Summary Wood density, a gross measure of wood mass rel- ative to wood volume, is important in our understanding of stem volume growth, carbon sequestration and leaf water sup- ply. Disproportionate changes in the ratio of wood mass to vol- ume may occur at the level of the whole stem or the individual cell. In general, there is a positive relationship between temper- ature and wood density of eucalypts, although this relationship has broken down in recent years with wood density decreasing as global temperatures have risen. To determine the anatomical causes of the effects of temperature on wood density, Eucalyp- tus grandis W. Hill ex Maiden seedlings were grown in con- trolled-environment cabinets at constant temperatures from 10 to 35 °C. The 20% increase in wood density of E. grandis seed- lings grown at the higher temperatures was variously related to a 40% reduction in lumen area of xylem vessels, a 10% reduc- tion in the lumen area of fiber cells and a 10% increase in fiber cell wall thickness. The changes in cell wall characteristics could be considered analogous to changes in carbon supply.
Lumen area of fiber cells declined because of reduced fiber cell expansion and increased fiber cell wall thickening. Fiber cell wall thickness was positively related to canopy CO2assimila- tion rate (Ac), which increased 26-fold because of a 24-fold in- crease in leaf area and a doubling in leaf CO2assimilation rate from minima at 10 and 35 °C to maxima at 25 and 30 °C. In- creased Acincreased seedling volume, biomass and wood den- sity; but increased wood density was also related to a shift in partitioning of seedling biomass from roots to stems as temper- ature increased.
Keywords: biomass partitioning, CO2assimilation, fiber cell, photosynthesis, xylem vessel.
Introduction
Density is an important attribute of wood that contributes to the quality and therefore economic value of timber (MacDon-
ald and Hubert 2002). Wood density, a gross measure of wood biomass relative to wood volume, is an integration of the pro- cesses involved in stem volume growth (i.e., cell division and cell enlargement) and the processes involved in biomass accu- mulation (i.e., cell wall thickening of the existing and newly formed cells). The arrangement of biomass in stems and the re- maining voids has important implications for both solute movement (Hacke et al. 2001, Roderick and Berry 2001) and inferring historical late-summer temperatures (Briffa et al.
1998).
Increased growth temperatures can be related to increased wood density of mature Eucalyptus dunnii Maiden trees (Muneri et al. 2004), but the anatomical causes responsible for the changes in wood density of eucalypts are unknown. Fur- thermore, seasonal variation in wood density has been used as a surrogate for temperature in interpreting historical climates.
However, in recent years, the general positive relationship be- tween temperature and wood density has broken down with wood density decreasing as global temperatures have risen (Briffa et al. 1998). This divergence is unusual because parti- tioning of biomass to shoots often increases as temperature rises (Wardlaw 1979, Bruhn et al. 2000, Weih and Karlsson 2001), a factor expected to increase wood density.
Plant biomass accumulation typically increases with in- creasing temperature before declining at supra-optimal tem- peratures. An increase in stem biomass would be expected to increase the biomass in fiber cell walls and, if this process oc- curs faster than increases in stem volume, then wood density will increase—a situation that can occur when carbon supply is increased (Richardson and Dinwoodie 1960, Larson 1964, Richardson 1964, Creber and Chaloner 1984, Conroy et al.
1990, Lindström 1996, Deleuze and Houllier 1998, Barber et al. 2000). However, when stem volume increases as fast as, or faster than, stem biomass then wood density may remain sta- ble or decline (Telewski et al. 1999, Ceulemans et al. 2002, Thomas et al. 2006). Therefore the independent effects of tem-
© 2007 Heron Publishing—Victoria, Canada
Temperature effects on wood anatomy, wood density, photosynthesis and biomass partitioning of Eucalyptus grandis seedlings
D. S. THOMAS,
1,2,3K. D. MONTAGU
4,5and J. P. CONROY
11Centre for Horticulture and Plant Sciences, University of Western Sydney, Locked Bag 1797 South Penrith, NSW 1797, Australia
2Present address: Forests NSW, Plantation Improvement, Land management and Technical Services, PO Box J19 Coffs Harbour Jetty, NSW 2450, Australia
3Corresponding author (danet@sf.nsw.gov.au)
4State Forests of NSW, Research and Development Division, PO Box 100 Beecroft, NSW 2119, Australia
5Present address: Cooperative Research Centre for Irrigation Futures, c/o School of Environment and Agriculture, University of Western Sydney, Locked bag 1797 South Penrith, NSW 1797, Australia
Received January 20, 2006; accepted March 11, 2006; published online November 1, 2006
perature on the processes of biomass accumulation of the stem and stem volume growth will determine the impact of temperature on wood density.
Stem volume is a function of the volumes of the various cell types (Kramer and Kozlowski 1979, Malan and Hoon 1992, Zhang and Zhong 1992, Zobel and Jett 1995, Denne and Hale 1999). If we assume that the stem comprises only fiber cells and xylem vessels, then the relative volumes of cell wall mate- rial and cell lumen will affect wood density (Roderick and Berry 2001). Temperature affects the relationship between fi- ber cell wall thickness and fiber cell lumen, and the number and size distribution of xylem vessels in seedlings and mature trees (Larson 1964, Richardson 1964, Roderick and Berry 2001, Thomas et al. 2004). In a simple model of wood produc- tion in gymnosperms, Deleuze and Houllier (1998) assumed temperature was the most limiting factor for initial fiber cell division. In their model, higher temperatures result in in- creased cell production, but the rate of cell production declines at supra-optimal temperatures (Deleuze and Houllier 1998).
These modeling results are supported by the finding that cambial activity, measured as rate of cell division in larch (Larix siberica Ldb.) and Scots pine (Pinus sylvestris L.), in- creased with increasing temperature, and then declined when mean temperatures were greater than about 20 °C (Antonova and Stasova 1993, 1997); the effects of temperature on stem biomass accumulation and wood density were not determined in these studies.
To test the hypothesis that Eucalyptus grandis W. Hill ex Maiden seedlings have higher wood density when grown at higher temperature, we grew seedlings in controlled-environ- ment cabinets at constant temperatures from 10 to 35 °C. Spe- cifically, we determined if the positive effect of temperature on wood density is correlated with an overall increase in seedling carbon supply and biomass allocated to stems. Additionally, we assessed if the positive correlation of temperature with stem wood density is associated with enhanced production of denser thick-walled cells, and decreased lumen areas of the fiber cells and xylem vessels.
Materials and methods
Plant culture
Soil was collected from the A horizon at Belanglo State Forest near Moss Vale, NSW (152°13′ E, 34°30′ S). The soil is podsolic derived from Triassic shales and sandstones. Total N is less than 0.09% and exchangeable Ca less than 0.4me%
(Turner 1982). Batches of soil were mixed with CaCO3(7 g kg– 1) and MgCO3(1.8 g kg– 1) to adjust the pH to 6.7 (1:5 w/v in 0.01 M CaCl2). Phosphorus as CaHPO4was added to the soil, which had an available P concentration of less than 1 mg k g– 1dry soil (Bray method 1, Bray and Kurtz 1954), at the rate of 1000 mg P kg– 1dry soil. Basal nutrients were added to the soil (mg kg– 1dry soil): K (90 + 360); B (5); Cu (5); Zn (10); Mo (0.1); Mn (50); and Fe (50) as K2SO4 + K2CO3; H3BO3; CuSO4; ZnSO4; Na2MoO4; MnSO4; and FeSO4. Nitro- gen was added weekly at 67 mg kg– 1 as KNO3, Ca(NO3)2,
Mg(NO3)2, (NH4)2SO4and NH4NO3in the ratio (by weight of salts) of 1:2:1:1:1.
Eucalyptus grandis seeds were germinated in a 1:1 (v/v) mix of perlite and amended soil at 25 °C in a Eurotherm Chessell 392 plant growth chamber (Thermoline, Australia) located at the University of Western Sydney, Hawkesbury campus, Richmond, NSW (150°45′ E, 33° 36′ S). The vapor pressure deficit in the chamber was maintained at 1 kPa. Pho- tosynthetic photon flux (PPF) was maintained at 1000 µmol m– 2s– 1at mid-canopy height with 1000-W metal halide lamps during the 12-h photoperiod.
When seedlings were 6 weeks old they were 2 ± 0.02 (SE) cm high, had a stem diameter at soil height of 0.4 ± 0.005 mm, a leaf area determined by a Delta T leaf area meter (Burwell, Cambridge, U.K.) of 3.5 ± 0.36 cm2and biomass after drying at 70 °C for 48 h of 0.04 ± 0.004 g. At this time, 75 seedlings were each transplanted to a 6.9-l pot (PVC pipe: diameter 15 cm, height 39 cm) containing 7.5 kg of air dried, amended soil. This allowed for five replicate seedlings in each of five growth temperature treatments at each of three harvests during the growth period. The growth temperatures were constant day/night temperatures of 10, 20, 25, 30 and 35 °C with vapor pressure deficit maintained at 1 kPa. Harvests were made when the seedlings were aged 11 weeks (Harvest 1), 15 weeks (Harvest 2) and 19 weeks (Harvest 3).
Seedlings were maintained at 25 °C for one week after transplanting, then allocated to one of the five temperature treatments in one of five replicated growth chambers. To re- duce temperature shock to the seedlings, the temperature within the growth cabinets was altered in daily steps no greater than 1 °C until the desired constant day/night temperatures were achieved. This meant that when seedlings were har- vested, the seedlings grown at 30 and 20 °C had experienced these constant temperatures for about 1 week less than seed- lings maintained at 25 °C, and seedlings grown at 35 and 10 °C experienced these constant temperatures for about 2 weeks less than seedlings maintained at 25 °C. Therefore, although the comparisons were made among seedlings of similar age (11 weeks at Harvest 1, 15 weeks at Harvest 2 and 19 weeks at Harvest 3), the seedlings were exposed to the treatment tem- peratures for different periods of time depending on the treat- ment.
During the period from seedling transplanting to harvest, seedlings were randomly repositioned within a growth cham- ber on a weekly basis. Growth chambers were randomly real- located to different treatments every 2 weeks. Temperature and vapor pressure deficit inside the growth chambers stabi- lized to the new conditions within 2 h of relocating the seed- lings. The seedlings were irrigated daily throughout the experiment.
Determination of CO2assimilation rate
Leaf CO2assimilation rate (Al) was measured on the youngest fully expanded leaf between Harvests 2 and 3 when the seed- lings were 18 weeks old. Measurements were made with an LI-6400 portable photosynthesis system (Li-Cor, Lincoln, NE) equipped with a light source (6400-02B LED, Li-Cor).
Measurements of Alwere completed between 0900 and 1500 h as it had previously been determined that Alshowed little diur- nal variation during this period. Measurements were made at the seedlings’ growth temperature and at a PPF of 1200 µmol m– 2s– 1to ensure the response was fully light saturated. Ambi- ent CO2was maintained at 360 µmol mol– 1during measure- ments of Al.
Canopy CO2assimilation capacity (Ac) was calculated by Equation 1 which is based on equations of gross plant photo- synthetic production (Atwell et al. 1999):
Ac= Al(1 – exp(–kSLAWl)) (1)
where the extinction coefficient, k, was assumed to be 0.5 (Atwell et al. 1999), and values of specific leaf area (SLA) and dry biomass of leaves (Wl) at the final harvest were used. Cal- culation of Acby Equation 1 was used as a representation of the seedlings’ carbon source strength at the final harvest, but it is acknowledged that this approach assumes Aldid not vary ei- ther over time or within the seedling, and does not account for carbon losses through respiration of the various organs.
Seedling growth and harvests
Seedling height from the stem–root interface to the apical meristem was measured every 2 weeks. At each harvest, stem diameter over bark and wood diameter, i.e., stem diameter un- der bark, were measured at the first internode. At Harvest 3 only, the next 5 mm of stem wood material was collected for microscopic examination of stem anatomy, and the next 20–40 mm of stem wood material was collected to measure stem wood density. The volume of the entire main stem wood was calculated from seedling wood diameter at the first internode and height assuming a conical stem form.
At each harvest, the areas of the two youngest fully ex- panded leaves on the main stem were measured with a Delta T leaf area meter (Burwell, Cambridge, United Kingdom) be- fore determining their dry mass. At the final harvest, the thick- ness of the youngest fully expanded leaves was measured with electronic callipers. The number and area of the remaining leaves were measured and their dry mass was determined after drying at 70 °C for 48 h. Specific leaf area (SLA) and mean area of a leaf were calculated from these data. The mass of the remaining main stem, branches and washed roots was deter- mined after drying at 70 °C for 48 h and seedling biomass was calculated after inclusion of the sample used to determine wood density. Leaf area ratio (LAR) was calculated from these data as the ratio of seedling leaf area to seedling biomass (cm2g– 1).
Wood density was calculated for the 20–40 mm stem seg- ment as dry mass/fresh volume. Fresh wood volume of the sample was measured by immersing the sample in a beaker of distilled water of known mass on a balance sensitive to 0.1 mg and observing the change in balance reading. The mass of wa- ter displaced in grams was taken to correspond closely to sam- ple volume in cm3. Each sample was lanced with a probe and placed just under the surface of the water. A clamp attached to a retort stand adjacent to the breaker was used to support the
probe and minimize movement of the sample. Measurements of volume were complete within 10 s of immersion to mini- mize entry of water into the cut ends of the wood sample. Mass of the sample was then determined after drying for 48 h at 105 °C.
Anatomy
Stem samples were fixed in 2.5% paraformaldehyde in 0.1 M Na-phosphate buffer (pH 7.0) for two days. Stem samples were washed in 0.1 M Na-phosphate buffer (pH 7.0), dehy- drated through an ethanol series and embedded in London White resin. Four of the five replicates per treatment collected at Harvest 3 were randomly selected for sectioning and ana- tomical observation. One µm thick sections of the sample were cut with glass knives with a Leica RM2165 microtome (Leica Microtomes, Nassloch, Germany). Mounted sections were stained with 0.5% (v/v) toluidine blue for 5 s before the digital images were captured with an Olympus BX60 compound mi- croscope (Olympus corporation, Tokyo, Japan) interfaced with an MTI 3CCD digital camera (DAGE MTI Inc., Michi- gan). All analyses of digital images were performed using Im- age Pro Plus (Media Cybernetics). The total cross-sectional area of wood, and the area of pith within the wood was calcu- lated. The number of fiber cells along radial sections (from the pith to the cambium), average fiber cell size, fiber cell wall thickness, the number of xylem vessels and the proportion of stem cross-sectional area occupied by xylem vessels were also determined. Over 75% of the growth in wood diameter and over 95% of the growth in biomass of the seedling and stem occurred in the final 8 weeks of growth, i.e., within the period during which constant growth temperatures were maintained.
The anatomical data were weighted to reflect the proportion of the stem’s entire cross-sectional area that the image repre- sented. Thus, the microscopy comparisons largely reflect the anatomical responses accountable for the treatment differ- ences of plant production.
The density of xylem vessels per area of wood and the size distribution of the xylem vessels were calculated. The diame- ters of fiber cell lumens were calculated from the difference between fiber cell diameter and fiber cell wall thickness. The cross-sectional areas of the fiber cell lumens (Fv) and of the fi- ber cell wall material (Fw) were calculated assuming the fiber cells were circular. The ratio Fv:Fwis therefore a measure of the amount of voids within a fiber cell in relation to the amount of carbon deposited in the fiber cell walls.
Statistical analysis
Where variance in data was heterogeneous, arcsin or log trans- formation were used, with data presented as non-transformed means. Effects of growth temperature were analyzed with the analysis of variance (ANOVA) algorithm of STATISTICA software (Version 6, StatSoft, Oklahoma, USA). Fishers LSD test was used for comparisons between treatments. The multi- ple regression algorithm of STATISTICA was used to deter- mine linear relationships between factors. The effect of plant ontogeny on biomass partitioning relationships was examined
by analysis of covariance (ANCOVA) of log-transformed data.
The slopes of these relationships were used to explore the par- titioning of biomass between different plant parts independent of plant size (Hunt 1990).
Results
Partitioning of seedling biomass – growth temperature or ontogeny?
We compared seedlings of a similar age but grown for periods ranging between 10 and 12 weeks at constant temperature (Harvest 3). Therefore, it is necessary to consider the effect of plant ontogeny on the results. Seedlings grown at different temperatures differed markedly in size; however, use of plants of similar size but different age did not alter the effect of tem- perature on wood density, namely, wood density increased with increasing growth temperature (Figure 1). Seedlings of similar size were grown for either 7–8 or 10–12 weeks at con- stant temperature (i.e., Harvest 2 and Harvest 3). In these anal- yses, we compared the slower-growing seedlings grown at 10 and 35 °C at Harvest 3 with the faster-growing seedlings grown at 20, 25 and 30 °C at Harvest 2. These groups of seed- lings were of similar biomass and had similar numbers of branches, indicating they were at an approximately similar de- velopmental stage. We were unable to make a similar compari- son between slower-growing seedlings at Harvest 2 with faster-growing seedlings at Harvest 1 because these groups of seedlings differed considerably in size and because of the rela-
tively short time the seedlings had grown at constant tempera- tures (between 4 and 6 weeks). Unfortunately, because Al, which is required for Ac, was measured only at Week 18, i.e., one week before Harvest 3, and data on stem anatomy at Har- vest 2 were not collected, we are able to compare the effects of growth temperature on Al, Acor stem anatomy only in seed- lings of similar age but not in seedlings of similar size.
The allometric relationships for seedlings of all harvests show that growth temperature influenced partitioning of seed- ling biomass between seedling organs (Figure 2). Thus we feel confident in exploring the relationships between biomass par- titioning and wood density. At Harvest 3 (10–12 weeks growth at constant temperature), partitioning of total seedling biomass in the roots declined from 35 to 20% as growth temperature in- creased (Figure 2). Partitioning of seedling biomass in stems increased with growth temperature from 0.5 g to a maximum of 23 g at 25 °C before declining, representing an increase from 4 to 13% of seedling biomass. Biomass partitioning to
Figure 1. Effects of growth temperature on: (a) wood density; (b) stem biomass; and (c) stem wood volume in 19-week-old seedlings grown at a constant temperature for the final 10–12 weeks (䊏). In (a) the ef- fects of growth temperature on wood density are also shown for 15-week-old seedlings grown at a constant temperature for the final 7–8 weeks (䊐). Each value represents the mean and SE of five repli- cates.
Figure 2. Effects of growth temperature on: (a) total seedling biomass;
(b) the proportion of total biomass in leaves; (c) wood in branches (䊐,䊏) and main stems (䊊,䊉); and (d) roots of 19-week-old seedlings (filled symbols) or of similar-sized seedlings (open symbols). Similar aged seedlings (19 weeks old) were grown at a constant temperature for the final 10–12 weeks. Seedlings, which were of similar size, were 19 weeks old if grown at 10 or 35 °C, but 15 weeks old if grown at 20, 25 or 30 °C. The 15-week-old seedlings were grown at a constant tem- perature for the final 7–8 weeks. Each value is the mean of five repli- cates. Bars represent SE.
woody branches increased more steadily from 15 to 22% as growth temperature increased (Figure 2). Leaf biomass re- mained constant at 46% of seedling biomass at all growth tem- peratures (Figure 2). Values were similar when seedlings of similar size were compared (Figure 2). Wood density was higher when a larger component of seedling biomass was par- titioned to the woody main stem and branches (r = 0.69; P <
0.05) (Figure 3).
Wood density and anatomy
Wood density increased 20% with increasing growth tempera- ture to 436 kg m– 3at 30 °C before declining to 410 kg m– 3 when seedlings were grown at 35 °C (P < 0.05) (Figure 1). The changes in wood density were related to changes in stem anat- omy (Figures 4 and 5). Wood density was positively correlated to fiber cell wall thickness (r = 0.62 ; P < 0.05), and negatively correlated to both the diameter of fiber cell lumen (r = –0.60;
P < 0.05) and the proportion of the stem occupied by xylem vessels (r = –0.76 ; P < 0.05) (Figure 4). Together these three variables explained 76% of the variation in wood density (r = 0.87). Seedlings grown at 10 °C had a larger proportion of stem allocated to xylem vessels and a higher abundance of xy- lem vessels than seedlings grown at higher temperatures (Ta- ble 1, Figure 5). Nearly 70% of xylem vessels in seedlings grown at 10 °C, like those present in seedlings grown at 35 °C, had diameters of less than 30 µm, and less than 5% of xylem vessels had diameters greater than 45 µm. In comparison, about 50% of xylem vessels in seedlings grown at tempera- tures between 20 and 30 °C had diameters less than 30 µm and 20% had diameters greater than 45 µm (Table 1). Temperature marginally decreased fiber cell diameter, and this, together with the changes in fiber cell wall thickness, reduced the fiber cell lumen diameter, and altered the spatial arrangement be- tween fiber cell lumen and fiber cell walls (Table 1). The ra- dius of the pith was similar in seedlings grown at 10 and 35 °C, and smaller than in seedlings grown at the other temperatures (Table 1); consequently, the proportion of stem area occupied by the pith was higher in seedlings grown at 10 and 35 °C. In seedlings in all treatments, however, the pith accounted for less than 3% of stem area (Table 1).
Seedling growth
The temperature optimum for maximum growth (i.e., biomass accumulation, stem volume, leaf area) of E. grandis differed from the temperature for maximum wood density (Table 1, Figures 1 and 6). Stem radial growth and stem volume were principally related to the production of new cells rather than cell expansion (Table 1, Figure 1). Fiber cell diameter varied by only 20% with changes in temperature, whereas fiber cell number increased by 290% (Table 1). Seedling height, diame- ter and biomass increased by up to 2000% with increasing growth temperature to a maximum at 25 °C before declining (Table 1, Figure 2). As a result, overall seedling size in the 10 and 35 °C growth temperature treatments were similar (Ta- ble 1, Figure 2). In contrast, leaf-scale parameters (number of leaves at final harvest, leaf area, leaf thickness, SLA, LAR, Al) had the same temperature optimum (30 °C) as observed for wood density (Table 1, Figures 1, 6). However, in all these cases, the 30% to greater than 80% decline in leaf scale param- eters at 35 °C relative to 30 °C contrast to the less than 10% de- cline in wood density as growth temperature increased from 30 to 35 °C (Figures 1 and 6). Leaf CO2assimilation rate in- creased with temperature to a maximum of 22 µmol m– 2s– 1at Figure 3. Wood density as a function of the proportion of total seed-
ling biomass partitioned to the woody stems and branches. Seedlings grown at a constant temperature for the final 2–4 weeks (䊏), 6–8 weeks (䊉) and 10–12 weeks (䉱) were 11, 15 and 19 weeks old, respectively. Each value represents the mean and SE of five replicates.
Figure 4. Relationships between wood density and (a) fiber cell wall thickness, (b) diameter of fiber cell lumen and (c) the cross-sectional area of the stem occupied by xylem vessels. Data are for 19-week-old seedlings grown for the final 10–12 weeks at constant temperatures of 10 °C (䊏), 20 °C (䊉), 25 °C (䉱), 30 °C (䉬) and 35 °C (夹). The re- gression line and correlation coefficient are for all data.
30 °C and declined sharply to 12 µmol m– 2s– 1at 35 °C (Fig- ure 6), whereas leaf respiration increased linearly from 1 µmol m– 2s– 1at 10 °C to 2 µmol m– 2s– 1at 30 °C then increased to 3 µmol m– 2s– 1at 35 °C. Although seedlings grown at 35 °C had greatly reduced growth but comparatively high wood den- sity, positive correlations existed between wood density and several leaf-scale parameters including leaf area, LAR and Al
(Figure 7).
Extrapolation of Alto Acbased on plant leaf area showed that Acpeaked at 25 and 30 °C and reached a minimum at 10 or 35 °C (Figure 8). As shown in Figure 8, Acwas positively cor- related with wood density (r = 0.63), wood volume (r = 0.85) and fiber cell wall thickness (r = 0.57).
Discussion
Wood density and anatomy
The increase in wood density of E. grandis seedlings grown at high temperatures was related to reductions in lumen trans- verse area of both xylem vessels and fiber cells and thicker fi- ber cell walls (Table 1, Figures 1, 4 and 5). Changes in these anatomical features accounted for 76% of the observed varia- tion in wood density.
The volume of pith could influence wood density, because the pith contains cells with thin walls and large lumens that af- fect overall stem density. However, the pith occupied less than 3% of stem area (Table 1) and thus contributed minimally to
wood density. The similar proportion of pith in stems of seed- lings grown at 10 and 35 °C and the large differences in wood density indicate that other factors play a more important role in the response of wood density to growth temperature (Table 1, Figures 4 and 5).
The decline in xylem vessel lumen diameter was most evi- dent in comparisons between seedlings grown at 10 and 35 °C (Table 1, Figures 1, 4 and 5). These seedlings were similar in other aspects of growth (stem volume, Al, leaf area and LAR) (Figures 1, 6 and 7), yet xylem vessel number, size distribu- tion, area of xylem vessels and wood density differed consid- erably (Table 1, Figures 1, 4 and 5). The importance of xylem vessel lumen diameter determining wood density of other an- giosperms has been shown in studies with East-Liaoning oak (Quercus liaotungensis Koidz) (Zhang and Zhong 1992) and other eucalypts (Malan and Hoon 1992, Thomas et al. 2004).
Xylem vessel size and number has a degree of plasticity that is dependent on environmental conditions. Why then do xylem vessels respond to growth temperature independently of evap- orative demand and water supply? Roderick and Berry (2001) postulated that changes in xylem vessel size in response to growth temperature are associated with the decrease in viscos- ity of water as temperature increases, making it unnecessary for a plant to have more or larger xylem vessels to maintain hy- draulic conductivity at high temperatures. This postulate is supported by the results of a study of E. camaldulensis Dehnh.
grown at different temperatures (Thomas et al. 2004), and by our finding that xylem vessel frequency and size of E. grandis Table 1. Characteristics of E. grandis seedlings aged 19 weeks and grown at constant temperature for the final 10–12 weeks. Values are means of five replicates for growth measurements and means of four replicates for anatomical determinations. Standard errors are shown in parenthesis.
Characteristic Growth temperature (°C) LSD0.05
10 20 25 30 35
Seedling morphology
Height (m) 0.14 (0.01) 0.61 (0.04) 0.99 (0.07) 0.70 (0.04) 0.21 (0.01) 0.15
Over bark stem diameter (mm) 6 (0.5) 14 (0.4) 16 (0.5) 13 (0.4) 5 (0.7) 1.5
Under bark wood diameter (mm) 4 (0.4) 11 (0.5) 12 (0.5) 10 (0.4) 3 (0.2) 1.5
Number of branches 9 (0.6) 15 (0.4) 18 (0.7) 17 (0.5) 11 (0.9) 1.8
Root dry matter (g) 4 (0.3) 31 (3.0) 36 (4.0) 30 (1.8) 2 (0.5) 6.8
Leaf dry matter (g) 6 (1.1) 61 (7.8) 79 (2.7) 67 (2.9) 5 (0.9) 13.9
Mean area per leaf (cm2) 30.7 (3.6) 37.6 (3.0) 40.2 (2.6) 25.4 (1.9) 7.0 (0.2) 8.4
Number of leaves on seedling at harvest 24 (2) 301 (44) 440 (37) 633 (46) 99 (14) 49
Leaf thickness (mm) 0.27 (0.01) 0.24 (0.02) 0.20 (0.01) 0.21 (0.01) 0.25 (0.02) 0.02
Wood anatomy
Pith radius (µm) 260 (68) 455 (39) 425 (28) 369 (37) 212 (37) 137
Pith area (% of stem area) 2.4 (0.6) 0.9 (0.2) 0.6 (0.1) 0.6 (0.2) 1.8 (0.4) 1.3
Number of fiber cells per radial transect 116 (13) 300 (16) 328 (13) 315 (17) 113 (10) 32 Average fiber cell diameter (µm) 15.1 (0.4) 15.6 (0.1) 16.0 (0.1) 14.6 (0.4) 13.4 (0.3) 0.9
Fiber cell wall thickness (µm) 1.8 (0.1) 2.1 (0.1) 2.3 (0.1) 2.3 (0.1) 2.0 (0.1) 0.3
Fiber cell lumen diameter (µm) 11.4 (0.4) 11.3 (0.3) 11.4 (0.1) 9.9 (0.4) 9.5 (0.5) 1.0 Cross-sectional areas of fiber cell lumen:fiber cell wall 1.31 (0.05) 1.12 (0.09) 1.05 (0.06) 0.87 (0.04) 1.03 (0.15) 0.23 Xylem vessel area (% stem area) 17.0 (0.3) 11.5 (0.5) 11.3 (0.4) 9.6 (0.2) 10.4 (0.4) 3.6 Xylem vessel number (vessels mm– 2stem ) 4 0 8 ( 6 3 ) 2 3 3 ( 2 1 ) 2 7 1 ( 3 3 ) 1 6 9 ( 1 4 ) 2 9 0 ( 4 8 ) 4 3 Proportion of xylem vessel < 30 µm diameter 0.68 (0.004) 0.54 (0.005) 0.52 (0.009) 0.47 (0.02) 0.88 (0.05) 0.05 Proportion of xylem vessel from 30 to 45 µm diameter 0.29 (0.006) 0.27 (0.02) 0.28 (0.03) 0.35 (0.03) 0.12 (0.05) 0.04 Proportion of xylem vessel > 45 µm diameter 0.03 (0.008) 0.19 (0.04) 0.20 (0.03) 0.18 (0.01) 0.0 (0.0) 0.04
decreased with increasing growth temperature (Table 1, Fig- ure 4).
The arrangement of mass to volume within the fiber cells decreased with growth temperature because of declining fiber cell expansion and increasing cell wall thickness (Table 1, Fig- ure 4). Both factors were important in explaining wood den- sity. These changes in fiber anatomy are analogous to the changes in tracheid cell anatomy of early and latewood in gymnosperms. The less-dense earlywood tracheid cells have thin walls and a large lumen, whereas the more dense latewood tracheids have thick walls and small lumens (Creber and Chaloner 1984, Zobel and Jett 1995). Wood density of gymno- sperms depends on the ratio of the lumen diameter of early- wood tracheids to latewood tracheids (Creber and Chaloner
1984). It appears that wood density of eucalypts depends on similar localized arrangements of fiber cell wall thickness to fiber lumen area, and these changes are influenced by growth temperature. The thickness of fiber cell walls was positively related to Ac (Figure 8), which was principally affected by changes in seedling leaf area. Cells with thicker walls had smaller lumen diameters, but the ratio of cell wall area to cell lumen diameter was also related to lumen fiber cell diameter (Table 1). Thus, growth temperature affects the diameters of fiber cells of E. grandis seedlings (Table 1) in a way that is similar to its effect on fibers of gymnosperm seedlings (Larson 1964, Richardson 1964). These data may not corre- spond with field data, because other growth effects, such as rainfall, elevation and soil conditions, may alter the general re- lationship between temperature and wood density (Wilkes 1987, Horacek et al. 1999) through independent effects on vol- ume growth and biomass accumulation or partitioning.
Figure 5. Photographs of 1 µm wood sections stained with toluidine blue obtained from 19-week-old seedlings grown at: (a) 10 °C; (b) 20 °C; (c) 25 °C; (d) 30 °C; and (e) 35 °C for the final 10–12 weeks.
The bar represents 100 µm.
Figure 6. Effects of growth at constant temperature for the final 10–12 weeks on: (a) leaf area; (b) leaf area ratio; (c) specific leaf area; (d) leaf CO2assimilation rate (Al); and (e) canopy CO2assimilation ca- pacity (Ac) of 19-week-old seedlings. Each value represents the mean and SE of five replicates.
Carbon supply and partitioning
In the stem, sink strength can be measured as the production of new cells and cell wall thickening. As hypothesized, tempera- ture affected sink strength by altering the overall biomass pro- duction and also by shifting the pattern of biomass partitioning from roots to shoots. The woody components of the shoots (stems and branches) rather than the leaves received a larger proportion of the extra biomass (Figure 2). The extra biomass contributed not only to greater cell production but also to thicker fiber cell walls and higher wood density (Table 1, Fig- ure 3). Similar changes in biomass partitioning in response to temperature have been observed in seedlings of beech (Fagus sylvatica L.) (Bruhn et al. 2000) and mountain birch (Betula pubescens Ehrh. ssp czerepanovii (Orlova) Hamet-Ahti) (Bruhn et al. 2000, Weih and Karlsson 2001).
The leaf-scale parameters contributing to Ac(i.e., Al, LAR and SLA) peaked at a growth temperature between 25 and 30 °C (Figure 6). The same parameters also increased in B. pubescens ssp czerepanovii seedlings grown at high air and soil temperatures (Weih and Karlsson 2001). Although the growth temperatures tested had no effect on biomass partition- ing to the leaves (Figure 2), LAR changed with growth tem-
perature because of changes in SLA, suggesting that the link to Alis through leaf morphology (Sefton et al. 2002). Both Aland leaf area contributed to changes in Ac(Figure 6), with changes in leaf area principally influencing these changes. Changes in Accontributed to changes in fiber cell wall thickness, wood volume and wood density of E. grandis (Figure 8), as found in other studies (Larson 1964, Creber and Chaloner 1984, Lindström 1996). Compared with seedlings grown between 20 and 30 °C, seedlings grown at 10 and 35 °C had low Aland very low leaf area resulting in low Ac, which was reflected in thin fiber cell walls (Table 1, Figures 6 and 8). Low photosyn- thate supply as a result of factors, such as higher night temper- atures leading to increased respiration, reduced solar irra- diance or shorter photoperiods, may result in reduced cell wall thickness of gymnosperm seedlings and saplings (Larson 1964, Richardson 1964).
Wood density: combining mass and volume
Wood density may be stable if Acand wood volume change by similar amounts. This occurred when growth temperature was increased from 20 to 25 °C (Figure 8). In contrast, wood den- Figure 7. Relationships between wood density and (a) leaf area, (b)
leaf area ratio and (c) leaf CO2 assimilation rate measured in 19-week-old seedlings. Each value represents the mean and SE of five replicates at each growth temperature. Seedling growth temperatures are indicated on graphs. Regression lines and correlation coefficients are for all growth temperatures (solid line) and for growth tempera- tures between 10 and 30 °C (dashed line).
Figure 8. Effects of canopy CO2assimilation capacity on (a) wood volume, (b) wood density and (c) fiber cell wall thickness of 19-week-old seedlings. Each value is the mean and SE of five repli- cates for each growth temperature. Seedling growth temperatures are indicated on graphs. Regression lines and correlation coefficients are for all growth temperatures (solid line) or for growth temperatures be- tween 10 and 30 °C (dashed line).
sity may increase if temperature affects wood volume inde- pendently of Ac. This occurred when growth temperature was increased from 25 to 30 °C (Figure 8), because fiber cell ex- pansion rather than cell division was affected (Table 1). When the temperature sensitivity of stem volume and stem biomass growth are similar, wood density is unaffected. Such coordina- tion of temperature responses occurred when growth tempera- ture increased from 20 to 25 °C. Similarly, elevated CO2
concentration, which can cause increases in Acand stem bio- mass, increased stem volume but not wood density of P. sylvestris and loblolly pine (Pinus taeda L.) (Telewski et al.
1999, Ceulemans et al. 2002).
In summary, temperature affected wood density by chang- ing the mass of the stem and the volume of the lumen within the stem. Stem mass changed because temperature altered both the seedling carbon source strength—by changing Al, and leaf area—and the partitioning of this carbon to the stems.
Temperature altered stem volume and stem radial growth by changing the rate of cell division more than the rate of cell ex- pansion. At a more detailed anatomical level, changes in wood density were driven by changes in the proportion of xylem lu- men volume. At the level of individual cells, the proportion of fiber cell lumen to fiber cell wall material altered because of declining fiber cell expansion and increasing cell wall thick- ness. The mechanism by which increasing temperature may reduce the total volume of xylem vessels per unit stem volume may be related to the decrease in the viscosity of water with increasing temperature, which allows fewer or smaller xylem vessels to transport similar volumes of water.
Acknowledgments
This research was funded by an ARC SPIRT grant in collaboration with State Forests of NSW (Grant No. C00001999). We thank Deborah Birch, Elizabeth Darley, Kate Düttmer, Mark Emanuel, Da- vid Giles, Matthew Searson and Christine Sefton for assistance during plant harvests and microscopy analysis, and Georgina Kelley for as- sistance with the preparation of the manuscript.
References
Antonova, G.F.V. and V.V. Stasova. 1993. Effects of environmental factors on wood formation in Scots pine stems. Trees 7:214–219.
Antonova, G.F.V. and V.V. Stasova. 1997. Effects of environmental factors on wood formation in larch (Larix Sibirica Ldb.) stems.
Trees 11:462–468.
Atwell, B., P. Kreidemann and C. Turnbull. 1999. Plants in action:
adaption in nature performance in cultivation. Macmillan Educa- tion Australia, Melbourne, Australia, 664 p.
Barber, V.A., G.P. Juday and B.P. Finney. 2000. Reduced growth of Alaskan white spruce in the twentieth century from temperature in- duced drought stress. Nature 405:668–673.
Bray, R.H. and L.T. Kurtz. 1954. Determination of total, organic and available forms of phosphorus in soils. Soil Sci. 59:39–45.
Briffa, K.R., F.H. Schweingruber, P.D. Jones, T.J. Osborn, S.G. Shi- yatov and E.A. Vaganov. 1998. Reduced sensitivity of recent tree-growth to temperature at high northern latitudes. Nature 391:
678–682.
Bruhn, D., J.W. Leverenz and H. Saxe. 2000. Effects of tree size and temperature on relative growth rate and its components of Fagus sylvatica seedlings exposed to two partial pressures of atmospheric [CO2]. New Phytol. 146:415–425.
Ceulemans, R., M.E. Jach, R. Van de Velde, J.X. Lin and M. Stevens.
2002. Elevated atmospheric CO2alters wood production, wood quality and wood strength of Scots pone (Pinus sylvestris L.) after three years of enrichment. Global Change Biol. 8:153–162.
Conroy, J.P., P.J. Milham, M. Mazur and E.W.R. Barlow. 1990.
Growth, dry weight partitioning and wood properties of Pinus radiata D. Don after 2 years of CO2enrichment. Plant Cell Envi- ron. 13:329–337.
Creber, G.T. and W.G. Chaloner. 1984. Influences of environmental factors on the wood structure of living and fossil trees. Bot. Rev.
50:357–448.
Deleuze, C. and F. Houllier. 1998. A simple process-based xylem growth model for describing wood microdensitometric profiles. J.
Theor. Biol. 193:99–113.
Denne, M.P. and M.D. Hale. 1999. Cell wall and lumen percentages in relation to wood density of Nothofagus nervosa. IAWA J. 20:
23–36.
Hacke, U.C., J.S. Sperry, W.T. Pockman, S.D. Davis and K.A. Mc- Culloh. 2001. Trends in wood density and structure are linked to prevention of xylem implosion by negative pressure. Oecologia 126:457–461.
Horacek, P., J. Slezingerova and L. Gandelova. 1999. Effects of envi- ronment on the xylogenesis of Norway spruce (Picea abies [L.]
Karst.). In Tree-Ring Analysis: Biological, Methodological and Environmental Aspects. Eds. R. Wimmer and R.E. Vetter. CABI Publishing,Wallingford, U.K., pp 33–54.
Hunt, R. 1990. Basic growth analysis: plant growth analysis for be- ginners. Unwin Hyman Ltd, London, U.K., 112 p.
Kramer, P.J. and T.T. Kozlowski. 1979. Physiology of woody plants.
Academic Press, New York, 811 p.
Larson, P.R. 1964. Some indirect effects of environment on wood for- mation. In Proc. The formation of wood in forest trees. Ed. M.H.
Zimmermann. Academic Press, New York, pp 345–365.
Lindström, H. 1996. Basic Density in Norway Spruce .1. A literature review. Wood Fiber Sci. 28:15–27.
MacDonald, E. and J. Hubert. 2002. A review of the effects of silviculture on timber quality of Sitka spruce. Forestry 75:
107–138.
Malan, F.S. and M. Hoon. 1992. Effect of initial spacing and thinning on some wood properties of Eucalyptus grandis. S. Afr. For.
J.163:13–20.
Muneri, A., T. Daido, M. Henson and I. Johnson, I. 2004. Variation in pulpwood quality of superior Eucalyptus dunnii families grown in NSW. In Proc. 58th Appita Annual Conference, Canberra, Austra- lia, pp 513–538.
Richardson, S.D. and J.M. Dinwoodie. 1960. Studies on the physiol- ogy of xylem development: The effect of night temperature on tracheid size and wood density in conifers. J. Wood Sci. 1:3–13.
Richardson, D.S. 1964. The external environment and tracheid size in conifers. In The formation of wood in forest trees. Ed. M.H. Zim- mermann. Academic Press, New York, pp 367–388.
Roderick, M.L. and S.L. Berry. 2001. Linking wood density with tree growth and environment: a theoretical analysis based on the motion of water. New Phytol. 149:473–485.
Sefton, C.A., K.D. Montagu, B.J. Atwell and J.P. Conroy. 2002. Ana- tomical variation in juvenile eucalypt leaves accounts for differ- ences in specific leaf area and CO2assimilation rates. Aust. J. Bot.
50:301–310.
Telewski, F.W., R.T. Swanson, B.R. Strain and J.M. Burns. 1999.
Wood properties and ring width responses to long term atmos- pheric CO2 enrichment in field grown loblolly pine (Pinus taeda L.). Plant Cell Environ. 22:213–219.
Thomas, D.S., K.D. Montagu and J.P. Conroy. 2004. Changes in wood density of Eucalyptus camaldulensis due to tempera- ture—the physiological link between water viscosity and wood anatomy. For. Ecol. Manage. 193, 157–165.
Thomas, D.S., K.D. Montagu and J.P. Conroy. 2006. Why does phos- phorus limitation increase wood density in Eucalyptus grandis seedlings? Tree Physiol. 26:35–42.
Turner, J. 1982. A long term superphosphate trial in regeneration of Pinus radiata at Blanglo State Forest, NSW. Aust. For. Res. 12:
1–9.
Wardlaw, I.F. 1979. The physiological effects of temperature on plant growth. Proc. N.Z. Agron. Soc. 9:39–48.
Weih, M. and P.S. Karlsson. 2001. Growth response of mountain birch to air and soil temperature: is increasing leaf-nitrogen content an acclimation to lower air temperature? New Phytol. 150:
147–155.
Wilkes, J. 1987. Review of the significance of variations in wood structure in the utilisation of Pinus radiata. Aust. For. Res. 17:
215–232.
Zhang, S.Y. and Y. Zhong. 1992. Structure-property relationships of woods in East-Liaoning oak. Wood Sci. Technol. 26:139–149.
Zobel, B.J. and J.B. Jett. 1995. Genetics of wood production.
Springer-Verlag, Berlin, 352 p.
