Introduction

Tree plantations have been identified as major contributors to sustainable supplies of wood products, for environmental amelioration and for carbon sequestration in tropical and sub tropical Australia (Keenan et al. 1998). In the past, tree planting in southeast Queensland has been limited to humid (>1,000 mm rainfall) coastal areas. However, with high land prices and low availability of suitable land within this zone resulting from the World Heritage Listing of large areas of coastal tropical Queensland, there is now an urgent need to explore the potential of establishing commercial plantations in marginal regions (i.e. 700–1,000 mm rainfall zone) (Loxton and Forster 2000). In these marginal forest regions, production is limited by low water availability and high summer vapour pressure deficit (Sands et al. 1999). The forest industries in both Queensland and sub tropical New South Wales have included E. cloeziana and E. argophloia for plantation establishment (Keenan et al. 1998). Knowledge of native eucalypt silviculture in these drought prone tropical and subtropical regions of Australia is limited to a few species and mainly to E. grandis (Doley 1978; Cromer and Jarvis 1990; Birk and Turner 1992) and E. camaldulensis (Blake 1980; Facanha et al. 1983). An understanding of the relationship between physiological responses of E. cloeziana and E. argophloia and the biophysical environment (particularly water stress) is necessary for the development of silvicultural regimes.

Plants possess three mechanisms that may counteract water deficit stress: (1) reduction of water loss through stomatal control and morphological changes; (2) increased extraction of water from the soil and; (3) increased ability to tolerate low tissue water potential (Osmond et al. 1980; Kramer 1983; Nilsen and Orcutt 1996). Stomatal closure is due directly to low turgor pressure in the guard cells (Collatz et al. 1991), which may result from root signals by abscisic acid (ABA) (Cowan 1982; Downton et al. 1988; Wartinger et al. 1990; Fort et al. 1997), or in response to increasing vapour pressure deficit (Turner et al. 1984). Avoidance of water stress by the development of a deep root system has been reported for E. camaldulensis and E. saligna (White et al. 2000), as well as for E. grandis (Dye 1996), and E. marginata (Doley 1967). Plants native to dry areas have also been reported to maintain photosynthesis at a lower leaf water potential than those native to humid areas (Kozlowski et al. 1991), favouring both survival and productivity (Ni and Pallardy 1991).

Eucalyptus cloeziana has a wide natural distribution in eastern Queensland, occurring in four disjunct geographical regions: southern coastal, southern inland, northern coastal and northern inland (Turnbull 1979). This distribution is characterised by large variations in rainfall, temperature, geology and soils, reflecting the high level of genetic variability reported by Turnbull (1979). Lee at el. (1997) reported satisfactory growth by both northern and southern coastal provenances tested on sites having an annual rainfall greater than 1,200 mm within Australia, Brazil and Congolese Peoples’ Republic. Nevertheless, estimates have not been made of the relative importance of genetically based adaptation and phenotypic plasticity to specific site conditions as components of E. cloeziana provenance variation in site performance (Turnbull 1979; Lee et al. 1997). E. argophloia is a potential commercial timber species with a narrow natural distribution northeast of Chinchilla in the Southern inland Queensland where it occurs under warm subhumid conditions (Boland et al. 1984). There is very limited information on the silviculture of this species. E. cloeziana and E. argophloia have excellent form and produce strong durable timber with many uses, and both are favoured for inclusion in forest plantations.

This study reports an investigation of physiological responses to water deficit stress of a dry inland provenance of E. argophloia and a dry inland, and humid coastal provenance of E. cloeziana. It was hypothesised that due to the disjunct nature of E. cloeziana populations, an inland provenance (Hungry Hills) would show greater drought resistance than the coastal provenance (Gympie), rendering it a better choice for drought prone sites. It was also hypothesised that the drought response mechanisms of the dry inland provenance of E. cloeziana provenance would be more comparable to those of E. argophloia than to the humid E. cloeziana provenance. Accordingly the objectives of the investigation were to determine for each of the three provenances: (1) responses of photosynthesis, stomatal conductance and leaf water potential to soil water availability, and (2) the rate of recovery of physiological processes when water stress is relieved.

Materials and methods

Seedling and growth conditions

Seedlings of a humid provenance of E. cloeziana from the Gympie, a dry zone provenance E. cloeziana from the Hungry Hills and of E. argophloia from Chinchilla (Table 1) were raised in pots containing a 50:50 mixture of peat and vermiculite at the Queensland Forestry Research Institute Gympie (QFRI). In June 2000, the 6-month-old seedlings were moved to a University of Queensland glasshouse (Brisbane) and transferred into plastic pots (175 mm diameter ×175 mm deep) lined with polythene and containing nursery top soil (loam-textured) mixed with vermiculite and perlite in the ratio of 4:1:1. Slow release fertiliser [Osmocote Plus (3–4 months) containing micronutrients] was mixed with the potting medium at a rate of 2 kg m−3. The pots were packed to a bulk density of approximately 1,100 kg m−3 and maintained in a naturally lit glasshouse with temperature control provided by evaporative coolers and electric fan heaters. The average daytime temperatures in the glasshouse over the experimental period (August and November 2000) ranged from 21 to 37°C.

Table 1 Seedlot information on the three Eucalyptus provenances used in the experiments (seed provided by Queensland Department of Primary Industries—Forestry)
Full size table

Response to variable watering

After 8 weeks of growth under favourable conditions, 60 healthy seedlings of uniform height were chosen from each provenance. Plastic beads were spread on the top surface of each pot to minimise evaporation from the soil (Myers and Landsberg 1989). All plants were watered to field capacity and arranged on a glasshouse bench using a completely randomised design with three water application treatments. The benches were rotated every 2 days. Each treatment comprised five plants of each provenance. A 2-day cyclical droughting treatment was applied to all plants. Plants in the control treatment (T100) were rewatered every 2 days with all the water lost through evapotranspiration. The loss was estimated by determining the weight change in the T100 pots of each provenance. Fresh weights of three plants treated in a similar manner and harvested every fortnight were used to adjust pot weight changes resulting from plant growth. Water stressed plants received 70% (T70) and 50% (T50) of water supplied to control plants. The experiment was continued for 14 weeks.

In the severe stress treatment (T0), water was withheld from 20 plants from each provenance until E. cloeziana seedlings had wilted overnight. During this period, sequential measurements of water potential and gas exchange were made on five plants in each of T0 and T100 treatments per provenance. The plants were then watered to field capacity and re-watered every 2nd day. Measurements of gas exchange and water potential were made on several occasions during the recovery periods until measured physiological variables of the stressed plants had recovered to the level of the control treatment (T100). E. argophloia seedlings were subjected to a second severe water deficit stress three weeks after the first stress cycle because they did not show visible signs of wilting at the wilting point of E. cloeziana seedlings.

Gas exchange and Leaf water potential

Photosynthesis (A), stomatal conductance to water vapour (g s), transpiration (E), and intercellular CO2 concentration (ci) were measured at ambient CO2 concentration (350–370 μmol mol−1) in the glasshouse. Measurements were taken at about midday using a portable photosynthesis system (Model LI-6200, Li-Cor, Lincoln, Neb., USA) equipped with a 250 ml cuvette. Gas exchange was measured on a young, fully expanded leaf under bright sunlight when photosynthetic photon flux density (PPDF) was greater than 700 μmol m−2 s−1. The mean midday vapour pressure deficit on the days when measurements were taken in the glasshouse ranged from 3.5 to 5.3 kPa, and mean midday air temperature measured in the leaf cuvette ranged from 30 to 39°C. Leaf conductance was partitioned into stomatal conductance (g s) and nonstomatal or residual conductance to diffusion of CO2 (g rc) according to Farquhar and Sharkey (1982) where g rc includes mesophyll conductance and carboxylation efficiency (Sun et al. 1995). Measurements of predawn (Ψpd) and midday (Ψmd) leaf water potential were made using a pressure chamber (PMS Instruments, Corvalis, Ore., USA). The measurements were made on a fully expanded leaf occupying the same developmental position as that used for gas exchange.

Measurement of stomatal properties

One fresh leaf from a well-watered plant of each provenance was cut into approximately 5 mm2 pieces with a sharp razor blade. The tissues were fixed in 4% glutaraldehyde in 0.05 M sodium phosphate buffer, pH 6.9, overnight at 4°C, post-fixed in 1% osmium tetroxide in 0.05 M sodium phosphate buffer of pH 6.9 for 1.5 h and then dehydrated through a graded ethanol series. The specimens were then critical point dried in a Polaron Critical Point Drier, mounted on stubs and gold coated in a BIO-RAD SC502 Sputter Coating System. Stomatal density (SD) on the adaxial and abaxial surfaces was assessed at ×400 and mean guard cell length (MGCL in μm) and total guard cell length per unit area (TGCL mm−2) at ×1,000 in a JEOL 6300 field emission electron microscope (FESEM) at 5 kV accelerating voltage.

Data analysis

Analyses of variance within and among provenances were undertaken using the PROC GLM procedure of the SAS statistical program (SAS 1985), while differences between means were segregated using Duncan’s multiple range tests. Water stress integral (SΨ) was used to derive an index of cumulative water stress (Myers and Landsberg 1989) for each treatment and provenance over the 14 weeks period. Regression analyses were undertaken using PROC REG and PROC NLIN of SAS software.

Results

Examination of stomata on young leaves showed that E. argophloia was amphistomatous, with more stomata on the lower leaf surface (483 mm−2) than on the upper surface (377 mm−2). E. cloeziana had stomata on the lower side only (hypostomatous) and there were no significant differences in stomatal density between the humid (674 mm−2) and dry provenance (629 mm−2). Moreover, there were no significant differences in mean guard cell length (MGCL) among provenances (Table 2).

Table 2 Stomatal dimensions from leaves of E. argophloia and E. cloeziana seedlings watered to field capacity every second day. The values shown are means±standard errors. ( SD stomatal density, MGCL mean guard cell length, TGCL total guard cell length)
Full size table

The patterns of Ψpd, which resulted from the three watering treatments imposed over the experimental period, are presented in Fig. 1. In E. cloeziana, stress treatments T70 and T50 resulted in distinct and consistent intensities of water deficit which were more pronounced in the humid than in the dry provenance. There were significant differences in Ψpd and Ψmd among T100, T70 and T50 in the humid provenance of E. cloeziana. In the dry provenance of E. cloeziana, there were significant differences in Ψpd between T100, T70 and T50, but not in Ψmd (Table 3). In E. argophloia, neither Ψpd nor Ψmd were significantly different among the treatments (Table 3). There were significant differences (P ≤0.05) between control and stress treatments of the three provenances in net photosynthesis (A). There were no significant differences between T70 and T50 in A and g s for any of the three provenances (Table 3).

Fig. 1
figure 1

Predawn water potential of E. argophloia, humid and dry provenances of E. cloeziana seedlings. Plants were watered every 2 days. Control plants (circles) received 100% of water lost through evapotranspiration and water stressed plants received 70% (triangles) and 50% (squares) respectively of water supplied to control plants

Full size image
Table 3 Net photosynthetic rate (A), stomatal conductance to diffusion of water vapour (g s ), predawn (Ψ pd ) and midday (Ψ md ) leaf water potential, and cumulative water stress (S Ψ ) for E. argophloia and E. cloeziana seedlings. Control plants (T 100 ) were watered to field capacity every second day, T 70 and T 50 were watered with 70% and 50% of water added to T 100 respectively. The values shown are means±standard errors over all days of measurements. For each variable within each provenance, means with the same subscript for x and row are not significantly different (P≤0.05) while within each column, similar variables with the same subscript for y are not significantly different ( P≤0.05)
Full size table

The cumulative water stress integrals (SΨ) for the three provenances showed that plants in T70 and T50 experienced greater water stress than the T100 plants (Table 3). However in all three treatments, the humid provenance of E. cloeziana experienced the greatest water deficit stress over the experimental period. Comparison among provenances in the three water treatments showed that E. argophloia had higher A, g s, and Ψpd than E. cloeziana provenances which did not differ from each other (Table 3). There were no statistically significant differences between E. argophloia and the E. cloeziana provenances in Ψmd (Table 3).

Regression analyses between Ψ pd and gas exchange variables showed that A, g s and residual conductance (g rc) declined with decrease in Ψ pd following an exponential trend of the form:

$${\text{Y}} = \beta _{0} {\text{e}}^{{\beta \;\Psi {\text{pd}}}}_{1} $$
(1)

where Y represents gas exchange variables (A, g s and g rc), β 0 is a constant, e denotes the base of the natural logarithm (2.178) and β 1 defines the rate of response. The response of E. argophloia to decreasing Ψ pd differed from that of E. cloeziana (Fig. 2). It maintained higher A and g s as the level of stress increased whereas in E. cloeziana provenances, rapid reduction in A occurred as Ψpd dropped from −0.3 to −1.2 MPa (Fig. 2). Generally there were limited differences between the responses of the humid and dry provenances of Eucalyptus cloeziana but the slope (β 1) for the relationship between Ψpd and A, g s and g rc for E. cloeziana provenances was greater than for E. argophloia (Table 4). In all three provenances, A increased with an increase in g s and there were minor differences among the slopes of the regression lines (Fig. 3a, Table 5). However the two species differed substantially in response of A to g rc (Fig. 3b) g rc to g s (Fig. 3c). The slope parameter for E. cloeziana provenances for this relationship was less than half that for E. argophloia (Table 5). In E. cloeziana provenances, the slope for the response of A to g s was 2–3 times greater than that for the response of A to g rc.

Fig. 2
figure 2

Response of a net photosynthesis (A), b stomatal conductance (g s) and c residual conductance (g rc) to decreasing predawn leaf water potential (Ψpd) for humid (squares, continuous line) and dry (triangles, dashed line) provenance of E. cloeziana and E. argophloia (circles, dotted line)

Full size image
Table 4 The slope (β 1) and coefficient of determination (r 2) obtained from exponential decay function fitted to the relation between predawn leaf water potential (Ψpd) and net photosynthesis (A), stomatal conductance (g s) and residual conductance to diffusion of CO2 (g rc) data from the four treatments (T100, T70, T50 and T0) for each provenance. The values shown are parameters ± standard errors
Full size table
Fig. 3
figure 3

Response of a net photosynthesis (A) to stomatal conductance (g s) and b residual conductance (g rc) and relationship between g rc and g s for humid (squares, continuous line) and dry (triangles, dashed line) provenance of E. cloeziana and E. argophloia (circles, dotted line)

Full size image
Table 5 Parameters for regressions analyses between net photosynthesis (A, μmol m–2s–1) and stomatal conductance (g s, mmol m–2 s–1) and residual conductance (g rc, mmol m–2 s–1) using the function y = β 1Ln(x)+ β 0 on combined data from the four treatments (T100, T70, T50 and T0) in each taxon. The values shown are parameters ± standard errors
Full size table

For all three provenances, withholding water resulted in a gradual decrease in both Ψpd and Ψmd (Fig. 4). Under continuous stress, E. argophloia maintained significantly higher A and g s at all levels of stress than either provenance of E. cloeziana, which showed no differences in their response to stress (Fig. 4). The estimates of Ψpd and Ψmd at which plants remained wilted overnight were respectively: −2.7 and −4.1 MPa for E. cloeziana (humid), −2.8 and −4.0 MPa for E. cloeziana (dry) and, −3.7 to −4.9 MPa, for E. argophloia. At these values of Ψmd, photosynthesis could not be detected. Plants stressed beyond these limits did not recover after watering. E. argophloia did not portray visible signs of leaf tissue damage at the wilting Ψpd whereas E. cloeziana (humid) suffered extensive leaf necrosis and E. cloeziana (dry) showed limited necrosis.

Fig. 4
figure 4

Response of a predawn water potential (Ψpd), b midday water potential (Ψmd), c net photosynthesis (A) and d stomatal conductance (g s) of E. argophloia (circles) and the humid (squares) and dry (triangles) provenance of E. cloeziana to severe stress and subsequent stress relief at day 17. The open symbols and hatched line represents the T100 (control), and closed symbols and continuous line represent the (T0) treatment

Full size image

It took approximately 10 days for water potential to recover to the level of the control treatment after re-watering in all provenances (Fig. 4a). Net photosynthesis and g s of E. argophloia recovered within 10 days, E. cloeziana dry provenance within 2 weeks and humid provenance within 3 weeks (Fig. 4c).

Discussion

Species differ widely in their responses to reduced soil water availability (Newton et al. 1986; Kozlowski et al. 1991; Blake and Tschaplinski 1992; Kramer and Boyer 1995), so several expressions of response may be investigated. The present study concentrated on simple water relations parameters, including shoot water potential, gas phase conductances of leaves, and carbon dioxide assimilation. Each of these parameters was compared between provenances and treatments, and while not every relationship was consistent, they indicated several associations between provenance and physiological characteristics that could be useful in site-species matching. Although application of these results to field-grown saplings and adult trees is not straightforward, the differences in water stress tolerance appear to be consistent with ecological observations (Turnbull 1979; Boland et al. 1984).

Distinct differences in Ψpd between treatments T100, T70 and T50 in E. cloeziana (Fig. 1) indicate that the plants were affected by the different levels of soil water availability in the manner indicated by Schulze (1986), Myers and Landsberg (1989), Wartinger et al. (1990) and Osório et al. (1998). Whilst the patterns of predawn water potential in well-watered plants were similar in all provenances (Fig. 1), E. argophloia showed overnight recovery under water deficit conditions whilst increasing water deficits resulted in progressively lower values of Ψ pd in E. cloeziana. These differences may be associated with the greater development of fine roots in E. argophloia than either provenance of E. cloeziana (Ngugi et al. 2003b). Similarly, White et al. (2000) reported that field grown trees of E. camaldulensis were more effective in exploitation of soil water than were E. leucoxylon or E. platypus. On the other hand, Colquhoun et al. (1984) found that E. wandoo, the most drought-tolerant but shallow-rooted of six species examined in the field, showed consistently lower Ψpd than the other species. It is potentially useful for rapid species evaluations that physiological differences that may be crucial to larger trees in the field may be reflected in pot-grown seedlings in the glasshouse.

Stomatal conductance (g s) and mesophyll conductance (g rc) for both provenances of E. cloeziana and for E. argophloia appeared to be very sensitive to small reductions in Ψpd between −0.2 and −0.8 MPa (Fig. 2). This relationship is quite distinct from those between gas exchange characteristics and instantaneous water potential, where linear regressions are more commonly observed (Ni and Pallardy 1991), or there may be a distinct threshold water potential above which there is little change in g s (Schulze 1986). However, there is a close similarity between the patterns recorded in the present study and those obtained in Eucalyptus globulus (Pereira et al. 1987) and Nothofagus solandri and N. menziesii (Sun et al. 1995). Both E. argophloia and N. solandri, which are relatively drought-resistant species, showed higher conductance and photosynthesis values at maximum Ψpd than did the more drought-susceptible E. cloeziana or N. menziesii. E. argophloia also showed higher values of g s at minimum Ψpd than did either provenance of E. cloeziana (Fig. 2). These differences indicate both the potential for high rates of gas exchange in E. argophloia when water is plentiful, and the maintenance of carbon gain under less favourable conditions.

The high maximum g s in well-watered E. argophloia (Table 3) may be related to its higher mean stomatal density and total guard cell length per unit area (cf. Table 2). When the ratio, maximum g s to mean total guard cell length (TGCL, mm−2) for both leaf surfaces is calculated, E. argophloia has a value of 0.027, whilst the dry and humid provenances of E. cloeziana have values of 0.015 and 0.013, respectively. This difference may reflect the mutual interference between the diffusion shells surrounding stomata when they occur at high density on only one surface of the leaf (Meidner and Mansfield 1968). Other situations in which higher gas exchange rates have been observed in drought-resistant than in mesic taxa include provenances of Azadarichta indica (Kundu and Tigerstedt 1998), Quercus stellata in relation to Juglans nigra (Ni and Pallardy 1991), and Eucalyptus camaldulensis in relation to E. leucoxylon and E. platypus (White et al. 2000). However, where root systems of trees have access to sufficient soil water, drought stress may be avoided (e.g. in E. marginata), even though other species regarded as more drought-resistant (e.g. E. wandoo) exhibited much lower water vapour conductances and leaf water potentials (Colquhoun et al. 1984).

Under severe water stress treatments, Ψpd decreased in a similar pattern for the three provenances (Fig. 4), with E. argophloia consistently showing a greater A and g s than E. cloeziana provenances (which did not differ much from each other) and a greater ability to sustain photosynthetic activity to a Ψ pd value of −3.7 MPa compared to −2.7 MPa for the E. cloeziana provenances. The results for E. argophloia were consistent with drought responses of xeric species from eastern North America (Abrams 1994) whilst the responses of E. cloeziana provenances are consistent with those of a glasshouse experiment in which E. cloeziana stomata closed at water potential of −2.7 Mpa compared to −1.2 MPa for E. grandis (Fisher 1991). As in E. cloeziana A and g s were close to zero at leaf water potential of −2.9 MPa in Quercus stellata, −2.6 MPa in Q. alba and Acer saccharium, and −2.2 MPa in Juglans nigra seedlings (Ni and Pallardy 1991). However E. behriana maintained photosynthesis at leaf water potential of −4.4 MPa (Küppers et al. 1987; Myers et al. 1987).

Almost identical non-linear relationships were found between A and g s for E. argophloia and the two provenances of E. cloeziana (Fig. 3a). Sun et al. (1995) fitted linear regressions to these parameters for the drought-resistant Nothofagus solandri and the drought-susceptible N. menziesii, but a non-linear relationship might appear to be more appropriate for N. menziesii. The differences between taxa were more distinct when A was compared with g rc in both the present study (Fig. 3b) and in N. solandri and N. menziesii (Sun et al. 1995). In both situations, the more drought-resistant species showed higher maxima for A and g s, but a similar range of g rc to the drought-susceptible taxa. For spring and summer conditions, A in E. globulus was linearly related to g s up to about 80 mmol m−2 s−1, after which there was no response (Pereira et al. 1987). The clear distinction between g s -limited and non-limited conditions observed in E. globulus was not obtained in the present study (Fig. 3a), indicating greater involvement of non-stomatal limitations to photosynthesis.

A capacity for substantial variation in stomatal opening would make the leaves very responsive to the ambient conditions, whereas g rc is likely to reflect internal structural and biochemical characteristics of leaves, and to be less variable than g s. The denser packing of chlorenchyma cells in leaves of E. argophloia might be expected to result in a shorter mean diffusion path length for CO2 and a higher g rc than in E. cloeziana, but this is not reflected in Fig. 3b, which shows a similar range of g rc in both species.

An insight into the variation in responses of A to g s and g rc may be gained by comparing these two conductances (Fig. 3c). For both E. cloeziana provenances, there was a reasonably close linear relationship between g rc and g s, but in E. argophloia high g s was often associated with low g rc and relatively low A. All the occasions on which A in E. argophloia exceeded the maximum value in E. cloeziana were associated with high values of g s rather than of g rc (Figs. 2, 3b). The cause of the varied g rc values is not clear, but the majority of occasions when low g rc was observed were days on which Ψpd was lower than −1MPa. In E. globulus, these occasions of low Ψpd were associated with low leaf conductance (Pereira et al. 1987), but in E. argophloia, some days of low Ψpd were associated with moderate to high g s. Kirschbaum (1987) identified photoinhibition as a contributor to decreased g rc in E. pauciflora, and it is possible that water deficits and the light regime of the present study could have resulted in patchy distribution of stomatal apertures (Downton et al. 1988; Beyschlag and Eckstein 1997) inducing this condition in E. argophloia and E. cloeziana.

Rewatering seedlings after an overnight wilting resulted in recovery of water potential of the three provenances over the next 10 days (Fig. 4). Gas exchange of E. argophloia also recovered within 10 days unlike that for the dry provenance of E. cloeziana and humid provenances of E. cloeziana that took more than 2 weeks. This difference in recovery was attributed to damaged photosynthetic tissues on the leaves. In humid provenance of E. cloeziana most of the leaves showed necrotic patches in regions away from the midrib during the recovery period, and many of the older leaves were shed. Similar observations have been reported for E. cloeziana seedlings under glasshouse conditions (Fisher 1991). Kirschbaum (1987) observed recovery of Ψpd in E. pauciflora within 7 h and in A within 1 day after rewatering. Although there were no significant differences in the response of the provenances of E. cloeziana to increasing water stress, the dry provenance recovered faster, indicating greater tolerance at low water potential.

Both E. cloeziana and E. argophloia developed fewer and smaller leaves in response to drought, but E. cloeziana also shed numerous lower leaves (Teskey and Hinckley 1986; Pereira and Pallardy 1987; Passioura et al. 1993; Ngugi et al. 2003a). However, these new leaves did not acclimate fully to the available water, as demonstrated by significant differences in gas exchange variables between the well-watered and water stressed treatments (Table 3).

This study confirms several aspects of the physiology of Eucalyptus species with respect to water deficits. It demonstrates that some degree of drought tolerance may occur (in E. argophloia), but that between provenances of another species, there may be no clear distinction between the drought tolerances of provenances from humid and sub-humid environments. In addition, E. argophloia demonstrates drought resistance through the possession of an extensive fine root system that exploits soil moisture more effectively than do either provenance of E. cloeziana. However, when water is freely available, the higher stomatal conductances and photosynthetic rates in the naturally restricted E. argophloia than in the much more widespread E. cloeziana indicate that the former species is likely to be suitable for planting on a greater range of sites than the latter. This finding indicates that some caution should be applied when matching species to sites simply on the basis of the climatic or distributional data associated with their natural occurrences.