cells in central cylinder of the needles was made as follows: no tanniferous cells, only a few tanniferous cells (approximately 10–20% of the cells in a given tissue) or frequent tanniferous cells, i.e. one-third or more of the cells in a given tissue showed accumulation of tannin.

Five current (C) year needles from the current whorl (the longest leader shoot candidates in the D trees) and five previous (C+1) year needles from the second year whorl were sampled (Table 1) from the middle of the shoots in the top section (1.5 m) of each tree. The needles were put in the same fixative solution as used for the buds and transported to the laboratory. The next day cross-section samples (size 0.5–1 mm) were cut approximately 3 mm from the tip of each needle, and subsequently fixed in the same way as for the buds. The needles born in 1999 are called C+1x and those born in 2000 Cx needles, “x” referring to the age of the needles in months (see Table 1). The time of needle birth was fixed to the end of May, i.e. the time of bud burst.

Cross-sections, about 2 μm thick, were cut for light microscopy from all the needles sampled and stained in the same way as for the buds. Quantitative analyses were made on the central cylinder and the tissues located there because this area is widely reported to react to nutrient disorders (Raitio 1983; Holopainen et al. 1992; Sutinen and Koivisto 1995; Fink 1999). The diameter of the vascular cylinder in needles of all ages was measured with the help of a scale in the microscope ocular, and the area of the cylinder was calculated by assuming that the cross-section of the vascular cylinder was a perfect circle. The number of sclerenchyma cells was counted and related to the area of the vascular cylinder. The fully matured needles, i.e. C6, C+111 and C+118, were digitally photographed and, in addition to the above-mentioned variables, the number of tracheids, the area of the xylem, the number of sieve cells in phloem rows (the height of the phloem) and the number of rows in the phloem (the width of the phloem), were determined from the digital pictures using Adobe Photoshop 6.

Statistical analysis

Statistical analyses were performed with SPSS statistical software package (SPSS for Windows, version 11.0, SPSS Inc., Chicago, IL). The data for microscopic variables were analysed separately for maturing needles (from C0 to C6) and for mature needles (C6, C+111 and C+118). Main effects (stand damage, needle age) were tested by General Linear Models (GLM, Univariate test) and differences among the stands by the Tukey’s test. The differences in nutrient concentrations of the needles and soil among the study stands were tested separately for C and C+1 needles by one-way Anova, and located among the stand means by the Tukey’s test. To avoid the possible effect of needle aging, as well as the known effect of starch accumulation on needle nutrient concentrations during summer (Linder 1995), the correlations (Pearson correlation test) between the needle nutrient concentrations and the structural variables were performed on C6 and C+111 needles, i.e. those sampled outside the growing season.

Results

Needle nutrient concentrations

In April only the N concentration of the C+111 needles differed significantly among the stands, being higher in the H stand than in the other stands. The N, P, and K concentrations of the previous-year needles decreased slightly towards July, and increased again in August. The B concentration in the previous-year needles were below 4 mg kg−1 in all the stands in the spring, but increased in the H and PD stands in July and August while remaining in the D stand at a very low level (Fig. 2). At the first sampling of current needles (7 June), the B concentration of the H stand (10.7 mg kg−1) differed significantly from that of the D (2.3 mg kg−1) and PD (3.7 mg kg−1) stands. Subsequently, the concentrations in the current needles from the H and PD trees first decreased, and then increased, but remained low in the D trees during the whole study period. The N, P, K, and Cu concentrations in the current needles were high in June and then decreased towards autumn, while the Ca and Mn concentrations followed an opposite trend (Fig. 2). No major changes were observed in the Mg and Zn concentrations. On 27 November, the nutrient concentrations in the C6 needles were, in general, higher than those in the C+111 needles in the spring. In addition to the N concentrations, there were also significant differences in the P, K, and B concentrations among the stands. The N concentration was highest in the H stand, and the P and K concentrations in the PD stand. The seasonal variations in the nutrient concentrations of all three spruce stands were relatively similar, apart from boron which showed a completely different behaviour (Fig. 2).

Fig. 2
figure 2

The development of needle dry weight (DW) and concentrations of boron, nitrogen, phosphorus, potassium and calcium in current (C) and previous (C+1) year’s needles of sample trees in damaged (D), partly damaged (PD) and healthy-looking (H) stands. Number of sample trees varied from 3 to 10 (see Table 1)

Full size image
Fig. 3
figure 3

Apex of Norway spruce bud from the visibly healthy looking terminal bud of a D tree in May. A Longitudinal section shows a round group of round cells rich in small lipid droplets in the upper part of the apex (A). Small distensions for bud scales (asterisks), as well as the uppermost developing needle (N) can be seen. The line in the base of the apex shows the cutting level for the cross-section in (B). B Group of collapsed dead cells in the meristematic area of the apex and cross-sections of developing needles (N) with tanniferous (T) epidermal cells can be seen. Bar = 100 μm for (A) and 20 μm for (B)

Full size image

The dry weight of the previous-year needles increased in all the stands up to mid-June (Fig. 2), and was significantly higher in the D trees than in the PD and H trees. In the current needles the dry weight increased from the end of August up until November: it was significantly higher in the D and PD trees than in the H trees (Fig. 2). The dry weight of the C6 needles showed negative correlations with the needle B (r=−0.475**) and Ca (r=−0.355*) concentrations, and positive correlations with the P concentrations (r=0.665**).

Bud emergence

The meristematic apex of emerging new buds had the same structure throughout the summer. In the middle of the ca. 0.2 mm long apex there was a round group of round meristematic cells that had a large nucleus, no vacuoles and a considerable number of small lipid droplets in the light cytoplasm. The diameter of this group was about 0.1 mm, and it was surrounded by two to three layers of meristematic cells without lipid droplets (Fig. 3a).

Bud and needle elongation started between 26 April and 22 May. The meristematic apex first produced the pith and the bud scales. The first signs of the bud scales were seen as small distension of the dividing meristematic cells at the base of the apex on 22 May (Fig. 3a). At that time also the first injuries, groups of dead cells characterized by precipitation of the cell contents, plasmolysis and collapse of the cells, were seen in these small distensions of the emerging bud scales and also in the outermost cells in the base of the apex (Fig. 3b) of the most healthy looking buds of the D trees. Bud scale formation continued up to 25 July at least. Whether the small distensions at the base of the apex on 25 July were for bud scales or already for needle primordia could not be determined. Injuries in the base of the apex of the emerging buds on the D trees, and in a few cases also on the PD trees, were subsequently seen at each sampling date in the same way as in the buds on 22 May (see Fig. 3b). In the end of July and August, five of the nine terminal buds sampled from the leader candidates of the D trees were fatally damaged, whereas all the buds sampled from the PD and H trees looked healthy.

Fig. 4
figure 4

Emerging terminal bud from the leader shoot of an H tree on 28 August. Upwards from fully developed needle primordia (Pn) the primordials are the smaller the higher in the bud they are situated, i.e. the emergence of the bud continues. A = apex. Arrows show living tanniferous cells. Cp = developing collenchyma plate. Bar = 100 μm

Full size image

On 28 August, a small primordial shoot, approximately 1 mm high (from the base of the collenchyma plate to the top of the meristematic apex) had emerged (Fig. 4). The uppermost, i.e. youngest needle primordia, consisted of a small group of cells, whereas the lowest needle primordia had a final size of about 0.1 mm thick and up to 0.2 mm long. The meristematic procambium extended from the earlier developed pith up to the lowest, first-developed needle primordia. The collenchyma plate was present, but still in contact with the earlier developed pith cells (Fig. 4). The primordial shoots had cells in which tannin had accumulated in the vacuoles of apparently living cells. The proportion of tanniferous cells (Fig. 4) per calculated area of the primordial shoots varied from 0 to 40% (mean 16%) in the D trees, and from 0 to 4% (mean 1.2%) in the PD and H trees. Bud development was complete, i.e. well-developed needle primordia could be seen up to the apex (see Fig. 1a), on 17 September.

The fresh cuttings of buds in November showed that fatally damaged buds only occurred in the D trees, with a minor exception of the lateral shoots of H the trees (Table 3). The mean number of lateral buds on the 5 cm section of both the leader and lateral shoots was the lowest in the D trees, but there were no significant differences compared to the PD or H trees (Table 3). In contrast, correlations were found between the B and also with the K concentrations of the C6 needles and the percentage of damaged terminal (B: r=−0.707**; K: r=−0.385*) and lateral (B: r=−0.677**; K: r=−0.460*) buds of the leader shoots. Corresponding correlations were also found between the B and K concentrations and the percentage of damaged terminal (B: r=−0.360°; K: r=−0.372*) and lateral (B: r=−0.390*; K: r=−0.473**) buds on shoots from the second whorl.

Table 3 Number of studied buds and proportion (%) of fatally damaged terminal and lateral buds, as well as the mean number of buds per 5 cm of branch from H, PD and D trees analysed under a light microscope from longitudinal fresh cuttings from material sampled in November
Full size table
Fig. 5
figure 5

Cross-sections of central cylinders from C6 needles of Norway spruce (A) from the healthy stand (H stand) and (B) from damaged stand (D stand) in November. Abbreviations in both pictures: X = xylem, P = phloem, S = sclerenchyma cell, T = tannin. Bars = 100 μm

Full size image

Needles

On 22 May, the differentiating new needles, still within the bud scales, already had resin ducts, the outlines of the central cylinder, sclerenchyma cells with very thin cell walls and the first tracheids with well-defined cell walls. The phloem tissue could not be discerned at the height where the cross-section was cut, i.e. at the level of the meristematic apex of the bud. In the needles from the D trees the accumulation of tannins in apparently living cells was frequent in epidermal (Fig. 3b) and mesophyll cells, whereas parenchyma cells in the central cylinder had only a few tanniferous cells. The H needles had a few tanniferous cells in the epidermis and mesophyll cells, and no tannin accumulation in the parenchyma cells. The PD needles were similar to the H needles, with the exception of the epidermis that had frequent tanniferous cells in about half of the studied needle cross-sections.

Table 4 Mean values (SD) of C+111 and C6 needle variables analysed under a light microscopy
Full size table

The C0 needles on 7 June were already well elongated and all the tissues were visible. During the maturation of the needles the area of the central cylinder significantly increased from 0.090±0.010 (mean±SD) mm2 in the C0 needles to 0.130±0.12 mm2 in the C2 needles. The number of sclerenchyma cells varied from 14.0±1.3 to 15.8±1.3, and did not show any changes during maturation. At the sampling on 25 July, i.e. in the C2 needles, the walls of the sclerenchyma cells had reached their final thick appearance, and the needle structure was similar to that of fully matured needles (Fig. 5a).

Because no significant (P<0.05) interactions between stand and needle age were found, the differences among the stands (D, PD and H stands) and between the needle age (C6, C+111 and C+118 needles), as well as between needle generations (C6 needles born in 2000 versus C+111 and C+118 needles born in 1999), are shown in Table 4. The area of the central cylinder was significantly larger in the D and in the PD trees than in the H trees, but the needle age or needle aging had no effect, whereas the number of sclerenchyma cells (Fig. 5b) differed significantly only between needle generations. Their number was also slightly less in the D trees than in the PD and D trees. The number of tracheids and area of the xylem were significantly larger in the D trees than in the H trees. The area of the xylem was significantly larger in the C+111 needles compared to the C6 or C+118 needles. The number of tracheids was also highest in the C+111 needles, but a significant difference was found only in the case of needle age. The number of sieve cells in a phloem row was higher in the PD and D trees than in the H trees, as well as in C6 needles born in 2000 compared to C+111 and C+118 needles born in 1999 (Table 4). The ratio of the number of sclerenchyma cells to the area of the central cylinder was highest in the H trees and in the 2000 needle generation. Quantitative analysis of tannin accumulation was not performed on mature needles, but it seemed to be more evident in D than in H needles (Fig. 5a and b).

Correlations (Table 5) for both C6 and C+111 needles with the corresponding needle B concentration (part of data shown in Fig. 2) were found in the area of the central cylinder and in the ratio of the number of sclerenchyma cells to the area of the central cylinder. In contrast, the area of the xylem and number of the tracheids showed correlations with the needle B concentrations only in the C+111 needles. In these cases, very similar correlations were also found with the needle N concentrations. However, the ratios of B to N, or to other nutrients, did not show higher correlations than B alone. In addition to B and N, the Cu and Fe concentrations showed some correlations with the needle structural variables (Table 5).

Table 5 Correlations between the needle structural variables with the needle nutrient concentrations in C+111 and C6 needles
Full size table

Discussion

In this study, cell death in the apical meristem of the emerging buds led to the cessation and death of a large number of new buds throughout the summer in the D trees. The number of fatally damaged buds, together with the extremely low B concentration in the current needles throughout the growing season, indicate an inadequate supply of boron to emerging buds in the D trees. Negative correlation with bud damage was found not only with the needle B concentration but also with the needle K concentrations, which may be due to enhanced leakage of K+ ions under B deficiency (Cakmak et al. 1995). The time frame within which the buds emerged, first the bud scales and then the primordial needles (cf. Bilková et al. 1999), was about 4 months in the spruce trees, as also reported in Picea sitchensis (Cannel and Willet 1975) in Scotland and in Picea abies (L.) Karst. (Svobodová et al. 2000) in two Czech mountainous regions. Thus, it is evident that an adequate supply/uptake of boron during the whole summer is an absolute necessity for the emergence of healthy buds.

The results of our structural studies indicate that tannins play a role in buds in at least two different ways. First, precipitation of the cell contents and consequently cell death in the meristematic apex is most probably due to the release of tannins after membrane rupture, similarly to that reported by Dordas and Brown (2005) in rose cell cultures. Second, the increase in the number of tanniferous cells, where tannin accumulates in the vacuoles of living cells, occurred in the differentiating needles and in the primordial shoots of the emerging buds. This type of tannin accumulation indicates a metabolic stress response, as noted e.g. in leaves of B-deficient Olea europaea (Liakopoulos and Karabourniotis 2005). Thus, the phenolic metabolism of the different organs of the plant, and different tissues of the same organ, may react differently to boron deficiency, as also reported in Dittrichia viscosa by Stavrianakou et al. (2006).

Differentiation of the needle tissues occurred before bud burst, as also noted earlier in spruce (de Faÿ et al. 2000). Needle maturation continued up to the end of July, when expansion of the central cylinder and thickening of the walls of the sclerenchyma cells, as well as shoot growth, were complete. Aging of the needles from 11 to 18 months was manifested as an increase in the sieve cells in the phloem rows, and as a decrease in the area of the xylem and in the number of tracheids, possibly due to the collapse of the outermost tracheids in aged needles. Secondary phloem growth, but little or no secondary xylem growth, have also been reported earlier in several coniferous trees (Gilmore et al. 1995 and references therein). The results indicate that needle age must be taken into consideration when comparing the structure of needles among given stress factors or treatments.

Due to the time frame of differentiation and maturation of the needles, the supply of B, as well as of other nutrients, must be sufficient from very early spring up to the end of July in order to ensure proper needle tissue development. In our case, the B concentrations in the current as well as in the previous-year needles of the D trees were around 2 mg kg−1, which is far below the value of 4–5 mg kg−1, considered to be the deficient level (Jukka 1988; Braekke 1994). In fact, the B concentrations of both 1999- and 2000-needle generations in the PD trees were also below the deficiency level up until the end of July. The levels of other foliar nutrients in each stand were, however, similar to those normally found in spruce stands in Finland (Tamminen and Saarsalmi 2004).

The observed “zigzag” pattern in the B concentrations through the summer seems to be connected to differences in the amount of precipitation during the days before sampling. For example, 6 days before sampling on 26 June, when the B concentrations were at their lowest, the amount of precipitation was only 0.2 mm, whereas before 25 July, when the B concentrations were relatively high, the precipitation during the six preceding days was 49.8 mm. In addition to the low precipitation and thus low soil moisture, a low soil temperature (Carter and Brockley 1990; Stone 1990; Goldberg 1997) and higher soil pH also negatively affect the supply of B (Goldberg 1997). In our case, the pH was below 4.6 in the mineral soil, and below 3 in the humus. The acidity of the mineral soil was the same order of magnitude as that reported for sites of this type in Finland (e.g. Tamminen 1991), and should not increase the adsorption of B onto soil particles (Goldberg 1997). The reason why the D trees did not show the “zigzag” pattern may be due to a reduced capacity to take up water as a result of reduced root growth caused by low soil B concentrations (Möttonen et al. 2001).

The increase in the size of the area of the central cylinder and xylem, seen in this study in the D and PD trees, are xeromorphic characters (e.g., Correia et al. 2001; Richardson et al. 2001) and in our case indicate morphological acclimation responses to drought during needle differentiation. Precipitation from 15 May to 22 April, i.e. during the period of tissue differentiation, totaled 38 mm in 2000 and only 20 mm in 1999, while during the 19-day period before sampling there was no rain. Thus, although these variables correlated negatively with the needle B concentrations, the increase in the areas cannot be attributed to the effect of B deficiency. This is supported by our experiment with pine and spruce seedlings, which were well watered with no B or adequate B, in which no differences were found in the area of the central cylinder or the xylem (Sutinen et al. unpublished).

The small number of sclerenchyma cells noted in our study has been reported earlier in Scots pine needles with low B, a high N/B ratio or with low K and P concentrations (Raitio 1983; Jokela et al. 1995; Sutinen et al. unpublished). The ratio of the number of sclerenchyma cells to the area of the central cylinder eliminates, or at least reduces, the effect of the size of the needle and that of the central cylinder and thus allows comparisons to be made between different sized needles. Our other studies on mature spruces in the field, as well as experimental studies with seedlings, have shown that the ratio in healthy spruce trees with a balanced nutrition can be as high as 150, or even more (unpublished data). Compared to these values, even the highest ratio of 119 in the H trees was rather low, which indicates that the needles even in the H trees were not completely normally developed. This is understandable when the B concentration in even the H trees (means for C+111 needles 3.4 mg kg−1 and for C6 needles 5.0 mg kg−1) is compared to the deficient levels of 4–5 mg kg−1 (Jukka 1988; Braekke 1994).

The variation in the dry weight in the previous-year needles during the study summer, with peak values in June and low values in the end of July, is similar to that reported by Linder (1995) and Raitio and Merilä (1998), and is due to the accumulation of starch grains that can increase the dry weight of the needles by up to 40% during the summer (Linder 1995). The significantly higher dry weight of the needles from the D trees (previous-year needles) and from the PD and D trees (current needles), compared to the H trees, may be connected to the significantly larger area of the central cylinder and the xylem tissues. Furthermore, the increase in the dry weight of the current needles in the D and PD trees, but not in the H needles, up to early winter may be connected, at least partly, to the continuing increase in tannins that was seen in the differentiating needles in the D and PD needles.

The reason for the fact that visibly damaged trees may grow side-by-side with healthy looking trees is unknown, but genetic differences could be one explanation (Rerkasem and Jamjod 1997). Although the origin of the seeds in the D and PD stands was unknown, it is most likely that the origin is within a distance of 300 km from the stand because this is the normal practice, and is the accepted rule for spruce seed transfer from south to north in Finland. Furthermore, similar visible and microscopic changes, together with low B concentrations in the needles, have also been reported in experiments with B-deficient seedlings (Sutinen et al. unpublished), as well as in forest stands established with seed of local origin (Saarsalmi and Tamminen 2005). Thus, an unsuitable origin cannot be a major reason of the observed changes. The heterogeneity in canopy health may also be partly due to small-scale variation in the properties of both the organic and mineral layers of boreal forest soil (Järvinen et al. 1993; Möttönen et al. 1999).

In conclusion, the results showed that boron deficiency had fatal effects on the buds, i.e. cell death, characterized by precipitation of the cell contents in the bud apex, starting at the time of onset of growth and continuing through the summer and finally leading to the death of numerous emerging buds, and consequently to a bushy appearance of the trees. To avoid this, the supply of boron should remain adequate throughout the vegetation period. The other changes were the B-related stress and/or defense response, seen in the needles and buds as an increase in the number of living cells that had accumulated tannins. An increase in needle xeromorphy, seen as an increase in the area of the central cylinder and the xylem, are due to the dry soil conditions during needle differentiation.