The coordination of veins and stomata during leaf acclimation to sun and shade can be facilitated by differential epidermal cell expansion so large leaves with low vein and stomatal densities grow in shade, effectively balancing liquid- and vapour-phase conductances. As the difference in vapour pressure between leaf and atmosphere (VPD) determines transpiration at any given stomatal density, we predict that plants grown under high VPD will modify the balance between veins and stomata to accommodate greater maximum transpiration. Thus, we examined the developmental responses of these traits to contrasting VPD in a woody angiosperm (Toona ciliata M. Roem.) and tested whether the relationship between them was altered. High VPD leaves were one-third the size of low VPD leaves with only marginally greater vein and stomatal density. Transpirational homeostasis was thus maintained by reducing stomatal conductance. VPD acclimation changed leaf size by modifying cell number. Hence, plasticity in vein and stomatal density appears to be generated by plasticity in cell size rather than cell number. Thus, VPD affects cell number and leaf size without changing the relationship between liquid- and vapour-phase conductances. This results in inefficient acclimation to VPD as stomata remain partially closed under high VPD.
One of the major questions in plant development is how plants efficiently construct leaves capable of supplying enough water to replace transpirational loss. Central to this is the balance between leaf vein density, which is a critical factor in hydraulic conductance and therefore water supply (Sack & Frole 2006; Brodribb, Feild & Jordan 2007), and stomatal density and size, which dictates maximum stomatal conductance and therefore maximum transpiration (Franks & Beerling 2009). Leaf vein density and stomatal density remain proportional during acclimation to sun and shade, both among populations and within individuals in a number of woody angiosperm species (Brodribb & Jordan 2011; Carins Murphy, Jordan & Brodribb 2012). This relationship reflects an efficient balance between investment in liquid and vapour conductances in the leaf. Thus, the most efficient utilization of vein and stomatal investment occurs when the supply of water to evaporative surfaces near the stomata is just enough to maintain fully open stomata under saturating light conditions in the field (Franks et al. 2012). If the maximum evaporative capacity of the epidermis is greater than the capacity of the vascular system to maintain leaf hydration, then stomata will be unable to remain open (Salleo et al. 2000; Brodribb & Holbrook 2003). Hence, exceeding a certain stomatal density will incur costs (in the construction and regulation of stomata) that are not matched by greater photosynthetic yield. Likewise, having excess venation will also be inefficient due to both the high carbon cost of synthesizing lignin (the main component of leaf venation) (Lambers & Poorter 1992) and the loss of photosynthetic potential resulting from the displacement of photosynthetic tissue by vascular tissue.
Coordination between water transport and stomatal systems allows leaves to maintain an efficient balance between water use and carbon acquisition while accommodating the different rates of photosynthesis and evaporation experienced under high and low irradiance (Brodribb & Jordan 2011; Carins Murphy et al. 2012). It remains to be seen, however, whether other environmental pressures that influence plant water use induce similar developmental responses. Although plasticity has been observed in both stomatal density and vein density in response to variation in humidity (Salisbury 1928; Torre et al. 2003), temperature (Ciha & Brun 1975; Luomala et al. 2005; Zhu et al. 2012) and carbon dioxide concentration (Woodward 1986), it is unknown how the balance between stomata and veins is maintained under such conditions.
While leaf anatomy (vein density and stomatal density/size) determines conductances to liquid water and water vapour, it is the leaf environment that translates these conductances into water fluxes. In particular, atmospheric humidity expressed as the ambient leaf to air vapour pressure difference (VPD) determines how much water is transpired at any given stomatal conductance. Thus, we would expect a leaf growing under high VPD conditions to require either a higher vein density or lower stomatal density than a leaf growing under low VPD conditions to maintain the same leaf water potential during midday transpiration. Published evidence of these responses, however, is somewhat contradictory. Studies of a range of herbs and trees have shown an increase in stomatal density with increasing VPD (Salisbury 1928; Leuschner 2002; Lake & Woodward 2008; Hovenden, Vander Schoor & Osanai 2012). However, the opposite response has been observed in tomato, capsicum, eggplant and Rosa (Bakker 1991; Torre et al. 2003). Furthermore, in the case of Rosa, the decrease in stomatal density was associated with an increase in vein density (Torre et al. 2003).
Assuming that the most efficient investment in stomata and veins occurs when stomata are able to open near to maximum aperture in hydrated soil and under saturated light (Franks et al. 2012), we would expect that increased VPD would induce a decrease in stomatal density/size or an increase in vein density. Thus, evaporative demand could be reduced to match water supply by reducing stomatal density or size, or alternatively vein density could be increased to improve hydraulic capacity, thus accommodating higher maximum transpiration. Therefore, we hypothesize that an increase in VPD would be associated with a commensurate change in the balance between vein and stomatal density.
In a previous study, we found that changes in epidermal cell size in response to differential leaf expansion under high and low irradiance allowed for coordinated development of leaf vein and stomatal density in the subtropical rainforest tree Toona ciliata M. Roem. (Carins Murphy et al. 2012). Leaves grown under high irradiance were significantly smaller and had higher vein and stomatal density than leaves grown under low irradiance (in which densities of veins and stomata were effectively ‘diluted’ by increased epidermal cell size). However, changes in VPD affect the transpiration rate per stomatal pore, potentially disrupting the balance between liquid- and vapour-phase conductances in the leaf. Thus, if vein and stomatal densities remain proportional under increased VPD, a reduction in stomatal aperture would be required to maintain the balance between water conductances. However, the costs associated with having closed stomata mean that the most efficient use of resources would occur if anatomical adjustments (i.e. a reduction in stomatal density or an increase in vein density) mitigated the need for stomatal closure. Here, we compare plasticity in the density of stomata and veins of T. ciliata individuals grown under contrasting VPD to those observed under different light intensities to test the aforementioned hypothesis that acclimation to VPD would involve modification of the relationship between vein and stomatal density.
Materials and Methods
The subtropical woody angiosperm T. ciliata was selected for this experiment because it is a long-lived pioneer species capable of growing under a range of conditions (Herwitz, Slye & Turton 1998) with known plasticity in leaf size, vein density and stomatal density (Carins Murphy et al. 2012). Four T. ciliata plants per VPD treatment were grown from commercially available seed (Seedlot 20189; Australian Tree Seed Centre, CSIRO, Highett, Vic., Australia) in a mixture of 76% composted pine bark and 24% coarse potting sand with fertilizer added. Plants were ∼3 years of age and 1 m in height at the time of the experiment.
All plants were transferred to a growth cabinet and grown under controlled conditions for 3 months, allowing a new cohort of leaves to initiate and expand. It was assumed that any new growth would respond to the conditions experienced during its development (Schoch, Zinsou & Sibi 1980). Growth conditions were 16 h days at 25 °C/20 °C day/night temperatures, with a maximum quantum photosynthetic photon flux density (PPFD) of 1000 μmol m−2 s−1. Light was provided by a mixture of fluorescent and incandescent lights. A high VPD treatment was maintained at approximately 15–20% relative humidity (or 2.5–2.7 kPa VPD) during the day by a commercial dehumidifier (Secco Ultra; Olimpia-Splendid, Gualtieri, Italy) and a vaporizer (TAAV Vaporaire steam vaporizer; Xidex Investments Pty. Ltd, Silverwater, NSW, Australia) maintained a low VPD treatment at approximately 70–80% relative humidity (or 0.6–1 kPa VPD) during the day. Temperature and humidity were monitored for the duration of the experiment using a HOBO Pro RH/Temp logger (Onset Computer Corporation Inc., Bourne, MA, USA).
Leaf gas exchange
Two leaves from each plant were used to determine the response of leaf gas exchange to variation in VPD. A portable infrared gas analyzer (Li-6400; Li-Cor Biosciences, Lincoln, NE, USA) was used to measure stomatal conductance (mol m−2 s−1), photosynthetic rate (μmol m−2 s−1) and transpiration rate (mmol m−2 s−1) between 1000 and 1300 h, when gas exchange was expected to be maximal. All variables within the leaf chamber of the Li-6400 were standardized during measurements (leaf temperature at 22 °C, CO2 concentration between 380 and 390 μmol mol−1; PPFD at 1000 μmol m−2 s−1, and VPD at approximately 2.0 kPa for the high VPD treatment and 1.0 kPa for the low VPD treatment). Two leaves per plant grown under low VPD were also measured at 2.0 kPa to observe the dynamic response of stomata to high VPD. The midday water potential (−MPa) of all leaves was measured using a pressure chamber.
Leaf hydraulic conductance
The same 2 leaves per plant were used to determine leaf hydraulic conductance (mmol m−2 s−1 MPa−1) under high and low VPD. Leaves were measured between 1000 and 1300 h using the evaporative flux method (Sack et al. 2002; Brodribb & Holbrook 2006). Leaves were excised and then immediately re-cut under water and attached to a flow meter (Brodribb & Holbrook 2006) (for construction details, see http://prometheuswiki.publish.csiro.au/tiki-index.php?page=Constructing+and+operating+a+hydraulics+flow+meter). Leaves were then placed in conditions favourable to transpiration (i.e. under a light source providing a PPFD greater than 200 μmol m−2 s−1 and heated evenly by a stream of warm air maintaining leaf temperature between 25 and 30 °C). Leaves were allowed to reach a transpirational steady state (less than 10% variation over 180 s) and the resulting transpirational flux was recorded. Leaf water potential was measured with a pressure chamber and leaf hydraulic conductance calculated using the following equation:
where KL is the leaf hydraulic conductance, F is the transpirational flux and ΨL is the leaf water potential at steady state. Leaf size was measured using a flatbed scanner in combination with ImageJ (National Institutes of Health, Bethesda, MD, USA). Leaf hydraulic conductance was normalized to leaf size and the viscosity of water at 20 °C using an empirical function based on data from Korson, Drost-Hansen & Millero (1969).
Vein density (total vein length per mm2 of leaf area) was determined from paradermal sections of each of the leaves used in gas exchange and leaf hydraulic conductance measurements. Sections were prepared and measured following the protocols described by Carins Murphy et al. (2012). In brief, this involved removing the adaxial epidermis and palisade tissue, clearing all pigment with bleach, and measuring vein density from slide mounts of these sections (2 sections per leaf) using image analysis of digital photomicrographs (5 fields of view per section). Stomatal density (total stomata per mm2 of leaf area), stomatal size (mm2) and epidermal cell size (mm2) were also determined from cuticles (2 per leaf and 5 fields of view per cuticle) prepared and measured following the protocols of Carins Murphy et al. (2012). Total vein length, number of stomata and number of epidermal cells per leaf were quantified by multiplying the density of veins, stomata and epidermal cells by leaf size.
The plasticity of leaf size, anatomical traits (vein density, stomatal density, epidermal cell size, stomatal size and total number of epidermal cells per leaf) and physiological traits (photosynthetic rate, stomatal conductance, transpiration rate, leaf hydraulic conductance and midday leaf water potential) were assessed by comparing the relative changes between plants grown under high and low VPD with unpaired t-tests. Dynamic changes to the stomatal conductance of plants grown under low VPD in response to short-term exposure to high VPD were also assessed by comparing initial and final stomatal conductance using a paired t-test and by comparing final stomatal conductance with that of plants grown under high VPD using an unpaired t-test.
To test whether variation in vein and stomatal density between plants was passively determined by leaf size, vein and stomatal density were plotted against 1/√leaf area and 1/leaf area, respectively. The coordination between vein and stomatal density and leaf area traits was then quantified as the deviation from a proportional relationship (i.e. one in which vein and stomatal density increase uniformly with leaf expansion). The relationships between vein density and stomatal density and between anatomical traits (total epidermal cells per leaf, total stomata per leaf and total vein length per leaf) and leaf size were tested for proportionality in the same way.
ancova analysis carried out using the R programming environment (R Core Team 2012) was used to test if the slope of the relationship between vein and stomatal density was the same regardless of whether changes were induced by VPD or by irradiance. R2 was also calculated for both relationships.
The relationship between transpiration and leaf hydraulic conductance under high and low VPD was plotted against the same relationship in plants grown under high and low irradiance to test whether transpiration was maintained at rates appropriate for the hydraulic capacity of the vascular system. R2 was calculated for the relationship between plants grown under contrasting irradiance.
Response of leaf size and anatomy to variation in VPD
In T. ciliata, VPD had a substantial effect on leaf size and a small but significant effect on vein and stomatal density (Fig. 1). Leaves from plants grown under high VPD were about one-third (38%) the size of those grown under low VPD (P < 0.001). However, leaves grown under high VPD had only 12% more densely packed veins and 33% more densely packed stomata (P < 0.05 in both cases) than those grown under low VPD. This response was considerably less than the proportional response expected if vein and stomatal density were controlled by the ‘dilution’ effect of leaf area (P < 0.001 and P < 0.01) (Fig. 2).
Despite this, vein and stomatal density remained almost proportional across high and low VPD, although plants grown under low VPD had 20% higher vein density than would be expected if vein and stomatal density were directly proportional (P < 0.05). Furthermore, the relationships between vein and stomatal density produced by changing VPD (this study) or irradiance (previous study) were not significantly different (P > 0.05) (irradiance data from Carins Murphy et al. (2012) (Fig. 3).
Epidermal cell size did not vary significantly between plants grown under high and low VPD (P > 0.05), although the stomata of plants grown under low VPD were slightly (14%) larger than those of plants grown under high VPD (P < 0.01; Table 1). In contrast to this lack of effect on cell size, the total number of epidermal cells per leaf varied more than twofold between treatments so that the total number of epidermal cells per leaf was proportional to leaf size (Fig. 4). The total number of stomata and length of veins per leaf were also related to leaf size, although these relationships were not directly proportional, as the total vein length per leaf in plants grown under low VPD was 9% less than expected if directly proportional to leaf area, and the total number of stomata was 24% less than would be expected if directly proportional to leaf area (P < 0.05 in both cases, data not shown).
Table 1. Comparison of anatomical and physiological traits between Toona ciliata plants grown under high and low vapour pressure difference (VPD) (values are means ± SE)
Means of the two treatment groups were compared using unpaired t-tests. P-values <0.05 are shown in bold.
ECA, epidermal cell area; SA, stomatal area; gs, stomatal conductance; KL, leaf hydraulic conductance; A, photosynthetic rate; E, transpiration rate; ΨL, midday leaf water potential.
2.19 × 10−4 ± 1.71 × 10−5
2.66 × 10−4 ± 2.19 × 10−5
1.29 × 10−4 ± 3.00 × 10−6
1.47 × 10−4 ± 2.69 × 10−6
mol m−2 s−1
0.09 ± 0.01
0.27 ± 0.03
mmol m−2 s−1 MPa−1
10.15 ± 0.39
11.07 ± 0.94
μmol m−2 s−1
8.47 ± 0.43
7.97 ± 0.61
mmol m−2 s−1
1.65 ± 0.28
1.87 ± 0.12
0.82 ± 0.09
0.79 ± 0.02
Response of gas exchange, leaf hydraulic conductance and midday leaf water potential to high and low VPD
VPD had a large and significant effect on stomatal conductance. Plants grown under low VPD had 215% higher stomatal conductance than those grown under high VPD (P < 0.001) (Table 1). However, variation in VPD did not induce a significant response in photosynthesis, transpiration, leaf hydraulic conductance or midday leaf water potential when measured under growth conditions (P > 0.05; Table 1) and transpiration was within the range of what would be predicted from the relationship between transpiration and leaf hydraulic conductance for this species under contrasting irradiance (Carins Murphy et al. 2012) (Fig. 5). Short-term exposure of plants grown under low VPD to high VPD conditions in the Li-Cor leaf chamber induced a 78% reduction in stomatal conductance (P < 0.01), so that the final stomatal conductance was not significantly different from that of plants grown under high VPD (P > 0.05).
Modification of stomatal conductance upholds transpirational homeostasis under high and low VPD
It was hypothesized that T. ciliata would accommodate the increased transpiration from stomatal pores at high VPD by changing the relationship between vein and stomatal density in the leaf. Contrary to this prediction, any change in this relationship was small, non-significant and in the opposite of the expected direction. Instead, we observed that the reduction in stomatal conductance necessary to maintain homeostasis was produced mainly by dynamic closure of stomata rather than a reduction in density. The small increase in stomatal density at high VPD was coordinated with a similarly small increase in vein density. Thus, T. ciliata appears to be incapable of sufficient flexibility in the rates of initiation of veins and stomata to compensate for the different transpirational demands under high and low VPD. Instead, this species adjusts stomatal conductance to maintain transpiration (and leaf water potential) at a static level that is constrained by the density of veins (hydraulic supply efficiency). The effectiveness of dynamic stomatal closure in response to VPD is apparent from the lack of variation in transpiration, photosynthesis and midday leaf water potential between T. ciliata plants grown under high and low VPD. However, the capacity of T. ciliata to modify the relationship between vein and stomatal density would be further clarified by investigating the developmental response to low irradiance under contrasting VPD.
Vein and stomatal allocation is not optimized under high and low VPD
The relationship between vein and stomata density under different VPD levels in T. ciliata was similar to that observed across different irradiance levels (Fig. 3). Plants grown under the same VPD but different levels of irradiance adjust vein and stomatal density so that water supply and transpirational demand remain proportional (Carins Murphy et al. 2012). However, if plants grown under the same light intensity but different VPD used the same strategy, water loss would increase in proportion to water transport capacity in those plants grown under high VPD and stomata would close in response to declining leaf water potential. If it is assumed that the most efficient stomatal density occurs when stomata are able to open to near maximum aperture in fully hydrated soil (Franks et al. 2012), then the only way to optimally adjust leaf anatomy to variation in VPD is to alter the coordination between vein and stomatal density. Apparently, T. ciliata is not capable of making sufficient independent modification of vein and stomatal traits to compensate for large differences in VPD, meaning that leaves grown under high VPD are forced to operate with stomata mostly closed.
As with plants grown under high and low irradiance, changes in VPD induced large differences in leaf size, so that leaves were much larger in plants grown under low VPD. Changes in vein and stomatal density, however, were largely independent of leaf size. This is in contrast to the veins and stomata of plants grown under high and low irradiance that are passively ‘diluted’ or ‘concentrated’ by differential leaf expansion (Carins Murphy et al. 2012). Despite this, veins and stomata were still proportional across high and low VPD treatments. This suggests that the coordination of vein and stomatal density was not maintained by passive ‘dilution’ during leaf expansion but by coordinated initiation and differentiation of vein and stomatal cells.
The large effects of VPD on the number of epidermal cells and minimal effect on cell size resulted in the total number of epidermal cells per leaf being proportional to leaf size. Thus, leaves expanded under high VPD were smaller mainly due to reductions in initiation rather than expansion of the epidermal cells. This contrasts with the effects of irradiance on leaf size, which was mainly the result of differential epidermal cell expansion and effectively regulated the density of veins and stomata in concert (Carins Murphy et al. 2012). However, the differences in vein and stomatal density between plants grown under high and low VPD were much less than would be predicted from the large differences in leaf size. This indicates profound differences in the way that leaves adapt hydraulic and stomatal configuration in response to VPD and irradiance (Fig. 6).
Response of vein density, stomatal density and leaf size to VPD
The few studies that investigate the intraspecific response of vein and stomatal density to variation in VPD provide conflicting results. There is disagreement about the direction of the response and there is no discussion of the coordination (or lack of coordination) of veins and stomata under varying VPD. There are, however, fairly well-documented global trends across species of increasing leaf size with mean annual temperature and rainfall (conditions congruent with decreasing VPD) (Webb 1968; Wilf et al. 1998). Glasshouse trials also report increased growth rates and leaf size within species with decreasing VPD (Mortensen 1986; Mortensen 2000; Hovenden et al. 2012). As vein and stomatal density have been observed to increase with decreasing leaf size in some cases (specifically when smaller leaf sizes are induced by high irradiance) (Carins Murphy et al. 2012), it might be expected that greater VPD may also increase vein and stomatal density. However, in this study anatomical traits were independent of leaf size in T. ciliata grown under high and low VPD with only a weak trend towards higher vein and stomatal density in the smaller leaves produced under high VPD. This suggests that under varying VPD (but constant irradiance), vein and stomatal density are regulated by fixed ratios of cell production between veins and stomata, and that the plant modifies the number of cells in the leaf to produce large differences in leaf size. This is supported by evidence that differences in leaf size between Nothofagus cunninghamii plants grown under high and low humidity are related to the number of epidermal cells rather than the size of epidermal cells (Hovenden et al. 2012). Furthermore, in a separate study encompassing almost 500 dicotyledonous species, total vein density was found to be independent of leaf area (Sack et al. 2012), while another study found that variation in vein and stomatal density between leaves of the temperate rainforest tree N. cunninghamii grown under high and low irradiance was only partially explained by differences in leaf size (Brodribb & Jordan 2011). Thus, active determination of vein and stomatal density by an environmental stimulus such as VPD rather than a passive response to leaf size may account for the independence of vein and stomatal density from leaf size observed in this and other studies.
Implications for vein and stomatal development
Epidermal cell size was not significantly affected by VPD in T. ciliata. However, the total number of epidermal cells per leaf was greater in plants grown under low VPD. Increases in epidermal cell size are associated with water movement into growing cells (Cosgrove 1986), such that epidermal cell expansion in plants experiencing high transpiration rates is limited by competition for water (Fricke 2002). Epidermal cell division, on the other hand, is limited by photosynthetic input (Pantin, Simonneau & Muller 2012). As epidermal cell size was not affected by VPD in T. ciliata, this implies that the water available for cell expansion was consistent across treatments. Thus, the homeostasis in transpiration observed in T. ciliata also brought about homeostasis in epidermal cell expansion. The total number of epidermal cells per leaf, however, varied with VPD so that plants grown under low VPD had more epidermal cells than those grown under high VPD. This difference suggests that epidermal cell initiation continued for longer in low VPD plants. There are several potential explanations for this: plants grown under low VPD may have been able to maintain higher net photosynthesis during the day (here, we only measured maximum assimilation in the morning), or it is possible that these plants invested less energy into roots in favour of leaf growth (Gislerød & Nelson 1989; Mortensen, Ottosen & Gislerød 2001). They may also have experienced lower nocturnal transpiration rates, extending the time period during which cell division was possible.
Thus, differences in leaf size between plants grown under high and low VPD were the result of variation in epidermal cell division (not epidermal cell expansion) with plants grown under low VPD initiating more cells. This accounts for the disassociation of vein and stomatal density from leaf size. As veins and stomata are initiated during the early stages of leaf development before leaf expansion is completed (Pantin et al. 2012), increased initiation of epidermal cells would also result in increased initiation of veins and stomata (assuming a constant relationship between veins and stomata). The small differences in vein and stomatal density that were observed between plants grown under high and low VPD may have been the result of a small (but not significant) variation in epidermal cell size.
Vein and stomatal density showed small but coordinated responses to growth under high and low VPD in T. ciliata, contradicting our hypothesis that the balance between vein and stomatal density would be modified to uphold transpirational homeostasis. Furthermore, the direction of this response was opposite to what was predicted (i.e. stomatal density increased with VPD). Instead, substantial changes in stomatal conductance maintained transpirational homeostasis in plants grown under contrasting VPD treatments. Thus, T. ciliata resorted to dynamic stomatal control rather than disruption of a fixed relationship between vein and stomatal tissue production. Consequently, the allocation of veins and stomata during leaf development must be tightly linked. Further research is needed to understand the developmental basis for this coordination.
We thank Michelle Lang and Tracey Winterbottom for glasshouse maintenance and Hugo de Boer for helpful suggestions regarding Fig. 6. This research was funded by an Australian Research Council grant to T.J.B. (DP120101868).