|Home | About | Journals | Submit | Contact Us | Français|
Submersed plants have different strategies to overcome inorganic carbon limitation. It is generally assumed that only small rosette species (isoetids) are able to utilize the high sediment CO2 availability. The present study examined to what extent five species of submersed freshwater plants with different morphology and growth characteristics (Lobelia dortmanna, Lilaeopsis macloviana, Ludwigia repens, Vallisneria americana and Hydrocotyle verticillata) are able to support photosynthesis supplied by uptake of CO2 from the sediment.
Gross photosynthesis was measured in two-compartment split chambers with low inorganic carbon availability in leaf compartments and variable CO2 availability (0 to >8 mmol L−1) in root compartments. Photosynthetic rates based on root-supplied CO2 were compared with maximum rates obtained at saturating leaf CO2 availability, and 14C experiments were conducted for two species to localize bottlenecks for utilization of sediment CO2.
All species except Hydrocotyle were able to use sediment CO2, however, with variable efficiency, and with the isoetid, Lobelia, as clearly the most effective and the elodeid, Ludwigia, as the least efficient. At a water column CO2 concentration in equilibrium with air, Lobelia, Lilaeopsis and Vallisneria covered >75% of their CO2 requirements by sediment uptake, and sediment CO2 contributed substantially to photosynthesis at water CO2 concentrations up to 1000 µmol L−1. For all species except Ludwigia, the shoot to root ratio on an areal basis was the single factor best explaining variability in the importance of sediment CO2. For Ludwigia, diffusion barriers limited uptake or transport from roots to stems and transport from stems to leaves.
Submersed plants other than isoetids can utilize sediment CO2, and small and medium sized elodeids with high root to shoot area in particular may benefit substantially from uptake of sediment CO2 in low alkaline lakes.
Photosynthesis of submersed aquatic plants is often limited by low availability of dissolved inorganic carbon (DIC; Madsen and Maberly, 1991). Although the concentration of CO2 in water can be orders of magnitude higher than in air due to CO2 accumulation from respiration or groundwater input, the availability in water is in general lower because molecular diffusion of dissolved gases is about 104 times slower in water than in air. Hence, the transport of CO2 through the diffusive boundary layer represents a bottleneck to the DIC supply of submersed plants (Maberly and Spence, 1989). The ability to utilize bicarbonate is probably the most important adaptation to overcome DIC limitation of underwater photosynthesis since bicarbonate is most often present in much higher concentration than free CO2 (Prins and Elzenga, 1989; Madsen and Sand-Jensen, 1991; Maberly and Madsen, 2002). In addition to bicarbonate utilization, crassulacean acid metabolism (CAM) and C4-like metabolism function as carbon-concentrating mechanisms facilitating inorganic carbon acquisition in some submersed plants (Maberly and Madsen, 2002). However, free CO2 in either the water column or the interstitial water of the sediment is a more favourable source if available and if it can be exploited by the plants (Maberly and Madsen, 2002; Jones, 2005).
Due to mineralization of organic matter within the sediment, the CO2 concentration in interstitial water is often 10–250 times higher than atmospheric equilibrium (approx. 16 µmol CO2 L−1; Wium-Andersen and Andersen, 1972; Pedersen et al., 1995). The use of sediment CO2 is well described for isoetids (Wium-Andersen, 1971; Boston et al., 1987; Madsen et al., 2002). Isoetids are small aquatic rosette plants with extensive root systems and well-developed air lacunae ensuring efficient gas transport between roots and shoot. Sand-Jensen and Søndergaard (1978) found that 95% of the CO2 fixed by the isoetid Lobelia dortmanna was obtained directly from the sediment, but the role of sediment CO2 has only been evaluated for a few non-isoetid species.
As sediment CO2 can exceed 5 mmol L−1, its use has a substantial potential to supplement water column-derived DIC and alleviate restrictions on CO2 supply to submersed plants. The role of sediment CO2 should, however, depend on large root surface areas relative to photosynthetic leaf area, gas-permeable root walls, proportionally high and well-connected root, stem and leaf aerenchyma, and moderate transport distances. Lobelia dortmanna, which relies almost exclusively on sediment CO2, has a very large biomass of highly gas-permeable roots, large lacunae connecting roots directly with leaves, and short leaves in a rosette ensuring short transport distance (Madsen et al., 2002). Hence, species with similar morphology and tissue characteristics should potentially be able to take substantial advantage of the high sediment CO2 concentrations. In some aquatic plant species, root walls have low gas permeability which minimizes radial O2 loss to the sediment and ensures effective O2 supply to the root tip (Colmer, 2003), and this feature is common among wetland species, floating-leaved plants and seagrasses growing in reduced sediments (Armstrong, 1979; Smits et al., 1990; Colmer, 2003; Jensen et al., 2005). If submersed freshwater species also produce almost impermeable roots, it could have adverse effects on root uptake of sediment CO2, although concentration gradients for CO2 can be much steeper than O2 gradients between roots and sediment (Greenway et al., 2006).
Evidence of sediment CO2 uptake already exists for several plant species, including but not limited to, the isoetids. The emergent Phragmites australis has the ability to use sediment CO2, although only 1% of the fixed carbon was derived from this source (Brix, 1990). The submersed Vallisneria americana is also able to mediate a substantial transport of CO2 from roots to leaves, with up to 85% of CO2 assimilation being supported by sediment CO2 (Kimber et al., 1999). The only published attempt to establish whether this phenomenon is important in situ was conducted by Loczy et al. (1983). They used sediments with [14C]DIC to test if Myriophyllum spicatum, Vallisneria americana and Heteranthera dubia were able to use sediment CO2, and concluded that none of these species used significant amounts of sediment CO2 in shoot photosynthesis. Accordingly, apart from the single record on Vallisneria (Kimber et al., 1999), documentation of significant supply of sediment CO2 has not been published for non-isoetid species.
The aim of this study was to examine whether the sediment can be a quantitatively important CO2 source for submersed aquatic plants other than isoetids. The hypothesis was that the importance of sediment CO2 uptake is affected by: (a) life form; (b) tissue porosity; and (c) area-based root to shoot ratio. The sediment CO2 uptake for whole shoot photosynthesis was assessed at varying CO2 concentrations around the roots in split chambers. Four macrophyte species of different morphology and life form, Ludwigia repens, Lilaeopsis macloviana, Hydrocotyle verticillata and Vallisneria americana, were used in the experiments and compared with the isoetid, Lobelia dortmanna (Fig. 1). In addition, root 14CO2 uptake experiments were conducted with V. americana and L. repens in order to illustrate spatial differences in carbon fixation and thereby identify bottlenecks to utilization of sediment CO2.
Five species with different life form and morphology were selected to represent potentially different abilities to take up sediment CO2 and transport CO2 by means of molecular gas-phase diffusion from roots to leaves. Lobelia dortmanna L. (water lobelia) is a small submersed isoetid with short, stiff leaves in rosettes. It inhabits shallow waters of low-alkaline, oligotrophic lakes across the Northern hemisphere. Ludwigia repens Forst., is a fast-growing perennial stem plant native to North America. It is found in both slow-flowing streams and stagnant waters. Lilaeopsis macloviana A.W. Hill has hollow, cylindrical leaves emerging directly from a horizontal rhizome. It originates from South America, from the highlands to brackish marshes. Hydrocotyle verticillata Thunb. (umbrella plant) is a small plant with single leaves on long petioles emerging directly from a horizontal rhizome. It is a cosmopolitan species usually growing in stagnant waters. Vallisneria americana Michx. is a perennial plant with long strap-shaped leaves in a rosette. It is an almost cosmopolitan species found in flowing waters and shallow lakes in America, Asia and Australia. Ludwigia, Lilaeopsis, Hydrocotyle and Vallisneria were all one genotype obtained as tissue culture-grown plants from Tropica Aquarium Plants, Denmark, whereas Lobelia was collected in Lake Skånes Värsjö, Southern Sweden.
All species, except Lobelia and Vallisneria, were grown under water from apical shoots of emergent plants. The plants were submersed in high-alkaline tap water (alkalinity = approx. 5 meq. L−1) at 23°C with 250 µmol photons m−2 s−1 (PAR) in a 12:12 h light:dark cycle. The water was fertilized with liquid fertilizer (Tropica AquaCare liquid + ) containing N and P plus micronutrients. The potting soil (washed river sand) was fertilized with 4–5 pellets per plant of a slow release fertilizer (Osmocote Extract-Scotts, standard). The pH was maintained at 6·8 in the water column using a pH stat (Dupla Aquaristik, Grafschaft-Gelsdorf, Germany) through periodically bubbling with CO2 providing approx. 1·2 mmol L−1 of free CO2 in the water. Water was changed twice a week to prevent excess accumulation of planktonic algae.
To estimate the potential for gas-phase diffusion of CO2 from roots to shoot, porosity (% gas space per unit tissue volume) was measured on samples of roots, leaves and stems by determining tissue buoyancy before and after vacuum infiltration of the gas spaces with water (Raskin, 1983), using the equations for gas volume as modified by Thomson et al. (1990).
Root area may serve as a bottleneck for root-mediated CO2 uptake and, thus, root area was calculated from measurements of the length and diameter of the main and lateral roots using a dissecting microscope. Leaf area was measured using a Li-Cor portable area meter model Li-3000 (Li-Cor, Lincoln, NE, USA). Tissue dry mass was determined after drying for 24 h at 105°C.
The aim of the experiment was to describe whole-plant photosynthesis as a function of varying CO2 concentration around the roots. Whole-plant photosynthesis was measured in two-compartment split chambers, where roots and shoot were hydraulically isolated into two separate water volumes using a Perspex cradle around the base of the plants tightened with hydrophobic lanolin and hardened coconut oil. In this way, CO2 availability could be manipulated around the roots without affecting the CO2 concentration in the shoot compartment.
Photosynthetic O2 production was measured using oxygen electrodes (Unisense OX-500 with a PA2000 pA meter connected to a laptop computer running Picolog for data-logging), which were inserted into both root and shoot compartments. The pH was continuously measured in the root chamber (Radiometer pH electrode with Radiometer pH meter PHM 92) and in the shoot compartment at the beginning and end of each experiment. The chamber was positioned horizontally at 20°C in a temperature-controlled water jacket of demineralized water. Two halogen light sources were used for Lobelia and Lilaeopsis providing 640 µmol photons m−2 s−1 (PAR). For Hydrocotyle, Vallisneria and Ludwigia, additional irradiance was needed to light saturate photosynthesis, and two extra fluorescent light sources increased the photon flux to 920 µmol m−2 s−1.
The shoot compartment was filled with artificial lake water without addition of HCO3− to keep DIC at a minimum (Smart and Barko, 1985). A phosphate buffer (1 mmol L−1) was used to stabilize the pH at 8 to make sure that DIC in the form of free CO2 entering the shoot compartment would mainly be converted to HCO3−. At pH 8 and concentrations of DIC <0·1 mmol L−1, 98% of the DIC occurs in the form of HCO3− and the concentration of free CO2 is ≤2 µmol L−1. The root compartment was filled with DIC-rich water (>10 mmol DIC L−1) made from tap water (approx. 5 mmol DIC L−1; pH 8·5) enriched with equal molar amounts of KHCO3 and NaHCO3.
The experiments were initiated at a pH of approx. 9 (adjusted by addition of 0·1 mol L−1 NaOH) in the root chamber to ensure low CO2 availability (<20 µmol L−1) for root uptake since most DIC occurs as HCO3− at this pH. The plants were pre-incubated in light until the oxygen concentration in both the root and shoot chamber decreased at a steady rate, reflecting that pools of DIC within the plant tissue were depleted and concentrations of free CO2 in the water were below CO2 compensation points. After the pre-incubation period, the pH was lowered stepwise in the root chamber by injecting small amounts of HCl (0·1 mol L−1), converting HCO3− to free CO2, and O2 evolution in both the root and shoot chamber was recorded at each CO2 concentration. When rates of O2 evolution in the chambers reached maximum, reflecting that photosynthesis was CO2 saturated, the experiment was stopped.
Control experiments were conducted to ensure that the plant was only taking up CO2 from the root chamber and not CO2 or HCO3− leaking from the root to the shoot chamber during the experiment. At the end of each experiment, the pH was raised to 9 in the root chamber by injecting NaOH to remove virtually all free CO2 from the root medium. If the oxygen concentration then declined, it was assumed that there had been no leakage of free CO2 into the shoot compartment. As an additional control, water samples were taken from the shoot compartment at the beginning and end of each experiment and analysed for DIC on an infrared gas analyser (ADC, 225-MK3) to verify that the DIC concentration in the shoot compartment remained low and did not increase during the experiment. After each experiment, leaf and root area and dry mass were quantified.
For all species, leaf photosynthesis based on water column CO2 was measured as a function of free CO2 to allow comparison with photosynthesis based on sediment CO2 and estimates of the relative importance of sediment and water column as sources of CO2 under different DIC availabilities. Leaf sections of approx. 10–20 mg dry mass were cleaned of epiphytes and mounted in a temperature-controlled Perspex chamber. Whole leaves of Lobelia and Lilaeopsis exhibited low photosynthetic rates, possibly due to the presence of cuticle or wax, and were therefore cut longitudinally to offset leaf surface resistance towards gas diffusion and enhance DIC transport to photosynthetic tissues. The medium was made from one part tap water and one part demineralized water, with a resulting alkalinity of 2·83 meq. L−1. A halogen light source provided 270 µmol photons m−2 s−1 of PAR, which should be sufficient to light saturate photosynthesis of individual leaves (Binzer et al., 2006). The availability of free CO2 was changed stepwise by changing the pH from a maximum of 9·2 through addition of 0·1 mol L−1 HCl. This procedure was repeated until the oxygen evolution reached a maximum, reflecting CO2-saturated photosynthesis. The oxygen concentration was measured and logged together with pH as described for whole-plant photosynthesis. After each experiment, leaf dry mass and area were measured as described earlier.
Root 14CO2 uptake experiments were conducted with V. americana and the elodeid species L. repens to describe spatial patterns in carbon fixation and identify potential bottlenecks in utilization of sediment CO2. The plants were mounted with their roots in closed PVC chambers and the shoots exposed to a common bulk chamber holding 13·5 L. A water-tight seal made of a Perspex cradle and lanolin was placed around the base of each plant. The root medium was made from tap water (alkalinity 5·7 meq. L−1) enriched with NaHCO3 and KHCO3 to raise the alkalinity to 7·8 meq. L−1. The pH was lowered to 6·1 using HCl to achieve a free CO2 concentration of 5 mmol L−1.14C was added as dissolved inorganic carbon to a specific activity of 1290 Bq mL−1. The shoot bulk chamber was filled with a DIC-poor medium (<0·1 mmol DIC L−1; Smart and Barko, 1985) at a temperature of 21°C.
The plants were incubated for 4 h, with one set of plants kept at 1490 µmol photons m−2 s−1 and another set in darkness. A sample of the shoot medium was analysed every hour to detect if 14C had leaked from the root chamber to the bulk water. To check further for any 14C leakage into the bulk water, shoots of Ludwigia with no access to the root medium were placed freely floating in the shoot medium and later analysed for 14C incorporation in the light. Both controls showed that no leakage of 14C had occurred. After incubation, the plants were harvested and cut into segments of stems or leaves, dried at 70°C for 24 h and homogenized with a pestle in a mortar. Approximately 1–8 mg of tissue was added to 20 mL plastic vials and 500 µL of ethylene glycol was injected to dissolve the plant tissue. After 24 h, 10 mL of Ultima Gold scintillation liquid were added (Perkin Elmer and Analytical Sciences). The vials were placed on a rotating wheel for 3 d to dissolve and homogenize the samples further, and then counted on a tri-carb 2800 TR scintillation counter. To compensate for differences in quenching with different amounts of 14C-labelled plant material, counting efficiency was estimated by adding inorganic 14C to variable amounts of non-labelled plant material dissolved in ethylene glycol and scintillation liquid.
Data were analysed statistically using Graph Pad Prism 5 and a 0·05 significance level. Some data were log transformed to ensure variance homogeneity (Bartlett's test). One- or two-way analysis of variance (ANOVA) with a Tukey post hoc test or Student's t-test were used to test for significant differences.
For each experiment, data on leaf and whole-plant photosynthesis were fitted to a modified Jassby and Platt (1976) model:
where Pmax is the maximum photosynthetic rate (μmol O2 g−1 d. wt leaf h−1), α is the initial slope of the curve [μmol O2 g−1 d. wt leaf h−1 (mmol CO2 L−1)−1] and x is the CO2 concentration (μmol CO2 L−1). Standard linear regression was used to estimate α because the model (eqn. 1) did not fit some of the replicates of each species properly. The α values were analysed by two-way ANOVA and a Bonferroni post hoc test which requires values for each replicate.
The shoot to root area ratio was highly variable among species, i.e. 0·76 for Lilaeopsis to 56·4 for Hydrocotyle. The shoot to root area ratio was 0·98 for Lobelia and 0·99 for Ludwigia, and only Vallisneria (3·1) and Hydrocotyle had ratios >1·0.
Porosity varied significantly among species, being higher in Lobelia, Vallisneria and Lilaeopsis than in Hydrocotyle and Ludwigia (Table 1). Porosity ranged from 4·2 to 62·2% in leaves and from 12·8 to 63·3% for roots. The highest leaf porosity was found in Lilaeopsis and the lowest in Hydrocotyle. Root porosity was significantly higher in Lobelia than in the other species. Root porosity exceeded leaf porosity in Lobelia, while the opposite was found for Lilaeopsis. No significant differences between leaf, root and stem porosity were found for the remaining species.
The split-chamber experiments showed that four of the five species were able to sustain photosynthesis based on sediment CO2 supply (Table 2). The response to increasing CO2 concentrations around the roots followed saturation curves (Fig. 2) very similar to that for leaf photosynthesis based on water CO2 supply (data not shown). Only Hydrocotyle did not follow this pattern as increasing CO2 around the roots did not stimulate photosynthesis at all. Maximum gross photosynthesis (Pmax (root CO2)) based on sediment CO2 varied from negative rates for Hydrocotyle to 432 µmol O2 g−1 d. wt h−1 in Lobelia (Table 2). Maximum gross photosynthesis based on water column CO2 supply (Pmax (leaf CO2)) varied only 2-fold from 303 to 630 µmol O2 g−1 d. wt h−1.
The initial slope of the relationship between photosynthesis and CO2 concentration reflects the efficiency with which the plant is able to utilize CO2 at low CO2 availability in the sediment or water column, with steep slopes reflecting low resistance to CO2 uptake (Table 2). The initial slopes were consistently steeper for water column CO2 supply, with no significant differences among species. When CO2 was taken up through the roots, the initial slopes varied significantly among species but with some overlap. Lobelia was the most efficient species to utilize sediment CO2 (i.e. exhibited the steepest slope), followed by Lilaeopsis and Vallisneria, while Ludwigia was rather inefficient. The initial slope could not be estimated for Hydrocotyle. For Lobelia, the initial slope was not significantly lower when CO2 was supplied by the roots even though leaves had been sliced open to facilitate uptake in the experiment with water column CO2 supply (Student's t-test).
Saturating concentrations for CO2 supplied by the roots varied from 1·3 mmol CO2 L−1 for Lilaeopsis to 9·5 mmol CO2 L−1 for Ludwigia. Since Hydrocotyle was unable to use root-supplied CO2, no saturating concentration could be estimated for root uptake (Table 2). The saturating CO2 concentrations estimated for water column uptake were much lower than for root-mediated uptake, except for Lobelia.
The role of sediment CO2 supply for whole-plant gross photosynthesis at different water column CO2 concentrations was estimated by combining data from the fits of photosynthesis of leaf segments supplied by CO2 from the water column with estimated maximum rates for whole shoots in split-chamber experiments (Table 3). For Lobelia, sediment CO2 supplied >85% of photosynthesis up to water column CO2 concentrations of 100 µmol L−1 even when water column uptake was facilitated by slicing leaves open, resulting in very conservative estimates of the role of sediment uptake. For Lilaeopsis, leaves were similarly sliced to facilitate water column CO2 supply, and roots supplied >50% of all CO2 as long as water column concentrations were <100 µmol CO2 L−1. For Vallisneria, sediment uptake supplied >34% of total DIC uptake at water column CO2 concentrations up to 100 µmol CO2 L−1, while sediment contributed <30% of total CO2 uptake in Ludwigia even at water column CO2 concentrations in air equilibrium (16 µmol L−1).
Using data for root area and Pmax (root CO2) for each of the five plant species, a maximum root CO2 uptake was calculated (Table 4). The highest root uptake was found in Lobelia of 123 pmol CO2 cm−2 root s−1, which was significantly higher than for the other species (Tukey post hoc). Lilaeopsis and Vallisneria had significantly higher CO2 uptake per root surface area than Ludwigia (Table 4).
The experiment with 14C supplied to the root medium showed that both Vallisneria and Ludwigia were able to take up inorganic 14C via the roots and incorporate it in the shoots (Fig. 3). Water samples from the shoot chamber and suspended shoots of Ludwigia with no access to labelled 14C in the root chamber showed no incorporation of 14C above background levels (data not shown), confirming that seals between root and shoot chambers were tight and that transport occurred only inside the plants.
For both species, 14C fixation decreased systematically with increasing distance to the roots (one-way ANOVA, P < 0·05; Fig. 3). In Vallisneria, the bulk of the 14C was assimilated in the lower 0–2 cm sections of the leaves, while transport to leaf tips (8–11 cm) was negligible. In Ludwigia, transport of 14C to stems and leaves followed a similar decline with distance to the roots (Fig. 3). 14C assimilation was much higher in stems of Ludwigia than in leaves, except in the shoot apex, where 14C fixation peaked immediately below the apex.
Four out of the five studied plant species were able to use CO2 taken up by the roots, although with highly variable efficiency. Lobelia was the most efficient to use CO2 supplied via the roots. It had the highest Pmax (root CO2) and the highest CO2 uptake per root area, rendering it extremely efficient at supplying CO2 from the root to the shoot. This agrees with the conclusion of Sand-Jensen and Søndergaard (1978) that 95% of the assimilated CO2 in Lobelia is usually derived from the sediment in accordance with its natural distribution in DIC-poor lakes. The results also showed that Lilaeopsis and Vallisneria are able to use sediment CO2, and Lilaeopsis in particular used sediment CO2 very effectively. Ludwigia was capable of sustaining low rates of photosynthesis with sediment CO2, but its Pmax (root CO2) was >40 times lower than that of Lobelia and ten times lower than in Vallisneria. Hydrocotyle was, however, unable to sustain photosynthesis with root-supplied CO2 even at very high CO2 concentrations around the roots.
The fact that both Vallisneria and Lilaeopsis could use CO2 supplied from the sediment challenges the general perception that only isoetids possess this capability (Loczy et al., 1983; Madsen and Sand-Jensen, 1991). Previously, it has been shown that species of Lilaeopsis and Vallisneria also have CAM metabolism similar to some isoetids (Webb et al., 1988), suggesting that these two non-isoetid plants have several mechanisms which can at least partly ameliorate inorganic carbon limitation in waters with low availability of free CO2. While this study concludes that some life forms such as Ludwigia and Hydrocotyle are able to use little, or no, sediment CO2, submersed plants other than isoetids are certainly able to cover a substantial part of their CO2 requirements by root-mediated CO2 uptake when the DIC availability in the water column is low. The variable ability among species to utilize sediment CO2 may depend on several factors: (a) the concentration of CO2 in the interstitial pore water relative to that of the water column; (b) root area relative to shoot area; (c) CO2 permeability of root walls; (d) root, stem and leaf porosity to support internal transport via gas phase diffusion; (e) internal diffusion barriers and morphological bottlenecks; and (g) distance from the roots to the leaves.
The specific CO2 concentrations in the root chamber at which leaf photosynthesis saturated varied greatly among the four species that were able to use sediment CO2. Photosynthesis of Lobelia and Lilaeopsis based on sediment CO2 saturated well below the sediment concentrations of 5 mmol CO2 L−1 reported for different types of sediment (Wium-Andersen and Andersen, 1972). For both species, even with the conservative estimates due to the sliced leaves during the measurements, sediment CO2 uptake played a major role at water column CO2 concentrations up to 100 µmol L−1, which is well above the atmospheric equilibrium of 16 µmol L−1. The photosynthesis of Vallisneria and Ludwigia saturated at somewhat higher root concentrations of 8–10 mmol L−1 and none of these species was able to achieve as high photosynthetic rates as when saturated with CO2 from the water column, indicating constraints on CO2 supply from the sediment other than the relative availability of CO2 from the two sources. For Vallisneria, the importance of water column DIC is further strengthened by its ability to utilize bicarbonate (Titus and Stone, 1982).
There was a negative correlation between Pmax (root CO2) and shoot to root area ratio for Lobelia, Lilaeopsis and Vallisneria, indicating that for these species a high root area relative to leaf area enhances the use of sediment CO2 (Fig. 4). In general, an area-based shoot to root ratio around or below 1·0 for these species seemed to be a suitable ratio for root uptake of CO2, as also suggested by Madsen and Sand-Jensen (1991). Ludwigia had a low shoot to root area ratio but low sediment-based photosynthesis, suggesting that the root supply of CO2 was limited by factors other than the area available for root CO2 uptake. In contrast, the fact that Hydrocotyle was unable to use CO2 taken up through the roots agrees well with its extremely high shoot to root area ratio.
The gas permeability of the root surface is essential for the ability of CO2 to diffuse into the roots, and therefore highly important for uptake of sediment CO2. Some plant groups (wetland species, seagrasses and aquatic plants with floating leaves) growing in anoxic soils or sediments exhibit low gas permeability in the roots to minimize radial O2 loss to the sediment and ensure oxygen supply to the root tips (Armstrong, 1979; Smits et al., 1990; Colmer, 2003). The same trait may be present in submersed freshwater plants and may also hold for exchange of CO2, although gradients in CO2 are much steeper than O2 gradients. Since Lobelia relies almost entirely on sediment CO2 and releases most of its photosynthetically produced oxygen to the sediment, it must have a high gas permeability (Sand-Jensen et al., 1982; Pedersen et al., 1995; Møller and Sand-Jensen, 2008). Since both Lilaeopsis and Vallisneria can effectively use root-mediated CO2, it is likely that their roots are also highly gas permeable although the estimated maximum CO2 uptake per root area was about half of that for Lobelia. However, as for Ludwigia, which exhibited very low CO2 uptake rates per root area, it is not possible to separate constraints on uptake from those of transport processes.
Root, stem and leaf porosity is obviously essential for efficient transport and use of sediment CO2. Once CO2 has penetrated into the roots, it needs to be transported by molecular diffusion in the air-filled lacunae to the photosynthetically active tissue (Brix, 1990; Madsen and Sand-Jensen, 1991). According to Fick's first law of diffusion, the rate depends on the cross-sectional area of the lacunae and the steepness of the concentration gradient determined by the difference in CO2 concentrations divided by the length of the lacunae. Hence, an estimate of the potential role of CO2 transport from roots to shoots of different lengths can be made by assuming no CO2 consumption along the diffusion path, as presented in a review by Madsen and Sand-Jensen (1991). Based on their calculations, Madsen and Sand-Jensen (1991) concluded that CO2 transport inside the lacunae of submersed plants would only be significant over short distances. However, applying the cross-sectional area of the lacunae in Ludwigia (0·93 mm2), which had a moderate stem porosity of 19%, revealed that diffusion could supply 40 pmol CO2 s−1 at a distance of 200 cm and 130 pmol CO2 s−1 at 60 cm if the root CO2 concentration was 5 mmol L−1 (Fig. 5). These supply rates can be compared with average photosynthetic CO2 requirements of between 40 and 125 pmol cm−2 leaf s−1 for elodeids as done by Madsen and Sand-Jensen (1991). Although the estimates and comparison are based on rather simple assumptions, the calculations show that root supply of CO2 by diffusion is certainly not trivial even over distances of ≥20 cm and especially not for species with high tissue porosity and large cross-sectional area of lacunae in stem and leaves as in Lobelia, Lilaeopsis and Vallisneria (Table 1).
Although tissue porosity was relatively high in Vallisneria, the 14C experiment and the role of sediment CO2 estimated from the split-chamber experiment showed a moderate CO2 contribution from the sediment, when water column CO2 was above air saturation. This could probably be due to the less favourable shoot to root area ratio, since Vallisneria had 3·1 cm2 of leaf to be supplied with CO2 per cm2 of root surface. Ludwigia had a low shoot to root area ratio and reasonable tissue porosity, but was nevertheless not very well supplied by sediment CO2, since <30% of maximum gross photosynthesis could be supported by root uptake even at very low water column CO2 availability. Either low gas permeability of roots or diffusion barriers between tissues could limit CO2 uptake or transport. For example, 14C fixation was much higher in the green Ludwigia stems than in leaves, suggesting transport constraints in the short petioles.
Hydrocotyle was virtually unable to utilize sediment CO2, and this agrees well with its morphology and life form. Although the plant is small with short transport distances, the use of sediment CO2 would be restricted by the very high shoot to root area ratio and the low tissue porosity. In addition, diffusion barriers may be present between the long petiole and the leaf.
All examined species except H. verticillata were able to utilize sediment CO2, although with variable efficiency. As expected, and as well described in the literature, the isoetid L. dortmanna was the most effective sediment CO2 user. Lilaeopsis macloviana was almost as efficient as Lobelia and it can be expected that quite a number of species with a life form and morphology involving low shoot to root area, high tissue porosity and close connection between leaves and roots can take considerable advantage of the high concentrations of free CO2 in the sediment. For such species, the diffusion model shows a potential for substantial CO2 supply over decimetres to even metres if the cross-sectional area of lacunae is large. If uptake or diffusion constraints occur, as in V. americana and L. repens, the role of sediment CO2 is smaller but still not trivial. Only species with poor root development and/or low tissue porosity may not benefit from sediment CO2 at all. Accordingly, sediment CO2 supply may play an important role for many small and medium-sized submersed plant species especially in low alkaline lakes with low CO2 concentrations in the water column.
We thank Kaj Sand-Jensen, Ole Pedersen, Theis Kragh and Claus Lindskov Møller for suggestions and constructive critique. We thank Tropica Aquarium Plants, Denmark for continuously supplying plant material. This study was part of the CLEAR-project (Lake Restoration Centre) funded by the Willum Kann Rasmussen Foundation.