Salt tolerance in Eucalyptus spp.: identity and response of putative osmolytes



    1. School of Plant Biology, Faculty of Natural and Agricultural Sciences, The University of Western Australia, Crawley, WA 6009, Australia,
    2. Centre of Excellence in Natural Resource Management, The University of Western Australia, Crawley, WA 6009, Australia and
    Search for more papers by this author

    1. Institute of Ecology and Conservation Biology, University of Vienna, Althanstrasse 14, A-1090 Vienna, Austria
    Search for more papers by this author

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

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

M. A. Adams, Centre of Excellence in Natural Resource Management, The University of Western Australia, 35 Stirling Highway, Crawley, WA 6009, Australia. Fax: + 61 86488 1001; e-mail:


In four species of salt-tolerant eucalypts (Eucalyptus raveretiana, E. spathulata, E. sargentii and E. loxophleba), we found substantial concentrations of quercitol – a cyclitol known for its accumulation in seeds of Quercus. Quercitol was absent in old foliage of E. globulus, a species noted for greater susceptibility to salinity, and also absent in the moderately tolerant E. camaldulensis, but, relative to other species, both had higher foliar concentrations of inositol. Simple sugars and cyclitols accumulated to osmotically significant concentrations in all species. The osmotic potential of expressed sap was always less than that of the external ‘soil’ solution and increasing salinity produced predictable reductions in growth and increases in ion concentrations in foliage of saplings of four eucalypt species. The more salt-tolerant species, E. spathulata, E. loxophleba and E. sargentii, were able to maintain well-regulated leaf Na+ concentrations even at 300 mol m−3 NaCl. These more salt-tolerant species also showed an apparent increase in net selectivity for K+ over Na+ as salinity increased, irrespective of the Na+ : Ca2+ ratio of the external medium (range 25 : 1 to 75 : 1; Ca2+ always ≥ 4.0 mol m−3). By contrast, E. globulus was unable to exclude Na+ when exposed to higher NaCl concentrations (e.g. 200 and 300 mol m−3). Carbon isotope signatures of foliage reflected imposed salinity but were not strongly enough correlated with growth to support previous suggestions that isotope discrimination be a means of evaluating salt tolerance. On the other hand, patterns of sugar and cyclitol accumulation should be further explored in eucalypts as traits contributing to salt tolerance, and with potential use as markers in breeding programmes.


In beginning a recent synthesis, Hare, Cress & Van Staden (1998) suggested: ‘many plants accumulate organic osmolytes in response to the imposition of environmental stresses that cause cellular dehydration’. However, these authors also concluded, in part, that we still do not fully appreciate the functional significance of osmolyte accumulation and that further work is needed to improve plant productivity in a constantly changing environment.

Amino acids, sugars, polyols and quaternary amines are the most widely examined classes of putative organic osmolytes in plants, and many of these compounds are found in appreciable concentrations in the tissues of woody genera (e.g. Pinus, Larix, Fraxinus, Acer, Olea, Quercus, Sequoia, Eucalyptus, Casuarina; Nguyen & Lamant 1988; Gucci, Lombardini & Tattini 1997; Popp et al. 1997). For example, accumulations of organic osmolytes in response to drought and salinity have been studied in selected species of Populus. Clones of Populus deltoides showed variable osmotic adjustment and accumulation of a range of low molecular weight carbohydrates and other potential osmolytes when grown with varying water availability in either glasshouse or field trials (Gebre, Kuhns & Brandle 1994; Gebre et al. 1998). The phreatophytic but desert-dwelling Populus euphratica accumulated cyclitols chiro-inositol and pinitol to significant concentrations when growing over saline groundwater in the Taklamakan desert in China (Arndt et al. 2004). Osmotic potential can account for a considerable proportion of total water potential in foliage of hardwoods (e.g. Tschaplinski, Gebre & Shirshac 1998) and the relative contributions of inorganic and organic solutes is the subject of continuing research (e.g. Patakas et al. 2002; Gebre et al. 1998).

In the Myrtaceae, a variety of sugars and nitrogenous compounds such as proline and betaines are found in foliage and are putative osmolytes. High concentrations of salt in the rooting medium lead to accumulations of: (1) nitrogenous solutes in Melaleuca spp. (e.g. Naidu et al. 1987; Poljakoff-Mayber et al. 1987; Naidu, Paleg & Jones 2000); and (2) a more broad range of compatible solutes in some eucalypts (Prat & Fathi-Ettai 1990). Foliage of mature Eucalyptus tereticornis growing in salt-affected eastern Victoria was enriched in proline ( Marsh & Adams 1995), especially the youngest fully expanded leaves. Direct measurement of the osmotic potential of expressed sap of eucalypts (e.g. Myers & Neales 1986; Ngugi et al. 2003) or indirect measurement of cellular osmotic potential via pressure-volume curves (e.g. Clayton-Greene 1983; White, Turner & Galbraith. 2000; Ngugi et al. 2003) indicate that osmotic potential is a significant component of overall water potential. Despite the general acknowledgement of the need for accumulation of compatible solutes in the cytoplasm to counter accumulation of monovalent ions in vacuoles (thus contributing to osmotic balance between compartments as well as to the overall turgor maintenance), we know little about the identity of such solutes for eucalypts. Moreover, as reported by Niknam & McComb (2000), the data available on salt accumulation or exclusion by eucalypts is restricted to a handful of species and to a wide range of different experimental or observational studies.

Eucalyptus is one of the largest and most diverse of the tree genera. Diversity of the genus has led to commercial exploitation, and eucalypts are now widely grown in plantations, mainly for pulp production, in more than 20 countries (Doughty 2000). The genus is also planted for a variety of other land protection purposes, including salinity control, in many more. Within Australia, there has been substantial recent expansion of eucalypt plantations, concentrated mainly in medium to high rainfall regions (> 700 mm annual rainfall). A more scattered development of smaller-scale plantings for remediation of environmental problems has increased the overall tree cover in lower rainfall areas (400–700 mm per year). Dry-land salinity already affects more than four million hectares of formerly productive agricultural land in mid- to lower-rainfall regions (e.g. Cullen 2003; Peck & Hatton 2003) and native trees are an integral component of most posed solutions. Given the diversity of the genus Eucalyptus and the world-wide commercial and environmental interest, the lack of knowledge of putative osmolytes is all the more surprising considering their importance to maintenance of metabolism and thus growth under drought or saline conditions.

More generally, the roles of organic osmolytes are still being vigorously debated, especially in relation to crop yield in water-limited environments (e.g. Serraj & Sinclair 2002). In a fashion similar to Hare et al. (1998) and after analysis of a large body of work, Hasegawa et al. (2000) suggested that ‘the purely osmotic contribution of these metabolites to stress tolerance may not describe their function completely . . .’ and that the ‘pathway leading to a particular osmolyte may be more important than accumulation per se . . .’. These authors provided several examples of other significant stress-relieving functions served by either the solute or the associated enzyme pathway; for example, in scavenging free radicals of oxygen.

Irrespective of the exact function of solutes or their pathway of synthesis, an important first step is identification and estimation of their significance to maintenance of turgor. Here we report on the identity in eucalypts of some putative osmolytes and their response to salinity in glasshouse and field experiments. The species examined were selected from the range of salt-tolerant and non-tolerant species (e.g. Niknam & McComb 2000). From a range of possible osmolytes, we found highly significant concentrations of cyclic and acyclic polyols (in addition to reducing sugars). One polyol, quercitol, has previously been reported in a single eucalypt species (Plouvier 1963), but is more well known for its accumulation in acorns of the genus Quercus. The present findings on organic osmolytes in eucalypts from field and glasshouse experiments, together with a range of further observations in relation to water use efficiency (δ13C), growth, and ion selectivity, are discussed in relation to salt tolerance and the possibilities to use this knowledge in further studies of this large and diverse genus.


Salt-tolerance amongst eucalypts has been assessed through glasshouse and field trials (reviewed by Niknam & McComb 2000). For the glasshouse trial, we selected two species (Eucalyptus loxophleba and E. sargentii) from among the group of species that Niknam & McComb (2000) suggested show consistently good tolerance of salinity in glasshouse and field trials and two (Eucalyptus globulus and E. spathulata) that show variable tolerance. Of the latter two, E. globulus was widely ranked by authors as of low to medium tolerance, whereas E. spathulata was more regularly ranked in the medium to high tolerance class. All are members of the subgenus Symphyomyrtus.

Glasshouse experiment

The four species of Eucalyptus used in the experiment were supplied as seeds by the Western Australian Department of Conservation and Land Management. Each species was grown from seed from a single provenance (and seed lot). Seeds were planted in potting mix in trays and raised for 7 months by a commercial nursery near Waroona, Western Australia. The trays containing seedlings were transported to a glasshouse in Nedlands, Western Australia and maintained in the trays for a period of 2 weeks prior to transplanting into a sand culture system.

Thirty-seven seedlings of each species were transplanted, each seedling into an individual pot containing 17 kg of washed white sand over a 40-mm layer of washed gravel. The potting mix used in the seedling trays was washed from the roots of each plant immediately prior to it being transplanted. Each pot was flushed with 0.2-strength nutrient solution prior to, and immediately after transplanting. The surface of each pot was covered with a 20–30 mm layer of white plastic beads in order to minimize evaporative losses from the sand. The composition of the nutrient solution at full strength was (mol m−3): K+, 1.22; Ca2+, 4.0; NH4+, 0.2; Mg2+, 0.2; Na+, 0.1; NO3, 1.4; SO42–, 0.2; H2PO42–, 0.02; Cl, 8.0; and micronutrients (mmol m−3): Fe-EDTA, 20; B, 25; Mn, 2; Zn, 2; Ni, 1; Cu, 0.5, Mo, 0.5. The pH of the solution was adjusted to 6.0 using KOH. Each pot was flushed with fresh solution every 3 d, and the nutrient concentrations in the solution were increased by 0.2-strength every 2 weeks until the full concentration was reached (a period of 56 d). The plants were maintained under these conditions for a further 69 d to allow recovery from transplanting. All the seedlings survived this transplanting procedure.

Salinity treatments were imposed 125 d after transplanting and comprised of selected combinations of NaCl and CaCl2 added to the full strength nutrient solution. The treatments enabled two clear contrasts: (1) the NaCl dose–response between 0.1 and 300 mol m−3 at a constant Ca2+ supply of 4 mol m−3 (i.e. increasing Na+ : Ca2+ ratio in treatments, from 25 : 1 to 75 : 1); (2) the NaCl dose–response at a constant Na+ : Ca2+ ratio of 25 : 1(i.e. increasing concentrations of both Na+ and Ca2+). A constant a Na+ : Ca2+ ratio was suggested by Greenway & Munns (1980) as the most appropriate design for salinity dose–response experiments; yet direct comparisons of dose–responses of plants to salinity with fixed and varying ratios of Na+ : Ca2+ are rare. The most uniform plants in 28 pots of each species were selected and then four pots of each species were randomly assigned to be replicates in each of the six treatments or the initial harvest. NaCl treatments were imposed as 50 mol m−3 increments every 3 d, and the CaCl2 treatments were applied from the first day of treatments. Plants were grown at the final treatment concentrations for a period of 80–82 d during which time each pot was flushed with the appropriate fresh treatment solution every second day.

At both harvests, the intact shoot of each plant was rinsed with de-ionized water to remove surface salts at least 3 h prior to the harvest. Leaves were excised from the stem and separated into three age classes: expanding, young fully expanded, and older fully expanded leaves. Leaf tissues were immediately frozen in liquid N2 and then stored at −70 °C prior to freeze-drying. Freeze-dried leaf samples were weighed and then stored in sealed containers above desiccant at −20 °C prior to chemical analyses. The stem and twigs were also excised at the soil surface, the roots were washed from the sand, and these tissues were oven dried at 70 °C. The dry weights of these tissues were also determined. Leaf samples were ground to a fine powder using a ball-mill prior to subsamples being used in the various chemical analyses.

Shoot water potential was measured between 1130 and 1430 h on a small sample of foliage excised from the crown of each plant after 80 d of treatments. Excised twigs were wrapped immediately in plastic film and placed in a pressure chamber (PMS 600; PMS Instrument Company, Albany, OR, USA) for the measurement. At the same time, leaf discs were cut from a young fully expanded leaf on another intact twig using a 6-mm-diameter punch. The samples were immediately transferred into a cryo-vial and frozen in dry ice and then stored at −70 °C. Samples were thawed while in the sealed vials and sap was then expressed from the leaf discs using a stainless steel press. The sap was immediately transferred into a microfuge tube, centrifuged at 12 900 g for 120 s, and the osmotic potential of a 10-µL subsample was measured using a freezing point depression osmometer (Model One-Ten; Fiske Associates, Norwood, MA, USA).

Water use by the plants was measured during the week prior to the final harvest. Pots were watered with the nutrient solution and left to drain for 2 h, after which the weight of each pot was recorded. Each pot was then re-weighed 24 h later. Replicated pots without plants were also watered with nutrient solution containing 0.1 to 300 mol m−3 NaCl and the changes in weight of these pots were also determined in order to estimate water losses via evaporation. Plant water use was calculated from the differences in the weight of each pot over the 24-h period minus the change in weight of pots without plants. These measurements were repeated over two separate 24-h periods.

Field study

We sought general confirmation of the results from our glasshouse experiment through the use of field trials of lines/clones/hybrids of eucalypts developed and selected specifically for salt tolerance (e.g. McComb & Bennett 1986; McComb, Hardy & Dell 1996; Niknam & McComb 2000). Samples of foliage were collected from trees planted between 1991 and 1995 at a range of sites in south-western Australia. Sites are named after their locality and varied in soil salinity (measured using EM38; Geonics, Mississauga, Ontario, Canada) and data on salinity at the sites when foliage samples were taken are reported in the heading of Table 4. The salinity associated with localities in this region is well described (e.g. Schofield & Ruprecht 1989; Peck & Hatton 2003). At each site, samples were collected from at least two individual trees of each clone/hybrid/line present, together with the recording of soil salinity adjacent to each tree using the EM38 device. Two clones of salt-tolerant E. camaldulensis were at three of the sites, whereas the other species were only present at one or two sites. Foliage samples were collected in March/April 2000 and were immediately frozen and processed as described above for foliage collected from the glasshouse experiment. All samples were analysed as described below.

Table 4.  Concentrations of quercitol and total organic solutes (as listed in Table 3) expressed as mol m−3 leaf water basis and as a proportion (%) of leaf sap osmotic (ψo) potential for young fully expanded leaves (Fig. 6 and Table 2)
E. globulusE. loxophlebaE. sargentiiE. spathulata
  1. Data calculated from those shown in Fig. 6 and Table 2, using ψo = (n/v×RT and assuming ideal solutions.

 (mol m−3) 0.5 65.3129.6118.1
 (% leaf sap ψo) 0.07  7.7 14.6 12.7
Total measured organic solutes
 (mol m−3)79.1129.7253.3228.5
 (% leaf sap ψo)11.6 15.2 28.5 24.5

Chemical analyses

Inorganic ions

Na+, K+, and Ca2+ were extracted from 0.1 g subsamples of freeze-dried and ground leaf tissues by shaking in 10 mL of 0.5 mol m−3 HCl at room temperature for 2 d ( Hunt 1982). Na+ and K+ concentrations in appropriate dilutions of the extracts were measured using a flame photometer (Model 410; Corning; Sherwood Scientific Ltd., Cambridge, UK), while Ca2+ was determined using atomic absorption spectroscopy (AAnalyst 300; Perkin Elmer, Wellesley, MA, USA). Cl was extracted from a 0.05 g subsample of tissue in 10 mL of hot water (60 °C for 3 h). Cl was determined using a chloridometer (Buchler-Cotlove model; Labconco Corporation, Kansas City, MO, USA).

Organic solutes

We used a modification of ‘hot water extraction’ for extracting organic solutes from eucalypt foliage. Several methods were tested and methanol–chloroform–water (MCW) extraction was ultimately chosen as it provided a means of removing co-extracted, water-soluble resins and gums (results not shown) that interfered with gas chromatographic (GC) analysis. Trimethylsilyl derivatives of sugars and sugar alcohols in purified extracts (prepared as described below), provided sharp peaks and clear resolution under chromatographic conditions similar to those described by Hoch, Richter & Körner (2003). GC-mass spectroscopy was additionally used to confirm the identity of pinitol and quercitol in gas chromatograms. Quercitol standards are not commercially available and were instead prepared from purified extracts of seeds of Quercus robur.

The extracts were prepared by shaking 40 mg samples of freeze-dried and ground foliage with 1 mL of hot (75 °C) MCW (12 : 5 : 3 v/v) for 30 min in an Eppendorf tube. After centrifugation (15 400 g, 1 min), 800 µL supernatant was removed and mixed thoroughly with 500 µL water and 200 µL chloroform. After standing for 10 min, 700 µL of the alcohol/aqueous phase was removed and added to approximately 400 µL of mixed bed resin (Dowex 50 W H+/Dowex 1 × 8 formate; Dow Chemical Co., Pevely, MA, USA) and 50 µL of internal standard in an Eppendorf tube. The extract was then shaken with the resin for 2 h at room temperature, centrifuged, and 400 µL of purified extract was removed and stored at <−20 °C until being analysed, using the methods and gas chromatograph as described in Hoch et al. (2003). Internal standards were 0.4% solutions of penta-erythritol and xylitol.

Carbon-isotope analysis

Carbon isotope composition has been used as a potential means of evaluating salt tolerance (Poss et al. 2000). In this study, freeze-dried foliage samples were subsampled (5.5–6 mg dry mass) and combusted to CO2 and N2 in the presence of O2 (Roboprep-CN; Europa Scientific, Crewe, UK) before passing into a mass spectrometer (Tracermass; Europa Scientific). δ13C (in ‰ units) was calculated with respect to the PDB standard {=[(13C/12Csample)/(13C/12Cstandard) − 1] × 1000}.

Statistical analyses

Data were analysed using Statview and SuperAnova (Abacus Concepts Inc.) for analysis of variance, while Genstat V was used for all other analyses.


Glasshouse experiment

As expected, the highest concentrations of NaCl (200 and 300 mol m−3) reduced growth (Figs 1a & b) of all species relative to that recorded at the lowest concentration (0.1 mol m−3). Seedlings of E. globulus died at 300 mol m−3 when Ca2+ was not supplied at a constant ratio. Data for E. globulus are thus generally presented for the 0.1, 100 and 200 mol m−3 NaCl treatments. The statistical significance of species (excluding E. globulus) and NaCl treatments were very strong (P < 0.01; Appendix 1) for all of the dependent variables: dry masses of roots, stems and old fully expanded leaves, irrespective of whether or not external concentrations of Ca2+ and Na+ were maintained at a constant ratio. The effects of species and treatment were less strong, but generally still significant (P < 0.05; Appendix 1) for dry mass of young leaves (fully expanded or still expanding).

Figure 1.

Effects of imposed salinity (NaCl) treatments on growth of four eucalypt species with: (a) variable Na+ : Ca2+ ratios (25 : 1 to 75 : 1, with concentration of Ca2+ constant at 4 mol m−3) or (b) constant Na+ : Ca2+ ratios (25 : 1 in three highest treatments). Data are means of four replicates for each plant fraction, the entire bar height gives total dry mass, with the standard error for total dry mass also shown. Results of two-way anova of effects of treatment (T) and species (S) and the interaction of these two factors (T × S) are shown for total dry mass (NS, not significant, *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001) and for components are given in Appendix 1.

While root and stem dry mass of all species declined with increasing NaCl concentrations from 0.1 to 300 mol m−3, the intermediate concentration (100 mol m−3) produced a slight but significant (P < 0.05; Appendix 1) increase in leaf and total dry mass of E. spathulata over that recorded at the lowest and highest external NaCl concentrations. The effect of NaCl on growth of E. spathulata was not significantly influenced by the external concentration of Ca2+ (detailed statistical analyses not shown). When the Na+ : Ca2+ ratio was maintained at 25 : 1, growth of all other species at higher concentrations of NaCl was significantly reduced (compare Fig. 1a & b). Shoot : root ratios were larger at higher NaCl concentrations owing to the stronger reductions in root growth than in shoot growth with increasing NaCl concentrations.

In all species, Na+ concentrations in foliage rose highly significantly (Fig. 2, Appendix 2; P < 0.001) from < 10 µmol g−1 dry weight (DW) at the lowest concentration of applied NaCl, to > 200 to > 1000 µmol g−1 DW in plants exposed to higher NaCl treatments, depending on leaf age, species, and external NaCl concentration, but in each case irrespective of external Na+ : Ca2+ ratio. Leaf age was included as a further independent variable in statistical analyses and had a highly significant effect (Appendix 2) on Na+ concentrations that were generally least in the expanding leaves and variably greatest in younger or older fully expanded leaves (Fig. 2). Concentrations of Na+ in foliage of E. loxophleba continued to increase with increasing external NaCl, irrespective of the concentration of Ca2+ in the external solution (Fig. 2) and reached > 600 µmol g−1 dry mass in expanding leaves, and > 800 µmol g−1 DW in the younger- or older-fully expanded leaves from the 300 mol m−3 treatment. Concentrations of Na+ in foliage of E. sargentii and E. spathulata reached a plateau at between 200 and 400 µmol g−1 DW, irrespective of leaf age and external Ca2+ concentrations. The greatest Na+ concentrations were measured in expanding and younger fully expanded leaves of E. globulus but these were also highly variable and associated, in some instances, with tissue death.

Figure 2.

Foliar concentrations of Na+ in four eucalypt species exposed to NaCl treatments. Data presented are for variable (0.025, 25, 50, 75) and constant (25) Na+ : Ca2+ ratios in external solution (see text for details). Data points are means of four replicates with standard errors. Results of three-way anova of effects of treatment (T), species (S) and leaf age (A) are shown (NS, not significant, *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001, details given in Appendix 2). Only the T × S interaction was statistically significant.

Chloride concentrations in foliage of all four species grown in non-saline conditions (i.e. controls) were generally greater than those of Na+ (compare Figs 2 & 3). For example, in expanding leaves of control plants, Cl concentrations were 89–267 µmol g−1 dry mass, as compared with Na+ concentrations of < 10 µmol g−1 DW. The relatively high Cl in foliage of control plants presumably results from greater concentrations of Cl compared with Na+ in the basal nutrient solution (namely 8.0 versus 0.1 mol m−3), and from active absorption of Cl by roots under non-saline conditions (White & Broadley 2001). Leaf Cl concentrations increased in foliage of plants exposed to the NaCl treatments (increases were 1.4- to 3.4-fold in expanding leaves, depending on species, Fig. 3). In contrast to control plants, leaf Cl concentrations in plants treated with NaCl were, on average, 0.88 of those for Na+ (data from all species, leaf ages, external NaCl concentrations, and external Na+ : Ca2+ ratios, Figs 2 & 3).

Figure 3.

Foliar concentrations of Cl in four eucalypt species exposed to NaCl treatments. Data presented are for variable (0.025, 25, 50, 75) and constant (25) Na+ : Ca2+ ratios in external solution (see text for details). Data points are means of four replicates with standard errors. Results of three-way anova of effects of treatment (T), species (S) and leaf age (A) are shown (NS, not significant, *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001). Only the T × S interaction was statistically significant.

Potassium concentrations in foliage generally fell as external NaCl concentrations increased, and the effects of all of: species (P < 0.0001), treatment (P < 0.0001) and leaf age (P < 0.01) were highly significant (see Appendix 3). Calcium concentrations in foliage declined, quite uniformly among all four species, with increasing concentrations of applied NaCl (Appendix 3). Calcium concentrations in the three classes of leaves in all four species could be reliably predicted by polynomial equations (not shown; r2 = 0.91–0.99) on the basis of concentrations of Na+ in the external nutrient solution. The combined effect of the species-dependent increases in Na+ concentrations in leaves and the species-independent reductions in leaf Ca2+ concentrations with increasing external NaCl concentrations, were sharp increases in tissue Na+ : Ca2+ ratios. Using as an example, expanding leaves of plants at constant external Na+ : Ca2+ ratio, tissue Na+ : Ca2+ ratios rose sharply from less than 1 in at the lowest concentration of applied NaCl to between 6 (E. sargentii) and 10 (E. spathulata, E. loxophleba) at the highest concentration (Fig. 2 and Appendix 3).

An index of selective cation uptake can be derived from comparison of internal and external concentrations of the ions: (K+int/Na+int)/(K+ext/Na+ext; Pitman 1976). Calculation of this index for younger fully expanded leaves (Fig. 4a) or older, fully expanded leaves (Fig. 4b), clearly indicates increased apparent net selectivity of uptake for K+ as external concentrations of NaCl increased, despite variable Na+ : Ca2+ in the external medium. The effect was most pronounced for E. spathulata, E. loxophleba and E. sargentii and was similar when external Na+ : Ca2+ was held constant (data not shown).

Figure 4.

Apparent net K+/Na+ selectivity (see text for details) for (a) younger and (b) older fully expanded leaves of eucalypts grown with varying NaCl and varying Na+ : Ca2+ ratios in external solution. Calculated from mean values (n = 4) in Fig. 2 and Appendix 3.

After 73 d of treatment, water use by E. spathulata, E. loxophleba and E. globulus was inversely related to concentrations of applied NaCl (results not shown). Water use fell, from between 6 and 8 mL g−1 shoot dry mass d−1 at 0.1 mol m−3 NaCl, to around 4 mL g−1 shoot dry mass d−1 at 300 mol m−3 NaCl. Maximum water use by E. sargentii (6 mL g−1 shoot dry mass d−1) was recorded at 100 mol m−3, consistent with data for leaf mass (and area, data not shown). Carbon isotope discrimination (δ13C) in the three most salt-tolerant species (Table 1) followed a pattern closely related to that of growth (Fig. 1a & b) and water use (Fig. 5). The effects of leaf age and salinity treatment on δ13C were remarkably consistent and highly significant (P < 0.0001) across the three most salt-tolerant species. δ13C was also moderately well related to plant water use (Fig. 5). The relationship was stronger for younger fully expanded leaves than for older leaves (Fig. 5) and carbon isotope discrimination increased (δ13C became more negative) as leaves aged and decreased as salinity increased. Providing increased Ca2+ to maintain Na+ : Ca2+ at 25 : 1, partially reversed the effect of increased NaCl concentrations on δ13C, at least in the three most salt-tolerant species (Table 1). In E. spathulata, E. loxophleba and E. sargentii, δ13C was independent of foliar concentrations of Cl or Na+ or Ca2+ tested using regression analysis. In E. globulus, a weak but significant dependence of δ13C on foliar concentrations of Cl was evident (P < 0.0001, r2 = 0.138, detailed results not presented).

Table 1.  Effects of species, leaf age and salinity treatment on foliage δ13C
Species and leaf ageTreatment (Na+ and Ca2+ concentrations in culture solution, mol m−3)
  1. Eucalyptus globulus is not included owing to shortages of tissue for analysis (see text for details). Means and standard errors of four replicates are given and all values are negative. From three-way anova, the effect of treatment (T) was highly significant (P < 0.001) as was the effect of leaf age (A) (P < 0.001) and species (S) (P < 0.001). There were no significant interactions (A ×T, P > 0.8; A ×S, P > 0.7; T ×S, P > 0.05; A ×T ×S, P > 0.9).

E. loxophleba
 Young fully expanded29.00.428.80.327.
 Old fully expanded29.70.328.70.329.00.327.
E. sargentii
 Young fully expanded27.90.427.90.427.
 Old fully expanded28.50.328.40.427.80.427.20.327.90.427.00.3
E. spathulata
 Young fully expanded27.90.627.70.327.40.326.
 Old fully expanded28.60.728.30.728.40.327.50.328.
Figure 5.

Relationship of carbon isotope signatures (δ13C) in older and younger fully expanded leaves of three eucalypt species (E. loxophleba, E. sargentii, E. spathulata) to whole plant water use. Data from all treatments (0–300 mol m−3 NaCl at constant or varying Na+ : Ca2+, see text for details) are included. OFE, older fully expanded leaves; YFE, younger fully expanded leaves.

Shoot water potential (ψw) and the osmotic potential of sap (ψo) expressed from younger-fully expanded leaves followed the same general patterns (Fig. 6). Both ψw and ψo were at least 1 MPa less than the calculated potential of the applied treatment solutions throughout the range of treatments (Fig. 6). Within each treatment, there were only small ranges (approximately 0.5 MPa) in ψw and ψo among the three most salt-tolerant species; E. spathulata, E. loxophleba and E. sargentii. Eucalyptus globulus tended to maintain rather higher (less negative) ψw and ψo than the other three species. Although species differed in leaf water content (P < 0.0001), there was no effect of treatment on the water content of leaves (Table 2). However, water content varied significantly with leaf age (P < 0.01), but the patterns were not consistent amongst the species (P < 0.0001 for species–leaf age interaction, Table 2).

Figure 6.

Shoot water (ψw) and leaf osmotic (ψo) potential in four eucalypt species grown in a glasshouse with varying salinity and constant or variable Na+ : Ca2+ ratios (see text). Data are means of four replicates and standard errors shown. ψw was measured on twigs excised from the crown and ψo on sap expressed from freeze/thawed young fully expanded leaves. Summaries of two-way anova for treatment (T), species (S) and the interaction (T × S) are shown (NS, not significant, *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001).

Table 2.  Ratio of fresh weight to dry weight for foliage of a range of eucalypt species grown under varying salinity regimes
Leaf ageSpeciesSignificanceS × A
E. globulusE. loxophlebaE. sargentiiE. spathulata
  1. Means (x) and standard errors (SE) of four replicates are shown, and so are the significance (*P < 0.01, **P < 0.001, ***P < 0.0001) of the main effects (S, species; A, leaf age) and the S–A interaction, from anova.

Young fully expanded3.850.343.110.122.380.092.580.03   
Old fully expanded2.830.082.460.   

Owing to limited sample sizes, organic solutes were not measured in expanding leaves. Furthermore, since these analyses are labour-intensive only two replicates of the young fully expanded and older fully expand leaves were analysed for sugar and cyclitols. Preliminary analysis of variance of available data showed a highly significant effect of salinity treatment on concentrations of individual cyclitols (for quercitol, P generally < 0.01); increasing NaCl concentrations in the external medium produced increasing concentrations of sugars and of cyclitols (e.g. Fig. 7a). However, statistical analyses also revealed singularities owing to the absence of data (death of E. globulus at higher NaCl concentrations) and, in subsequent analyses, the data for different salinity treatments were pooled within species and leaf age. In the younger fully expanded leaves, concentrations of compatible solutes were least for E. globulus and greatest and similar for E. sargentii and E. spathulata(Table 3). With the exception of E. globulus, total concentrations in older, fully expanded leaves were less than in younger, fully expanded leaves. Among the specific sugars or cyclitols, only quercitol varied significantly (P < 0.01) between younger and older leaves (Table 3) and concentrations were always less in older leaves (usually by at least 30%). The effect of species was highly significant (P < 0.01) for each measured sugar or cyclitol. Most notable among the species differences were the absence (or very low concentrations) of quercitol in old foliage of E. globulus and the abundance of this solute in the other three species in which it comprised about half (115–186 µmol g−1 DW) of the total concentration of sugars plus cyclitols. In the three species with high concentrations of quercitol, it was present at 65–129 mol m−3 on a tissue water basis and on average this solute accounted for 8–15% of the measured leaf sap ψo(Table 4). Conversely, myo-inositol was relatively abundant in E. globulus vis-a-viz the other three species (Table 3). Fructose (25–75 µmol g−1 DW) and sucrose (60–100 µmol g−1 DW) were the two most dominant sugars in all species.

Figure 7.

(a) Concentrations of quercitol in young fully expanded foliage of three salt-tolerant eucalypts grown in the glasshouse with varying salinity and constant or variable Na+ : Ca2+ ratios (see text). (b) Relationship between soil salinity and concentrations of sugars plus cyclitols in fully expanded leaves of lines of Eucalyptus camaldulensis (for a complete list of solutes see Table 5). Samples were collected from lines selected for their apparent salt tolerance and planted on sites with varying soil salinity (Table 5). At least one pair of adjacent trees of each line or clone was sampled at each site. Soil salinity is given as uncorrected EM38 readings owing to within-site variation in soil properties and is as recorded within the proximity of each pair of sample trees (see text for details).

Table 3.  Mean concentrations (µmol g−1 DW) of sugars and cyclitols in foliage (young fully expanded or old fully expanded leaves) of four eucalypt species grown with salinity treatments as described in Table 1
CompoundLeaf ageSpeciesSignificanceS × A
E. globulusE. loxophlebaE. sargentiiE. spathulata
  1. Means (x) and standard errors (SE) of three independent analyses are shown together with the significance (*P < 0.05, **P < 0.01, ***P < 0.001) of the main effects (S, species; A, leaf age) and the S ×A interaction, from anova.

QuercitolYoung  1.3 0.7144.413.2175.517.9186.121.2*****NS
Old  0.0 0.0115.911.9114.614.6143.711.9   
myo-InositolYoung 27.8 5.6  3.3 0.6  2.8 0.6  3.3 0.6***NSNS
Old 25.0 3.9  5.0 1.1  4.4 1.1  3.9 0.6   
FructoseYoung 43.8 5.6 26.1 3.9 53.3 7.2 74.910.5***NSNS
Old 51.113.3 30.0 5.0 46.1 5.6 74.4 7.2   
GlucoseYoung 31.6 5.0 17.2 2.2 22.8 2.2 25.0 2.8***NSNS
Old 36.6 5.6 20.5 2.2 22.8 1.7 30.5 2.2   
SucroseYoung100.919.0 86.310.2 73.7 5.6 59.9 6.1**NSNS
Old 96.512.6 81.610.5 74.9 7.0 69.9 5.8   
RaffinoseYoung  8.7 1.0  8.7 0.7 12.8 1.3  7.9 1.0**NSNS
Old  9.3 1.2  8.2 0.7 10.1 1.0  7.2 0.8   
StachyoseYoung  0.5 0.3  0.2 0.2  2.0 0.5  3.0 0.5***NSNS
Old  0.8 0.5  0.5 0.3  1.8 0.6  1.8 0.3   

Field trial

Data from the field trial provided important confirmation of results from the glasshouse experiment, even though E. globulus was the only species from the glasshouse study also available at the field sites. The spectrum of compatible solutes found in each of the tested species from across all field sites is given in Table 5. For ease of comparison, data are summarized for each ‘species’ (including E. camaldulensis×E. globulus hybrids) at each site.

Table 5.  Concentrations (µmol g−1 DW) of soluble sugars and cyclitols in fully expanded leaves of Eucalyptus spp. growing on four sites in south-west Western Australia
CompoundE. camaldulensis
E. camaldulensis
E. camaldulensis
Mt Barker
Quercitol  0.0 0.0  0.0 0.0  0.0 0.0
Pinitol  5.7 0.5  6.2 1.0  8.8 1.0
myo-Inositol 33.3 2.8 33.3 1.7 40.0 4.4
Fructose 17.8 1.1 31.1 5.6 38.3 6.1
Glucose 16.7 1.1 21.1 2.2 25.0 2.2
Sucrose 66.4 6.4142.1 6.4155.0 8.8
Raffinose  2.7 0.3  2.5 0.3  6.6 1.3
CompoundE. globulus
Mt Barker
E. globulus
E. raveretiana
Quercitol  0.0 0.0 0.00.0 30.5 2.6
Pinitol  0.0 0.0 0.00.0 28.4 2.1
myo-Inositol 31.7 4.419.40.6 16.7 1.1
Fructose 17.8 58.3 7.2
Glucose 16.7 1.710.01.7 40.0 4.4
Sucrose 98.0 7.651.50.3 29.8 4.1
Raffinose  3.9 0.2 1.00.2  2.4 0.2
CompoundE. globulus × E. camaldulensis
PinjarraMt Barker
  1. Means (x) and standard errors (SE) are shown (n ≥4). Data for each species are presented with sites in order of increasing mean soil salinity (n = 10–28; average of vertical and horizontal measurements using EM38). Darkan (39.5 dS m−1) > Mt Barker (27.3 dS m−1) > Serpentine (21.7 dS m−1) > Pinjarra (4.6 dS m−1).

Quercitol  0.00.0  0.0 0.0
Pinitol  0.00.0  0.0 0.0
myo-Inositol 25.04.4 31.7 2.8
Fructose 13.90.6 18.3 3.3
Glucose 11.71.1 12.8 1.1
Sucrose 75.72.6137.4 7.6
Raffinose  1.70.2  5.9 0.8

Quercitol was again found in appreciable concentrations, but only in E. raveretiana where it comprised about 15% of the total pool of sugars and cyclitols. Pinitol and quercitol were absent from E. globulus and these cyclitols were also absent in the E. camaldulensis×E. globulus hybrid (Table 5). Total concentrations of soluble sugars and cyclitols ranged between 97 and 273 µmol g−1 dry mass, with the lowest values in E. globulus. With the exception of E. raveretiana, sucrose again dominated among the low molecular weight organics in all species (Table 5), comprising approximately 50% or more of the total pool of sugars and cyclitols.

As the numbers of lines or clones for each species varied, as did their presence at each of the sites, the capacity for analysis of the effects of salinity on compatible solutes (and on growth) is limited. We contrasted the concentration of compatible solutes in foliage with soil salinity for the species best represented in the field trials, E. camaldulensis. Accumulation of sugars and cyclitols by E. camaldulensis across all sites and for two clones, was strongly related (P < 0.01, r2 = 0.75) to soil salinity (Fig. 7b).


Quercitol has hitherto been found in appreciable concentrations in tissues of Quercus species; for example, at 1.7–3.3% of leaf dry mass (Popp et al. 1997). Our finding of high concentrations of quercitol (average values of 65–129 mol m−3 on a tissue water basis; contributing 8–15% of the measured leaf sap ψo) in the three most salt-tolerant eucalypts used in the glasshouse experiment (E. spathulata, E. loxophleba, E. sargentii) as well as in the highly salt-tolerant E. raveretiana (sensuNiknam & McComb 2000) in the field trial, coupled with its absence from other, less tolerant eucalypts in the present study (e.g. E. globulus), suggests that this organic osmolyte is of some significance for salt tolerance. Confirmation from analysis of foliage of field-grown plants (Table 5, Fig. 7a), should not be underestimated for, as demonstrated by Niknam & McComb (2000), there have been numerous examples of conflicting results in responses of eucalypts to salinity from field and glasshouse trials.

Although quercitol has previously only been documented at osmotically significant concentrations in Quercus species (Popp et al. 1997), the present data on cyclitols in Eucalyptus species are the first to implicate quercitol in salt tolerance. A proposed function of quercitol as an organic osmolyte in some Eucalyptus species, is consistent with the view that other cyclitols play key roles in salt tolerance of some halophytes. Pinitol is a major organic solute in leaves of mangroves (Popp 1984), and myo-inositol and its derivatives (ononitol and pinitol) accumulate in the common ice plant, Mesembryanthemum crystallinum (Nelson, Rammesmayer & Bohnert 1998). In addition to cyclitols having roles in osmotic relations and scavenging of radicals (Orthen, Popp & Smirnoff 1994; Popp & Smirnoff 1995), Bohnert and coworkers have speculated that myo-inositol might also be involved in regulation of Na+ transport (Nelson, Koukoumanos & Bohnert 1999) and act as an osmoprotectant for photosynthesis (Nelson et al. 1998), in M. crystallinum. The putative roles of cyclitols in salt tolerance has also led to some of these compounds being targeted for bioengineering approaches to improve salt tolerance in non-halophytes (Sheveleva et al. 1997). As noted by Popp et al. (1997), myo-inositol is only two reaction steps removed from glucose-6-phosphate; the short synthesis pathway of cyclitols from primary metabolites is a further argument in support of their major role in osmotic adjustment and enhances the prospects for genetic engineering of these pathways.

In proposing cyclitols, acyclic polyols and other carbon-based compounds as osmolytes in eucalypts and other native Australian tree species (e.g. Casuarina spp.), we recognize that alternative nitrogenous osmolytes seem less likely owing to the generally short supply of N in many habitats. Accumulation of simple organic solutes is a parsimonious mechanism by which salt-tolerant, long-lived woody species cope with accumulation in foliage of Na+ and Cl. Similarly, shortages of water (without salinity) generally produce significant decreases in osmotic potential (i.e. more negative) of foliage of eucalypts and other trees (e.g. White et al. 2000). While evidence for nitrogenous solutes as osmotica in Myrtaceous species has been found (e.g. various betaines in Melaleuca spp.; Naidu et al. 1987, 2000), we found no such evidence for the species studied here. We found no glycinebetaine (analysed as per Colmer et al. 2000) and in other studies (Adams et al. unpublished results), we found only very low concentrations of proline and related compounds. High concentrations of nitrogenous solutes in foliage have also been cited as causes of increased insect herbivory on stressed trees (White 1984; Cockfield 1988; Marsh & Adams 1995) and, thus, accumulation of non-nitrogenous organic solutes might be advantageous for plants in N-deficient ecosystems.

Coupled with the identification of significant accumulation of potential osmolytes, the present study also showed that the more salt-tolerant species, E. spathulata, E. loxophleba and E. sargentii were able to maintain well-regulated leaf Na+ concentrations even at 300 mol m−3 NaCl. These more salt-tolerant species also showed an apparent increase in net selectivity for K+ over Na+ as salinity increased, irrespective of the Na+ : Ca2+ ratio of the external medium. Provision of adequate Ca2+ in the external medium is crucial for control of Na+‘exclusion’ and K+ : Na+ selectivity in roots (Greenway & Munns 1980). That the responses of the three tolerant Eucalyptus species to increasing NaCl were essentially the same whether the Na+ : Ca2+ ratio was maintained at 25 : 1 or allowed to increase to 75 : 1, suggests that Ca2+ was adequate for these species in all treatments (Ca2+ was supplied at 4 mol m−3 or above in all treatments). By contrast, E. globulus was unable to maintain Na+ exclusion when exposed to the higher NaCl concentrations used in the present study (i.e. 200 and 300 mol m−3). There were also clear symptoms of ‘ion toxicity’ including leaf necrosis and then the sudden death of the whole shoots of this relatively salt-sensitive species (cf. Munns 1993).

Chloride accumulation in foliage, in addition to Na+, has been implicated as the cause of ‘ion toxicity’ in some woody species, most notably citrus (Banuls et al. 1997; Romero-Aranda et al. 1998). In the present study, Cl concentrations in leaves of the four Eucalyptus species exposed to ≥ 100 m m NaCl were, on average, 0.88 of those for Na+ (calculated from data in Figs 2 and 3). Thus, whether Na+ or Cl causing tissue damage (i.e. salt toxicity) in leaves of the Eucalyptus species can not be distinguished based on tissue concentrations of these ions; both presumably contribute to ion toxicity, although it remains a possibility that cells differ in their capacity to compartmentalize, or tolerate, Na+ and Cl.

Following earlier studies suggesting carbon isotope discrimination (δ13C) might be used to distinguish amongst eucalypt clones with varying water use efficiency (e.g. Bond & Stock 1990; Orsório et al. 1998), Poss et al. (2000) recently proposed that δ13C, standardized using minimum and maximum observed values, might be used to evaluate salt tolerance of eucalypts. Although our work, like that of Poss et al. (2000) is restricted to a glasshouse experiment, the patterns and statistical strength of the treatment differences shown in Table 1 (e.g. the consistent 1–2°/00 differences between the 100/4 and 300/12 treatments for all species) argue in support of the proposition. However, we remain somewhat sceptical owing to the carefully controlled conditions required to eliminate the many other confounding influences on δ13C (such as light, humidity, nutrition) that are often species-dependent and difficult to control in the field.

Serraj & Sinclair (2002) concluded that osmotic adjustment might be of little benefit to yields of annual crops affected by drought. Serraj & Sinclair (2002) also made the point that maintenance of turgor and stomatal conductance may well worsen any soil drought experienced by crop plants and that ‘water conservation is the appropriate response’ for crop monocultures.

The situation for woody perennials is somewhat different. First, many trees are ‘drought avoiders’ inasmuch as they maintain turgor and stomatal conductance and access water well beyond the root zone of competing, and frequently herbaceous, species (e.g. Nothofagus spp., Read & Farquhar 1991; Pinus spp., Warren, McGrath & Adams 2001; Eucalyptus spp., White et al. 1999). Osmotic adjustment is also required (see Introduction), owing to the inability to maintain water supply at rates sufficient to prevent short periods of partial dehydration. Secondly, numerous eucalypts and other trees from arid regions grow well under moderately saline conditions and show osmotic adjustment. The possible benefits of the process of synthesizing and accumulating osmolytes, in addition to the end-products being osmotica (Hare et al. 1998; Hasegawa et al. 2000), merits further investigation. The present study showed that salt-tolerant eucalypts accumulated quercitol to osmotically significant concentrations, but further work is required to elucidate whether patterns of accumulation of cyclitols are also related to taxonomy. The genus (including Corymbia) comprises more than 800 species, many endemic to saline and/or arid areas, and is thus a major resource for further studies of drought and salt tolerance in woody plants.


M.A.A. thanks the Alexander von Humboldt Foundation for financial support and the opportunity to work in Vienna where Professor Marianne Popp provided wonderful hospitality. The authors thank A/Professor Jen, McComb for samples and additional data from the field trial.



Table 6. Summary of two-way anova of effects of salinity and species on growth of eucalypts (Fig. 1). Eucalyptus globulus was excluded from the analysis owing to death of plants at the high salt concentrations (see text).
Source of variationd.f.MSF-valueP-value
(a) Variable Na+ : Ca2+
 Treatment 3 447.3 7.30.001
 Species 21021.516.6< 0.001
 Treatment × Species 6  93.9 1.50.200
 Residual33  61.4  
 Treatment 3  97.710.8< 0.001
 Species 2 181.620.1< 0.001
 Treatment × Species 6  18.7 2.10.083
 Residual33   9.0  
Old fully expanded leaves
 Treatment 3 119.8 4.70.008
 Species 2 138.9 5.40.009
 Treatment × Species 6  42.6 1.70.162
 Residual33  25.6  
Young fully expanded leaves
 Treatment 3  26.5 2.80.054
 Species 2   5.4 0.60.570
 Treatment × Species 6  10.7 1.10.366
 Residual33   9.4  
Still expanding leaves
 Treatment 3   2.9 3.30.003
 Species 2   5.2 5.90.006
 Treatment × Species 6   1.7 2.00.096
 Residual33   0.9  
(b) Constant Na+ : Ca2+
 Treatment 3 421.4 7.20.001
 Species 2 544.0 9.30.001
 Treatment × Species 6  94.2 1.60.175
 Residual33  58.4  
 Treatment 3  80.510.2< 0.001
 Species 2 231.829.4< 0.001
 Treatment × Species 6  10.9 1.40.248
 Residual33   7.9  
Old fully expanded leaves
 Treatment 3 133.7 4.80.007
 Species 2  88.4 3.10.055
 Treatment × Species 6  21.1 0.80.609
 Residual33  27.9  
Young fully expanded leaves
 Treatment 3  28.2 3.70.021
 Species 2   9.3 1.20.306
 Treatment × Species 6   7.8 1.00.425
 Residual33   7.6  
Still expanding leaves
 Treatment 3   2.7 3.20.038
 Species 2   3.8 4.50.018
 Treatment × Species 6   1.6 1.90.103
 Residual33   0.8  


Table 7. Summary of three-way anova of effects of salinity treatments, species and leaf age on concentrations of Na+ in foliage of four species of eucalypts. Data shown in Fig. 2; see caption of Fig. 2 for further explanation.
(a) Variable Na+ : Ca2+
Treatment (T) 31555321103.8< 0.001
Species (S) 2 408416 27.2< 0.001
Leaf Age (A) 2  78498  5.20.001
A × T 6  12522  0.80.545
A × S 4   6655  0.40.777
T × S 6  92028  6.1< 0.001
A × T × S12   6640  0.40.942
Residual95  14992  
(b) Constant Na+ : Ca2+
Treatment (T) 31168859 95.8< 0.001
Species (S) 2 564815 46.3< 0.001
Leaf Age (A) 2  98587  8.10.001
A × T 6  11817  1.00.451
A × S 4  15651  1.30.283
T × S 6  67883  5.6< 0.001
A × T × S12   2685  0.20.997
Residual94  12203  


Table 8. Concentrations (µmol g−1 dry weight) of Ca2+ and K+ in foliage of different ages and developmental stage, for four species of eucalypt as seedlings grown in a glasshouse with differing external concentrations of Na+ and Ca2+.
Nutrient, species and tissueTreatments (Na+ and Ca2+ concentrations in culture solution, mol m−3)
  1. Means of four replicates (x) and standard errors (SE) are shown. Data presented are for expanding leaves, younger fully expanded leaves, and older fully expanded leaves of plants grown with variable (0.025, 25, 50, 75) or constant (25) Na+ : Ca2+ ratios in the external solution (see text for details).

  2. Summary of three-way anova

  E. globulus
  Still expanding254451435114250nd nd nd 
  Young fully expanded418632453625713nd 262 60nd 
  Old fully expanded489513382529335nd 429134nd 
 E. loxophleba
  Still expanding2122913339 5823 7039 8122 4416
  Young fully expanded357401582315822 7212116 712637
  Old fully expanded381242114014019 97121811019534
 E. sargentii
  Still expanding1845010416118 7 5810 8719 3612
  Young fully expanded2283015126137126918 9519 6122
  Old fully expanded294141872621531204171794318735
 E. spathulata
  Still expanding21619 9716 4818 4618 7531 95 9
  Young fully expanded23426108 85083354450 94 5 9517
  Old fully expanded258 92141633964484532111021420
 E. globulus
  Still expanding412403472630318nd nd nd 
  Young fully expanded482523311217812nd 29638nd 
  Old fully expanded42545225 517219nd 24376210 
 E. loxophleba
  Still expanding5582544132334493255433732296 7
  Young fully expanded6013447031404 9391153445935132
  Old fully expanded563254713246835449273965348654
 E. sargentii
  Still expanding3622436052289252851628728291 5
  Young fully expanded397133943827816282263081433320
  Old fully expanded334233563731550312263022037733
 E. spathulata
  Still expanding436272951823431270212871726329
  Young fully expanded517193102323421258193301925318
  Old fully expanded573283442030441291282993931926
 TreatmentSpeciesLeaf age
  1. NS, not significant, *P < 0.05, **P < 0.01, ***P < 0.001, P < 0.0001). The treatment by species interaction was always highly significant (P < 0.001), other two-way and three-way interactions were always non-significant.

Calcium, variable external Na+:Ca2+****NS****
Calcium, constant external Na+:Ca2+**********
Potassium, variable external Na+:Ca2+**********
Potassium, constant external Na+:Ca2+***********