Surviving floods: leaf gas films improve O2 and CO2 exchange, root aeration, and growth of completely submerged rice


  • Ole Pedersen,

    Corresponding author
    1. School of Plant Biology, Faculty of Natural and Agricultural Sciences, The University of Western Australia, 35 Stirling Highway, Crawley, WA 6009, Australia, and
    2. Freshwater Biological Laboratory, Institute of Biology, University of Copenhagen, Helsingørsgade 51, DK-3400 Hillerød, Denmark
    Search for more papers by this author
  • Sarah Meghan Rich,

    1. School of Plant Biology, Faculty of Natural and Agricultural Sciences, The University of Western Australia, 35 Stirling Highway, Crawley, WA 6009, Australia, and
    Search for more papers by this author
  • Timothy David Colmer

    1. School of Plant Biology, Faculty of Natural and Agricultural Sciences, The University of Western Australia, 35 Stirling Highway, Crawley, WA 6009, Australia, and
    Search for more papers by this author

(fax +45 3232 1901; e-mail


When completely submerged, the leaves of some species retain a surface gas film. Leaf gas films on submerged plants have recently been termed ‘plant plastrons’, analogous with the plastrons of aquatic insects. In aquatic insects, surface gas layers (i.e. plastrons) enlarge the gas–water interface to promote O2 uptake when under water; however, the function of leaf gas films has rarely been considered. The present study demonstrates that gas films on leaves of completely submerged rice facilitate entry of O2 from floodwaters when in darkness and CO2 entry when in light. O2 microprofiles showed that the improved gas exchange was not caused by differences in diffusive boundary layers adjacent to submerged leaves with or without gas films; instead, reduced resistance to gas exchange was probably due to the enlarged water–gas interface (cf. aquatic insects). When gas films were removed artificially, underwater net photosynthesis declined to only 20% of the rate with gas films present, such that, after 7 days of complete submergence, tissue sugar levels declined, and both shoot and root growth were reduced. Internal aeration of roots in anoxic medium, when shoots were in aerobic floodwater in darkness or when in light, was improved considerably when leaf gas films were present. Thus, leaf gas films contribute to the submergence tolerance of rice, in addition to those traits already recognized, such as the shoot-elongation response, aerenchyma and metabolic adjustments to O2 deficiency and oxidative stress.


Flooding has an adverse impact on plants in many ecosystems worldwide (Jackson, 2004). Submergence tolerance can be achieved by two contrasting strategies, the ‘quiescence response’ or the ‘escape response’; both strategies involve a number of traits (Bailey-Serres and Voesenek, 2008). The present study demonstrates the importance of leaf gas films in the submergence tolerance of rice; a feature, in addition to those already identified by Bailey-Serres and Voesenek (2008), that contributes to submergence tolerance in some wetland plants.

Water-repellent (i.e. hydrophobic) cuticles are common in insects and plants (Neinhuis and Barthlott, 1997; Vogel, 2006), and some of these surfaces retain a microlayer of gas when submerged. In aquatic insects, the gas layer (plastron) enables continued respiration underwater because of the enlarged water–gas interface between the tracheary system and surrounding water (Thorpe and Crisp, 1949; Hebets and Chapman, 2000; Vogel, 2006). Leaf gas films also improve underwater O2 and CO2 exchange (Raskin and Kende, 1983; Colmer and Pedersen, 2008a), and have recently been termed ‘plant plastrons’ (Raven, 2008). In plants, gas films on submerged leaves also enlarge the water–gas interface (cf. plastrons of insects, Hebets and Chapman, 2000), which otherwise would be restricted to localized areas adjacent to stomata. Moreover, as stomata are hypothesized to close upon submergence of leaves without gas films (Mommer and Visser, 2005), if gas films enable stomata to remain open, at least during light periods, uptake of CO2 via stomata (suggested by Colmer and Pedersen, 2008a) would bypass the cuticle resistance that is a major impediment to CO2 entry in submerged leaves that do not possess gas films (Mommer et al., 2004). Improved gas exchange by submerged leaves can enhance CO2 uptake for underwater photosynthesis during light periods and O2 entry during dark periods, with benefits to whole-plant internal aeration (Sand-Jensen et al., 2005; Pedersen et al., 2006).

Gas films increased underwater net photosynthesis by 1.5–6-fold in leaves of four wetland species (at 50 mmol m−3 CO2; Colmer and Pedersen, 2008a). In rice, relative 14C incorporation was 9–10-fold greater in leaves with gas films present than when these were removed (Raskin and Kende, 1983). Underwater photosynthesis can enhance the tolerance of plants to complete submergence (Mommer and Visser, 2005), as it provides O2 for internal aeration (Waters et al., 1989; Colmer and Pedersen, 2008b) and photosynthates (Setter et al., 1989) for respiration/fermentation and growth. The rice cultivars that are most tolerant of complete submergence during flash floods maintain higher tissue sugar concentrations than less tolerant cultivars, and this is attributed to a combination of higher initial carbohydrate concentrations and slower consumption rates during submergence (Ram et al., 2002; Jackson and Ram, 2003); the slower sugar consumption largely results from lack of a shoot-elongation response (Setter and Laureles, 1996; Jackson and Ram, 2003; Xu et al., 2006). In addition to differences in sugar consumption, underwater photosynthesis can also influence the survival of submerged rice (Setter et al., 1989; Ram et al., 2002), although if the floodwaters are turbid, the importance of underwater photosynthesis is diminished.

The present study demonstrates that gas films on leaves of completely submerged rice greatly improve entry of O2 from floodwaters when in darkness, and of CO2 for photosynthesis when in light. Root aeration was improved both in darkness and when shoots were in light, and tissue sugar levels and growth of submerged rice were also improved. These findings demonstrate the beneficial role of leaf gas films for rice when completely submerged, such as occurs during flash-flooding of lowland rice, a condition that is very different to that studied previously (Raskin and Kende, 1983; Beckett et al., 1988) with gas films along partially submerged leaves of deepwater rice.


Gas films enhance the internal O2 status of submerged rice in darkness and in light

Gas films can facilitate O2 and CO2 exchange of completely submerged leaves, with benefits for leaf respiration and net photosynthesis (Colmer and Pedersen, 2008a). However, the influence on whole-plant aeration had not been evaluated.

Rice plants were transferred into a two-chamber system, so that the roots were in stagnant de-oxygenated 0.1% agar and shoots were initially in air. The shoot was then submerged in medium with 200 mmol m−3 CO2 and 236 mmol m−3 O2 (ionic composition given in Experimental procedures). The O2 level was air saturation, whereas the CO2 level was 12.5 times the air saturation, as floodwaters in rice-growing areas of Asia such as Thailand (Setter et al., 1987) and India (Ram et al., 1999) are often supersaturated with CO2. CO2 accumulates in flooded soils (Ponnamperuma, 1984; Greenway et al., 2006), and the higher solubility of CO2 compared with O2 results in a high level of dissolved CO2 even when the floodwater O2 level remains close to air saturation. When submerged, gas films were evident along all shoot tissues, presumably because the surfaces are hydrophobic; the contact angles (°) of water droplets placed on horizontally held leaf blades, determined as described by Neinhuis and Barthlott (1997), were 148 ± 2 (adaxial) and 149 ± 1 (abaxial). Using O2 microelectrodes, the pO2 dynamics within adventitious roots were measured while shoot conditions were manipulated. Recordings are shown for plants with shoots in light (Figure 1a) and in darkness (Figure 1b), and the quasi steady-state values for replicated experiments are given in Table 1.

Figure 1.

 Dynamics of O2 partial pressures (pO2) in adventitious roots of 4-week-old rice (Oryza sativa L.) in light (a) or darkness (b), before and after submergence, and with and without removal of gas films (lamina and sheath exteriors), at 30°C.
O2 microelectrodes were inserted 20–25 mm behind the tip of intact adventitious roots of plants with roots in deoxygenated 0.1% agar medium. Quasi-steady state pO2 in the root was established with the shoot in air, then submerged (200 mmol m−3 CO2), then the shoot was gently lifted out of the water and brushed with 0.1% Triton X-100 so that gas films were absent when re-submerged, and finally de-submerged so that the shoot was again in air. The small peak in pO2 in the trace taken in darkness (b) during gas film removal was due to entry of atmospheric O2 during brushing. Tissue porosities (%) were: leaf blades, 19.2 ± 1.2; sheaths, 30.7 ± 2.9; adventitious roots, 26.7 ± 3.4. Means of replicated measurements of pO2 taken from different plants are given in Table 1.

Table 1.   Influence of leaf gas films on pO2 in roots of rice (Oryza sativa L.) when submerged with shoots in light or darkness
TreatmentRoot pO2 (kPa)
  1. Four-week-old plants were mounted in a two-compartment chamber with the roots in deoxygenated 0.1% agar medium and the shoot in air and then in artificial floodwater with 200 mmol m−3 CO2. An O2 microelectrode was inserted into the cortex of an adventitious root (8–12 cm long), 20–25 mm behind the root tip. Root pO2 was recorded with the shoot in light or in darkness (both at 30°C). Values are means ± SE (= 4–7, each replicate was a different plant). Different letters indicate significant differences (Tukey test, < 0.05).

Shoot in light
 Shoot in air (= 7)11.8a,c ± 0.2
 Shoot completely submerged (= 4)13.5b ± 0.5
 Gas films removed, shoot submerged (= 4)10.5a ± 0.1
 Shoot de-submerged (= 4)12.8b,c ± 0.2
Shoot in darkness
 Shoot in air (= 4)9.9a,c ± 0.5
 Shoot completely submerged (= 7)3.4d ± 0.8
 Gas films removed, shoot submerged (= 7)0.1e ± 0.04
 Shoot de-submerged (= 4)10.8a,b,c ± 0.9

For plants with shoots in light and initially in air, root pO2 was approximately 12 kPa (Figure 1a and Table 1); as expected, the pO2 was lower than atmospheric partial pressure (approximately 20.6 kPa) as O2 will have been consumed along the diffusion path within the aerenchymatous roots (see Armstrong, 1979). Upon submergence of the shoots in light, root pO2 increased (Figure 1a and Table 1), presumably reflecting an increase in shoot pO2 due to impeded outward diffusion of photosynthetically produced O2 when tissues are surrounded by water (Colmer and Pedersen, 2008b). Removal of leaf gas films resulted in a decline in root pO2 to just below the initial level (Figure 1a and Table 1), as O2 production via photosynthesis declined due to impeded entry of CO2 (see section on underwater photosynthesis). De-submergence of the shoot resulted in a rapid return of root pO2 to near the original levels prior to shoot submergence (Figure 1a and Table 1), indicating that removal of leaf gas films with Triton X-100 had no irreversible effect on shoot gas exchange when in air.

For plants submerged in darkness (Figure 1b), root pO2 declined to about 25% of that when submerged in light (Table 1). Removal of leaf gas films caused the root pO2 to decline to very low levels (approximately 0.1 kPa) when shoots were in darkness (Figure 1b and Table 1), as diffusion of O2 from the surrounding floodwater into the leaves is impeded. De-submergence of the shoot resulted in a rapid increase in root pO2 to the level prior to submergence (Figure 1b and Table 1).

Gas films: volume, dimensions and diffusive boundary layers

When submerged, O2 and CO2 exchange between leaves and the surrounding water is impeded due to a 104-fold slower diffusion of gases in water compared with air (Armstrong, 1979; Vogel, 2006). Diffusive boundary layers, therefore, are a major component of the resistance to uptake of dissolved gases, and so were quantified for O2 uptake by leaf segments with or without gas films (Figure 2 and Table 2).

Figure 2.

 Examples of diffusive boundary layers adjacent to completely submerged leaf segments of rice (Oryza sativa L.): lamina segments 60 mm long and 8–10 mm wide either with gas films (a) or without gas films (0.1% Triton X-100-treated) (b), in darkness in a flume with mean bulk flow velocity of 15 mm sec−1, at 30°.
Profiles were measured using an O2 microelectrode (tip diameter 10 μm). Dimensions of the true, and effective, diffusive boundary layers were estimated according to Jørgensen and Revsbech (1985), and are indicated in (a) but not in (b) as the extension of the two zones is roughly the same in the example shown in (b). Means of replicate values taken from leaf segments from different plants are given in Table 2.

Table 2.   Leaf gas film volume and dimensions, surface pO2, O2 consumption rates, leaf porosity, and boundary layer condition of rice (Oryza sativa L.) leaf segments with or without gas films at 30°C
ParameterMean ± SE (n)
  1. Values are means ± SE with the number of replicates in parentheses (each replicate comprised a leaf segment from different plants). For explanations of true and effective diffusive boundary layers, and the transition zone, see the legend to Figure 2. P values refer to comparisons between with or without gas films, within each parameter measured.

True diffusive boundary layer (μm) (< 0.05)
 With gas film266 ± 38 (6)
 Without gas film124 ± 16 (4)
Effective diffusive boundary layer (μm) (< 0.05)
 With gas film360 ± 35 (6)
 Without gas film156 ± 17 (4)
Transition zone between turbulent flow and viscous laminar flow (μm) (< 0.05)
 With gas film373 ± 53 (6)
 Without gas film198 ± 42 (4)
Leaf surface pO2 in darkness (kPa) (< 0.05)
 With gas films18.2 ± 0.1 (6)
 Without gas films13.9 ± 0.1 (4)
Leaf O2 consumption (μmol O2 m−2 sec−1) calculated from O2 profiles (< 0.05)
 With gas films0.21 ± 0.01 (6)
 Without gas films0.60 ± 0.21 (4)
Leaf O2 consumption (μmol O2 m−2 sec−1) determined using respiration chambers (> 0.05)
 With gas films0.55 ± 0.01 (8)
 Without gas films0.63 ± 0.04 (8)
Leaf gas film thickness (μm) (> 0.05) determined by the buoyancy method
 Adaxial67 ± 4 (4)
 Abaxial57 ± 12 (4)
 Average62 ± 4 (8)
Leaf gas film thickness range (μm) determined by the microelectrode methodAdaxial<10–140 (6)
Leaf porosity (% gas volume per unit tissue volume) 19.2 ± 1.2 (4)
Gas film volume relative to internal leaf gas volume 3.8 ± 0.4 (4)

Figure 2 shows typical examples of diffusive boundary layers, measured using O2 micro electrode profiling, adjacent to leaf segments with (Figure 2a) or without (Figure 2b) gas films when in darkness and with air-saturated medium flowing across the segments fixed in a flume tank. The diffusive boundary layers in the liquid phase (both ‘true’ and ‘effective’; see Figure 2) were thicker on leaf segments with gas films compared to those without gas films (examples in Figure 2; means in Table 2). Furthermore, the transition zone between turbulent and laminar flows was also wider on leaf segments with gas films (Figure 2 and Table 2). We tested whether these effects could be attributed to Triton X-100 per se by conducting measurements on the aquatic leaves of Haloragis brownii (no natural gas film) with or without pre-brushing with Triton X-100; the Triton X-100 treatment had no effect on the diffusive boundary layers (control 118 ± 2 μm; Triton X-100-treated 116 ± 2 μm; mean ± SE, = 5). The results for rice leaves demonstrate that the better aeration in plants with gas films was not related to lower liquid-phase diffusive boundary layer resistances.

The O2 concentration gradient within the true diffusive boundary layer can be used to calculate net O2 fluxes. The gradient at the position measured was larger for leaves without gas films, implying that these had faster rates of net O2 uptake. However, when the actual O2 consumption by whole leaf segments was determined by depletion in well-mixed closed chambers, the rates for leaf segments with or without gas films at atmospheric saturation did not differ (pooled mean = 0.6 μmol m−2 sec−1, Table 2). The net fluxes calculated from the O2 profiling in the flume system for leaf segments without gas films were also 0.6 μmol m−2 sec−1, whereas for those with gas films it was three times lower (Table 2). We hypothesize that the discrepancy between the two techniques for leaves with gas films could result from position effects. With gas films present, longitudinal O2 diffusion would be rapid, so if O2 uptake was enhanced at the leading edge of the leaf segment, due to eroded diffusive boundary layers at that edge, then uptake at the position measured (midpoint of the 60 mm leaf segment) would be lowered, as O2 supply to tissue at this point would also be via longitudinal O2 diffusion within the gas layer. This possibility is supported, assuming equal tissue O2 demands within the two treatments (cf. respiration chamber results), by the finding of a higher pO2 of 18.2 kPa at the surface of leaves with gas films, compared with 13.9 kPa on those without (Table 2).

The O2 profiles were also used to estimate the thickness of gas films on the adaxial side of leaf segments, as high diffusivity in air means O2 concentration gradients are not established over short distances (see vertical portion of the profile immediately adjacent to the leaf surface in Figure 2a), whereas the diffusive boundary layer profiles were clearly evident in water (Figure 2a,b). Gas film thickness varied from <10 to 140 μm (mean 89 ± 23 μm; Table 2), a mean value that is not significantly different from that obtained using a buoyancy method, on segments before and after gas film removal, to determine gas film volumes per unit area, from which the mean gas film thickness was calculated (62 ± 4 μm; Table 2). The leaf blade thickness was 156 ± 16 μm (mean ± SE, n = 12), and so the gas film volume (adaxial plus abaxial sides) was roughly equal to the tissue volume (Table 2). Consequently, the volume of gas within the films was almost four times higher than the internal leaf tissue gas volume (Table 2).

Gas films enhance underwater net photosynthesis, tissue sugar levels and growth of completely submerged rice

The experiments described above established that leaf gas films enhance the internal O2 status of submerged rice. In the experiments described in this section, we evaluated whether gas films are also beneficial to underwater net photosynthesis, tissue sugar status, and ultimately plant growth when completely submerged.

Gas films enhanced underwater net photosynthesis in leaf segments of rice across a wide range of external CO2 concentrations (Figure 3). At low-to-medium CO2 availability (15–180 mmol m−3), underwater net photosynthesis was 4–4.9-fold higher in leaf segments with gas films, whereas the rates in leaf segments with or without gas films were equal at high CO2 supply (2000 mmol m−3) (Figure 3). The low-to-medium CO2 concentrations are of relevance to some field conditions such as in Thailand (Setter et al., 1987) or India (Ram et al., 1999), whereas the high CO2 levels were used to investigate whether the rates for leaf segments with or without gas films converge. The apparent resistance to CO2 entry was five times less in leaf segments with gas films (legend to of Figure 3), such that underwater net photosynthesis was enhanced at lower CO2 availability. By increasing the external CO2 supply, however, the higher apparent resistance to CO2 entry in leaf segments without gas films could be overcome (Figure 3). Nevertheless, the maximum rate of 4 μmol m−2 sec−1 under water was below that of leaves in air (11.3 ± 0.3 μmol m−2 sec−1 with 380 μl L−1 CO2 and 17.8 ± 0.7 μmol m−2 sec−1 with 760 μl L−1 CO2; PAR 350 μmol m−2 sec−1; mean ± SE, = 4).

Figure 3.

 Underwater net photosynthesis in leaf segments from 4-week-old rice (Oryza sativa L.) either with gas films or without gas films (0.1% Triton X-100-treated), as a function of CO2 in the medium. Lamina segments (30 mm) were incubated in glass cuvettes attached to a rotating wheel within a water bath at 30°C, and net photosynthesis was measured as O2 evolution over 60–90 min. The data were fitted to the equation of Jassby and Platt (1976), with Pmax estimated to be 4.00 ± 0.08 μmol m−2 sec−1 (r2 = 0.93) with gas films and 4.66 ± 0.30 μmol m−2 sec−1 (r2 = 0.97) without gas films, the two means not being significantly different. The apparent resistance to CO2 uptake was fivefold lower with gas films present (57 850 sec m−1) than without (274 375 sec m−1). The apparent resistance to CO2 uptake was determined as 1 divided by the slope of a linear regression for the 15–180 mmol m−3 range. Values are means ± SE (= 5, each replicate being a leaf segment from a different plant).

The greater rates of underwater net photosynthesis by leaves with gas films (Figure 3) not only enhanced the internal O2 status of completely submerged rice (Figure 1 and Table 1), but also enhanced tissue sugar concentrations (Figure 4a) and growth (Figure 4c,d). After 7 days of complete submergence, the total sugar levels in leaves of plants without gas films had declined to 66% of the level in those with gas films; moreover, total sugar levels in submerged plants with gas films did not differ from levels when the shoot was in air (Figure 4a). In roots, however, although total sugar levels in plants without gas films were 63% of those with gas films (Figure 4b), the variation was large and so this difference was not significant at the 5% level. The growth of submerged plants without gas films was impeded; the shoot dry mass was 49% of the values for submerged plants with gas films, and the root dry mass was 33%. The beneficial effect of gas films during complete submergence is further demonstrated by the submerged plants with gas films not differing with respect to shoot growth from the controls with shoots in air (Figure 4c), although root growth was reduced in comparison with controls (Figure 4d).

Figure 4.

 Total sugar levels in (a) leaves (lamina only) and (b) roots, and dry mass of (c) shoots and (d) roots of rice (Oryza sativa L.) with shoots in air (with or without 0.1% Triton X-100 treatment) or when completely submerged with or without gas films.
Fourteen-day-old rice with roots in containers of stagnant, de-oxygenated 0.1% agar with nutrient solution were either kept for an additional 7 days with shoots in air within transparent Perspex tanks (shoot in air), brushed with 0.1% Triton X-100, rinsed and retained in air (shoot in air control), completely submerged in a medium with 200 mmol m−3 CO2 (sub with gas films), or leaves and sheaths were brushed with 0.1% Triton X-100 to prevent gas films when completely submerged (sub without gas films). Samples were also taken when treatments were imposed (initial). Gas films on submerged leaves had positive effects on growth (Tukey test, < 0.05, indicated by different letters). Assuming exponential growth, whole plant relative growth rates were: shoot in air = 0.184 day−1; shoot in air control = 0.155 day−1; sub with gas films = 0.181 day−1; sub without gas films = 0.071 day−1. Gas films on submerged leaves had positive effects on total sugar concentration in leaves (Tukey test, < 0.05, indicated by different letters), but differences were not significant for roots. Values are means ± SE (= 4, each replicate being a different plant).

For plants with shoots in air, those treated with Triton X-100 showed a decline in root growth (Figure 4d), but not in shoot growth (Figure 4c). A second experiment established that net photosynthesis in air following Triton X-100 treatment and rinsing with submergence medium was initially depressed to 60% of that in untreated plants, and the rate recovered by 60 min (data not shown). Plants without Triton X-100 treatment were also rinsed; the solution was instantly shed from these hydrophobic leaf surfaces, and so net photosynthesis in these plants did not differ before and after rinsing (data not shown). The inhibitory effect of surface wetness on leaf gas exchange, as probably occurred in the Triton X-100-treated plants, has also been reported for terrestrial species with naturally low leaf water repellency (Smith and McClean, 1989). Thus, a transient decline in photosynthesis is the likely cause of the lower growth in ‘shoot in air controls’ (i.e. in air with Triton X-100 treatment) in comparison with plants in air without Triton X-100 treatment. This effect, however, would not have influenced the submerged plants, as Triton X-100 per se did not influence boundary conditions when under water (see results in preceding section for the check of Triton X-100 effects on aquatic leaves of Haloragis brownii).


This study demonstrates that leaf gas films facilitate underwater net photosynthesis, with resultant improvements in tissue pO2, sugar levels and growth of rice when completely submerged. Enhanced growth during submergence, due to leaf gas films, was evident for both shoots and roots (Figure 4), and the higher pO2 in roots (Figure 1 and Table 1) would also have promoted radial O2 loss from root tips, which is of importance for growth into anaerobic substrates (see Armstrong, 1979; Colmer, 2003). The beneficial effect of gas films on submerged leaves (present study) reinforces the view that underwater photosynthesis is a key determinant of the survival and growth of plants during complete submergence (Mommer and Visser, 2005; Bailey-Serres and Voesenek, 2008).

Underwater photosynthesis is a major source of O2 for submerged wetland plants (e.g. rice; Colmer and Pedersen, 2008b; Waters et al., 1989), and also provides sugars. For rice that had been completely submerged for 7 days, underwater net photosynthesis was enhanced by leaf gas films, and this enabled maintenance of tissue sugar levels and growth (Figure 4). The importance of underwater photosynthesis for maintenance of sugar levels and growth of completely submerged rice has also been demonstrated by CO2 enrichment of the submergence water (Setter et al., 1989). The present finding of the benefit of leaf gas films for the whole-plant carbon balance of submerged rice extends an earlier prediction of improved leaf-level carbon balance of submerged Phragmites australis when gas films were present (Colmer and Pedersen, 2008a) and an earlier observation of enhanced 14CO2 uptake by submerged rice leaf segments with gas films present (Raskin and Kende, 1983). Even with gas films present, however, the maximum rate of net photosynthesis under water was approximately 40% of that in air with ambient CO2, and 22% of that when high CO2 levels were supplied in air. Nevertheless, underwater photosynthesis facilitated by leaf gas films (present study), as well as the consumption of sugars during submergence (Ram et al., 2002; Jackson and Ram, 2003; Bailey-Serres and Voesenek, 2008), could both contribute to the carbon balance of rice when completely submerged, and so determine survival. In turbid waters, however, low light would restrict photosynthesis.

The proposed mechanisms by which gas films on submerged leaves facilitate underwater net photosynthesis include the larger gas–water interface for CO2 uptake from the floodwater than if restricted to localized areas adjacent to stomata (cf. plastrons of insects, Hebets and Chapman, 2000), and the possibility that stomata may remain open for gas movements (suggested by Colmer and Pedersen, 2008a). Although stomata are hypothesized to close upon submergence of leaves without gas films (Mommer and Visser, 2005), data on this subject are lacking, and reliable measurements to compare leaves with or without gas films are not available, so the suggestion by Colmer and Pedersen (2008a) remains speculative. Nevertheless, the present study clearly demonstrates that apparent resistance to CO2 entry at environmentally relevant concentrations was fivefold less in leaves of rice with gas films compared to those without. Similarly, the presence of gas films on leaves of Phragmites australis also reduced the apparent resistance to CO2 uptake 5.2-fold (calculated from Figure 2 in Colmer and Pedersen, 2008a). In the case of rice, the higher apparent resistance to CO2 entry in leaves without gas films was, as expected, overcome by increasing the external CO2 supply (Figure 3), whereas, unexpectedly, this was not the case for Phragmites australis (Figure 2 of Colmer and Pedersen, 2008a). The enhanced underwater gas exchange would also promote outward O2 diffusion during light periods, so that the level of photorespiration would presumably also be lower; photorespiration can be substantial in submerged rice (Setter et al., 1989).

Diffusive boundary layers in the liquid phase remain a major resistance component to dissolved gas exchange between the floodwater and leaves even when gas films are present, as these layers were thicker for leaf segments with gas films than for those without (Figure 2 and Table 2). We speculate that the elastic properties of the gas films, which fluctuate between a compressed and extended state as high-density packets of water pass along the leaves, would dissipate energy and so decrease erosion of the diffusive boundary layer (Vogel, 2006). The fluctuating gas layer would also have resulted in a wider transition zone between laminar and turbulent flows adjacent to the leaf segments when in flowing water (Figure 2). Despite the thicker diffusive boundary layers (determined from O2 profiles; Figure 2), the leaves with gas films had enhanced CO2 uptake (determined from underwater photosynthesis; Figure 3), so decreases in other resistance components (see above) outweighed the change in diffusive boundary layers. Understanding of gas transport processes across water–air interfaces in natural environments is an active area of study (Banerjee, 2007); knowledge from interface transport (Turney and Banerjee, 2008) as well as bubble gas-exchange modeling (Woolf et al., 2007) should both aid understanding of the physics of leaf gas films, although the gas films would be considered ‘stationary bubbles’ and are not spherical.

In summary, leaf gas films make a substantial contribution to the submergence tolerance of rice; internal aeration, sugar status and growth were all enhanced by the films. Whether leaf gas films differ amongst rice cultivars should be evaluated. Some other wetland plants also possess gas films, with benefits for underwater net photosynthesis (Colmer and Pedersen, 2008a), and some species that lack gas films can acclimatize so that newly produced leaves have improved gas exchange underwater (e.g. thin cuticles, thin leaves, chloroplast re-orientation; Mommer et al., 2004, 2005, 2007). In the wider context of submergence tolerance in plants, as reviewed recently by Bailey-Serres and Voesenek (2008), leaf gas films and associated higher rates of underwater net photosynthesis, and thus sugar and O2 supply, would be beneficial for species/cultivars with a shoot elongation ‘escape‘ response or a non-elongation ‘quiescence’ response.

Experimental procedures

Plant material

Oryza sativa L. var. Amaroo was raised in aerated nutrient solution (Colmer and Pedersen, 2008b), and then pre-treated for the final 7 days in stagnant 0.1% w/v agar nutrient solution, with shoots remaining in air, prior to use in experiments. The ages of plants used and the numbers of replicates (different plants, or tissues from different individuals), in each experiment, are given in of the table and figure legends.

Morphological, chemical and biophysical properties of leaves

The most recent fully expanded leaves were characterized for selected properties. Water repellency of leaf surfaces was assessed by measuring the contact angle of a 5 mm3 droplet of water on the leaf surface (Adam, 1963; Brewer and Smith, 1997; Neinhuis and Barthlott, 1997) as described by Colmer and Pedersen (2008a). Leaf thickness was measured on transverse sections of lamina, using a microscope equipped with a digital camera.

Surface gas film volume and tissue porosity (% gas volume per unit tissue volume) were measured by determining tissue buoyancy before and after vacuum infiltration of the gas spaces with water (Raskin and Kende, 1983), using equations as modified by Thomson et al. (1990). Triton X-100 (0.1% v/v) was used to remove surface gas films on the tissues.

Diffusive boundary layers adjacent to submerged leaf segments

For each replicate, two lamina segments of approximately 60 mm length and 8–10 mm width were taken halfway up the blade of the most recently fully expanded leaf. One was used as a control (with gas films) and the other was used for gas film removal treatment. Segments used for gas film removal treatment were brushed on both sides with a fine paintbrush soaked in 0.1% v/v Triton X-100 in incubation medium (composition given below in the section on underwater net photosynthesis), and then washed for 5 sec, three times, in fresh incubation medium. This treatment prevented formation of gas films on the lamina surfaces when submerged (Raskin and Kende, 1983), and did not cause any ‘flooding’ of internal gas spaces (Colmer and Pedersen, 2008a). The control segments (with gas films) were also washed in incubation medium prior to use.

Lamina segments were mounted on double-sided adhesive tape in a custom-built flume tank with a mean bulk flow velocity of 15 mm sec−1 parallel to leaf segments. The flume allowed O2 microeletrodes (tip diameter = 10 μm, OX-10, Unisense A/S, to penetrate the lamina segments from below, with the tip eventually protruding into the air-saturated incubation medium flowing above the segments. Thus, diffusive boundary layers were measured without the microelectrode itself affecting flows, which would not have been the case if the electrode had been moved towards the leaf surface from above instead of from within and away (see Glud et al., 1994). The O2 microelectrodes were connected to a pA meter (PA2000, Unisense A/S), and electrodes were advanced in steps of 10 μm every 5 sec using micromanipulators (MM33, Unisense A/S). All measurements were performed in darkness at 30°C. Diffusive boundary layer thickness was calculated from the O2 gradients, as described by Jørgensen and Revsbech (1985).

Dark respiration by leaf segments

Dark respiration of lamina segments of approximately 30 mm length was measured in a microrespiration system. Measurements were taken using 4 ml glass chambers with a capillary hole in the glass stopper (MR Ch-4000, Unisense A/S), through which an O2 microelectrode (OX-MR, Unisense A/S) was inserted. The medium used was identical to that used for measurement of underwater net photosynthesis (see below), with a CO2 concentration of 50 mmol m−3 at 30°C (for further details, see Colmer and Pedersen, 2008a).

Underwater net photosynthesis by leaf segments

For each replicate leaf, two lamina segments of approximately 30 mm length were taken halfway up the blade of the most recently fully expanded leaf. One was used as a control (with gas films) and another was used as the treatment in which gas films were removed by brushing with 0.1% v/v Triton X-100 in incubation medium, and then washed in incubation medium without Triton X-100. Underwater net photosynthesis by the lamina segments was measured using the method described by Colmer and Pedersen (2008a). Measurements commenced within 15 min of excision. Glass cuvettes (35 ml) with stoppers contained individual lamina segments in incubation medium and two glass beads for mixing as the cuvettes rotated on a wheel within an illuminated water bath at 30°C. Photosynthetically active radiation (PAR) was 350 μmol m−2 sec−1, measured inside the submerged glass cuvettes (using a 4π US-SQS/L quantum sensor, Walz,

The incubation medium contained 0.50 mol m−3 Ca2+, 0.25 mol m−3 Mg2+, 1.00 mol m−3 Cl, 0.25 mol m−3 SO42− and 5.0 mol m−3 MES, with pH adjusted to 6.00 using KOH. The dissolved O2 concentration in the incubation medium was set at 50% of air saturation by bubbling in a 1:1 ratio (by volume) of N2 and air (prior to adjustment of dissolved CO2); this avoided build-up of high O2 levels during the measurements that might have led to photorespiration and thus decreased net photosynthesis, as previously described for submerged rice leaves (Setter et al., 1989). As flasks were incubated in light immediately after adding the lamina segments, and as the segments produce O2 when in light, there was no risk of tissue hypoxia. Dissolved CO2 treatments were imposed by adding specific concentrations of KHCO3 to the incubation medium, with pH always adjusted to pH 6.00 using various amounts of KOH depending on the amount of KHCO3 added, to provide a range of CO2 concentrations from 15 (air saturation) to 2000 mmol m−3 (Stumm and Morgan, 1996). K2SO4 was added as required in the various treatments so that the K+ concentration was equal across treatments.

Following incubations of known duration (60–90 min), dissolved O2 concentrations in the bottles were measured using a Clark-type O2 mini-electrode (OX-500, Unisense A/S) connected to a pA meter (PA2000, Unisense A/S). The electrode was calibrated immediately before use. Dissolved O2 concentrations in bottles prepared and incubated in the same way as described above, but without lamina segments, served as blanks. The projected area of each lamina segment was measured using a leaf area meter (Li-Cor LI-3000,, and the fresh and dry masses (after freeze drying) of each segment were determined.

Net photosynthesis in air

Net photosynthesis in air at a PAR of 350 μmol m−2 sec−1 and ambient (380 μl L−1) and high (760 μl L−1) CO2 were measured at 30°C on intact leaves (most recent fully expanded) using a flow-through leaf cuvette connected to an infra-red gas analyser (LI-6400, Li-Cor).

Root aeration of whole plants

Four-week-old plants were transferred into a horizontal chamber (length 0.8 m) made from 150 mm diameter PVC pipe sliced lengthways in half, with caps at each end and a divider fitted 300 mm from one end (Colmer and Pedersen, 2008a). The root–shoot junction was positioned 10 mm below the divider, so that the roots were submerged in incubation medium (see net photosynthesis measurements for composition) that also contained 0.1% w/v agar to prevent convective movement and had been deoxygenated by prior flushing with N2. Initially, shoots were always in air. The root compartment was covered with glass plates and a small opening enabled insertion of an O2 microelectrode. The shoot base was held in place within the divider using BluTac putty (Bostik, The shoot compartment was filled with incubation medium (without agar) containing 200 mmol m−3 free CO2 and O2 in air equilibrium (236 mmol m−3). Any possible contact between leaves and air was prevented by covering the submergence solution with transparent polyethylene sheeting. Experiments were performed in a room kept at 30°C. During light periods, PAR (350–400 μmol m−2 sec−1) was provided by halogen spotlights.

Clark-type O2 microelectrodes with a guard cathode and tip diameter of 25 μm (OX-25, Unisense A/S) (Revsbech, 1989) were used. A microelectrode was inserted into the root cortex 20–25 mm behind the root tip using a micromanipulator (MM5, Märzhäuser, The microelectrode was connected to a pA meter (PA8000, Unisense A/S) and the outputs were logged every 10 sec on a computer using an analog to digital converter (ADC-16, Pico Technology, O2 concentrations in the shoot submergence solution and in the 0.1% stagnant agar solution in the root compartment were both monitored using Clark-type O2 minielectrodes (OX-500, Unisense A/S); in all experiments, the solution in the shoot compartment remained close to air saturation and O2 in the bulk medium of the root compartment increased to a maximum of 0.5 kPa. The water temperature was recorded using type-K thermocouples connected to a resistance converter (TC-08, Pico Technology), and remained at 30 ± 0.5°C.

Growth experiment for whole plants

To investigate the influence of leaf gas films on sugar status and growth, 14-day-old plants were completely submerged for 7 days. The experiment was a 2 × 2 factorial design, and consisted of controls with shoots in air, a second control with shoots in air but brushed with 0.1% Triton X-100, submerged with gas films, and submerged and Triton X-100-treated to remove gas films on leaves and sheaths. In all treatments, shoots were also rinsed with incubation medium. The experiment was performed in a constant-environment room (30°C, 12 h light/12 h darkness, 80% relative humidity, PAR of 500–600 μmol m−2 sec−1).

All plants were placed with their roots in 500 ml darkened bottles containing stagnant 0.1% agar nutrient solution (Colmer and Pedersen, 2008b), and then into transparent acrylic plastic cylindrical tanks. Half of the Perspex tanks did not contain any solution; those assigned to the submergence treatment were filled with 12 L of incubation medium (see net photosynthesis measurements for composition, but without MES). Free CO2 was kept at 200 mmol m−3 (Stumm and Morgan, 1996) using a pH controller (α-control, Dupla Aquaristik, connected to a cylinder with pressurized CO2. Shoots treated with Triton X-100 were rinsed with incubation medium prior to insertion into the Perspex tanks. Leaves formed during the 7-day treatment period were brushed with Triton X-100 and rinsed. Plastic mesh held approximately 50 mm below the top of each Perspex tank prevented emergence of any leaves. This mesh was present on all Perspex tanks (i.e. both air controls and the submergence treatments).

Initial and final harvests were carried out to quantify leaf, sheath and root dry masses (freeze-dried) and tissue sugar concentrations. Plants were sampled as blocks during a 2 h window 8–10 h into the light period. Plants were retrieved from the Perspex tanks, rinsed with deionized water and blotted to remove surface water, then tissues were excised and wrapped in aluminium foil, frozen in liquid N2, freeze-dried and dry mass was recorded, before pulverisation in a ball mill. Sugars were extracted from tissue samples boiled twice in 80% ethanol with reflux for 20 min. Total sugar levels in the extracts were measured using anthrone (Yemm and Willis, 1954) and a spectrophotometer (UV-240, Shimadzu, The reliability of the method was verified by checks on recovery of glucose spiked into some samples immediately prior to extraction.

Data analyses

GraphPad Prism 5.0 (GraphPad Software Inc., was used to fit a Jassby and Platt (1976) model to the CO2 response curves. This program was also used for one- or two-way anova (with Tukey or Bonferoni post hoc tests) and Student’s t tests to compare means.


We thank the Faculty of Natural and Agricultural Sciences at the University of Western Australia for supporting Ole Pedersen under the Distinguished Visitors Programme, CLEAR (a Villum Kann Rasmussen Centre of Excellence at the University of Copenhagen), Imran Malik for his assistance with the Li-Cor, Ray Scott at the University of Western Australia Combined Workshop for his valuable assistance, and Hank Greenway for comments on a draft manuscript.