Anatomical Features and Its Radial Variations among Di ﬀ erent Catalpa bungei Clones

: Research highlights: Annual wood anatomy (xylem) aids our understanding of mature wood formation and the growth strategies of trees. Background and Objectives: Catalpa bungei is an important native species in China that produces excellent quality wood. Herein, we clariﬁed the e ﬀ ects of the genetic origin and cambial age on the anatomical characteristics of C. bungei wood. Materials and Methods: Six new 13-year-old C. bungei clones: ‘1-1’ (n trees = 3), ‘1-3’ (n trees = 3), ‘2-7’ (n trees = 3), ‘2-8’ (n trees = 3), ‘8-1’ (n trees = 4), and ‘9-1’ (n trees = 3) were removed for study from a plantation in Tianshui City, Gansu province, China. Xylem features were observed and the anatomical variables were manually measured via image analysis on (macro- micro-, and ultra-) features cut from radial increments of earlywood and latewood sampled at breast height. Results: Between the age of 1 and 2 years, wood was di ﬀ use-porous; between the age of 3 and 9 years, wood was semi-ring-porous; and between the age of 10 and 13 years, wood was ring-porous. The e ﬀ ect of clones on anatomical characteristics was signiﬁcant except for the microﬁbril angle in latewood and ring width. The transition between juvenile and mature wood was between 7 and 8 years based on patterns of radial variation in ﬁber length (earlywood) and microﬁbril angle. From the pith to the bark, ﬁber length, double wall thickness, ﬁber wall: lumen ratio, vessel diameter in earlywood, proportion of vessel in earlywood, and axial parenchyma in latewood increased signiﬁcantly, whereas ring width, earlywood vessels, and the proportion of ﬁber decreased signiﬁcantly. In addition, other features, such as vessel length, microﬁbril angle, and ray proportion, did not di ﬀ er signiﬁcantly from the pith to the bark. Conclusions: Breeding program must consider both clone and cambial age to improve the economic proﬁtability of wood production.


Introduction
Catalpa bungei is a woody plant belonging to the genus Catalpa, from the Bignoniaceae family [1], commonly used to make furniture, coffins, musical instruments, and boats in ancient China (since 600 BC) for its high dimensional stability, resistance to deterioration, and mechanical performance [2,3]. Because natural plantations of C. bungei have decreased sharply in their abundance over the last century and are now protected, the utilization of C. bungei in the wood industry has decreased over the last two decades. Fast-growing C. bungei plantations have become increasingly cultivated to address the competing needs of the demand for wood products and the protection of natural forest. Much work in plant husbandry has been conducted to select genotypes that can grow fast, adapt to growing in various habitats, and are resistant to damage inflicted by pathogens and herbivores. The selection of genotypes has been successful given that several clones of C. bungei planted in central and northern parts of China now show many of these desired properties [4][5][6][7][8][9][10][11]. Aside from the survival and growth breast height and the straightness of their stems showing no apparent defects. At the time of harvesting, the height (11.4 ± 1.5 m) and stem diameters (15.3 ± 2.2 cm) of trees were measured (Table 1), the north direction was also marked. The average height of trees was between 8.8 and 12.5 m, the average diameter at breast height was between 11.7 and 16.9 cm, height and DBH were the smallest in 1-3. A 3-cm-thick cross-sectional disc from each tree was sampled at breast height for further analyses of wood anatomy. Table 1. Means and standard deviations (SD) in height and DBH of six C. bungei clones. Age, years 13 13 13 13 13 13 Height, m 11.9 ± 1.6 a 8.8 ± 1.8 b 11.9 ± 0.1 a 11.6 ± 0.3 a 11.8 ± 0.6 a 12.5 ± 0.6 a Diameter at breast height, cm 16.9 ± 1.9 a 11.7 ± 2.4 b 15.7 ± 1.2 ab 14.8 ± 0.6 ab 16.4 ± 1.8 a 15.9 ± 1.1 ab Note: n is the number of trees within a clone. Statistical differences between clones are denoted using lowercase letters, with different letters corresponding to significant differences (p < 0.05). Diameter at breast height (bark included) was measured along the north-south direction. Different letters corresponding to significant differences, ANOVAs were followed by Tukey's tests at the 5% significance level to compare means using SPSS 19.0 software.
The samples for anatomical properties analyses were taken from the north side of each trunk so that the growth conditions of the studied wood were as similar as possible. Age was rechecked by counting the rings from the anatomical crosscuts at breast height; the length of strips of north side and annual ring width (mm) in the four directions (east, west, south, and north) was measured for each sample tree. Sapwood and heartwood were separated on the basis of color [40,41]. Growth rings were distinct; the boundary between earlywood (EW) and latewood (LW) was not distinct at the macroscopic scale. The unusual narrower rings occurred from years 10 to 13 in 1-3; it was difficult to distinguish the boundary of the last four rings with the naked eye. Observations of the wood features of C. bungei wood were made under a microscope.

Measurements of Anatomical Characteristics
For anatomical characterization, wood macerations were conducted using glacial acetic acid and 30% hydrogen peroxide (1:1 ratio) at 80 • C for 4 h. The macerated tissue was washed in running water until the tissue was free of all acid traces so that the lengths of the fibers and vessel elements could be determined. The fibers and vessel elements were observed using a depth of field microscope (Keyence VXH-600E, Keyence Corporation, Tokyo, Japan), and images were taken at 200× magnification for analysis. In total, 15 measurements were taken for vessel length and 50 measurements were taken for fiber length of both EW and LW. These measurements were carried out for each annual ring from the pith to the bark.
Wood cubes (2 × 2 × 2 cm 3 ) were prepared annually along radial positions from the pith to the bark. The wood blocks were softened by boiling in water with a microwave for approximately 15-20 min, transverse, radial and tangential microscopic sections (approximately 25-30 µm thick) were prepared with a rotary microtome Leica RM 2265 (Wetzlar, Germany). The sections were stained with safranin (5%) for approximately one minute, dehydrated with an ethanol series (50%, 70%, 80%, 90%, and 100%) for a few seconds, washed with 50% xylene and pure xylene, and mounted with a drop of Canada balsam. After drying, sections were observed using a light microscope (Nikon ECLIPSE Ni, Nikon Corporation, Tokyo, Japan) with a DS-Ri2 camera and Image NIS 5.10 software, and images were taken at 2× and 4× magnifications for analysis. ImageJ (National Institutes of Health, Bethesda, MD, USA) was used to analyze all images.
The Nikon microscope was used to measure the double thickness of the fiber wall. Fiber lumen diameter in the tangential direction was measured randomly by selecting 50 cells in each ring. Furthermore, the fiber diameter was calculated (double cell wall + lumen diameter). Fiber wall: lumen ratio was calculated using the double thickness of the cell wall and fiber lumen diameter. As vessels became rarer in LW, it became increasingly difficult to locate 50 vessel cells; in such cases, measurements of vessel diameter were reduced to 30. In addition, the number (No.) of vessels was counted at 4× magnification for each ring per mm 2 . The average vessel number was calculated from 20 vessels per slide.
The proportion of tissue types was calculated in the transverse section on 30 randomly selected areas using the NIS 5.10 image analysis system coupled to a Nikon microscope. A grid of 25-points was placed over each image and tissue types (i.e., fiber, axial parenchyma, vessel, and ray) were counted and converted into a percentage of the total area following Quilho et al. [14].
For scanning electron microscopy (SEM), 1-mm thick wood slices in the radial and tangential directions were prepared and oven-dried at 60 • C for 24 h. The slices were subsequently coated with gold at 15 mA in an anion sputter coater (KYKY SBC-12, KYKY, Ltd., Beijing, China) for 20 s and were observed via SEM (TESCAN VEGA 3 SBH, TESCAN, Ltd., Kohoutovice, Czech Republic) under a high vacuum (10 kV accelerate voltage, 10 mm or 15 mm working distance, SE detector).
The demarcations of the juvenile and mature wood regions were determined through a visual analysis of graphs obtained from the fiber length and MFA, as recommended by Zobel et al. [20], Mansfield et al. [42], Palermo et al. [37], Gorman et al. [43], and Dobrowolska et al. [44].
Wood slices with 1 mm (R) × 10 mm (T) × 30 mm (L) were prepared and equilibrated for 1 week in a constant temperature and humidity chamber where the temperature was 20 • C and the relative humidity was 65%. MFA was measured by X-ray diffraction (XD-3, Beijing Persee Instrument Co., Ltd., Beijing, China) with the following parameters: 36 kV voltage, 20 mA electricity, and 1.54 wavelength. The data were then calculated using the 0.6 T method [45] with Origin 2019b software.

Statistical Analyses
The means of anatomical properties were calculated from the individual samples and ANOVAs were performed using SPSS 19.0 software IBM (USA) to determine whether there were significant differences between clones. ANOVAs were followed by Tukey's tests at the 5% significance level to compare means. Variation in traits with cambial age was fitted with a multiple linear regression equation using Origin 2019b software (USA). Models of the regression curves were constructed using a 5% significance level. R square was also calculated to assess the robustness of the correlations.
Descriptive anatomical terminology followed the IAWA list of microscopic features for hardwood identification [46].

Xylem Anatomical Features
The sapwood and heartwood of C. bungei was clearly distinguished (brightness value was 64.2 in sapwood and 62.3 in heartwood), as sapwood was light brown while heartwood was dark brown ( Figure 1). Sapwood was narrow from 10 to 13 years and regular along the stem circumference. The length of strips (from pith to the bark) was different among clones; the length was longest in 8-1 and shortest in 1-3, indicating that the growth rate varied among clones.
Between the age of 1 and 2 years, the wood was diffuse-porous; between the age of 3 and 9 years, the wood was semi-ring-porous; and between the age of 10 and 13 years, the wood was ring-porous ( Figure 2). The latewood percentage became smaller from the pith to bark. The vessels were solitary, in short radial multiples of 2-3 cells, in clusters common (more frequent in LW) and occasionally filled with thin-walled tyloses (Figure 3d). They were almost circular to oval in outline with an average tangential diameter of 148 µm (184 in EW and 105 in LW) and an average vessel element length of 278 µm (200 in EW and 378 in LW). The transition from EW to LW was gradual. Vessels occupied 18% (28% in EW and 9% in LW) of the ring area, corresponding to an average of 14 (12 in EW and 19 in LW) vessels/mm 2 .
Fibers represented approximately 63% (53 in EW and 71 in LW) of the wood cross-sectional area and were 851 µm (816 in EW and 886 in LW) in length, 18.6 µm (18.9 in EW and 18.4 in LW) in width, and 4.6 µm (4.1 in EW and 5.0 in LW) in double wall thickness.
Rays occupied 10% of the wood cross-sectional area. The rays were 2-3 cells wide ( Figure 2f) and heterocellular with one row of upright and square marginal cells (Figure 3h).
Intervessel pits were alternates for C. bungei (Figure 3a  Between the age of 1 and 2 years, the wood was diffuse-porous; between the age of 3 and 9 years, the wood was semi-ring-porous; and between the age of 10 and 13 years, the wood was ring-porous ( Figure 2). The latewood percentage became smaller from the pith to bark. The vessels were solitary, in short radial multiples of 2-3 cells, in clusters common (more frequent in LW) and occasionally filled with thin-walled tyloses ( Figure 3d). They were almost circular to oval in outline with an average tangential diameter of 148 μm (184 in EW and 105 in LW) and an average vessel element length of 278 μm (200 in EW and 378 in LW). The transition from EW to LW was gradual. Vessels occupied 18% (28% in EW and 9% in LW) of the ring area, corresponding to an average of 14 (12 in EW and 19 in LW) vessels/mm 2 .  Between the age of 1 and 2 years, the wood was diffuse-porous; between the age of 3 and 9 years, the wood was semi-ring-porous; and between the age of 10 and 13 years, the wood was ring-porous ( Figure 2). The latewood percentage became smaller from the pith to bark. The vessels were solitary, in short radial multiples of 2-3 cells, in clusters common (more frequent in LW) and occasionally filled with thin-walled tyloses ( Figure 3d). They were almost circular to oval in outline with an average tangential diameter of 148 μm (184 in EW and 105 in LW) and an average vessel element length of 278 μm (200 in EW and 378 in LW). The transition from EW to LW was gradual. Vessels occupied 18% (28% in EW and 9% in LW) of the ring area, corresponding to an average of 14 (12 in EW and 19 in LW) vessels/mm 2 . Figure 2. Sections of xylem in C. bungei in 1-1: (a) between the age of 1 and 2 years, wood was diffuseporous; (b) between the age of 3 and 9 years, wood was semi-ring-porous, ring width was approximately 5 mm, and the latewood percentage was 80%; (c) between the age of 10 and 13 years, wood was ring-porous, ring width was approximately 2 mm, and the latewood percentage was 30%; axial parenchyma (P) with 5-15 cells per parenchyma strand; (d) vessels solitary (V1), radial multiples of 2-3 cells (V2), and clusters common (V3) in wood diffuse-porous; (e) radial section of the second year, vasicentric axial parenchyma (P); (f) tangential section of latewood areas from the 8th year, fusiform ray (R) width was 2-3 cells, axial parenchyma (P) in paratracheal-zonate. Sections of xylem in C. bungei in 1-1: (a) between the age of 1 and 2 years, wood was diffuse-porous; (b) between the age of 3 and 9 years, wood was semi-ring-porous, ring width was approximately 5 mm, and the latewood percentage was 80%; (c) between the age of 10 and 13 years, wood was ring-porous, ring width was approximately 2 mm, and the latewood percentage was 30%; axial parenchyma (P) with 5-15 cells per parenchyma strand; (d) vessels solitary (V1), radial multiples of 2-3 cells (V2), and clusters common (V3) in wood diffuse-porous; (e) radial section of the second year, vasicentric axial parenchyma (P); (f) tangential section of latewood areas from the 8th year, fusiform ray (R) width was 2-3 cells, axial parenchyma (P) in paratracheal-zonate.

Variation in Anatomical Characteristics Between Clones
Significant differences were found among the clones for all characteristics except annual ring width and MFA in LW (Table 2).

Variation in Anatomical Characteristics Between Clones
Significant differences were found among the clones for all characteristics except annual ring width and MFA in LW (Table 2).

Radial Variation in Anatomical Characteristics
Each of the six clones showed the same pattern (Figure 4) of ring width which significantly decreasing with cambial age (R 2 = 0.89). From the pith to the eighth year, ring width decreased rapidly. After the eighth year, changes in ring width stabilized.

Radial Variation in Anatomical Characteristics
Each of the six clones showed the same pattern (Figure 4) of ring width which significantly decreasing with cambial age (R 2 = 0.89). From the pith to the eighth year, ring width decreased rapidly. After the eighth year, changes in ring width stabilized. In the radial direction, EW fiber length showed a rapid and nearly linear increase (R 2 = 0.88) until the age of 7-8 years after the pith ( Figure 5). From years 7 to 8 to the periphery, the rate of increase in fiber length decreased and tended to stabilize to a constant value. In LW, fiber length showed slight increases (R 2 = 0.35) from the pith to the bark, and there was no distinct demarcation between juvenile  Note: EW, earlywood; LW, latewood; MFA, microfibril angle; n, number of samples; CV%, coefficient of variation. Statistical differences between clones are denoted using lowercase letters, with different letters corresponding to significant differences (p < 0.05). Different letters corresponding to significant differences, ANOVAs were followed by Tukey's tests at the 5% significance level to compare means using SPSS 19.0 software.
In the radial direction, EW fiber length showed a rapid and nearly linear increase (R 2 = 0.88) until the age of 7-8 years after the pith ( Figure 5). From years 7 to 8 to the periphery, the rate of increase in fiber length decreased and tended to stabilize to a constant value. In LW, fiber length showed slight increases (R 2 = 0.35) from the pith to the bark, and there was no distinct demarcation between juvenile and mature wood. MFA of EW and LW showed similar trends from the pith to the bark. In EW and LW, the angle increased from the pith until the 8th year, and then the angle decreased to the periphery. The effect of age on MFA was not significant in EW.
The demarcation between juvenile and mature wood based on fiber length and MFA was calculated by two-segment regression. The juvenile wood zone consisted of the first 7 to 8 growth rings. Juvenile wood width ranged between 4.2 and 5.8 cm, and the proportion of juvenile wood ranged between 82.7 and 88.3% (Table 3). The juvenile wood widths were significantly different among clones, but the effect of clone on the proportion of juvenile wood was not significant. Note: n is the number of trees within a clone. Statistical differences between clones were denoted using lowercase letters, with different letters corresponding to significant differences (p < 0.05). MFA of EW and LW showed similar trends from the pith to the bark. In EW and LW, the angle increased from the pith until the 8th year, and then the angle decreased to the periphery. The effect of age on MFA was not significant in EW.
The demarcation between juvenile and mature wood based on fiber length and MFA was calculated by two-segment regression. The juvenile wood zone consisted of the first 7 to 8 growth rings. Juvenile wood width ranged between 4.2 and 5.8 cm, and the proportion of juvenile wood ranged between 82.7 and 88.3% (Table 3). The juvenile wood widths were significantly different among clones, but the effect of clone on the proportion of juvenile wood was not significant. Table 3. Means and standard deviations (SD) in six C. bungei clones. Juvenile wood width, cm 5.8 ± 0.8 a 4.2 ± 0.9 b 5.4 ± 0.4 ab 5.3 ± 0.1 ab 5.7 ± 0.6 ab 5.5 ± 0.2 ab Proportion of juvenile wood, % 82.7 ± 3.7 a 88.3 ± 1.9 a 84.0 ± 2.9 a 85.8 ± 2.0 a 83.8 ± 2.4 a 82.7 ± 2.6 a Note: n is the number of trees within a clone. Statistical differences between clones were denoted using lowercase letters, with different letters corresponding to significant differences (p < 0.05). Different letters corresponding to significant differences, ANOVAs were followed by Tukey's tests at the 5% significance level to compare means using SPSS 19.0 software.
The fiber diameter of LW significantly decreased with cambial age until the 8th year and then increased, whereas diameter in EW was largely independent of cambial age (R 2 = 0.08) ( Figure 6). Forests 2020, 11, x FOR PEER REVIEW 3 of 18 Figure 6. The fiber characters in both earlywood (EW) and latewood (LW): fiber diameter, double thickness of fiber wall, and fiber wall: lumen ratio for the six C. bungei clones versus cambial age.
There were pronounced effects of age on the double thickness of cell wall (R 2 = 0.44 for EW and 0.48 for LW). Thickness significantly increased with cambial age in EW. In contrast, thickness increased until the eighth year and then remained stable in LW. The fiber wall: lumen ratio increased from the pith to the eighth year and then remained stable in EW and decreased from the eighth year to the bark in LW. Radial variation in LW (R 2 = 0.69) was more pronounced than in EW (R 2 = 0.31).
Vessel length in LW, vessel diameter in EW, and the number of vessels in EW varied significantly from the pith to the bark (Figure 7). Variation in vessel length in EW, vessel diameter in LW, and the number of vessels in LW was not radially significant. Variation in the radial pattern was characterized by increases in vessel diameter in EW with cambial age, vessel length increases and then decreases in LW, and a decreasing trend in the No. of vessels of EW. There were pronounced effects of age on the double thickness of cell wall (R 2 = 0.44 for EW and 0.48 for LW). Thickness significantly increased with cambial age in EW. In contrast, thickness increased until the eighth year and then remained stable in LW. The fiber wall: lumen ratio increased from the pith to the eighth year and then remained stable in EW and decreased from the eighth year to the bark in LW. Radial variation in LW (R 2 = 0.69) was more pronounced than in EW (R 2 = 0.31).
Vessel length in LW, vessel diameter in EW, and the number of vessels in EW varied significantly from the pith to the bark (Figure 7). Variation in vessel length in EW, vessel diameter in LW, and the number of vessels in LW was not radially significant. Variation in the radial pattern was characterized by increases in vessel diameter in EW with cambial age, vessel length increases and then decreases in LW, and a decreasing trend in the No. of vessels of EW. Proportion of fiber in both EW and LW decreased significantly until the eighth year ( Figure 8). Proportion of vessel increased with cambial age in EW until the eighth year and then stabilized. Proportion of axial parenchyma in LW also increased until the eighth year and then stabilized similar to the pattern observed for the proportion of vessel. Overall, the proportion of ray, proportion of vessel in LW, and proportion of axial parenchyma in EW did not change radially.

Xylem Anatomical Features
The wood anatomical features changed along radial direction, which was consistent with the proposed hypotheses. Woody angiosperms are widely divergent in the cellular makeup of their xylem [47]. Ring-porous wood is usually widespread in temperate climates [48]. According to our study, there were three different types of radial porous distributions: diffuse-porous, semi-ring-porous, and ring-porous. These results were not consistent with those of Li et al. [39], who studied mature C. bungei and concluded that C. bungei was largely a ring-porous tree species. This difference may stem from the fact that the most obvious characteristics of immature xylem consisted of the gradual transition between EW and LW [49].
Sapwood and heartwood can be clearly distinguished based on their color and the presence of tyloses [50]. Some authors [51,52] have reported that heartwood permeability is low, suggesting that tyloses or other deposits obstruct vessels. In our study, the observation of decreasing amounts of deposits from heartwood to sapwood in C. bungei was consistent with the observations of previous studies. Tyloses, warty layers, and cell walls with helical thickenings tend to be more effective structures for maintaining moisture and thereafter resist microorganisms and embolism [53].

Variation in Anatomical Characteristics between Clones
The effect of clones on most of the anatomical characteristics was significant (Table 2), which verified our hypothesis. The six clones in this study were selected by clonal breeding, meaning that the goal of this previous selection was to increase the commercial production of wood. With the development of the wood industry, increasingly competitive market pressures are encouraging forest industries to shift their focus from wood volume to wood quality [54]. Wood anatomical traits related to wood quality include the proportion of juvenile wood, wood density, fiber length, and MFA. These anatomical traits underly the physical and mechanical properties of wood, such as its shrinkage behavior, bending, stiffness, and strength [55]. For example, juvenile wood is typically characterized as being less dimensionally stable and having weaker mechanical properties than mature wood [43]. The quality of fibers has a pronounced effect on wood strength and shrinkage [56,57]. The fiber wall: lumen ratio has a major effect on wood density [58,59]. Extractives in axial parenchyma commonly increase the resistance of wood to decay [60]. MFA has been shown to directly impact the shrinkage behavior and flexural properties of wood, and these changes can alter several physical and mechanical properties of wood [55,[61][62][63][64].
In addition, these results were similar to those found by Ma et al. [38], who studied the wood properties of 11 three-year seedling trunks of C. bungei clones and showed that fiber length, fiber percentage, and vessel percentage were highly significantly different between clones with high repeatability and strong genetic control.
Some clones showed higher performance in terms of wood quality characters relative to others. For example, if fiber quality is considered, 8-1 was the best clone. Thus, selection of this clone may result in the greater resistance of wooden beams to buckling [57]. Other important data for the commercialization of wood products may be the fiber wall: lumen ratio, primarily because it can be used to predict density [58]. Further, 1-1, 8-1, and 9-1 showed clear advantages relative to the other clones in this trait. In addition, 2-7 had a small MFA, wide vessels, and the highest proportion of axial parenchyma and thus selection of this clone could produce wood suitable for use in manufacturing high-end, value-added products, such as furniture and decorative materials [16]. Knowledge of the intrinsic anatomical characteristics of C. bungei as a raw material for solid wood could facilitate the selection of desired traits in future breeding programs.

Radial Variation in Anatomical Characteristics
There was difference in anatomical characteristics among different cambial ages, indicating that we proved our hypothesis. The ring widths of six C. bungei clones were 2-4 rings per cm and these results were consistent with Chinese Catalpa from Yunnan [2]. The annual ring width of C. bungei decreased with cambial age [65] and all six clones showed highly similar radial changes, which might be attributed to environmental factors [17,66].
Juvenile wood can be defined as being close to the pith and it differs from mature wood in several ways [20,67]. Maeglin [68] reported shorter fibers with generally thinner walls at or near the pith. Plomion et al. [32] found that juvenile wood often has lower density and larger MFAs. Historical and contemporary studies have identified juvenile and mature wood by density, cell length, MFA, and longitudinal shrinkage [20,42,43,69]. In our study, fiber length and MFA were used to distinguish between juvenile and mature wood. The transition age between juvenile and mature wood was between 7 and 8 years; the width of juveniles ranged between 4.2 and 5.8 cm. These results were similar to those found by Palermo et al. [37], who reported that mature wood occurred between the age of 8 and 13 years. In addition, Ferreira et al. [36] found that the juvenile wood of Hevea brasiliensis occurred approximately between 40 and 55 mm from the pith.
According to our results, fiber length, double thickness of cell wall, and the fiber wall: lumen ratio increased significantly with cambial age [30,70,71]. Fiber diameter varied inconsistently radially [39,72]. MFA has generally been found to be below 20 • in hardwoods and does not show clear decreasing trends in juvenile and mature wood [73]. Vessel diameters in EW increased rapidly from the pith [74][75][76]; in LW, vessel diameter was not significantly related with cambial age. The number of vessels decreased with cambial age in EW [71], but in LW there was no radial change. Vessel diameter increased and the number of vessels decreased with cambial age in EW consistent with the porous distribution, which changed from diffuse-porous to semi-ring-porous and then to ring-porous. From the pith to the bark, the porous distribution was also accompanied by decreases in the proportion of fiber and increases in vessel and axial parenchyma proportion; fiber abundance was negatively correlated with both vessel and parenchyma abundance [77].

Conclusions
The impacts of the genetic origin on anatomical properties were significant on most of anatomical properties. Radial variations of anatomical properties were distinct. The transition between juvenile and mature wood could be observed between the age of 7 and 8. A similar transition age was observed in the variation curves for other properties such as ring width, double wall thickness in LW, and the fiber wall: lumen ratio. Additional work is needed to characterize the steps that take place during this transition age. Significant differences among clones and patterns of radial variation in anatomical features suggest that care is required in designing selection regimes to obtain wood with desired properties.