Introduction

Leaf hydraulic conductance (K leaf) is an important physiological parameter reflecting the ability of plants to transport water, since the resistance of leaves to water transport accounts for 30–80% of the total plant hydraulic resistance (Sack and Holbrook 2006). K leaf is measured as a rate of water flow through the leaf expressed on the unit of driving force for water movement (usually pressure) and scaled by leaf area (Sack et al. 2005). Leaf water transport likely has important ecological implications and may determine the ecological strategies of plant species (Siso et al. 2001; Sack and Holbrook 2006; Sinclair et al. 2008). For different species, maximum K leaf often correlates with photosynthetic capacity, maximum stomatal conductance and stomatal pore area per leaf area (Aasamaa et al. 2001; Sack et al. 2003; Brodribb et al. 2005).

Owing to complex leaf anatomy, the separation of leaf water transport into component pathways (apoplastic and cell-to-cell) and their contributions to short-term changes in K leaf is challenging. Inside leaves, water flows through a hierarchy of veins diminishing both in conduit number and size (midrib or primary vein, secondary veins, tertiary veins, etc., and ending in minor veins) and leaves the xylem through pit membranes. After leaving the leaf xylem, water moves across the bundle sheath cells, then through cell walls and may cross membranes to enter cells inside leaves (Sack and Holbrook 2006). Finally, inside the leaves, water evaporates and then exits the leaf through the stomata. Brodribb et al. (2007) have shown that K leaf is physically controlled by the length of post-venous hydraulic pathway. Changes in the plasma membrane water permeability brought about by aquaporins have been proposed to modulate the rapid short-term changes in K leaf (Nardini et al. 2005; Cochard et al. 2007). In the leaf apoplast, rapid changes in pH and ion composition can take place (Shabala and Newman 1999) and ion-mediated changes of xylem hydraulic conductivity have been documented in twigs (Zwieniecki 2001). One factor contributing to the age-related decline in K leaf is the accumulation of emboli in the vein xylem and their subsequent blockage by tyloses (Salleo et al. 2002; Nardini et al. 2003).

Since leaf hydraulic conductance can be regulated by changes in cell-to-cell and apoplastic pathways, it is a highly dynamic parameter. At a small temporal scale, K leaf can rapidly respond to environmental factors including temperature (Sack et al. 2004). During the season, K leaf typically declines during senescence resulting in the loss of photosynthetic capacity (Brodribb and Holbrook 2003a). In some plant species, K leaf has been shown to be sensitive to irradiance (Sack et al. 2002, 2003; Scoffoni et al. 2008; Tyree et al. 2005; Voicu et al. 2008; Voicu and Zwiazek 2010). However, the processes contributing to this response are still poorly understood. We have previously shown that in both bur oak (Quercus macrocarpa Michx.) and trembling aspen (Populus tremuloides Michx.), leaf lamina hydraulic conductance (K lam) is sensitive to chemical inhibitors (Voicu et al. 2008), while it is sensitive to irradiance only in bur oak, but not in trembling aspen (Voicu et al. 2008). In contrast to bur oak (Knapp 1992; Hamerlynck and Knapp 1994), trembling aspen is an early successional species and its stomata tend to remain open under fluctuating light conditions (Tobiessen and Kana 1974; Roden and Pearcy 1993), while leaf stomatal conductance in bur oak rapidly responds to changes in light (Knapp 1992; Hamerlynck and Knapp 1994).

In the present study, we examined the diurnal and seasonal effects on the responses of leaf hydraulic conductance components (leaf lamina conductance and petiole conductance) to irradiance in irradiance-sensitive (bur oak) and irradiance-insensitive (trembling aspen) species to understand if these responses were regulated by similar mechanisms. It has been previously shown that K leaf can vary diurnally and seasonally (Aasamaa and Sõber 2005; Lo Gullo et al. 2005; Nardini et al. 2005), although the exact regulation mechanisms are poorly understood. We hypothesized that the irradiance response of K lam in bur oak would not be present at night and that it would decline toward the end of the growing season. We also expected aspen K lam to be irradiance insensitive, regardless of the time of measurements and to decline during the growing season.

Materials and methods

Plant material collection

Shoots or individual leaves were collected from mature (20–30 years old) aspen (Populus tremuloides Michx.) and oak (Quercus macrocarpa Michx.) trees growing at the University of Alberta campus (Edmonton, AB, Canada) from the mid part of the canopy. Three individual trees of each species, 4–10 m tall were used for sampling. Excised shoots (about 1–2 years old) or leaves were immersed with their cut end in water and transported to the laboratory where the sun leaves were used immediately for K lam measurements (within 15–20 min of shoot sampling).

Leaf lamina hydraulic conductance (K lam), petiole hydraulic conductivity (K pet) and leaf-specific petiole conductivity (k p) determination

Petioles were re-cut under water and leaves were attached to the high-pressure flow meter (HPFM, Tyree et al. 1995) using compression couplings. Degassed distilled water was then forced inside leaves at constant pressure between 0.3 and 0.45 MPa depending on the leaf size. A computer recorded the water flow rate (Q, kg s−1) and the applied pressure (P, MPa) and computed the absolute leaf hydraulic conductance (K L, kg s−1 MPa−1) as Q/P. The values were determined every 2 s and saved as means every 60 s. Depending on the experiment, K L was continuously measured for 60 or 90 min in ambient laboratory light (<15 μmol m−2 s−1 PPFD) or under high-irradiance white light (about 1,000 μmol m−2 s−1 PPFD at the leaf level). Light was supplied by LED lamps (BL-300 series, Lamina Ceramics, Westhampton, NJ, USA). During K L measurements, leaves were immersed in a glass container filled with distilled water to buffer temperature fluctuations. At the end of the K L measurement, the leaf lamina was detached and its area (A) measured with a leaf area meter (LI-3000, Li-Cor Biosciences, Lincoln, NB, USA). Following lamina detachment, the water flow across the petiole open vessels stabilized within 1–2 min and was recorded for another 6–10 min. Petiole conductance (K P, kg s−1 MPa−1) was calculated as the mean of the last six stable readings. Five to six leaves were used to measure K L and K p.

The absolute hydraulic conductance of the leaf lamina (K L, kg s−1 MPa−1) was calculated as K lam = 1/[(1/K L) − (1/K P)] (Sack et al. 2002) and was divided by A to obtain leaf lamina hydraulic conductance (K lam, kg s−1 MPa−1 m−2) (Sack et al. 2002). Since the temperature of the water bath was relatively stable (20 ± 0.8°C for high-irradiance treatments), no corrections of K L for temperature were performed.

Petiole hydraulic conductivity (K pet, kg s−1 MPa−1 m) was calculated as K P multiplied by petiole length (Sack et al. 2002). Leaf-specific petiole hydraulic conductivity (k p, kg s−1 MPa−1 m−1) was calculated as K pet divided by leaf area (Sack et al. 2002).

Klam diurnal and seasonal variation

In June 2006, bur oak and aspen leaves were collected and K lam determined as above at 06:00, 10:00, 14:00, 18:00 and 24:00 h. Air temperature and irradiance in the vicinity of the trees were also measured at the time of leaf collection. Leaf hydraulic conductance appeared stable throughout the day, with the least variation in the morning. Therefore, in August and September 2006, leaves were collected between 08:00 and 12:00, in the morning, and air temperature, irradiance and K lam measured. The values were then compared with the ones obtained from the leaves collected at 10:00 h in June 2006. For bur oak, an additional set of leaves that were brown in color (dead leaves), but still attached to stems, were collected and measured at the end of August to the beginning of September.

Stomatal movement in detached leaves

Since K leaf determination was performed in detached leaves under laboratory conditions, we wanted to check the extent of stomatal movement under similar conditions to those used for K leaf measurements. Freshly collected leaves of aspen and oak trees had their petioles re-cut under water and then they were connected through plastic tubing to containers filled with distilled water and placed on the laboratory bench. Stomatal conductance (g s) was measured with a steady-state porometer (LI-1600, LI-COR, Lincoln, NE, USA) every 5 min for 30 min each in ambient laboratory fluorescent light (<15 μmol m−2 s−1 PPFD), followed by high irradiance and then again ambient light. High irradiance (approximately 1,200 μmol m−2 s−1 PPFD) was supplied by white LED lamps as described above. The experiments were performed in August 2006. Six leaves for each tree species were used to assess stomatal movement.

Leaf anatomy

Leaf segments (1.5 × 0.5 cm) were cut with a razor blade from the mid part of the leaves of oak and aspen trees. The segments were fixed in FAA (formalin–ethanol (70%)–acetic acid, 90:5:5). Subsequently, tissue samples were dehydrated in increasing ethanol concentrations and embedded in paraffin using the Fisher Histomatic Tissue Processor Model 166 (Fisher Scientific, Pittsburgh, PA, USA). Leaf sections, 10-μm thick, were cut with a microtome, stained with Safranin O and Fast Green FCF, mounted in DPX media and viewed under the light microscope. The images were captured using a Leica DMRXA light microscope (Meyer Instruments, Houston, TX, USA) with an Optronics MacroFire Digital Camera (LM-MFCCD) (Optronics, Goleta, CA, USA). Seven leaves were sampled from three bur oak and four aspen trees. Six leaf sections from each leaf were examined.

Statistical analysis

The data were analyzed using SAS 9.1 (SAS Institute, Cary, NC, USA). Residuals were checked for normality and homogeneity of variances. The daily and annual variation of K lam data were analyzed using repeated measures analysis of variance performed by Proc Mixed procedure. A Kenward-Roger option was used to calculate denominator degrees of freedom. Stomatal conductance was analyzed by paired t test or non-parametric paired test (sign rank). Least-square means or means were calculated for each treatment and tested for significance at P = 0.05. Parametric (r p) and Spearman rank (r s) correlation analysis was performed among air temperature and the irradiance recorded at the time of leaf sampling and K lam recorded at times 1, 30, 60 and 90 min (t1, t30, t60 and t90) during the HPFM K lam determination, K pet and k p. Lamina hydraulic conductance was also correlated with K pet and k p.

Results

Klam irradiance dependency and its diurnal variation in leaves from oak and aspen trees

Bur oak

Klam abruptly decreased for the first 10 min of HPFM measurements and reached a quasi-stable value after about 30 min in ambient light (Fig. 1a). After 30 min of high-irradiance exposure, Klam increased by about seven times. Upon returning to ambient light, Klam decreased slowly and did not return to the initial values in ambient light after 30 min (Fig. 1a). Diurnal variations in Klam irradiance response were examined every 4 h between 06:00 and 24:00 h (Fig. 1b). When measured in ambient light, the differences in Klam at different times of a day were not statistically significant; however, Klam was slightly higher in leaves collected at 06:00 and 10:00 h (0.62 and 0.64 ± 0.09 × 10−4 kg s−1 MPa−1 m−2, respectively) compared with the leaves collected at 18:00 and 24:00 h (0.52 and 0.50 ± 0.09 × 10−4 kg s−1 MPa−1 m−2, respectively) (Fig. 1b). When leaves were exposed to 30 min of high irradiance during HPFM measurements, Klam increased by about seven times, regardless of the leaf collection time, with the lowest Klam values attained in leaves collected at midnight (Fig. 1b). After 30 min of ambient light following high-irradiance exposure, Klam in leaves collected at 06:00 h remained significantly higher than those of leaves collected at 24:00 h. The variation shown by Kpet and kp was not statistically significant; however, the lowest values were recorded at 06:00 and 24:00 h (Table 1).

Fig. 1
figure 1

The light response of leaf lamina hydraulic conductance (K lam) and its diurnal variation in a, b bur oak (Quercus macrocarpa), and c, d aspen (Populus tremuloides) trees. a, c During HPFM measurement of K lam, leaves were exposed to 30 min of ambient light (PPFD <15 μmol m−2 s−1), followed by 30 min of high irradiance (1,000 μmol m−2 s−1 PPFD), and by another 30 min of ambient light. b, d Leaves were collected at 06:00, 10:00, 14:00, 18:00 and 24:00 h in June 2006, and K lam values at the end of different irradiance treatments (t30, t60 and t90) are shown. Least-square means or means ± SE are presented (n = 5–6). Bars with different letters are significantly different (P ≤ 0.05) from the corresponding values for different times of the day

Full size image
Table 1 Petiole hydraulic conductivity (K pet, kg s−1 m MPa−1) and leaf-specific petiole hydraulic conductivity (k p, kg s−1 m−1 MPa−1) of bur oak (Quercus macrocarpa) and aspen (Populus tremuloides) leaves as measured in different experiments
Full size table

Aspen

After connecting to the HPFM, K lam in leaves exposed to ambient light increased for the first 10 min followed by a decline for the next 20 min and then another increase (Fig. 1c). For aspen leaves, there was no difference in K lam, K pet and k p between ambient light and high-irradiance treatments (Fig. 1c and Table 1). The values of K lam were highest in the morning and decreased throughout the day reaching a minimum at 18:00 h (Fig. 1d) regardless of the irradiance level during HPFM measurements. At night, K lam increased again, as shown by the higher values measured at 24:00 h (Fig. 1d). Changing the irradiance level during K lam measurements did not significantly affect K lam in aspen throughout the day (Fig. 1d). Petiole hydraulic conductivity (K pet) and leaf-specific petiole conductivity (k p) did not significantly change throughout the day (Table 1). Their highest values were measured at 06:00 h.

Seasonal changes in K lam of oak and aspen trees

Bur oak

The initial K lam values measured in ambient light for the first 1–2 min of HPFM measurements were highest in June followed by August and September 2006. For leaves collected in June and August, 2006 K lam increased by about seven times following exposure to high irradiance (Fig. 2a). Leaves that were measured in June and August maintained green color (Fig. 2b). The leaves collected in September were mostly yellow with some green patches (Fig. 2b). K lam response to irradiance was still present, although the magnitude of the response was lower than the magnitude of the response for the leaves measured in June or August (Fig. 2a). The leaves that were brown at the time of collection showed no responses of K lam to irradiance (Fig. 2a, b). The differences in K pet and k p were not statistically different (Table 1).The greatest values of K pet and k p were recorded in June and August 2006, respectively, and were followed by a decline through the rest of the sampling dates (Table 1).

Fig. 2
figure 2

Seasonal variation of leaf lamina hydraulic conductance (K lam) in a bur oak (Quercus macrocarpa) and c aspen (Populus tremuloides) trees. Leaves were collected in June, August and September. a Bur oak leaves K lam was measured for 30 min in ambient light (PPFD <15 μmol m−2 s−1), followed by 30 min in high irradiance (1,000 μmol m−2 s−1 PPFD), and by another 30 min in ambient light. Brown leaves were collected in August and September. b Oak leaf appearance: green, senescing and brown leaves. c K lam for aspen leaves was measured in ambient light for 90 min; in June and August K lam was measured in green leaves. d Aspen leaf appearance: green and senescing leaves. For a and c least-square means or means ± SE are shown (n = 6)

Full size image

Aspen

Seasonal changes in aspen K lam were examined only in ambient light, since aspen K lam did not change with the irradiance level during HPFM measurements (Fig. 1c). When measured at different times during the season, K lam showed the same pattern of variation during the 90 min of HPFM measurements. Lamina hydraulic conductance was significantly greater in June than in August (Fig. 2c). Both June and August measurements were carried out with healthy green leaves (Fig. 2d). Yellow, senescing leaves that were measured in September 2006 had the lowest K lam throughout the measurements. However, the K lam of senescing leaves was not significantly different from the K lam of leaves measured in August 2006 (Fig. 2d). There were no significant changes in K pet and k p (Table 1) during the season. However, K pet was highest in senescing leaves (Table 1).

Irradiance level and g s in detached oak and aspen leaves

Detached leaves from bur oak trees had relatively low stomatal conductance when measured after 30 min of exposure to ambient laboratory light (Fig. 3). After 30 min of high-irradiance exposure, g s increased from 4.55 ± 0.28 to 96.99 ± 17.14 mmol m−2 s−1 (P = 0.003); within 30 min of returning the leaves to ambient light, the g s returned to 4.48 ± 0.28 mmol m−2 s−1 (P = 0.003, Fig. 3).

Fig. 3
figure 3

Stomatal conductance (g s, mmol m−2 s−1) in detached leaves from aspen (Populus tremuloides) trees and bur oak (Quercus macrocarpa) trees. To avoid dehydration, leaves were connected to containers filled with distilled water. Stomatal conductance was measured for 30 min in ambient laboratory light (PPFD < 15 μmol m−2 s−1) followed by 30 min of high-irradiance white light (1,000 μmol m−2 s−1 PPFD) and by 30 min of ambient light. Mean ± SE are shown

Full size image

In aspen leaves, g s decreased from 119.35 ± 23.27 to 57.72 ± 15.65 mmol m−2 s−1 (P = 0.015) after 30 min in ambient light, and increased to 121.30 ± 21.59 mmol m−2 s−1 (P = 0.003) following 30 min of high-irradiance exposure (Fig. 3). Upon returning the leaves to ambient light, g s decreased to 84.93 ± 13.81 mmol m−2 s−1.

Correlation analysis of K lam, K pet and k p with environmental characteristics

For the diurnal variation, the ambient air temperature and irradiance at the time of leaf sampling were generally not correlated with the K lam values that were recorded at different times during the measurement procedure (t1, t30, t60, and t90) or with K pet and k p (Table 2). The exceptions were aspen K lam at t1 and bur oak K lam at t90, which weakly correlated with ambient air temperature (Table 2).

Table 2 Correlation coefficients of leaf lamina hydraulic conductance (K lam, as determined after 1, 30, 60 and 90 min of HPFM measurements; t1, t30, t60 and t90), petiole hydraulic conductivity (K pet) and leaf-specific petiole hydraulic conductivity (k p), with temperature (T) and irradiance (PPFD) recorded at the time of leaf sampling, in bur oak (Quercus macrocarpa) and aspen (Populus tremuloides) trees
Full size table

Lamina hydraulic conductance of aspen, but not bur oak, leaves that were measured throughout the season was positively correlated with the air temperature and irradiance at the time of leaf sampling (Table 2). Petiole hydraulic conductivity was negatively correlated with the environmental variables for aspen trees (Table 2).

Correlation analyses were also performed for K pet and k p with the different diurnal and seasonal values of K lam (Table 3). Diurnal values of K pet and k p were correlated with K lam recorded at t30 (after 30 min of ambient light HPFM measurements) for both species (Table 3). During the annual variation, K pet and k p were correlated with K lam values recorded at t60 (30 min of ambient laboratory light followed by 30 min of high irradiance during the HPFM measurements) for both species (Table 3).

Table 3 Correlation coefficients of K lam (as determined after 1, 30, 60 and 90 min of HPFM measurements: t1, t30, t60 and t90) with K pet and k p in bur oak (Quercus macrocarpa) and aspen (Populus tremuloides) trees
Full size table

Leaf anatomy of bur oak and aspen

Bur oak leaves were characterized by a thick adaxial epidermal layer, two layers of thin and elongated palisade cells and the presence of bundle sheath extensions throughout the mesophyll (Fig. 4a). Smaller vascular bundles were placed closer to the abaxial epidermis, and the lower epidermal cells were much smaller compared with the adaxial epidermis (Fig. 4a). Stomata were present on the leaf abaxial side. Major veins typically contained sclerenchyma tissue that was more developed in the lower part of the vein, while the midrib had sclerenchyma equally developed around the vein (Fig. 4b).

Fig. 4
figure 4

Cross section of bur oak (Quercus macrocarpa) leaf blade (a) and midrib (b). Different structures are indicated by arrows and letters: UE upper epidermis, LE lower epidermis, P palisade cells, SM spongy mesophyll, BSE bundle sheath extensions, X xylem, SC sclerenchyma, C colenchyma. One division on the scale bar represents 10 μm

Full size image

The cross sections of aspen leaves revealed two layers of palisade cells and a thinner layer of spongy mesophyll (Fig. 5a). The midrib was surrounded by sclerenchyma tissue (Fig. 5b). Major veins also contained upper and lower sclerenchyma caps (Fig. 5a). Bundle sheath extension from smaller veins to the adaxial epidermis had the characteristics of parenchyma cells and extended all the way to the adaxial epidermis. In aspen leaves, the stomata were present on the leaf abaxial side (Fig. 5b).

Fig. 5
figure 5

Cross section of leaves from aspen (Populus tremuloides) trees: a leaf blade and b midrib. Different anatomical features are indicated by arrows and letters: UE upper epidermis, LE lower epidermis, P palisade cells, SM spongy mesophyll, VB vascular bundle, X xylem, S stomata, SC sclerenchyma. One division on the scale bar represents 10 μm

Full size image

Discussion

Leaf hydraulic conductance has been shown to vary diurnally in various angiosperm species (Lo Gullo et al. 2005; Nardini et al. 2005) implying circadian regulation. Diurnal depression of K leaf may be due to cavitation of the xylem in leaf veins (Nardini et al. 2001, 2003; Brodribb and Holbrook 2004). During HPFM measurements, water is pushed under pressure inside leaves and likely dissolves any preexisting embolism (Tyree et al. 1995; Nardini et al. 2001). Therefore, vein embolism was likely not a factor in the diurnal variation of K lam in aspen and bur oak leaves reported here and measured using the HPFM technique.

Irradiance-induced increases in K leaf were previously reported in red oak, walnut and bur oak leaves (Tyree et al. 2005; Cochard et al. 2007; Voicu et al. 2008). The present study showed that the irradiance response of K lam in bur oak (Fig. 1a) was present throughout the day (Fig. 1b) and that K lam (at t60 or t30) showed a small decline from morning to midnight. Irradiance-dependant increases of K lam appears to be an intrinsic property of bur oak leaves, while part of K lam reduction in darkness was likely due to energy deprivation of leaf cells, which may affect membrane transport processes (Aasamaa and Sõber 2001).

In aspen leaves, K lam was not sensitive to the irradiance level during the HPFM measurements (Fig. 1c), regardless of the time of the day when the leaves were sampled (Fig. 1d), and was higher than K lam of bur oak leaves. Lamina hydraulic conductivity tends to be higher in fast-growing species than in slow-growing ones (Aasamaa and Sõber 2005). Aspen is also known to maintain high stomatal conductance during the day irrespective of irradiance (Tobiessen and Kana 1974; Roden and Pearcy 1993) and, therefore, leaves may lose large quantities of water. Since in July, days are quite long in Edmonton, it is possible that as a consequence of water loss aspen K lam in the present study declined from early morning to 18:00 h (Fig. 1d). Daily changes in K leaf usually parallel changes in leaf water potential (Brodribb and Holbrook 2003b; Johnson et al. 2009). However, further studies would help elucidate the relationship between K lam and diurnal changes of water potential in aspen. The recovery of aspen K lam occurring at night appears to be independent of the energy-exchange processes happening during the day and may be a consequence of re-hydrated leaves. Maintenance of a high K lam throughout the day irrespective of the irradiance level in aspen leaves may be part of the strategy to better exploit rapid changes in irradiance (sunflecks). Other components of this strategy are the maintenance of a high stomatal conductance irrespective of irradiance and a faster photosynthetic induction response (Tobiessen and Kana 1974; Roden and Pearcy 1993). A high K lam is consistent with an apoplastic route as the dominant pathways for water movement is aspen leaves (Voicu and Zwiazek 2010). However, this may have negative consequences under water shortage. Therefore, the question of how aspen leaves would respond to water stress in terms of adjusting K lam through the contribution of different pathways should be further addressed.

Diurnal variation of K lam may arise from changes in the apoplastic or cell-to-cell pathways within a leaf. In leaves, diurnal variation in the aquaporins transcript and protein levels have been reported (Moshelion et al. 2002; Hachez et al. 2008) and K lam irradiance-dependant increase correlated with the expression of aquaporins in walnut leaves (Cochard et al. 2007). However, the contribution of different pathways to K lam appears to be species specific (Aasamaa and Sõber 2005; Sack et al. 2005) and K lam is likely differentially regulated. Changes in the xylem permeability of tree branches can be brought about by solute-mediated changes in xylem permeability (Zwieniecki 2001) mediated by the phloem (Zwieniecki et al. 2004) or by xylem parenchyma cells (Fromard et al. 1995). These effects still remain to be proven in leaves. However, these processes, if existent in leaves, would have been likely suppressed by the HPFM measurements in which the leaves are always flushed with the same solution. Interestingly, in both species, the diurnal variation of K pet, largely disappeared when k p was calculated, suggesting that the ability of a petiole to supply water to a given leaf area remained relatively unchanged throughout the day. This may also be indicative of a stronger effect of the leaf lamina tissues on the leaf hydraulic conductance.

A positive correlation was previously demonstrated between diurnal K leaf values, measured by the vacuum technique, and the PPFD values at the time of leaf sampling (Lo Gullo et al. 2005). In our study, PPFD and air temperature recorded at the time of leaf collection were largely not correlated with K lam, suggesting that the HPFM measurement conditions may override the influence of some of the environmental variables on K lam. However, we still detected changes in K lam, which likely reflected the transport properties of apoplast and symplast within the leaves and not the changes in water viscosity that affect the hydraulic conductivity (Sack et al. 2004).

In the present study, aspen and bur oak showed different K lam and K pet seasonal variations. For bur oak, K lam irradiance response declined only in senescing (yellow) leaves and the decline was coordinated with the reduction in K pet. At the same time, ambient light K lam values for leaves sampled in June, August and September were similar to those of brown leaves (Fig. 2a). Since water may move mostly through apoplast under low irradiance conditions (Cochard et al. 2007), it is tempting to speculate that brown leaf K lam may be a measurement of the leaf apoplast, as the brown leaves did not likely contain much living tissue. However, under normal circumstances, there is no water flow in dead leaves and xylem is blocked by tyloses and air bubbles. In aspen leaves, K lam decreased by August (Fig. 2c), which was correlated with the decline in air temperature (Table 2). This was not likely due to leaf senescence, since the leaves were still green and showed no signs of senescence. Seasonal changes of K lam are related to changes in the properties of both protoplasts and apoplasts (Aasamaa and Sõber 2005) and include loss of membrane function, reduction in the permeability of cell walls (Fleet 1950), and the accumulation of tyloses in previously embolized xylem conduits (Salleo et al. 2002). For aspen petioles, K pet was negatively correlated with air temperature during the season, in concurrence with an increase in K pet in senescing leaves. Increased temperature usually results in increased hydraulic conductivity due to a reduction in water viscosity (Sack et al. 2002), but in the present study, the measurements were carried out at the same temperature. Since petioles may be subjected to cavitation in growing trees and are known to be more vulnerable to cavitation than stems (Hacke and Sauter 1996), the increase in aspen K pet during a growing season may be a consequence of cavitation fatigue in a xylem weakened by cycles of embolism and refilling (Hacke et al. 2001) that take place during the growing season.

Concerns have been raised that the irradiance response of K lam during HPFM measurements may be dependent on stomatal opening (Sack et al. 2002). In the present study, except for the initial decrease of K lam in ambient light, g s responses in detached oak leaves mirrored K lam responses when measured under the same irradiance conditions (Figs. 1a, 3), consistent with bur oak leaves within the tree crown for which g s can rapidly decrease in transition from sun to shade (Knapp 1992; Hamerlynck and Knapp 1994). In detached leaves of aspen, g s response to changing light conditions followed a different trend than K lam in variable light (Figs. 1c, 3). However, ABA-treated leaves of Juglans regia failed to open the stomata in high-intensity light, while K lam was still increasing (Tyree et al. 2005). More recently, (Cochard et al. 2007) showed by rapidly freezing water-infiltrated leaves that water conductance through stomata was five to seven orders of magnitude higher than K leaf, as predicted by theoretical calculations (Tyree et al. 2005). Therefore, in the present study, stomatal opening probably did not influence the K lam values. Moreover, other methods of assessing K leaf were also used to show irradiance-induced increases in K leaf (Scoffoni et al. 2008). However, a possible impact of very tight stomatal closure on K lam (Tyree et al. 2005) should be addressed by future studies.

Leaf hydraulic conductance is correlated with the anatomical characteristics of xylem, mesophyll and epidermis (Aasamaa et al. 2001) and is positively related to palisade mesophyll thickness and to the ratio of palisade to spongy mesophyll (Sack and Frole 2006). In the present study, the anatomy of leaves from aspen and bur oak was consistent with type 2 hydraulic design as described for Quercus rubra and Populus nigra (Zwieniecki et al. 2007). These features include bundle sheath extensions that may conduct water from vascular bundles to the epidermal cells and disconnect photosynthesizing mesophyll cells from the pathway of rapid water movement (Wylie 1952). A higher K lam in aspen likely reflected a shorter path for water flow from veins to the stomata (Brodribb et al. 2007), which is consistent with the anatomy of aspen leaves showing the veins surrounded by less compact tissues than in oak leaves and in close proximity to the stomatal opening.

In conclusion, our study showed that the response of K lam to irradiance in bur oak was present in leaves sampled at different times during the day and that it declined in senescing leaves. In aspen leaves, K lam was not sensitive to irradiance during HPFM measurements and showed both diurnal and seasonal variations. Therefore, leaves of different study species should be first examined for possible light sensitivity and diurnal variations before conducting the measurements of K leaf or K lam. Also, K lam and K pet should be separated in all studies, since K pet can undergo changes that are independent of K lam.