|Home | About | Journals | Submit | Contact Us | Français|
Although the lateral movement of water and gas in tree stems is an important issue for understanding tree physiology, as well as for the development of wood preservation technologies, little is known about the vascular pathways for radial flow. The aim of the current study was to understand the occurrence and the structure of anatomical features of sugi (Cryptomeria japonica) wood including the tracheid networks, and area fractions of intertracheary pits, tangential walls of ray cells and radial intercellular spaces that may be related to the radial permeability (conductivity) of the xylem.
Wood structure was investigated by light microscopy and scanning electron microscopy of traditional wood anatomical preparations and by a new method of exposed tangential faces of growth-ring boundaries.
Radial wall pitting and radial grain in earlywood and tangential wall pitting in latewood provide a direct connection between subsequent tangential layers of tracheids. Bordered pit pairs occur frequently between earlywood and latewood tracheids on both sides of a growth-ring boundary. In the tangential face of the xylem at the interface with the cambium, the area fraction of intertracheary pit membranes is similar to that of rays (2·8 % and 2·9 %, respectively). The intercellular spaces of rays are continuous across growth-ring boundaries. In the samples, the mean cross-sectional area of individual radial intercellular spaces was 1·2 µm2 and their total volume was 0·06 % of that of the xylem and 2·07 % of the volume of rays.
A tracheid network can provide lateral apoplastic transport of substances in the secondary xylem of sugi. The intertracheid pits in growth-ring boundaries can be considered an important pathway, distinct from that of the rays, for transport of water across growth rings and from xylem to cambium.
Water in tree stems is transported through the xylem predominantly in the axial direction, but it also must move radially across short distances, in particular to supply the cambium and differentiating xylem cells during wood formation. The radial movement of water in xylem is also demonstrated by the fact that axial flow can occur throughout many growth layers within tree stems while the sap is directly pulled by transpiration of leaves predominantly through the outermost growth rings of sapwood (Gartner and Meinzer, 2005). Studies of the mechanism of uptake of water from xylem to phloem in leaves and roots of angiosperm plants have shown that there are both apoplasmic and symplasmic pathways (for reviews, see van Bel et al., 1994; van Bel and Hafke, 2005; Botha, 2005; Thorpe et al., 2005). There is limited evidence showing that vessel networks extend across growth-ring boundaries in hardwood xylem (Fujii et al., 2001; Kitin et al., 2004) but more research is needed to clarify the anatomical pathways for water or gas movement in the secondary xylem from one growth ring to another and from the xylem to the cambium.
The wood of gymnosperms is less variable than that of angiosperms with respect to the cell type and growth-increment structure, and gymnosperm species may be easier models for studying the relationship between the structure and radial permeability of wood. For example, the xylem of conifers has features, such as bordered pits in tangential walls of tracheids and ray tracheids in some species that are apparently designed to facilitate the radial movement of substances (Liese and Bauch, 1967; Laming and ter Welle, 1971; Flynn, 1995). Also, leaves in some conifer species can live for many years, and leaf traces have been assumed to function as pathways for the radial movement of sap (MacDougal et al., 1929; Maton and Gartner, 2005). However, why and how the leaves in some evergreen conifers pull water from several xylem growth layers remains unclear (Maton and Gartner, 2005).
Axial permeability (conductivity) of softwood xylem to water and gas varies at different radial positions within the stem (growth ring number) and depends on the inter-tracheid pit structure and frequency (Mark and Crews, 1973; Milota et al., 1995; Fujii et al., 1997; Domec and Gartner, 2003; Gartner and Meinzer, 2005; Choat et al., 2008), but the factors that determine the radial permeability of wood are little studied. It is generally known that there is a radial pathway for water and nutrients through the living rays (van Bel, 1990). Also, in dried wood, ray tracheids or ray cells (Liese and Bauch, 1967; Matsumura et al., 1998; Usta and Hale, 2003) have been shown to conduct liquids. Despite these findings, the relative physiological role of other xylem features in the lateral transport of solutes remains unknown. In particular, in conifers, the radial continuity of tracheid networks, as well as the frequency of bordered pits in tangential walls of tracheary elements and the area fractions of ray cells and intercellular spaces in tangential sections, may be closely related to the radial permeability of xylem to water and gas.
In the current studies, the xylem in sugi (Cryptomeria japonica, Taxodiaceae) was qualitatively and quantitatively analysed to determine (a) the tracheid-to-tracheid pit connections between earlywood and latewood within a growth ring and across growth rings; (b) the area fractions of intertracheid pit membranes and ray cells in the tangential face of a growth-ring boundary; (c) whether the intercellular spaces are continuous across growth-ring boundaries; and (d) the relative importance of intercellular spaces as radial passages for water or gas in comparison to the rays and the bordered pit membranes. As far as is known, the occurrence of bordered pits in tangential walls of conifer tracheids has not yet been quantitatively analysed. These studies should improve our understanding of tree hydraulics in conifers. Modelling the radial transport in xylem should not only be helpful to tree physiologists, but it may also have implications in the woodworking industry because the radial permeability of xylem to fluids is an important factor for the preservation of wood materials, in particular long and narrow boles (Siau, 1984). Sugi was selected for these investigations because it is a major timber species in Japan and because a great deal of research has focused on the improvement of the drying and preservation of sugi wood (Fujii et al., 1997; Kobayashi et al., 1998a, b; Kawai et al., 2003).
Wood samples from sugi (Cryptomeria japonica D. Don) were collected during the winter dormant season (February 2005) from a 30-year-old tree that was growing near Futatsui city, Akita Prefecture, Japan. The samples were obtained from the stem at the height of 1·30 m and included cambium and adjacent bark plus several growth rings of sapwood. The tissues were fixed and stored in a 18:1:1 (v/v/v) 50 % ethanol/acetic acid/formaldehyde.
Tangential, radial and transverse sections were hand-cut with a razor blade or with a sliding microtome equipped with a freezing stage (HM400R; Microm Laborgeräte GmbH, Walldorf, Germany). The sections were mounted in glycerol without staining and observed with an epifluorescence microscope (Olympus BX60; Olympus, Tokyo, Japan) with excitation by incident blue light from a mercury lamp. Images were captured with a digital CCD camera (DP70; Olympus).
Some of the sections were observed with a CLSM system (MRC 1024 T; Bio-Rad Microscience Ltd, Hemel Hempstead, UK) mounted on the Olympus BX60 microscope following safranin staining or by autofluorescence induced by incident-light excitation from an argon ion laser (488 nm) with a long-pass filter (585 nm). The safranin-stained sections were thoroughly washed and mounted in water or in immersion oil (Donaldson and Bond, 2005).
Wood samples were trimmed into small blocks with dimensions 5 × 5 × 5 mm and clean surfaces were cut with a razor blade or split in the longitudinal radial or tangential direction. The blocks were passed through a tertiary-butyl alcohol series and freeze-dried with a JFD-300 freeze-drier (Jeol Co. Ltd, Tokyo, Japan). The blocks were then mounted on specimen stubs and coated with gold using an ion-sputter coater (JFC-1200; Jeol Co.) at 1·5 keV for 1·5 min. The samples were observed with a JSM-5310LV scanning electron microscope (Jeol Co.) at an accelerating voltage of 10–15 keV.
Small blocks containing cambium and the adjacent phloem and xylem were cut into slabs with dimensions approx. 4 × 10 × 5 mm (radial × axial × tangential). The slabs were treated with pectinase to macerate the cambium as described by Kitin et al. (1999). Next, the phloem and cambium were removed from the samples, and the exposed tangential surfaces of the xylem boundary at the interface between cambium and xylem were observed by SEM as described above. In addition, some wood blocks were used to study the distribution of pits in subsequent rows of earlywood and latewood tracheids on both sides of a growth-ring boundary. For this purpose, cross-sections and obliquely cut tangential faces of xylem that included adjacent regions of earlywood and latewood at the previous-year growth-ring boundary were prepared and observed by SEM.
Dimensions and area fractions of pit membranes and ray cells were measured with Image J (Wayne Rasband, National Institute of Health, Bethesda, MD, USA) on exposed faces of wood at the xylem–cambium interface. Images of random areas with dimensions 630 × 480 µm were captured by SEM as described above, and 20 such images (total area of approx. 8·6 mm2) were used for analysis of each wood feature. Because of the comparatively low contrast of SEM images, a semi-automatic approach was used: the areas of pit membranes and rays were first extracted from the images using the manual selection tool in Adobe Photoshop, after which their dimensions and corresponding area fractions were measured with Image J. The intercellular spaces in rays were more clearly visible in tangential-cut faces of xylem; therefore, tangential-cut xylem (total area of approx. 8·6 mm2) were used to determine the area fraction of intercellular spaces. First, the average cross-sectional area of intercellular spaces was determined at higher magnification images. Then, the average area was multiplied by the number of spaces in each randomly captured image.
The occurrence of pits in tangential walls of tracheids was analysed in cross-sections and tangential faces of wood (Figs 11–5). Obliquely cut longitudinal tangential sections were also prepared to allow simultaneous visualization of tangential walls of several adjacent tracheids in a radial row (Fig. 6). Bordered pit pairs between latewood and earlywood tracheids on both sides of a growth-ring boundary were infrequently observed in single cross-sections (Figs 1 and and4E,4E, D). Therefore, cross-sections may create the impression that such pits are rare; however, exposed tangential faces of growth-ring boundaries revealed that pits in tangential walls that face the growth-ring boundary were common in both the latewood (Fig. 2) and earlywood sides (Fig. 4A, B). Pits connecting earlywood and latewood tracheids on both sides of a growth-ring boundary were mostly located in tangential walls (Figs 1, ,22 and and4)4) but also occurred in radial walls or in cell corners as demonstrated in Fig. 4D. In a single latewood or earlywood tracheid, bordered pits were present in both radial and tangential walls as seen in Z-series images by CLSM or SEM cross-sections (Figs 1 and and4A).4A). The location of pits in both radial and tangential walls of a single tracheid was clearly visualized in longitudinal-oblique sections or split faces of wood (Figs 3A and and66A).
Apart from the growth-ring boundary at the earlywood side, the presence of bordered pits in tangential walls of earlywood tracheids was rarely observed and appeared limited to small areas of several adjacent tracheids (Fig. 6). In contrast to earlywood, bordered pits in tangential walls of latewood tracheids were common not only at the growth-ring boundary but also in the tracheid layers apart from it (Fig. 3), as well as in the tracheid layers of the transition zone between earlywood and latewood. The growth rings of sugi are characterized by a gradual transition from earlywood to latewood. Xylem cells at the transition zones, which were ‘earlywood tracheids’ according to the Mork's definition (radial diameter of the cell lumen larger than radial thickness of the double cell wall; see Denne, 1988), also had pits in both tangential and radial walls. In addition, such earlywood tracheids at the transition zone appeared to have thicker walls than the typical earlywood tracheids from the initial portions of growth rings.
Close-up views of pits in radial and tangential walls of tracheids are shown in Figs 33–5. The inner walls of pit chambers of intertracheid and tracheid-to-ray pits were covered with warts (Figs 3B, B,4C4C and and5B),5B), similar to those of the S3 layer of secondary walls of the same tracheids (Fig. 5D, H, I). Aspirated and unaspirated pits were present in the radial walls of adjacent earlywood tracheids (Fig. 5J) in agreement with the observations of sugi sapwood by Fujii et al. (1997). The pits in radial walls had torus and margo with loosely arranged fibrils. In contrast, the pit membranes of the pits in tangential walls of both earlywood and latewood tracheids had no pores and no torus–margo structure was visible (Figs 3C and and55F).
Although the sizes of pits in tangential walls of both earlywood and latewood tracheids were variable and tended to be smaller near cell tips (Fig. 5E), typically, the diameter of pit membranes was approx. 7 µm, and they had elliptical or slit-like canals with a small diameter approx. 2 µm (Fig 3A, B and 5D). Occasionally, there were pits positioned near cell corners that were smaller in diameter than the typical pits in radial and tangential walls (Fig. 5H).
The bordered pits in radial walls of earlywood tracheids were much larger (approx. 2-fold wider pit membranes) than those in the tangential walls (Fig. 5C, F, I, J). In the latewood tracheids, however, pits in radial and tangential walls were similar in size (Fig. 6A). Furthermore, in split tangential faces of xylem, the structure of pit chambers in tangential walls appeared similar in both earlywood and latewood tracheids (see Figs 3B and and44C).
The present observations of radial longitudinal sections showed that the tips of earlywood tracheids, which are normally rich in bordered pits, deviate from the axial to the radial direction and form a radial grain (Fig. 7). A series of optical sections by confocal microscopy demonstrated that each individual earlywood tracheid from a certain radial row directly contacts via pits with several subsequent tracheids from the adjacent radial rows. Such connections are established by the radial inclination of tracheid tips and the different angle of the radial grain in adjacent radial rows of cells (Fig. 8). Hence, tracheids of sequential tangential layers of the earlywod appear to be directly connected via radial wall pit contacts.
Radial intercellular spaces consistently occurred in the xylem rays (Figs 9 and and10).10). The average size of individual intercellular spaces and the fraction of the area occupied by spaces in tangential section are shown in Table 1. Longitudinal split faces of wood revealed that the intercellular spaces form continuous canals across growth-ring boundaries and from earlywood to latewood and that they do not visually change in size (Fig. 10). Therefore, determining the area fraction of radial intercellular spaces in a tangential section allows for calculation of their volume in the particular wood sample. For example, it was estimated that the radial intercellular spaces accounted for 2·07 % of the volume of rays (including cell walls) and 0·06 % of the entire volume of the wood sample. Although how the xylem ray proportion can vary in different parts of the plant is unresolved, the pattern and spacing of rays is species-specific (Larson, 1994); therefore, the present results may be fairly representative for the species. For instance, 1 m3 of wood with the characteristics of the current samples (Table 1) would have 600 cm3 of radial intercellular spaces and 29 000 cm3 of ray cells.
As described earlier, exposed tangential faces of xylem boundary were examined at the interface between cambium and xylem by SEM to analyse the distribution of pits facing the growth-ring boundary. The method of pectinase treatment and removal of cambium exposed the outer wall of the last-formed layer of tracheids. This allowed the area fraction of tangential wall pit membranes that face the dormant cambium, which is also representative of the structure of the latewood side of growth-ring boundaries, to be measured. A close-up view of an exposed xylem boundary is shown in Fig. 2B and C. The areas of intertracheid pit membranes and ray cells were extracted from the images using the manual selection tool of Adobe Photoshop (Fig. 11). The dimensions and area fractions of pit membranes, ray cells, and intercellular spaces in tangential sections are displayed in Table 1.
Intertracheid pits were observed in tangential walls of tacheids in all layers of latewood as well as in the transition zone from earlywood to latewood. Tangential pitting in earlywood tracheids has been noted only in some species and was considered rare or anomalous (Laming and ter Welle, 1971). However, in the earlywood tracheids of sugi located at growth-ring boundaries, pits were common in both tangential and radial walls (Figs 2 and and44).
In addition to pit frequency, the permeability of softwood depends on pit structure, particularly the diameter and structure of pit cavities as well as the pore size of pit membranes (Mark and Crews, 1973; Bolton and Petty, 1975; Ohgoshi et al., 1982b; Siau, 1984; Sano and Nakada, 1998; Sperry and Hacke, 2004; Pittermann et al., 2005; Domec et al., 2006; Choat et al., 2008). In sugi, it is known that extraction of substances that cover pit membranes or removal of pit membranes significantly improves the permeability of logs to fluids and gas (Kobayashi et al., 1998a, b). Wood permeability also depends on the ratio between aspirated and non-aspirated pits in earlywood (Comstock and Côté, 1968; Milota et al., 1995; Fujii et al., 1997).
Ohgoshi et al. (1982a) reported quantitative data on the shapes and sizes of pores in radial-wall pit membranes in sugi and hinoki (Chamaecyparis obtusa). Using TEM, they determined the average Feret's diameter and the diameter equivalent to a circle in pit-membrane pores to be 0·107 µm and 0·073 µm, respectively. The pores in earlywood pits were approx. 1·2–1·5 times larger than those in latewood pits, and their mean sizes in hinoki were only slightly larger than those in sugi. These remarkable TEM studies provide valuable information but they did not address the structural variation according to radial position and developmental changes (sapwood–heartwood transition) of wood. Furthermore, the results may be dependent on the chemical treatments that accompany sample preparations for electron microscopy (for discussion, see Choat et al., 2008; Jansen et al., 2008).
Using field-emission scanning electron microscopy, Sano et al. (1999) found that most, but not all, of the radial wall pit membranes in latewood tracheids of Abies sachalinensis had small pores. The degree of development of pores in the radial wall pit membranes of Abies sachalinensis appeared to decrease from the inner to the outer side of latewood. Also in Douglas fir, Domec et al. (2006) reported a variation in the pore-size of pits in radial walls of tracheids in different plant parts.
Previous reports show that the bordered pits in tangential walls of conifer tracheids have a structure distinct from those in the radial walls (Fengel, 1968; Bauch et al., 1972; Fujikawa and Ishida, 1974; Koran 1977). Higher resolution studies using carbon replicas and TEM found that the tangential wall pit membranes did not have pores, whereas pit membranes in radial walls of latewood tracheids had pores, although they were smaller than those in the earlywood tracheids (Fujikawa and Ishida, 1973, 1974). Similarly, no pores were visible in the membranes of tangential wall pits in the present study, even though some of the samples were cleaned with ethanol to remove extraneous materials (Fig 3C and and5F).5F). In addition, the pit membranes in tangential walls appeared homogeneous with no apparent torus–margo structure. In contrast, torus and margo with fibrillar structures and wide pores were visible in non-aspirated pit membranes in radial walls of earlywood tracheids (Fig. 5J), which agrees with earlier observations of sugi wood (Bauch et al., 1972; Fujikawa and Ishida, 1973, 1974; Ohgoshi et al., 1982a; Sano and Nakada, 1998).
It has already been shown in some species that the ultrastructure of pit membranes in xylem conduits may depend on the ontogenic and physiological status of the xylem tissue (Mark and Crews, 1973; Sano and Nakada, 1998) as well as their permeability may depend on the chemical composition of the sap (Zwieniecki et al., 2001). None of the studies on tangential wall pits to date has considered physiological, positional or age factors that may affect their structure. Sano and Nakada (1998) found that incrusting materials in radial wall pit membranes of tracheids in sugi started to appear in the middle layers of sapwood and accumulated in stages during several subsequent years. Therefore, it can be expected that the variation in structure of tracheid pits which was observed in sugi, and the corresponding functionality for water transport will be further affected by the gradual deposition of incrusting materials in the pit membranes in subsequent years.
In the latewood, inter-tracheid pits frequently occurred in tangential walls (Figs 2B and and3A),3A), which suggests that solutes can potentially spread radially through adjacent tangential rows of cells via a tracheid network pathway. In the earlywood, however, pits in tangential walls were restricted to the initial tracheids of earlywood and those in the transition zone, and, therefore, could play a role in connecting earlywood and latewood but not in the radial movement of substances within earlywood. If water and solutes can spread in the radial direction across growth rings, a radial pathway within earlywood must also exist. As discussed earlier, rays can serve as such a radial pathway (van Bel, 1990; Utsumi et al., 2003). In addition, we suggest that radial transport in earlywood can take place via a tracheid network because of the occurrence of a radial grain (Figs 7 and and88).
The radial grain would provide axial and radial transport of water within earlywood so that tangential-wall pitting can allow transfer of water from earlywood to latewood and across growth-ring borders. Such transport surely occurs through the last-formed growth ring as demonstrated by dye flow experiments in many conifer species (Tyree and Zimmermann, 2002; Maton and Gartner, 2005) and by cryo-SEM (Utsumi et al., 2003). In older growth rings, however, cavitations may occur in certain layers of tracheids (Sperry and Tyree, 1990), which may interrupt the radial flow via tracheid networks. Utsumi et al. (2003) showed in three conifer species in a cold temperate climate (Larix kaempferi, Abies sachalinensis and Picea jezoensis) that tangential layers of tracheids were cavitated in the transition zone from earlywood to latewood of each growth ring of the sapwood. In that case, functional connection between water-filled portions of xylem can be provided by the rays (Fig. 5A). In the last-formed growth ring, the cavitation of tracheids occurred near the end of the growing season presumably caused by a drought-induced air-seeding. The cavitations further progressed during the winter months, but many tracheids of the earlywood, as well as the latewood tracheids near growth-ring boundaries, remained constantly filled with water in all growth rings of the sapwood. Utsumi et al. (2003) explained the tangential spread of cavitation in the transition zone from earlywood to latewood by the peculiar structure of transition zone tracheids, namely, the occurrence of bordered pit pairs with rigid pit membranes in radial walls, which cannot become aspirated (see also, Domec and Gartner, 2002; Domec et al., 2006). While latewood compared with earlywood is less resistant to drought-induced embolism (Domec and Gartner, 2002), some latewood tracheids at growth-ring boundaries always remain filled with water at the physiological range of water potentials. This can be explained with decreased porosity of the margo in radial-wall pit membranes of the outermost layers of latewood tracheids compared with the inner layers of latewood tracheids (Sano et al., 1999; also for discussion, see Utsumi et al., 2003; Domec et al., 2006). In addition, the outermost layers of latewood tracheids have very narrow lumens which make them more resistant to freezing-induced embolism (Tyree and Zimmermann, 2002; Pittermann and Sperry, 2006). Utsumi et al. (2003) suggested that the narrow lumina of the latewood tracheids at growth-ring boundaries create strong capillary pressure which helps retain water in those tracheids. Furthermore, wide earlywood and narrow latewood tracheids are connected via pit pairs across the growth-ring boundaries (Fig. 1) and water could be drawn from earlywood to latewood due to capillary pressure difference. The present observations of frequent pit-contacts between earlywood and latewood tracheids in Cryptomeria japonica support such an idea and it would be interesting to study it further. Also, further research should address the intertracheid pit membrane ultrastructure in tangential walls and the occurrence of aspirated pits and cavitated tracheids as factors that can affect the functionality of the tracheid network.
It has been suggested that intercellular spaces in wood can participate in gas exchange or water conduction (Hook and Brown, 1972; Hook et al., 1972; Bolton et al., 1975; Kučera, 1985; Sun et al., 2004; also for a review, see Larson, 1994) and can be important for capacitance (Tyree and Yang, 1990; Tyree and Zimmermann, 2002). Furthermore, intercellular spaces can serve as an extracellular diffusion pathway for heartwood substances in wood (Zhang et al., 2004).
Previous reports indicate that the xylem intercellular spaces of various species are connected via blind pits with ray cells (review in Bolton et al., 1975; Zhang et al., 2004). The intercellular spaces in xylem rays were particularly large in pneumatophores and they were connected via blind pits not only with ray cells but also with the adjacent vessels (Sun et al., 2004). Axial intercellular spaces were not apparent in the present samples of sugi but they do exist in the normal wood of various species and they appear to be connected with the radial intercellular spaces of rays (Bolton et al., 1975; Sun et al., 2004; Zhang et al., 2004).
The available information on the structure and dimensions of intercellular spaces in xylem of hardwoods as well as softwoods is still limited. The present study showed that in the wood of sugi, the intercellular spaces of rays form continuous and uniform radial canals across growth-ring boundaries. This allows calculation of the volume of the radial intercellular spaces on the basis of their area in the tangential face of wood. The volume of radial intercellular spaces in the samples was 0·06 % of the total volume of the xylem. Future studies on the xylem physiology could clarify whether the size of intercellular spaces is large enough to be a factor in the hydraulics and metabolism of the xylem. To understand better the mechanism of their function, further research should address more aspects of the anatomy of intercellular spaces, such as the pit connections between xylem cells and intercellular spaces, and the continuity of these spaces between xylem and phloem.
The present observations of tangential sections of sugi wood revealed bordered pit pairs in the tangential walls not only in latewood but also in earlywood tracheids that are positioned near growth-ring boundaries and in the transition zone between earlywood and latewood. In addition, this study showed that pits in tracheid walls that face a growth-ring boundary occur frequently in both latewood and earlywood sides. The area fraction of intertracheary pit membranes in growth-ring boundaries was statistically the same as that of the tangential walls of ray cells (2·8 % and 2·9 %, respectively). In reality, the total area of intertracheid membranes at growth ring boundaries must be even greater than that shown in Table 1, because, it was not possible by the present method to measure the area of pit membranes in radial walls and cell corners (see Fig. 3D). The tangential walls of ray cells and the intertracheid pit membranes must have different mechanisms of conductivity in the living plant (i.e. predominantly symplastic vs. apoplastic flow); therefore, no straightforward comparison can be made between them on the basis of area fractions. Nevertheless, the fact that the intertracheary pit area is statistically the same as that of the tangential walls of ray cells suggests that not only rays but also the tracheid network is involved in the transfer of substances across growth rings and from xylem to cambium.
As discussed earlier in this report, the difference in structure between intertracheary pits in radial and tangential walls, in particular the absence of pores in the pit membranes in tangential walls, needs careful consideration. It is well established that the earlywood tracheids are responsible for most of the axial water flow within a single ring and the permeability of earlywood to water in the axial direction is 11 times greater than that of latewood (Domec and Gartner, 2002). Although the frequent occurrence of intertracheary pits in tangential walls seems to clearly indicate a radial path for water, more research on the pit ultrastructure and the actual permeability of membranes is needed to clarify this finding. The pores of the pits in radial walls of earlywood tracheids provide rapid axial transport which is required for the transpiration stream. By contrast, we suggest that the absence of apparent pores in the pit membranes in tangential walls of tracheids might be related to a slow movement of water through tangential walls over short radial distances in the stem (rapid apoplastic transport across growth rings may not be necessary for the plant). The anatomical structure of xylem provides slow rates of radial movement of water in comparison to the axial transport; nevertheless, radial transport is important for vital processes such as the growth of cambium and differentiating wood, as well as possibly for capacitance.
It can be concluded on the basis of this study that radial transport in xylem can occur via tracheid network in a distinct pathway from that of the rays. This tracheid network has different patterns in earlywood (radial grain) and latewood (tangential-wall pitting). Further research should clarify the functionality of the radial connections between tracheids, in particular, the permeability of pit membranes in tangential walls. This study also showed that the intercellular spaces of rays form continuous radial canals which can serve for transport of substances across growth rings in wood.
Funding for this work is gratefully acknowledged from the Japan Society for the Promotion of Science, and a Marie Curie Outgoing International Fellowship of the European Commission (MOIF 040400 to P.K.).