Introduction

Vascular bundles in tracheophytes are composed mostly of xylem and phloem, the two principal long-distance transport pathways. Xylem serves as a conducting tissue primarily for transferring water and nutrients from the soil to leaves, whereas phloem serves as the pathway for transporting photosynthates, genes, and hormones from leaves or storage organs (source) to regions of active growth or storage organs (sink). The loss of secondary phloem due to wounding can, therefore, be detrimental to overall tree vitality (Aulen 1991; Rawlinson et al. 2004). Numerous studies have reported that the manipulation of source–sink relationships by stem girdling can effectively kill trees, although the time taken for trees to die can vary between species (Negreros-Castillo and Hall 1994; Wilson and Gartner 2002; Taylor and Cooper 2002; Pariona et al. 2003).

Stem girdling has been shown to lead to an accumulation of non-structural carbohydrates (NSC) in leaves and stems above the girdle (Myers et al. 1999; Murakami et al. 2008) and to the gradual depletion of carbohydrate reserves in roots and, subsequently, to starvation (Weaver and McCune 1959). At the same time, removal of the root sink generally results in decreased stomatal conductance (López et al. 2015), in an accumulation of abscisic acid in leaves (Setter et al. 1980; Williams et al. 2000) and to a consequential decline in photosynthesis through end-product inhibition (Myers et al. 1999; Urban and Alphonsout 2007; Murakami et al. 2008). Lower transpiration in girdled trees due to stomatal closure leads to a reduction in leaf hydraulic conductance (Domec and Pruyn 2008; Sellin et al. 2013) and to an increase in leaf water potential (Williams et al. 2000).

Girdling affects annual ring width, the duration of cambial growth, as well as the start time and duration of latewood production (Wilson and Gartner 2002; Domec and Pruyn 2008; Maunoury-Danger et al. 2010). As a consequence, the hydraulic properties of wood produced after girdling differ to those of unwounded trees (Cernusak and Marshall 2001; Domec and Pruyn 2008). However, no studies have assessed the effects of stem girdling on leaf and needle xylem structure. Leaves and needles typically comprise a substantially smaller proportion of the path length in plant vascular systems than stems, yet they represent a disproportionately large fraction (> 30%) of whole-plant hydraulic resistance (Yang and Tyree 1994; Nardini and Salleo 2000). The current consensus is that the hydraulic resistance of leaf/needle xylem is about equal to that of extra-xylem pathways (Nardini et al. 2005; Sack and Holbrook 2006). Thus, needle xylem structure is a key component for understanding tree hydraulic architecture. The other interesting question is how long-term phloem transport disruption influences the correlation between needle structural parameters given that anatomical traits can differ dramatically from needle to needle along a canopy (Gebauer et al. 2015). These features are likely regulated by an interplay between endogenous and environmental factors (Bongard-Pierce et al. 1996). For example, leaf lamina growth requires coordination between the epidermis and the mesophyll layers (Kawade et al. 2013). Thus, leaf size is expected to be a complex trait (Zhao et al. 2011) and any disturbance such as the accumulation of abscisic acid and carbohydrates above the girdle ultimately interferes with normal leaf development.

The current study focused on Scots pine (Pinus sylvestris L.) as it is the most widely distributed pine species and one of the most important timber species in Eurasia (Poyatos et al. 2007). The aim of this study was to increase our knowledge of the effects of stem girdling on needle xylem structure adaptation of Scots pine trees 8 and 14 months after girdling (hereafter referred to as 227 and 411 DAG—days after girdling, respectively). We tested the following three hypotheses: (1) girdling will change the correlations between needle structural parameters as a result of sink–source disturbance; (2) xylem area, tracheid dimension, the number of tracheids and theoretical hydraulic conductivity will increase as a result of carbohydrate accumulation in needles 227 DAG; and (3) the opposite will be observed 411 DAG, as root starvation will lead to decreased sap flow and water shortage in the needles.

Materials and methods

Study site and experimental design

The study site is an even-aged mixed coniferous/deciduous forest stand located in the Training Forest Enterprise Masaryk Forest in Křtiny (with site coordinates: 49°15′39″N, 16°36′20″E), an organizational unit of Mendel University in Brno (Czech Republic). The altitude of the site is 340 m a.s.l. Long-term average annual air temperature and precipitation are 8.7 °C and 490 mm, respectively, based on daily weather data from 1961 to 1990 from the nearest meteorological station at Brno–Tuřany. At this location, the mean air temperature for the whole of 2014 was 11.3 °C and the sum of precipitation was 609 mm. The stand was planted in 1928, and the abundance of Scots pine (Pinus sylvestris L.) is higher than 70%. The soil type is mesotrophic Cambisol. Two experimental plots (each measuring 2500 m2, situated 200 m apart) were established within the stand in March 2013. Eleven co-dominant Scots pine trees, all 86 years old, with breast height diameters ranging from 29 to 36 cm were selected from each experimental plot for detailed analysis (22 trees in total). On each experimental plot, three trees were selected as controls, four trees were girdled on July 15, 2013, and additional four trees were girdled 6 months later on January 15, 2014 (Fig. 1). Girdling consisted in removing 7–10-cm-wide bands of bark (periderm, cortex, secondary phloem and cambium) from around the entire stem at the height of 1.3 m above the soil surface, thus exposing the outer surface of the sapwood and disrupting phloem transport. Stem tissue regeneration was regularly checked to avoid phloem sap flow reconnection. The girdled area was completely covered with resin at the end of experiment.

Fig. 1
figure 1

Diagram illustrating individual actions (bottom axis labels and black vertical bars) applied to control or girdled trees during experiment duration (upper axis labels). Two groups of trees were girdled separately (period after girdling in gray) on July 15, 2013, and January 15, 2014. Sap flow and soil water potential were measured from April 1 to August 30, 2014. Needle samples for non-structural carbohydrates (NSC) analysis were taken on April 1 and August 30, 2014. Needle samples for anatomical analysis were taken on August 30, 2014. DAG days after girdling

Full size image

Sap flow and phenology

Sap flow was monitored from April 1 to August 30, 2014 in all experimental trees (Fig. 1), using EMS 51 sap flow meters (EMS Brno, Czech Republic) on the basis of the trunk heat balance principle (Čermák et al. 1973; Kučera et al. 1977). Sap flow sensors were located above the girth. To account for variability in tree size, sap flow was standardized per unit of a circumference and expressed in kg cm−1 of the tree circumference. Soil water potential was measured at three depths (15, 50, and 90 cm) from April 1 to August 30, 2014, with gypsum blocks (GB1, Delmhorst, USA) connected to a MicroLog SP3 datalogger (EMS Brno, Czech Republic). Phenological events (bud burst, beginning of flowering and end of flowering) were observed for each tree in the year 2014.

Carbohydrate analysis

Current-year needles from sun-exposed branches were used to analyze levels of non-structural carbohydrates (NSC). Needles were collected from the crowns (at least twenty from each tree) on two separate dates (April 1 and August 30, 2014; Fig. 1). Samples were immediately frozen and later dried in a laboratory oven at 65 °C until constant weight was achieved. We followed the protocol devised by Poorter and Berkel (2010) in measuring concentrations of soluble and insoluble NSC in the needles. Soluble NSC (total sugars) was extracted from 50 mg needle samples using 80% ethanol and centrifugation. Lipids and chlorophyll were removed from the supernatant using chloroform. Insoluble NSC was extracted by boiling the centrifuged sample in HCl. Complex saccharides were broken down to hexoses using the Anthrone. NSC concentrations were estimated spectrophotometrically using Spekol 1300 (Chromatographic Specialties, Brockville, Canada) operating at a wavelength of 625 nm.

Sampling for needle anatomical analysis

Five current-year shoots (from each tree) with fully developed needles were collected from branches on the south-facing azimuthal side of the canopy growing in the middle part of the canopy on August 30, 2014 (i.e., 227 and 411 days after girdling; hereafter referred to as DAG) (Fig. 1). In the field, the samples were fixed in FineFIX solution (Milestone, Bergamo, Italy). Subsequently, four randomly selected fascicles (with two needles per fascicle) were removed from each of the collected shoots and excised. One needle from each fascicle was used for xylem structure analysis (i.e., 20 needles per tree), and a second needle for measuring the projected needle surface area.

Needle xylem structure

Hand-cut cross sections for anatomical analyses were taken at the needle base and treated with phloroglucinol and hydrochloric acid staining to show lignin in red. The stained sections were examined under an Olympus BX51 light microscope (Olympus Corporation, Tokyo, Japan) at a magnifications of up to 400×, and photographed using a Canon EOS 1100D digital camera (Canon Inc., Tokyo, Japan) connected to a computer by QuickPhoto Micro 3.1 software (Promicra, Prague, Czech Republic). Tracheid lumens were manually colored using Adobe Photoshop 9.0 (Adobe Systems, San Jose, CA, USA) (Fig. 2). The phloem area (A p), xylem area (A x), mean tracheid lumen area (A lum), and maximum and minimum lumen diameters (d max and d min, respectively) were measured using ImageJ software (The University of Texas Health Science Centre, San Antonio, TX, USA). Tracheid roundness (T r) was defined as the ratio of d max to d min, with a ratio of 1 denoting tracheids with a perfectly round cross section and larger ratios denoting more or less ellipsoidal tracheids. We also quantified the number of tracheids per needle (T num), total lumen area per needle (A n_lum), and xylem to phloem area ratio (R a).

Fig. 2
figure 2

Detailed micrograph of the vascular cylinder of a Scots pine needle. Xylem area is shown encircled by the white line (a). Tracheid lumens were filled in with white using Adobe Photoshop 9.0 (b) and measured using ImageJ software. The black bar represents 50 µm

Full size image

Theoretical hydraulic conductivity

The tracheid theoretical hydraulic conductivity (K th_t) was calculated according to the Hagen–Poiseuille law (Eq. 1) (Cruiziat et al. 2002). Since the tracheid lumen is not usually circular, we applied a minor modification to the formula, as recommended by Martre et al. (2000) (Eq. 2).

$${K_{{{\rm th}\_{\rm t}}} = \frac{\pi}{8\eta}r_{\rm lum}^{4}},$$
(1)
$$r_{\rm lum}^{4} = \frac{{d_{\max}^{3} d_{\min}^{3}}}{{8d_{\max}^{2} + 8d_{\min}^{2}}},$$
(2)

where r lum is the tracheid lumen radius and η is the viscosity of water at 20 °C (1.002 e−9 MPa s).

The needle theoretical hydraulic conductivity (K th_n) was calculated as the sum of all K th_t per needle. Then, the xylem-specific hydraulic conductivity (K s_xylem) was estimated as K th_n divided by A x, and needle-specific hydraulic conductivity (K s_leaf) was estimated as K th_n divided by the projected needle surface area. To determine the projected needle surface area, 20 fresh needles per tree were scanned. Further, all K th_t in each needle were sorted in descending order and cumulatively summed (starting from the largest tracheid) until the sum equals to 95% of K th_n. The mean diameter (D 95) was then calculated for this subset of tracheids, as recommended by Tyree and Zimmermann (2002).

Statistical analysis

A linear mixed effect model (LME) with restricted maximum likelihood method (REML) was used, where individual trees were specified as a random effect (Zuur et al. 2009). The simple linear model was used for K s_leaf. Subsequently, Tukey’s honest significant difference (HSD) test was used to differentiate between treatments (significance level 0.05). We used the open source R software environment (R Core Team 2015) for all statistical analysis. LME values were fitted using the lme function in the nonlinear mixed effect (NLME) library, and Tukey’s test was performed using the generalized linear hypothesis test (GLHT) function in the multcomp library. Final empirical equations were derived using a generalized linear model with gamma distribution and identity link since variance was increasing with the predictor. Mean values are presented with ± standard errors unless stated otherwise.

All acronyms, abbreviations, and symbols are defined in Table 1.

Table 1 Overview of the needle parameters, their abbreviations, definitions, and units used throughout this study
Full size table

Results

Sap flow and phenology

Sap flow was highest in the control trees and lower in both groups of girdled trees (p < 0.001). There was no difference between the trees 227 DAG and 411 DAG (p = 0.147). Sap flow accumulated within the period April 1 to August 30, 2014 was 23.5 ± 2.6, 11.3 ± 5.3, and 7.2 ± 3.8 kg cm−1 of tree circumference in the control trees, trees 227 DAG and trees 411 DAG, respectively.

There were no phenological differences between girdled and control trees. Budburst occurred between April 23 and 24 , flowering started between May 7 and 9, and flowering ended between May 23 and 26 .

Carbohydrates

The NSC concentration in needles varied between the two sampling days, in April and in August (Fig. 3). In April, insoluble NSC concentrations were higher in both groups of girdled trees than for the control trees (p < 0.001) (Fig. 3a), while soluble NSC concentrations were the same for all three groups of trees (p = 0.19) (Fig. 3a). In August, needles 227 DAG had more insoluble NSC (p < 0.001) than needles 411 DAG and control trees, which did not differ from each other (p = 0.55) (Fig. 3b). The concentrations of soluble NSC were higher in needles 227 DAG (p < 0.001) and lower in needles 441 DAG (p < 0.001) compared to the control trees (Fig. 3b).

Fig. 3
figure 3

Non-structural carbohydrate concentrations (NSC; soluble—black column; starch—gray column) for April 1 and August 30, 2014. Vertical bars indicate standard errors. Asterisks indicate a significant difference from control

Full size image

Correlation between needle parameters

Several needle parameters were correlated with each other in control needles (16 out of 45 possible correlations), in needles 227 DAG (12 correlations), and needles 411 DAG (22 correlations) (Fig. 4). From all observed correlations, 11 were identical for all three treatments. The strongest relationships were found for K th_n, d min, and K s_xylem (Fig. 4). The less well-correlated parameters were T r and R a. In practical terms, the correlations between A x and K th_n and between A x and T num in all treatments allowed us to devise empirical equations to calculate the more difficult-to-measure parameters K th_n and T num from the more easily measured A x (Fig. 5).

Fig. 4
figure 4

Correlation matrices for needle parameters of control trees (upper panels), trees 227 days after girdling (DAG; middle panels), and trees 411 DAG (lower panels). Pearson’s correlation coefficient (numbers in boxes) and significance level (asterisks in boxes) are distinguished in background color: black box (r ≥ 0.6; p < 0.05); white box (r < 0.6; p > 0.05). P values are denoted as p ≤ 0.001 (***), p ≤ 0.01 (**), p ≤ 0.05 (*), p ≥ 0.05 (–)

Full size image
Fig. 5
figure 5

Linear regressions of theoretical needle hydraulic conductivity (K th_n) and a number of tracheids (T num) per needle with needle xylem area (A x). The p value was lower than 0.001 for each regression line

Full size image

Needle xylem structure 227 DAG

On average, A lum, d max, and d min increased in needles 227 DAG by 8, 9, and 16%, respectively, compared to the control needles (Fig. 6a, d, e). It resulted in as much as 9% higher D95 in needles 227 DAG than in the control needles (Fig. 6f). The increased tracheid dimensions in needles 227 DAG also corresponded to 27% higher K th_n and 23% higher K s_xylem compared to the control needles (Fig. 6g, h). While A n_lum, A x, and K s_leaf in needles 227 DAG were also higher than in the control needles (Fig. 6b, c, h), these differences were not significant (Table 2). Other studied needle parameters (R a, T num, and T r) 227 DAG showed similar levels to those for the control needles (Fig. 6j, k, l and Table 2).

Fig. 6
figure 6

Means of needle parameters from Scots pine control needles and needles 227 and 411 days after girdling (DAG). Vertical bars are 95% confidence intervals derived from the appropriate model. Parameters that are significantly different at the 5% significance level are indicated by different letters

Full size image
Table 2 P values from the Tukey’s honest significant difference (HSD) test showing the probability of differences in parameters between control needles and needles 227 or 411 days after girdling (DAG) and between needles 227 and 411 DAG
Full size table

Needle xylem structure 411 DAG

Patterns of needle xylem structure and development in needles 227 DAG were different to those observed in needles 411 DAG (Fig. 6). Tracheid dimensions, K s_leaf, and K s_xylem were similar in both 411 DAG and control needles (Fig. 6d, e, h, l and Table 2); however, K th_n was as much as 39% lower than in control needles (Fig. 6g and Table 2). Lower K th_n can be explained by the 36% lower level of T num in needles 411 DAG (Fig. 6k), which also corresponded with lower A n_lum and A x (both by 36%) in needles 411 DAG (Fig. 6b, c). In contrast to their levels, 227 DAG, R a, and T r increased in needles 411 DAG when compared to the control needles (Fig. 6j, l and Table 2). This indicated that decreased d min in needles 411 DAG caused tracheids to be more ellipsoidal.

Discussion

For both ecologists and physiologists, foliar morphology and anatomy are important for understanding the physiology of the whole tree (Yan et al. 2012). The control of leaf size, and, by extension, leaf form, is a principle way in which plants adapt to environmental conditions. These features are likely regulated by the interplay between several endogenous and environmental factors (Bongard-Pierce et al. 1996). Our study on Scots pine clearly shows that stem girdling in this species triggers significant structural changes in needle xylem, which corresponds to previously measured morphological and physiological changes observed after stem girdling in several tree species (e.g., Myers et al. 1999; Rivas et al. 2007; Urban and Alphonsout 2007; Murakami et al. 2008). We found evidence of complex relationships related to growth, as several needle xylem parameters correlated with one or more of other parameters (we found 11 correlations between needle parameters common for all three variants of experiment out of 45 possible combinations). Only T r and R a were developed independently of the other studied parameters and did not correlate with anything else. The number of intercorrelated parameters decreased in trees 227 DAG to 12, but increased in trees 411 DAG to 21 as compared to 16 in control trees. These findings support our first hypothesis that change in the physiology will affect on the correlation among needle xylem anatomical traits. We expected to find strong correlations between K th_n and d max, d min and T num, as K th_n should theoretically be more or less dependent on these parameters. However, such a correlation was weak in both the control needles (weak correlation with T num) and the needles of girdled trees (weak correlations with d max or d min). This resulted especially from weak correlations of T num with the dimensions of tracheids (d min, d max, D 95), even though one could expect a negative correlation here. Such a weak correlation suggests that cell expansion takes place, at least to some extent, aside from the development of a number of tracheids, as some tracheids could already be developed in dormant buds; however, their dimensions are thereafter affected by the conditions of the following growing season. Nevertheless, the strong correlation of T num and K th_n with A x facilitated the development of empirical equations in the present study, which enabled K th_n and T num to be calculated from the more easily measured A x.

Several needle xylem parameters were influenced by girdling 227 DAG. Most importantly, tracheid diameters increased which brought along increased values of K th_n and K s_xylem as hydraulic conductivity increases with the fourth power of the conduit diameter, according to Hagen–Poiseuille equation. In spite of increased needle hydraulic conductance, the sap flow decreased. It indicated reduced stomatal conductance (Meinzer et al. 1997; Phillips and Oren 1998; Ewers et al. 2007; Kučera et al. 2017), which was also observed in other studies on girdled trees (De Schepper and Steppe 2011; López et al. 2015). The reason for stomatal closure after stem girdling could be the feedback inhibition of photosynthesis due to the accumulation of NSC in leaves (Fig. 2) (López et al. 2015), which leads to a more favorable leaf water status (Williams et al. 2000; Sellin et al. 2013). Such conditions could have lead to the increase in tracheid dimensions 227 DAG, as cell expansion is positively affected by higher water potential (Dale 1988), and a higher NSC concentration helps to maintain turgor pressure, which plays a key role in cell growth (Morgan 1984; De Schepper and Steppe 2011). Since NSC content is also the main source of energy driving cell division and expansion, we also expected increase in A x and T num 227 DAG. However, such an increase was not observed in our study. The reason could be that needles were already partly developed in dormant buds, which were established before the girdling treatment. Thus, the second hypothesis was only partially supported.

In contrast to trees 227 DAG, prolonged exposure to stem girdling in trees 411 DAG seemed to bring more unfavorable conditions concerning tree survival. Considering the more favorable leaf water status in 227 DAG needles, the leaf water potential of girdled trees should later decrease as the roots are gradually depleted of their carbohydrate reserves causing them to suffer from starvation (Weaver and McCune 1959). This scenario led to very low transpiration rates in trees 411 DAG compared to control trees and consequently to the death of the whole tree (Negreros-Castillo and Hall 1994; Pariona et al. 2003), which in our study, was observed to occur in all 411 DAG trees. It is difficult to assess the contribution of plant water and carbon balance to tree death as they are interdependent (McDowell 2011; Hartmann et al. 2013a). For example, Hartmann et al. (2013a) showed that drought killed trees much faster than carbon starvation; however, they also showed that drought caused phloem pathway disturbance leading to carbon starvation in roots, similar to girdling (Hartmann et al. 2013b). Poor leaf development and smaller leaf size were previously observed in three deciduous tree species during the second growing season following stem girdling (Baldwin 1934). This corresponds to our observations of decreased A x levels in Scots pine needles 411 DAG, which were as much as 36% lower than A x levels in the control needles. Interestingly, the tracheid dimensions were nearly the same as for control trees. However, a significant reduction in the number of tracheids in needles (T num) was observed 411 DAG. Such a reduction resulted in lower K th_n. On the other hand, both K s_leaf and K s_xylem 411 DAG were similar to control needles, as there are several close interconnections between needle parameters. The decrease in K th_n was coupled with a reduction in A x (used for K s_xylem calculation) and also with a reduction in the projected needle surface area (used for K s_leaf calculation) leading to only marginal changes in K s_xylem and K s_leaf. These observations partly support the third hypothesis and confirm that several leaf parameters are closely linked to each other, as previously proposed by Sack and Holbrook (2006).

In general, K s_leaf is a parameter representing the capacity of a shoot to supply water to distal parts (i.e., leaves) (hydraulic sufficiency; Tyree and Ewers 1991), and K s_xylem is a useful parameter for gauging the hydraulic efficiency of a particular tree segment (Cruiziat et al. 2002). According to the studies above, the needle hydraulic sufficiency observed here did not depend on how much time passed after stem girdling, in contrast to needle hydraulic efficiency, which increased in needles 227 DAG. It should also be noted that calculated values of hydraulic conductivity can differ from actual measured values. Measured hydraulic conductivity, for example, was found to be only 30–50% of the calculated conductivity in conifers due to the high fraction of pits and pit membranes hydraulic resistance (Froux et al. 2002; Sperry et al. 2006). Nevertheless, for tracheid diameters below 20 µm, which was the case here, lumen accounts for a much greater proportion of hydraulic resistance than pits (Hacke et al. 2005). Thus, in our study, the calculated hydraulic conductivity could have been close to measured values. A new and promising technique was recently developed which enables direct measurements of needle hydraulic conductivity using the flow meter technique (Charra-Vaskou and Mayr 2011; Charra-Vaskou et al. 2012). With this innovation, the relationship between calculated and measured needle hydraulic conductivity can be studied in future research to evaluate resistance partitioning within needles.

In addition to stem girdling, other factors such as physical site characteristics, climate conditions, tree age, nutrition, and overall tree health also play important roles in needle development. The d max in our trees was ~ 7.8 µm, which was 4 µm less than in Scots pine trees growing in optimal conditions in central France (Cochard et al. 2004) and similar to the d max of Scots pine trees growing in Belgium in soil with low levels of magnesium and phosphorus and high levels of nitrogen (Neirynck et al. 2008; Gebauer et al. 2015). The T num in control needles (~ 110) also resembled that for Scots pine trees growing in nutrient-imbalanced soil in Belgium (Gebauer et al. 2015), but was substantially lower for needles 411 DAG. These lower tracheid dimensions and numbers may have been the result of drought stress that occurred in spring at the beginning of needle development. While total precipitation for 2014 was above normal, these rain events were not evenly distributed throughout the season. Lower precipitation (69% of normal) occurred in winter and spring, from January to June, which resulted in soil drought conditions at the study site (Supplementary Figure S1, Fajstavr et al. 2017).

Conclusions

The anatomical parameters of Scots pine needles were affected both by summer and winter girdling. The date of girdling and time elapsed after girdling were important factors in how the hydraulic parameters of needles were affected. Needles which were sampled 227 days after girdling (DAG) were girdled in winter, and therefore, their basic structure was already present in buds which were formed in the summer of the previous year. Tracheid lumen area, theoretical needle hydraulic conductivity, and xylem-specific hydraulic conductivity were by 18, 39, and 28%, respectively, larger in needles 227 DAG than in control needles. Nevertheless, xylem area of needle and number of tracheids per needle 227 DAG were similar to control needles. Thus, hypothesis 2 was only partially supported. In contrast, trees girdled in summer (July) and sampled 411 DAG developed their buds after girdling. Moreover, 411 DAG provided enough time to deplete carbohydrate reserves in the roots. It corresponded with very low transpiration rates measured in trees 411 DAG compared to control trees. Therefore, total tracheid lumen area per needle, xylem area of the needle, theoretical needle hydraulic conductivity, and number of tracheids per needle in trees 411 DAG were by 48, 42, 56, and 39% smaller than in control needles. Nevertheless, tracheid dimensions 411 DAG were similar to control needles. Thus, hypothesis 3 was only partially supported. Girdling changed the internal balance during the development of needle xylem and phloem. We identified 16 correlations between the studied needle parameters in control trees, 11 in the trees 221 DAG, and 21 in trees 411 DAG (hypothesis 1 was supported). Two empirical equations were constructed, which enabled theoretical needle hydraulic conductivity and number of tracheids per needle to be calculated from an easily measurable xylem area of the needle. These equations could be useful tools for understanding whole-tree hydraulic architecture, as needles represent a disproportionately large fraction of the whole-plant hydraulic resistance (Nardini and Salleo 2000; Sack and Holbrook 2006). Understanding of this balance between various needle parameters will help us to understand mechanisms of regulation of stomatal conductance, photosynthesis, and transpiration.