The metabolism of chlorophylls and related tetrapyrroles directly influences, and is influenced by, the proteins and cell structures with which they are associated. During net accumulation, de-greening and at the steady state, chlorophyll and its derivatives are important elements in the post-translational regulation of the expression of genes for chloroplast proteins. At the same time, they represent potential photodynamic hazards against which green cells need to have protective mechanisms. This review deals with genetic, chemical and environmental perturbations of chlorophyll biosynthesis that impact on protein stability, membrane organization and susceptibility to photodamage. NADPH-protochlorophyllide oxidoreductase is considered in detail as a pigment-protein regulating, and regulated by, chlorophyll metabolism. The question of the extent and significance of chlorophyll turnover at the steady state is addressed, with particular emphasis on the dynamics of the photosystem II reaction centre. The pathway of chlorophyll catabolism is described, along with its interrelationship with protein mobilization in chloroplast senescence. Finally, the structural basis of pigment–protein interaction and stability is examined, and the discussion ends by expressing some general thoughts about the control of protein lifetimes in the living cell.

Ion exchange chromatography of water extracts was used to classify the fructans in leaves of 40 temperate dicots and 110 temperate grass species. Of the 150 species studied, only Poa ampla (Merrill) contained a single β-2,6 linked fructan series. All other species grown at 10/5°C day/night temperatures (16 h day) contained some β-1,2 linked oligofructans, even when the major fructan series was β-2,6 linked. Leaves of P. ampla plants grown at 15/10°C day/night temperatures contained some β-1,2 linked fructans and raffinose in addition to the dominant β-2,6 series, but did not contain significant amounts of 1-kestose or bifurcose (1&6-kestotetraose). Leaves of Poa arctica R.Br. and Poa bulbosa L. contained significant amounts of raffinose and various β-1,2 linked oligomers but no 1-kestose. In vivo synthesis of 6-kestose in leaves of P. ampla grown at 10/5°C proceeds in the absence of significant amounts of either 1-kestose or bifurcose. Thus P. ampla probably has a specific 6-SST (6-sucrose-sucrosefructosyl transferase) and might provide a ready source of β-2,6 linked grass fructan.

Thioredoxins are ubiquitous small-molecular-weight proteins (typically 100–120 amino-acid residues) containing an extremely reactive disulphide bridge with a highly conserved sequence -Cys-Gly(Ala/Pro)-Pro-Cys-. In bacteria and animal cells, thioredoxins participate in multiple reactions which require reduction of disulphide bonds on selected target proteins/enzymes. There is now ample biochemical evidence that thioredoxins exert very specific functions in plants, the best documented being the redox regulation of chloroplast enzymes. Another area in which thioredoxins are believed to play a prominent role is in reserve protein mobilization during the process of germination. It has been discovered that thioredoxins constitute a large multigene family in plants with different subcellular localizations, a unique feature in living cells so far. Evolutionary studies based on these molecules will be discussed, as well as the available biochemical and genetic evidence related to their functions in plant cells. Eukaryotic photosynthetic plant cells are also unique in that they possess two different reducing systems, one extrachloroplastic dependent on NADPH as an electron donor, and the other one chloroplastic, dependent on photoreduced ferredoxin. This review will examine in detail the latest progresses in the area of thioredoxin structural biology in plants, this protein being an excellent model for this purpose. The structural features of the reducing enzymes ferredoxin thioredoxin reductase and NADPH thioredoxin reductase will also be described. The properties of the target enzymes known so far in plants will be detailed with special emphasis on the structural features which make them redox regulatory. Based on sequence analysis, evidence will be presented that redox regulation of enzymes of the biosynthetic pathways first appeared in cyanobacteria possibly as a way to cope with the oxidants produced by oxygenic photosynthesis. It became more elaborate in the chloroplasts of higher plants where a co-ordinated functioning of the chloroplastic and extra chloroplastic metabolisms is required.

More than 200 angiosperms, distributed in 25 genera, develop root nodule symbioses (actinorhizas) with soil bacteria of the actinomycetous genus Frankia. Although most soils studied contain infective Frankia, cultured strains are available only after isolation from root nodules. Frankia infects roots via root hairs in some hosts or via intercellular penetration in others. The nodule originates in the pericycle. The number of nodules in Alnus is determined by the plant in an autoregulated process that, in turn, is modulated by nutrients such as nitrogen and phosphate. Except in the genera Allocausarina and Casuarina, Frankia in nodules develops so-called vesicles where nitrogenase is localized. Sporulation of Frankia occurs in some symbioses. As a group, actinorhizal plants show a large range of anatomical and biochemical adaptations in order to balance the oxygen tension near nitrogenase. In symbioses with well aerated nodule tissue like Alnus, the vesicles have a multilayered envelope composed mainly of lipids, bacterio-hopanetetrol and their derivatives. This envelope is assumed to retard the diffusion of oxygen into the nitrogenase-containing vesicle. In symbioses like Casuarina, the infected plant cells themselves, rather than Frankia, appear to retard oxygen diffusion, and high concentrations of haemoglobin indicate an infected region with a low oxygen tension. At least in Alnus spp., ammonia resulting from N2 fixation is assimilated by glutamine synthetase in the plant. The carbon compound(s) used by Frankia in nodules is not yet known. Nitrogenase activity decreases in response to a number of environmental factors but recovers upon return to normal conditions. This dynamism in nitrogenase activity is often explained by loss and recovery of active nitrogenase and has been traced to loss and recovery of the nitrogenase proteins themselves. Recovery is partly due to growth of Frankia and to development of new vesicles in the Alnus nodules. In the field, varying conditions continuously affect the plants and the measured rate of N2 fixation is a result not only of the conditions prevailing at the moment but also of the conditions experienced over preceding days. N2 fixed by actinorhizal plants is substantial and actinorhizal plants have great potential in soil reclamation and in various types of forestry. Several species are also useful in horticulture.

Sequential harvests of cotton seedlings grown in soil cores enabled the quantification of the density of arbuscular mycorrhizal fungi to detect the effects of time, cultivation and periodic wetting of the soil. Cotton seedlings grown in soil cores from three locations formed arbuscular mycorrhizas at similar rates when cores were stored dry for up to 18 months. Disturbance of dry cores followed by dry storage for 18 months did not reduce the rate of establishment of mycorrhizas. Periodic wetting and drying of the cores, especially if the cores had first been disturbed, significantly reduced the rate of establishment of mycorrhizas. We suggest that long fallow disorder is possibly caused by falls of rain in clay soils of eastern Australia used to grow cotton. The proportion of the root with mycorrhizas at 3 wk was strongly correlated with the infection at 8 wk. We also suggest that it might be possible to predict maximum levels of infection and early uptake of phosphate of seedlings by determining the proportion of roots that are mycorrhizal 3 wk after emergence of cotton seedlings.

The developmental regulation and organ-specific expression of shikimate dehydrogenase was studied in seedlings of Capsicum annuum L. during their early development. The results obtained suggest that there is both a developmental regulation and an organ-specific expression of shikimate dehydrogenase in seedlings of C. annuum, seen mainly in the cotyledons. For example, shikimate dehydrogenase isoenzyme 3 was differentially expressed in cotyledons, but not at all in roots or hypocotyls. When the amounts of shikimate dehydrogenase were compared with the end-products of the shikimate pathway, chlorogenic acid and lignins, it was observed that the fall in shikimate dehydrogenase activity in cotyledons coincided with a decrease in the concentrations of phenolics during primary leaf development. However, this did not seem to be the case with the hypocotyls, where the fall in phenolics was not related to the amount of shikimate dehydrogenase but rather to an increase in lignin deposition, supporting the view that chlorogenic acid might act as a precursor of the aromatic moiety of cinnamoyl alcohols during cell wall lignification.

To test the hypothesis that fructan structure can be used as a taxonomic character for the Poaceae we examined the accumulation of the linear, β-2,6-linked fructan series with a terminal glucose residue within the tribes Aveneae and Poeae. Only Dactylis glomerata L. (Poeae) has been shown to contain this series, making it unique among fructan structures and a potential taxonomic marker. To this end, 13 members of the tribes Aveneae and Poeae were surveyed for the presence of this fructan series by comparing extracts of water-soluble carbohydrate by TLC with an extract of D. glomerata. The fructans of four species were similar to those of D. glomerata, as determined by TLC. Further comparison by anion-exchange chromatography and linkage analysis demonstrated that Phalaris aquatica L., Puccinellia stricta (Hook. f.) C. H. Blom and Holcus lanatus L. contained a linear, β-2,6-linked series of fructan with a terminal glucose residue but that Lagurus ovatus L., although containing predominantly β-2,6-linked fructan was composed of more than one series of fructan, one with a terminal glucose residue and one with an internal glucose residue. A linear β-2,6-linked series of fructans with a terminal glucose residue was the dominant fructan accumulated by some species of genera belonging to both the Aveneae (Phalaris and Holcus) and the Poeae (Dactylis and Puccinellia). As both of these tribes also contain genera that accumulate fructans with an internal glucose residue, e.g. Avena and Lagurus (Aveneae) and Lolium (Poeae), structure of fructan cannot be used to distinguish these tribes. Review of the literature, however, showed that it is possible to separate the supertribes Triticodae and Poodae on the basis of the structure of the fructan that they accumulate.

Observations were made on the vertical distribution of colonies of Aphanizomenon flos-aquae for 9 d at a drift-station east of Bornholm Island in the Baltic Sea. The buoyant colonies were dispersed in the upper layers of the water column during periods of wind-induced mixing but floated up during calm periods. From measurements of the vertical light extinction, surface irradiance and the photosynthesis versus irradiance curve, calculations were made of the changes in the daily integral of photosynthesis with respect to time and depth throughout the water column. From these calculations it is demonstrated that net photosynthesis by the population of Aphanizomenon flos-aquae increased nearly threefold by floating up after a deep mixing event. It is estimated that, averaged over alternating periods of calm and mixing, the buoyancy provided by gas vesicles in this organism will result in a nearly twofold increase in photosynthesis. A quantitative analysis has been made of the relationship of the daily integral of photosynthesis by the Aphanizomenon population with the mean depth of the population in the water column and the daily insolation. The analysis shows that the integral decreases linearly with respect to mean depth.

The daily integral of photosynthesis for a population of planktonic cyanobacteria in the Baltic Sea is calculated by numerical analysis based on the photosynthesis–irradiance curve of the organism, the vertical distribution of the population and details of the underwater light field. Variations in the light field are calculated from continuous recordings of surface irradiance and measurements of vertical light attenuation, with corrections for losses by reflection at the water surface that depend on the sun's elevation and roughening by wind. Integrals calculated using models in which simplifying assumptions are made about the surface irradiance overestimate the calculated value by 21% on a cloudless day and by 75%, or more, on an overcast day. Even larger discrepancies (>100%) result when the population, which is concentrated in near-surface layers, is assumed to be evenly distributed with depth. Errors of up to 5% result from neglect of surface reflection, and up to 12% from neglect of changes in phytoplankton distribution over 24 h. Details are given of the method of numerical analysis, using standard computer spreadsheets. Versions of the spreadsheet are described in which effects of changes in the vertical distribution of the phytoplankton and light attenuation can be calculated and the effects of water temperature and photoinhibition can be modelled. The method will improve estimates of primary productivity in natural waters. It can also be used for calculating values of the euphotic depth integrated over 24 h. The spreadsheets are available on http://www.bio.bris.ac.uk/research/walsby/integral.htm.

Safflower (Carthamus tinctorius L. cv. S555) and spring wheat (Triticum aestivum L. cv. Anza) were grown with or without the arbuscular mycorrhizal fungus Glomus etunicatum Becker & Gerd., under environmentally controlled conditions. Soil phosphate concentrations were adjusted before planting to produce mycorrhizal (M) and non-mycorrhizal (NM) plants that had similar leaf areas and root length densities at the same stage of development before initiating drought stress treatments. Drought did not affect the amount of mycorrhizal infection in safflower or wheat. Interactions between water stress treatments and mycorrhizal infection on plant growth and phosphorus uptake were limited and only occurred in wheat. NM wheat plants had 28% greater shoot d. wt, slightly greater root length densities, and 39% greater P acquisition than M plants when grown under well watered conditions, but under droughted conditions plant size and tissue P contents of M and NM wheat plants were similar. Mycorrhizas did not affect stomatal behaviour during drought stress in either safflower or wheat, i.e., transpiration and stomatal conductance declined independently of infection as soil water was depleted and leaf water potentials declined. Therefore, mycorrhizal infection did not alter the intrinsic hydraulic properties of the plant/soil system. Whilst wheat maintained turgor of recently expanded leaves during severe drought and safflower did not, mycorrhizal infection had no effect on leaf turgor during drought in either plant species.

Fructans are linear or branched polymers containing a single sucrose and repeating fructose residues. An early model for fructan biosynthesis in higher plants suggested that partial synthesis of the polymer occurred in the cell cytosol. The current model suggests that synthesis requires the interaction of two separate fructosyltransferases located in the vacuole. Tobacco lines containing a chemically induced promoter, directing expression of the Bacillus amyloliquefaciens SacB gene in the present study, provided an opportunity to regulate and target fructan synthesis to the cytosol of transgenic plants. Induced expression of the gene led to rapid destruction of leaf tissue. Amino acid substitution at a highly conserved site (Arg331) in the SacB gene reduced the fructosyltransferase efficiency without reducing the invertase activity of the enzyme. Expression of the mutant gene in transgenic tobacco also resulted in leaf damage. However, the appearance of necrotic tissue was greatly delayed. The results suggest that the phenotype is due to accumulation of fructan in the cytosol. Fructan metabolism in the cytosol of potato tubers was also detrimental to development. Tuber size and starch synthesis was significantly reduced in lines containing the untargeted gene. Transgenic tobacco and potato containing the SacB gene offer an opportunity to study the metabolism of fructan and the effect of accumulation on plant cell development.

In contrast to that of Anabaena sp. PCC 7120, the fdxN gene in Anabaena variabilis ATCC 29413 is not interrupted by a 55-kb DNA element, making this strain more suitable for genetic analysis of fdxN independent of the developmentally regulated excision during heterocyst formation. As a basis for mutational analysis, the fdxN gene of A. variabilis was cloned and sequenced. The deduced FdxN protein sequence was highly homologous to the Anabaena 7120 fdxN gene product including eight cysteine residues that are known to be conserved among ferredoxins containing two [4Fe-4S] clusters. The fdxN gene of A. variabilis was disrupted by insertion of an interposon within the fdxN coding region resulting in mutant strain KG29. Diazotrophic growth and in vivo nitrogenase activity of KG29 were similar to those of the wild-type, indicating that FdxN was not essential for N2 fixation in A. variabilis.

Fructans (polyfructosylsucrose) are synthesized by a number of plants and micro-organisms. Plant fructans are localized in the vacuole and have a low degree of polymerization (DP), whereas the fructans synthesized by micro-organisms are usually much bigger. There is an increasing interest in fructans for both food and non-food applications. In order to accumulate fructans of high DP in the plant vacuole, the levansucrase protein of Bacillus subtilis was fused to the vacuolar targeting sequence of sporamin and expressed in plants. Transgenic tobacco plants in which this fusion gene is expressed accumulate fructans to levels up to 21% of the d. wt. They showed a reduced translocation of carbohydrates, bleaching of the leaves, stunted growth and increased levels of hexoses and starch. The levansucrase protein was not translocated to the plant vacuole, but retained in the endomembrane system, even though the same targeting signal was able to translocate the E. coli GUS protein to the plant vacuole.

The objective of this study was to determine if infection by arbuscular mycorrhizal fungi alters water uptake by roots under well watered to severely droughted conditions. Safflower and wheat plants were grown with and without the mycorrhizal fungi, Glomus etunicatum or G. intraradices in nutrient-amended soil under environmentally controlled conditions to yield mycorrhizal and non-mycorrhizal plants with similar leaf areas, root length densities, d. wt, and adequate tissue phosphorus and nitrogen. Specific water uptake rates (cm3 of water cm−1 root length d−1) were estimated non-destructively at various depths in the soil from changes in the soil water content measured using a gamma attenuation method. When soil water was severely depleted, changes in soil water potentials were also measured with soil psychrometers. Roots from both plant species extracted water at the fastest rate from the upper soil layers when the soil water content was high, and later, extracted water primarily from deeper depths as water in the upper soil layers was depleted. Mycorrhizal infection did not affect the rates at which roots extracted water from soil whether soil moisture conditions were at their wettest condition, at container capacity, or at the driest extreme when soil water potentials ranged from −1·5 to −2·0 MPa and the plants were completely wilted. Plant water relations were also largely unaffected by infection. Mycorrhizal infection did not alter the ability of plants to extract water from soil even during extreme drought.

To study the relationship between copper sensitivity and membrane permeability in Arabidopsis, the seedling growth rates (measured as root extension and f. wt increase) of 10 ecotypes with differing copper sensitivities were compared with their patterns of potassium efflux and copper uptake. At 36 h, inhibition of root extension in 40 μM Cu correlated well with K+ efflux (r=0·96) and a decrease in osmolality (r=0·96). By contrast, neither tissue osmolality nor K+ efflux correlated significantly with growth inhibition after 4 h Cu treatment. In detailed time-course studies of two of the less sensitive ecotypes (Shadhara and Ws) and two of the more sensitive ecotypes (Berkeley and Limeport) in 20 μM Cu, growth, K+ retention and osmolality decreased rapidly during the first 4 h. Only the two less sensitive ecotypes recovered during the following 32 h. A similar pattern was observed with 40 μM Cu, except that growth inhibition was more marked. Interestingly, after 4 h the less sensitive ecotypes exhibited greater inhibition than the more sensitive. In time-course measurements, the Cu contents of the four ecotypes were not significantly different when normalised on a f. wt basis. We conclude that variations in the metal sensitivities of Arabidopsis ecotypes are not due to constitutive differences in membrane permeability but, rather, to differences in the ability to reverse K+ efflux. The possibility that short-term Cu-induced K+-leakage is regulated by channels rather than lipid oxidation is discussed.

Changes in leaf water content, night-time accumulation of malic (Δ-malate) and citric acid (Δ-citrate) and phosphoenolpyruvate carboxylase (PEPC, EC 4 . 1 . 1 . 31) activity were followed for 60 d after germination in well watered and salt-stressed plants of the facultatively halophytic ephemeral Mesembryanthemum crystallinum L. To separate the effects of development, salt stress and water deficit on crassulacean acid metabolism (CAM) induction plants were stressed initially 10 d after germination and then successively at 1-wk intervals (five sets). Related to dry mass or organic matter (i.e. dry mass corrected for the mass of inorganic ions) water content started to decrease during the late embryonal phase of the life cycle. Water content on a dry mass basis was always lower in salt-stressed than in well watered individuals. However, on an organic matter basis no difference was detectable. This indicated that salt treatment did not reduce leaf water content but falsified the basis (dry mass). Increases in leaf succulence and in pressure potential prevented long-term water deficit in well watered and in salt-stressed plants. Instead, these changes displayed enhanced vacuolisation, which is an essential prerequisite for the development of CAM. The end of that differentiation process might allow the initiation of nocturnal malic acid accumulation in a threshold response. At the onset of each salt treatment, short-term water deficits occurred due to an incomplete osmotic adaptation independent of plant age. As Δ-malate only appeared when plants were c. 35 d old this water deficit was unlikely to be a decisive CAM-inducing factor. About 2 wk after germination water content began to decline during the light periods in plants of all treatments. This pattern disappeared again when CAM had been fully established. Daytime transpirational water loss is therefore unlikely to be the decisive factor because it failed to induce the metabolic shift in young plants. Environmental stress (e.g. salt or drought) can therefore only induce Δ-malate when leaf and plant differentiation has reached a certain stage.

The resurrection plant Chamaegigas intrepidus Dinter (Scrophulariaceae) grows as a typical hydrophyte in shallow rock pools on granitic outcrops in arid areas of Namibia. During the rainy season, the rock pools are temporarily filled with water. When the pools dry up, C. intrepidus desiccates and survives in an air-dry condition for at least 8 months. After rewatering, the plants regain their metabolic activity in under 2 h. The desiccation of the vegetative organs is accompanied by a dramatic accumulation of abscisic acid (ABA). Beyond this, desiccation of roots is accompanied by the occurrence of specific dehydration-related proteins, whereas the leaves of C. intrepidus show high levels of dehydrins in the dehydrated as well as in the hydrated state.

Investigations in Namibia showed drastic diurnal fluctuations in the pH of the rock pools. The pH value increased from slightly acidic or neutral conditions during the morning to alkaline conditions (up to pH 12) during late afternoon. Since compartmental ABA distribution depends strongly on pH gradients across membranes, the external pH would be expected to affect the ABA relations in the plant. According to the anion trap concept, an alkaline pH in the surrounding medium should cause a release of ABA from the roots, although C. intrepidus appeared to release less ABA than the terrestrial rosettes of Valerianella locusta.

Sucrose[ratio ]sucrose fructosyltransferase (SST) activity was partially purified from whole shoots of Lolium rigidum by a combination of affinity chromatography, gel filtration and anion-exchange chromatography. The SST activity co-eluted with some fructan[ratio ]fructan fructosyltransferase (FFT) and invertase activities and consequently the partially purified preparation was termed the fructosyltransferase (FT) preparation.

The SST-like activity in the FT preparation was purified 214-fold and had an apparent molecular mass of 84000. The FT preparation contained several peptides with an apparent pI of 4·6–4·7. When assayed with sucrose concentrations up to 600 mM, the FT preparation synthesized 1-kestose at all concentrations, and synthesized 6-kestose at concentrations of 150 mM and greater. The Km of 1-kestose production was 0·2 M. When the FT preparation was assayed at a concentration of activity approximately half that measured in fresh tissue with 100 mM sucrose, 1-kestose, or 6G-kestose as substrates, fructans of degree of polymerization (DP) [les ]5 were synthesized.

A partially purified FFT activity, free of SST and invertase activities, which synthesized β-2,1-glycosidic linked oligofructans of DP [les ]6, was combined in vitro with the FT preparation (FFT-FT preparation) to give a ratio of SST[ratio ]FFT activities similar to that measured in crude enzyme extracts from L. rigidum. The FFT-FT preparation synthesized oligofructans when assayed with 100 mM concentrations of sucrose, 1-kestose or 6G-kestose, but was not able to synthesize fructans of DP [ges ]6 even after extended assays of up to 10 h. The FFT-FT preparation was also assayed with 100 mM sucrose with small amounts of concentrated sucrose added periodically during the assay to maintain the substrate concentration. In this assay, the FFT-FT preparation synthesized fructans up to an apparent DP of 17 or greater. The fructans of DP [ges ]6 synthesized in the assay appeared to form two molecular series containing both β-2,1- and β-2,6-glycosidic linked fructosyl residues with terminal or internal glucosyl residues. The apparent rate of SST activity in the assay of the FFT-FT preparation was greater than that measured in a similar assay of the FT preparation alone which did not result in fructans with DP [ges ]6. It was concluded that the FFT-FT preparation, when assayed with a continual supply of sucrose, contained a factor which promoted synthesis of fructans of DP [ges ]6 and synthesis of β-2,6-glycosidic linkages between fructosyl residues.

Spontaneous mutation of the cyanobacterium Nostoc muscorum to a caesium resistant (Cs+-R) phenotype resulted in severe impairment of its nitrogenase activity, oxygenic photosynthesis, chlorophyll a content and osmotolerance. Among alkali cations only Cs+ or Rb+ restored these physiological processes to an almost normal level. Parent and mutant were similar with respect to Cs+ or Rb+ uptake and accumulation and also to regulation by NH4+. The ions Na+ or K+, at much higher concentration, significantly influenced Cs+ uptake and accumulation in both the mutant and the parent. These findings have important implications for cyanobacteria growing in Cs+-polluted habitats.

The ability of diffusive gas transport and pressurized, convective flow to satisfy internal oxygen demands was examined for an aquatic sedge, Eleocharis sphacelata R. Br. Resistances to convection and diffusion through the plant were quantified from anatomical studies of the airspace dimensions, and these were used in mathematical models to calculate the fluxes required to satisfy oxygen demands measured in the tissue. The greatest resistance to diffusion in the underwater tissue was the submerged culm between the waterline and sediment surface (1560 Ms m−3 per m culm length). Resistances of the nodal intercalary meristem (52 Ms m−3) and rhizome internode (34 Ms m−3) were minor. In contrast, resistances to convection were low in the culms (38 MPa s m−3 per m culm length), and higher in the nodal meristems (93 MPa s m−3). The rhizome internodes had large cortical canals with a low convective resistance (0·75 MPa s m−3), and a parallel spongy pith with a very high resistance (518 MPa s m−3) that is probably short-circuited by convection. The resistance of the submerged culm means that diffusion is inadequate to satisfy oxygen demands in plants growing in >10 cm of water, and that convection is therefore essential in the natural habitat of this species (water to c. 2 m depth). However, a convective oxygen influx as low as 2·8×10−8 mol s−1 per culm (equivalent to a gas flow rate of 3 μl s −1 per culm) could satisfy the entire oxygen demand of the underwater tissue; this value is well below actual rates. At this flow rate, the spongy pith in the rhizome would also remain aerobic: it has a low resistance to diffusion (73 Ms m−3) and could receive sufficient oxygen by diffusion from the node. The data agree well with previous empirical measurements of convection in this species and show that diffusion and convection are both important processes for its aeration.