Tree Physiology 20, 953–963
© 2000 Heron Publishing—Victoria, Canada
Tolerance of salinized floodplain conditions in a naturally occurring
Eucalyptus hybrid related to lowered plant water potential
TATIA M. ZUBRINICH,1 BETH LOVEYS,1 SONYA GALLASCH,1 JACK V. SEEKAMP1 and
STEPHEN D. TYERMAN1,2
1
School of Biological Sciences, Flinders University of South Australia, G.P.O. Box 2100, Adelaide, South Australia 5001, Australia
2
Author to whom correspondence should be addressed
Received September 3, 1999
Summary Rising saline groundwater and reduced flooding
frequency are causing dieback of Eucalyptus largiflorens
F. Muell. along the Murray River in Australia. A green-leaved
variant of E. largiflorens, which is probably a hybrid with a local mallee species (E. gracilis F. Muell.), tolerates saline conditions better than the more common grey-leaved variant. The
green variant exhibited more negative water potentials than the
grey variant, and comparison with soil water potential profiles
indicated that the green variant extracted water from slightly
higher up the soil profile where the salt content was lower but
the soil was drier. However, the stable isotopes of water (2H
and 18O) in the xylem did not differ significantly between
paired green and grey trees, suggesting that both variants used
the same water source. The green variant may be able to extract
water for a longer period from a given point in the soil profile
and tolerate a higher salt concentration around its roots than the
grey variant. Predawn leaf water potentials of both variants decreased with increasing salinity of groundwater and decreasing
depth to the groundwater, probably because the roots were being progressively confined to soil with lower matric potential
as groundwater discharge through transpiration progressively
salinized soil up the profile. The green variant had a lower assimilation rate and stomatal conductance than the grey variant,
although the differences were not statistically significant during most of the year. Discrimination of 13C indicated that the
green variant had a higher leaf internal CO2 concentration than
the grey variant, indicative of a greater biochemical limitation
on photosynthesis, perhaps resulting from the effects of operating at lower water potentials. The green variant had significantly lower stem hydraulic conductivity than the grey variant,
probably because of its smaller xylem vessel diameter and
higher degree of embolism. The more conservative water use
of the green variant and its ability to operate at lower water potential than the grey variant appear to underlie its ability to tolerate conditions of reduced useable water above the saline
groundwater. This advantage appears to outweigh the costs of
increased xylem embolism and reduced assimilation.
Keywords: assimilation, Eucalyptus largiflorens, groundwater, hydraulic conductance, salt tolerance, stable isotopes,
stomatal conductance, water stress, xylem embolism.
Introduction
Dieback of vegetation on riverine floodplains as a result of rising saline water tables is a common problem worldwide and is
caused by the combined influences of altered river flow regimes, decreased flooding frequency, land clearance and irrigation. Revegetation with more salt-tolerant and deep-rooted
species, groundwater pumping, and provision of environmental river flows are often considered as potential solutions in remediation of degraded areas. A significant area of naturally
vegetated floodplain in Australia occurs along the Murray
River in the Chowilla anabranch system, occupying an area of
1,650 km2 in southwestern New South Wales and eastern
South Australia. The region overlies a natural, highly saline
groundwater that has risen by about 3 m since 1930 as a result
of the construction of a series of weirs and widespread land
clearance throughout the floodplain (Simpson and Herczeg
1991). The most common tree in the Chowilla floodplain is
Eucalyptus largiflorens F. Muell., with stands covering approximately 38% of the region (Noyce and Nicholson 1992).
A variant of E. largiflorens has been identified in scattered
locations throughout the Chowilla floodplain. It is referred to
as the green variant because of its bright, glossy, green leaves
that contrast to the normal, glaucous, grey-green leaves. The
green variant is considered to be a hybrid between E. largiflorens and a mallee species, E. gracilis F. Muell., which
grows on sandy soils adjacent to the floodplain (Zubrinich
1996). The green variant is intermediate between E. gracilis
and E. largiflorens for several characteristics including fruit
and flower morphology, cotyledon morphology and leaf
characteristics (wax and color) (Zubrinich 1996). The green
variant exhibits a similar time to germination of seeds, germination percentage and overall tree structure to that of E. largiflorens, whereas the frequent presence of an outer ring of
floral staminodes, leaf color and xylem vessel size are more
similar to that of E. gracilis. Compared with trees of the grey
variant, trees of the green variant are generally larger, visibly
healthier and exhibit significantly less dieback, indicating that
they are coping better with the rising, highly saline groundwater in the region.
Trees of the grey variant of E. largiflorens obtain water
954
ZUBRINICH, BYRNE, GALLASCH, SEEKAMP AND TYERMAN
mostly from groundwater (Thorburn et al. 1993). The groundwater is saline, and although transpiration rates are low
(Thorburn et al. 1993), root water extraction will salinize the
zone of extraction such that roots will be “salted out” and confined to extraction of water from drier regions in the soil profile (Thorburn et al. 1995). A front of soil salinization will
move up the profile and eventually restrict the tree to soil that
may be too dry for water extraction. A model based on this
process showed that predicted groundwater uptake closely
matched observed uptake, and that groundwater depth and salinity were the main determinants of the degree of groundwater uptake (Thorburn et al. 1995). Based on these observations,
we postulated that the apparent salt tolerance of the green variant may be related to a different pattern of water extraction
from the soil, perhaps related to a lower rate of water use and
hence a lower rate of soil salinization, or to the ability to extract water from drier but less saline regions in the soil. These
possibilities have links to a variety of physiological processes
including: (1) the ability to lower water potential and the consequent impact on xylem cavitation; (2) the degree of control
exerted by the stomata and the implications for water-use efficiency; and (3) the degree of salt uptake that could impact on
photosynthesis and osmotic adjustment. We compared several
physiological characteristics of the green and grey variants,
and we also examined soil salinity and soil water potential profiles. We found that the greater salt tolerance of the green variant in the field was associated with its ability to extract water
from drier or more saline regions of the soil.
Methods
Study sites
The Chowilla floodplain spans part of the South Australia–New South Wales border within the Murray Darling Basin
in Australia. The region exhibits a semi-arid/arid climate,
characterized by mild winters and hot summers with an annual
rainfall of 248 mm (Bureau of Meteorology 1994). Trees from
four study sites within the Chowilla floodplain were regularly
monitored for physiological characteristics in 1992, 1993 and
1994. The sites contrasted in degree of dieback, which generally correlated to the depth and salinity of the groundwater
(Table 1). Chowilla Driveway (140°49.0′ E, 33°58.6′ S) is
considered the healthiest site, because it shows few signs of
episodic dieback. It comprises a deep water table (450 cm)
with a moderate groundwater salinity (conductivity = 42 dS
m –1). Monoman Island (140°52.7′ E, 33°57.6′ S) is located approximately 5 km from the Chowilla Driveway site and is
surrounded by anabranches of the River Murray. The trees at
this site show large amounts of dieback and the water table occurs at 400 cm with a groundwater conductivity of 60 dS m –1.
The Coppermine North (140°51.5′ E, 33°57.6′ S) and Coppermine South (140°49.4′ E, 33°58.4′ S) sites also contrasted in
groundwater salinity and depth to water table. Coppermine
North has a shallow water table (300 cm) that is saline
(58 dS m –1), whereas Coppermine South has a deeper water
table (400 cm) that is moderately saline (45 dS m –1). Some
measurements were also made at a site where extreme dieback
of the grey variant has occurred (Tareena, 140°02.5′ Ε,
33°59.0′ S). At this site the groundwater salinity is 68 dS m –1,
and the depth to groundwater is only 2.6 m.
Field measurements
Leaf conductance to water vapor, leaf water potential and leaf
ion concentration were measured monthly at the Chowilla
Driveway and Monoman Island sites during a 12-month period from January 1992 until January 1993. The Coppermine
North and South sites were monitored between January 1993
and February 1994. Leaf water potential was also monitored
between March 1994 and June 1994 at all sites. Assimilation
rates were measured at Chowilla Driveway and the Coppermine North and South sites in January 1993. Collections of
stems for determination of hydraulic properties were made on
three occasions in 1994. Additional measurements of leaf water potential and xylem sap osmotic potential were made at
each site during 1997.
For each of the physiological parameters, a pair-wise sampling and analysis procedure was adopted whereby a tree of
the green variant and its closest E. largiflorens (grey variant)
neighbour of similar estimated age were examined. This
method of sampling was employed because the green variant
trees are relatively rare throughout the study region. The distance between trees within a pair varied depending on the age
and density of trees at each site, but was always less than 5 m
(measured between tree bases).
Plant and soil water relations
Leaf water potential (Ψl ) was measured with the Scholander
pressure chamber, taking precautions to minimize transpiration from leaves during measurement. Twig samples were
taken from between 1.5 and 2.5 m aboveground from grey and
green variant trees at each site. Three replicate twigs were ana-
Table 1. Groundwater conditions and health index of green and grey variant trees at study sites throughout the Chowilla anabranch system. Health
index (mean and SEM) was determined as described by Jolly et al. (1993) (nd = not determined).
Site
Chowilla Driveway
Coppermine South
Monoman Island
Coppermine North
Tareena
Depth to groundwater (cm)
Groundwater conductivity (dS m –1)
Green variant health index
Grey variant health index
450
42
4.4 (0.1)
3.9 (0.3)
400
45
nd
nd
400
60
4.1 (0.2)
3.1 (0.3)
300
58
4.1 (0.1)
2.9 (0.3)
260
68
3.8 (0.2)
1.9 (1.1)
TREE PHYSIOLOGY VOLUME 20, 2000
SALT TOLERANCE IN A EUCALYPTUS LARGIFLORENS HYBRID
lyzed per tree. Measurements were recorded before dawn and
at midday. A full 24-hour set of measurements was recorded at
the Coppermine North site on February 23–24, 1994 to investigate if trees of either variant were equilibrating during the
evening. Coppermine North was chosen because trees at this
site had the most negative midday Ψl values among the study
sites. Three replicates were recorded per tree at each time interval.
Measurements of xylem sap osmotic potential were made
with a Wescor vapor pressure osmometer (Wescor, Inc., Logan, UT) after collection of expressed sap with the pressure
chamber. The bark was peeled back from the exposed end of
the twig and the wide end of a plastic pipette tip was fitted over
the exposed wood. This minimized evaporation of the sap
during collection and prevented the sap from being dispersed
by bubbles. The osmometer chamber was cleaned thoroughly
and calibrated in the low measurement range (less than
100 mosmol kg –1) so that reliable measurements of xylem sap
could be obtained.
Soil profiles from the four sites were analyzed for matric potential, osmotic potential and, by addition, total soil water potential. Hand-augured samples were taken as a function of
depth from a single hole dug at a position approximately halfway between a green variant and grey variant tree at each site,
and usually no further than 3 m from an individual tree. Roots
were often observed up to 400 cm in the soil profile. The soil
was quickly removed from the auger at each depth and placed
in double-sealed glass jars for transport to the laboratory.
Matric potential was measured by the filter paper method
(Greacen et al. 1989). To measure osmotic potential, a 1:5
(w/v; 20 g soil to 100 ml distilled water) soil extract was prepared by shaking the extract in bottles on a rotating soil mixer
for 24 h. Osmolality of each extract was measured with a
Wescor vapor pressure osmometer and corrected for the dilution factor and the known soil water content. The chloride
concentration of each soil extract was measured with a
Corning Eel 920 chloride titrator (Evans Electroselenium Ltd.,
Halstead, U.K.) and corrected for dilution.
Leaf conductance to water vapor and CO2 assimilation
Leaf conductance to water vapor (g) was monitored with an
automatic cycling porometer (Delta-T AP4, Delta-T Devices
Ltd., Cambridge, U.K.). Trees at one site were measured over
1 day, with the other sites monitored on consecutive days to
minimize variations in climatic conditions. At each site, one
tree of the green variant and the closest grey variant tree of
similar age were measured at approximately 60-min intervals
from dawn until dusk. Measurements were made on five
leaves per tree on the north side of the tree. Leaves were standardized for age and health by sampling the fourth healthy leaf
from the terminal end of a branch. Abaxial and adaxial leaf
surfaces had similar g and were averaged. The same leaves
were used for an entire day; however, new leaves were tagged
and monitored at each time interval throughout the year. The
same pair of trees at each site was monitored at each time interval during the year.
955
Assimilation of CO2 (A) and conductance to water vapor
was measured with an LI-6200 photosynthesis system (LiCor, Inc., Lincoln, NE) equipped with a 1-dm 3 chamber and
configured to measure absolute CO2 concentrations. The humidity of the chamber was set to reflect the ambient humidity
at the time of measurement. Measurements were made in January and July at the Coppermine North and South sites and in
January 1993 at the Chowilla Driveway site.
Mean daily conductance and assimilation rates were calculated by measuring the area under a curve generated from
time-interval measurements throughout a day. The area measurements were made by the trapezoid method that matched
the area under a fitted polynomial between fixed times. This
procedure was necessary to obtain accurate mean daily conductance and CO2 assimilation rate values, because the measurements were not always made at equally spaced time
intervals throughout the day.
Leaf ion concentrations
Leaf samples were collected from five green variant trees and
their closest grey counterparts at the Chowilla Driveway and
Monoman Island sites. Two replicates were assayed per tree.
Leaves were standardized for age and health by sampling the
fourth healthy leaf from the terminal end of a branch. Approximately 12 leaves were collected per sample for each tree. The
leaves within a sample were pooled and the middle 2 cm of
each leaf was cut into thin sections. Approximately 0.5 g of the
sectioned leaf material was weighed into a preweighed glass
vial and oven dried for 24 h at 80 °C. The samples were then
reweighed, and 5 ml of glass-distilled water was added to each
sample. The vials were placed in a 100 °C water bath for 6 h.
Chloride was measured with a chloride titrator (Corning Eel
920) and sodium and potassium were assayed by flame photometry (Corning Eel 430, Evans Electroselenium Ltd.). Results are expressed as µmol g –1 of dry leaf matter.
Hydraulic properties of stems
Two pairs of green and grey variant trees were selected from
each of the Monoman Island, Chowilla Driveway and Coppermine North sites. Two stems of less than 12 mm in diameter and 400 mm in length, without secondary branches, were
cut from each tree for stem hydraulic conductivity measurements. Samples were transported to the laboratory in sealed
polyethylene bags and stored in a refrigerator. The method and
apparatus described by Sperry et al. (1988) were used to quantify xylem embolism. Stem segments were submerged in tap
water for 15 min before cutting to allow relaxation of the tension in the xylem. A segment 7–15 cm in length was cut from
the central region of the stem. These segments were individually attached to the apparatus between two pieces of silicon
tubing and made airtight by the use of hose clamps. Filtered
(0.22-µm) distilled water was then gravity fed from a measured height into the stem for 3 min. Water that dripped out of
the stem was collected in a beaker and weighed. Hydraulic
conductivity (kg s –1 mm MPa –1) was measured as the rate of
fluid flow (kg s –1) times the length of the segment (mm) di-
TREE PHYSIOLOGY ON-LINE at http://www.heronpublishing.com
956
ZUBRINICH, BYRNE, GALLASCH, SEEKAMP AND TYERMAN
vided by the pressure gradient (MPa) (Tyree and Ewers 1991).
The pressure gradients used were around 9 kPa. The procedure
was duplicated and the stem was then flushed with a degassed
10 mM oxalic acid solution at a pressure of 0.2 MPa to remove
embolisms. Hydraulic conductivity was measured after the
oxalic acid flush to obtain the maximum hydraulic conductivity. Repeated flushes of oxalic acid showed that a single flush
yielded the maximum hydraulic conductivity. Embolisminduced loss in hydraulic conductivity was estimated as the
initial hydraulic conductivity as a percentage of the maximum
hydraulic conductivity. Hydraulic conductivity divided by the
leaf area terminal of the selected segments gave leaf specific
conductivity (kg s –1 mm –1 MPa –1).
Xylem vessel diameters were determined in wood samples
obtained from randomly chosen 1-cm diameter branches located throughout the canopy of six individuals of each
eucalypt type. Three replicate wood samples per tree were
placed in 75% ethanol to prevent contamination by fungi and
microorganisms and subsequent decay of the wood sample.
Preparation of the wood involved removing the bark from the
twig and trimming the transverse section with a microtome to
obtain a smooth surface. The samples were then reduced in
size to a 5-mm cube (encompassing both the pith and outer
edge of each wood sample) and mounted on SEM stubs. Samples were dried and gold coated with a Polaron SEM Autocoating Unit E5200 (VG Microtech, West Sussex, U.K.) and
examined with the aid of a scanning electron microscope.
Three photographs of each wood sample were taken of the
outer region of wood adjacent to the bark layer, at each 120°
rotation of the sample. The diameters of 10 randomly chosen
xylem vessels per photograph were calculated by averaging
the vessel diameter measured in two directions.
Carbon isotope discrimination
Three replicate sets of mature leaves that were at least 1 year
old were collected from two green variant individuals and
their closest grey variant neighbors at both the Monoman Island and Chowilla Driveway sites. The leaves were dried and
shipped to the Department of Botany, University of
Queensland, Australia, where they were analyzed for carbon
isotope discrimination, courtesy of Professor George Stewart.
Isotopic compositions were converted from δ13C (relative to
PeeDee Belemnite) to discrimination (∆) taking the δ13C of air
as –8‰ (Farquhar et al. 1989):
∆ = (δ air − δ plant ) (1 + δ plant ).
Stable isotopes of water
Twigs of 1-cm diameter were collected from paired green and
grey variant trees at each of the four regular study sites. Three
extra pairs of trees from the Monoman Island site were also
sampled. The samples were immediately cut into 5-cm lengths
and all lateral branches and leaves were removed. Bark was
scraped away to eliminate enriched water in the bark contaminating xylem water. The samples were immersed in kerosene
in water-tight jars for transport to the laboratory. Water was
extracted from the samples by azeotropic distillation (Revesz
and Woods 1990). Isotopic analysis was conducted at the
CSIRO Division of Water Resources in Adelaide, Australia,
under the guidance of Ms. Keryn McEwan. Isotopic concentrations are given based on the standard delta notation and are
relative to Vienna standard mean ocean water.
Statistical analysis
Means and standard errors were calculated to show seasonal
trends in g, Ψl and foliar ion concentration, and diurnal trends
in g and Ψl. To examine the difference between variants for
water potential and stomatal conductance, the main study
years (1992–1994) were divided into cool (May–August) and
warm (October–March) periods based on a mean day temperature threshold of 20 °C. Paired t-tests were then performed
for the mean values for each of the two periods for paired
green and grey trees, with sites providing the replicates. To determine if osmotic potential of xylem sap was significantly
different between variants, paired t-tests were used for data
obtained during the warm months of 1997. Paired sites were
analyzed separately by three-way analyses of variance
(ANOVA) to determine differences in foliar ion concentration
between variants, between sites and over time. Two-way
ANOVA was used to determine the difference in ∆ between
variants and sites. One-way ANOVA was used to compare diameter of xylem vessels, conductivity of stem segments and
leaf specific hydraulic conductivity between grey and green
variants. Log transformations were used where appropriate.
To determine if g was correlated with solar irradiance, leaf
temperature or Ψl for either variant, Pearson correlations were
conducted and regression analyses were used to assess the relationship between g and Ψl at both midday and predawn.
Multi-linear regression was used to determine the relationship
between predawn Ψl and groundwater salinity and depth to the
water table. Analyses were performed with SPSS V 8.0 (SPSS
Inc., Chicago, IL) or Prism V 3.0 (GraphPad, Inc., San Diego,
CA) software.
Results
Water potential and osmotic potential of xylem sap
At all sites, predawn Ψl was always more positive in the grey
variant compared with the green variant (Figure 1a). For comparison, leaf conductance to water vapor (g) averaged over the
warm and cool periods is shown in Figure 1b. At the Coppermine North site, Ψl reached a steady value in both variants
well before dawn (Figure 2), indicating that both variants
equilibrated overnight. Mean Ψl for both variants was about
0.4 MPa more negative at midday than at predawn. Both variants exhibited similar patterns of change in predawn and midday Ψl throughout the year.
Across all sites, mean osmotic potential of xylem sap was
–0.22 ± 0.09 MPa and –0.09 ± 0.02 MPa for green variant and
grey variant trees, respectively. There was no significant difference in xylem sap osmotic potential between green and
grey variants (paired t-test, P = 0.3).
TREE PHYSIOLOGY VOLUME 20, 2000
SALT TOLERANCE IN A EUCALYPTUS LARGIFLORENS HYBRID
957
Soil profiles and stable isotopes of water
Figure 1. (a) Predawn water potential (Ψl) and (b) stomatal conductance (g) of green and grey variant trees on the Chowilla floodplain.
Four replicate pairs of green and grey variant trees were examined
over 2 years. The years were split into cool and hot periods based on a
temperature threshold of 20 °C (see text). The mean values of each individual tree were calculated for the two periods from several separate
field trips in each period. Within each period, values followed by different letters are significantly different at P < 0.05 using paired t-tests.
Error bars indicate SEM.
Figure 3 shows soil depth profiles of matric potential, osmotic
potential, total soil water potential (Ψsoil) and chloride ion
(Cl – ) concentration (right axis). Also shown (horizontal lines)
are the predawn Ψl data obtained on the same field trip (January 1993). Correspondence between predawn Ψl and Ψsoil indicates approximately where in the soil profile the roots may
be extracting water. Based on the correspondence between
predawn Ψl and Ψsoil, the difference in depth of soil water extraction between variants was only about 10–20 cm. At
Monoman Island, Coppermine North and Coppermine South,
water extraction by trees seemed to be confined to a shallow
depth of less than about 50 cm. At Chowilla Driveway, water
extraction by trees appeared to be occurring at greater depth,
because Ψl matched Ψsoil at 280 cm. Soil Cl – concentration increased with depth. Comparison of soil Cl – concentration with
the apparent depths of root water extraction indicated that
water extraction occurred up to a soil Cl – concentration of
50 mol m –3 at each site (arrows in Figure 3).
To determine whether trees growing at saline sites with
shallow water tables are forced to extract water higher in the
profile where the soil is generally drier, a multilinear regression was performed to assess the relative contribution of salinity and depth to water table to predawn Ψl. Figure 4 shows that
for both variants there were good correlations between predawn Ψl (yearly mean) and salinity of the groundwater and
depth to the water table. Increasing salinity correlated with decreasing mean Ψl, and the Pearson correlation coefficients for
both variants (green = –0.901, grey = –0.859) were significant
(P = 0.001 and P = 0.003, respectively). The effect of depth to
the water table on predawn Ψl, although in the direction predicted (green = 0.545, grey = 0.595), was not significant (P =
0.081 and P = 0.060, respectively). Linear regression of Ψl
versus salinity of groundwater indicated that the two variants
did not differ in slope (green = –0.0487, grey = –0.0455, P =
0.83), but differed significantly in intercept (green = –0.552,
grey = –0.271, P = 0.01).
Neither δ18O nor δ2H differed between variants across sites
using paired green and grey individuals (paired t-test; δ18O,
P = 0.516; δ2H, P = 0.345). For green and grey variants, the
mean values of δ18O were –2.828 ± 0.493‰ and –3.070 ±
0.471‰, respectively, and for δ2H they were –29.13 ± 0.87‰
and –27.70 ± 1.21‰, respectively.
Leaf gas exchange
Figure 2. Leaf water potential (Ψl) measured at approximately 3-h intervals in trees at the Coppermine North site. Measurements were recorded over a 24-h period for a green and grey variant pair in close
proximity to each other. Each value is the mean of three twig samples.
Error bars indicate SEM.
Leaf conductance differed significantly between variants only
in the cooler months, reflecting a greater seasonal variation in
g for the grey variant (Figure 1b, Table 2), with similar values
of g for the two variants during the warmer months.
Figure 5 illustrates the three patterns of diurnal variation in
g observed at different times during the year. The pattern of
early opening and progressive closure throughout the day
(Figure 5a) occurred in warm, dry conditions. Mean leaf temperature for this curve was 26 °C. It was under these circumstances that green and grey variants showed similar mean
daily g (Figure 1b). Under cool and dry conditions, a clear
TREE PHYSIOLOGY ON-LINE at http://www.heronpublishing.com
958
ZUBRINICH, BYRNE, GALLASCH, SEEKAMP AND TYERMAN
Figure 3. Soil water potential
and Cl – concentration profiles
with depth at the four main
study sites. The left axis of
each profile gives matric (䉱),
osmotic (䊏) and total water
potential (䉲) of soil. The right
axis gives the soil Cl – concentration (ⵧ). Soil samples from
each site were collected on
consecutive days in January
1993. Dashed and solid horizontal lines indicate predawn
Ψl on the day the soil samples
were collected for the grey
(gy) and green (gn) variant, respectively. The dashed vertical
arrows show the depth at
which the Cl – concentration
was 50 mol m –3. This value
correlated approximately to
the depth of water extraction
as judged by the intercept of
the relationship between predawn Ψl and total soil water
potential.
midday depression in g was observed for both tree variants
(Figure 5b). Mean leaf temperature on this day was 15 °C.
Also illustrated is a summer diurnal trend (Figure 5c) observed
after significant rainfall in the preceding 4 months, when g was
at a maximum for both variants.
Although g showed less variation over the year in the green
variant than in the grey variant, g of the green variant was significantly correlated with light intensity, predawn Ψl and leaf
temperature (all sites pooled, Table 3). The grey variant did not
show significant correlations with these parameters (Table 3),
although the slopes of the regression lines were similar for
both variants (see also Figure 7a). More scatter was observed
in the g data for the grey variant than for the green variant.
Diurnal CO2 assimilation rates (A) were measured at Coppermine North and Coppermine South in January and July
1993 and Chowilla Driveway in January 1993. Although
mean A was lower in the green variant than in the grey variant,
the difference was significant only in July (P = 0.046). For
January, mean daily A values were 4.9 ± 0.6 µmol m –2 s –1
(green) and 6.1 ± 0.9 µmol m –2 s –1 (grey). For July, the rates
were 2.4 ± 0.5 µmol m –2 s –1 (green) and 3.6 ± 0.6 µmol m –2 s –1
(grey). Both variants showed a relatively poor but significant
correlation between A and g (Figure 6). Although there was no
significant difference in the slopes of A versus g between the
variants, there were large differences in the range of A and g
values.
Figure 4. Dependence of predawn Ψl on soil conditions at
sites across the Chowilla
floodplain. The mean annual
predawn Ψl for (a) grey and
(b) green variant trees is plotted against ground water electrical conductivity (EC) and
depth to ground water. Two
separate data collection periods were used for each variant:
1992–1994 and 1997. The
1997 data included additional
sites to the four main study
sites. The fitted surfaces are
given by the following equations: Ψl = –1.996 + 0.0036D + –0.0392EC (grey); Ψl = –1.785 + 0.0025D + –0.0443EC (green), where D = depth. The adjusted r 2 values are
0.72 (grey) and 0.78 (green). Analysis of variance showed that the regressions were significant (P = 0.02 (grey) and P = 0.01 (green)).
TREE PHYSIOLOGY VOLUME 20, 2000
SALT TOLERANCE IN A EUCALYPTUS LARGIFLORENS HYBRID
959
Table 2. Maximum, minimum and range of leaf conductance (g) at
each site. Data from Chowilla Driveway and Monoman Island were
collected between January 1992 and January 1993. Data from Coppermine North and South were collected between January 1993 and
February 1994.
Site
g (mmol m –2 s –1)
Variant
Max.
Min.
Range
Chowilla Drive
Green
Grey
109.88
183.01
30.22
30.64
79.66
152.37
Monoman Island
Green
Grey
87.62
184.08
17.65
36.09
69.97
147.99
Coppermine North
Green
Grey
69.37
110.82
12.72
42.30
56.65
68.52
Coppermine South
Green
Grey
104.14
199.33
34.38
42.99
69.66
156.34
Carbon isotope discrimination
Values of ∆ differed significantly between the two sites examined (Chowilla and Monoman; P = 0.04) and were highly significant between variants (P = 0.0007) with no significant
interaction (P = 0.73). There was a significant correlation between ∆ and mean predawn Ψl taking both variants together
(Figure 7a), such that more negative predawn Ψl correlated
with larger ∆ (r 2 = 0.94, P < 0.001). Figure 7b shows mean annual g plotted against mean annual Ψl for each site and variant.
For both variants together, there was a significant correlation
between mean annual g and mean annual predawn Ψl, with g
declining as Ψl became more negative (r 2 = 0.617, P = 0.02).
The higher ∆ in the green variant than in the grey variant indicates a higher long-term mean internal CO2 concentration.
Based on Equation 8 of Farquhar et al. (1989) (see also
Ehleringer et al. 1992), which relates ∆ to the ratio of internal
to external CO2 concentrations (ci /ca), the green variant operates on average with a ci /ca of 0.62 compared with a ci /ca of
0.58 for the grey variant. These ratios were calculated based
on fractionation caused by diffusion in air equal to 4.4‰ and
fractionation caused by carboxylation equal to 27‰ (Farquhar
et al. 1989).
Leaf ion concentrations
+
+
–
Table 4 presents mean Na , K and Cl concentrations for two
sites and variants for 1992. Mean foliar ion concentrations
were significantly different between variants (P < 0.01). Concentrations of Na+ and Cl – were higher in the green variant
than in the grey variant, whereas K+ concentration was similar
between variants. In addition, there was a significant decrease
in foliar Na+ and Cl – concentrations between January 1992 and
November 1992 for both variants (P < 0.001). The foliar
Na +:K + ratio also declined throughout the year (P < 0.001).
Foliar Cl – was significantly higher in trees at Monoman Island than at Chowilla Driveway for both variants (P < 0.001).
Although foliar Na+ and Na+:K + ratio were higher for both
variants at Monoman Island than at Chowilla Driveway, the
difference was not statistically significant.
Figure 5. Representative daily patterns of stomatal conductance (g)
measured on a pair of green and grey variant trees at the Chowilla
Driveway site. The time course in January 1992 is representative of a
hot period when the green and grey variant trees had similar g. The
time course in August 1992 is representative of a cool period. The
time course in December 1992 was observed after summer rains and
represents the maximum conductance likely to be observed for both
variants on the floodplain. Each value is the mean of five measurements on each of a green and grey pair of trees. Error bars indicate
SEM.
Stem hydraulic conductivity and xylem vessel radius
Perfusion of stem segments under pressure resulted in increased hydraulic conductivity in both variants. Hydraulic
conductivity steadily declined over time with pure water per-
TREE PHYSIOLOGY ON-LINE at http://www.heronpublishing.com
960
ZUBRINICH, BYRNE, GALLASCH, SEEKAMP AND TYERMAN
Table 3. Pearson correlations of leaf conductance to solar irradiance,
leaf temperature and predawn xylem water potential for the green and
grey variants. Number of cases is denoted in parenthesis, ns = not significant, ** = P < 0.01, and * = P < 0.05.
Factor
Green
Grey
Solar irradiance
0.4323
(29)
**
0.2596
(29)
ns
Leaf temperature
0.3898
(29)
*
–0.3453
(29)
ns
Predawn Ψl
0.4890
(23)
**
0.3944
(23)
ns
Figure 6. Assimilation plotted against stomatal conductance for green
(䊊) and grey (䊏) variant trees from two sites measured during summer and winter. The regression lines shown (grey = solid line, green =
dashed line) were forced through zero. Dotted lines indicate 95% confidence limits for the slopes. The slopes of the regression lines are:
green variant = 0.073 µmol (CO2 ) mmol –1 (H2O); grey variant =
0.054 µmol (CO2 ) mmol –1 (H2O).
Figure 7. (a) Mean annual stomatal conductance and (b) 13C discrimination plotted against mean annual predawn Ψl for green and grey
trees. Discrimination was measured on mature leaves at the Chowilla
Driveway and Monoman Island sites. Stomatal conductance is an annual mean for each of the four study sites (mean of average daily g).
Error bars indicate SEM. Regression lines shown are through all the
data (green and grey pooled). The slopes and intercepts of the lines,
respectively, are: (a) 44.63 mmol m –2 s –1 MPa –1, 203 mmol m –2 s –1;
r 2 = 0.617, P = 0.021; (b) –1.395‰ MPa –1, 14.13‰; r 2 = 0.941, P <
0.0001.
Discussion
fusions, whereas hydraulic conductivity was maintained over
time with oxalic acid perfusions. Oxalic acid perfusions were
therefore routinely used for measuring the maximum conductivity of the stems.
Initial and maximum conductivities were significantly
higher in the grey variant than in the green variant (Table 5).
However, there was no significant difference in leaf specific
conductivity (Table 5). Both variants showed a high degree of
embolism; 81% of the xylem in the green variant was nonfunctional, compared with 65% for the grey variant.
Mean xylem vessel diameter was significantly larger in grey
variant trees than in green variant trees (46 versus 38 µm;
t-test, P < 0.01).
The most obvious difference between green and grey variants
was in leaf water potential, which was significantly lower in
the green variant at all sites and at all times. The difference between variants in water potential was observed both at midday,
when water potential was close to its minimum daily value,
and just before dawn, when it reached its maximum daily
value. The more negative water potentials in the green variant
compared with the grey variant cannot be attributed to xylem
osmotic potential, because it was similar in the two variants.
Furthermore, the difference was not caused by a lower stem
hydraulic conductivity in the green variant, because both the
change in water potential from predawn to midday and leaf
specific hydraulic conductivity were similar for both variants.
Myers (1995) found that the lignotuber of a mallee eucalypt
TREE PHYSIOLOGY VOLUME 20, 2000
SALT TOLERANCE IN A EUCALYPTUS LARGIFLORENS HYBRID
961
Table 4. Concentrations of Na+, K + and Cl – (µmol gdw–1; mean and SEM) and Na+:K + ratio for green and grey variants at Chowilla Driveway and
Monoman Island (n = 5 trees of each variant).
Site and variant
Na+
K+
Cl –
Na+:K +
Chowilla Driveway
Grey
Green
135.0 (12.9)
168.7 (13.2)
120.8 (6.0)
111.7 (8.3)
68.8 (5.3)
105.5 (6.1)
1.2 (0.1)
1.6 (0.1)
Monoman Island
Grey
Green
158.8 (14.9)
186.7 (17.7)
121.5 (5.4)
114.4 (7.3)
91.4 (7.4)
152.9 (9.5)
1.3 (0.1)
1.5 (0.1)
Table 5. Stem hydraulic conductivity in the green and grey variants, measured during 1994. Initial conductivity is the pre-flushed value, maximum conductivity is the value after flushing. Leaf specific conductivities were calculated from the initial conductivity and the leaf area distal to
the stem segment. Within each measurement group, values followed by different letters are statistically different (t-test; P < 0.05). Values are
means (SEM), n = 6.
Variant of E. largiflorens
Initial conductivity
(kg s –1 mm MPa –1)
Max. conductivity
(kg s –1 mm MPa –1)
Leaf specific conductivity
(kg s –1 mm –1 MPa –1) × 10 –6
Grey
Green
0.070 (0.019) a
0.025 (0.004) b
0.203 (0.028) a
0.130 (0.014) b
0.91 (0.14) a
1.20 (0.25) a
was the site of low hydraulic conductance, which could result
in the low water potentials often observed for mallees. Because Eucalyptus gracilis (a mallee) is the putative paternal
parent of the green variant (Zubrinich 1996), it is possible that
the more negative water potentials of the green variant are associated with a more developed lignotuber than in the grey
variant. If so, the difference in predawn Ψl between variants
may not reflect different zones of water extraction from the
soil. For both variants, Ψl equilibrated overnight, so the lower
Ψl of the green variant cannot be attributed to incomplete
equilibration. Seedling experiments have also demonstrated
that both variants equilibrate with the soil (Zubrinich 1996). If
significantly different sources of water were being used by the
two variants, a difference in the stable isotopes of water between paired green and grey variant trees should have been detected; however, no difference was observed, perhaps because
the difference was small, as indicated by the steep water potential gradient where the trees were extracting water (Figure 3
and Mensforth et al. 1994). Alternatively, the grey and green
variants may be using the same water source (groundwater-derived), but the green variant may be extracting more water
from a given point in the soil profile, resulting in a greater
buildup of salt around its roots. Consequently, the green variant trees would be exposed to a higher soil salt concentration
than the grey variant trees, which is consistent with the finding
of a higher salt concentration in leaves of green variant trees
than in leaves of grey variant trees. We conclude, therefore,
that the green variant is able to extract water from a drier or
more saline region of the soil than the grey variant.
Under conditions where the groundwater is salinized, the
upward movement of groundwater is combined with movement of solutes into the plant root zone (Thorburn et al. 1995).
Uptake of water by plants concentrates solutes in the soil until
a threshold, which differs for different plant species, is
reached (Jolly et al. 1993). Above this threshold, plants can no
longer take up water because of the salt and osmotic effects
that are imposed by the concentrated solutes in the soil. It
would appear that the green variant can either: (1) extract water slightly higher up in the profile above the saline front
where the total water potential is lower because of a lower
matric potential; or (2) extract water for a longer period from a
saline region because it has a higher threshold for salt buildup
around its roots. Both possibilities would extend the period of
water extraction by the green variant, and combined with more
conservative water use, would enable the green variant to outperform the grey variant on saline sites.
The range of Ψl values for both variants was generally lower
than values reported for eucalypt species inhabiting semi-arid
environments (Sinclair 1980, Myers and Neales 1984, Crombie and Milburn 1988, Davidson and Reid 1989). For
E. microcarpa (Maiden) Maiden and E. behriana F. Muell.,
Myers and Neales (1984) noted a threshold in diurnal leaf conductance trends between trees exhibiting leaf water potentials
above and below –2.5 MPa. Leaf water potentials below
–2.5 MPa were correlated with a leaf conductance pattern that
was characteristic of water-stressed trees. In our study, predawn and midday water potentials were lower than –2.5 MPa
at all sites except Chowilla Driveway, where grey variant trees
had water potentials greater than –2.5 MPa throughout the
year. Trees at this site, which had the least saline and deepest
water table and was therefore considered to be the most favorable site, were visibly healthier than trees at Monoman Island.
Although Myers and Neales (1984) reported that the Eucalyptus species with the highest water potential were healthiest, we
TREE PHYSIOLOGY ON-LINE at http://www.heronpublishing.com
962
ZUBRINICH, BYRNE, GALLASCH, SEEKAMP AND TYERMAN
found that green variant trees exhibited little dieback, even
though they had lower water potentials than grey variant trees.
The low hydraulic conductivity of both variants supports
the view that arid land trees and shrubs generally have low hydraulic conductivity (Klob et al. 1996). The lower initial and
maximum hydraulic conductivities of stems of the green variant compared with the grey variant can be attributed mostly to
smaller vessel diameter and increased embolism in the green
variant. The small diameter of xylem vessels in the green variant may be a characteristic inherited from E. gracilis, which
has xylem vessels of smaller diameter than either the green or
grey variants (Zubrinich 1996). Water stress may also affect
the development of xylem vessels, resulting in a decrease in
xylem diameter (February et al. 1995). Generally, xylem vessels with a larger diameter are more likely to cavitate with increasing salinity or water deficit than small-diameter vessels
(Cochard and Tyree 1990, Lo Gullo et al. 1995). However, we
found a greater proportion of embolized vessels in the green
variant than in the grey variant, indicating that the smalldiameter xylem vessels did not fully compensate for the
greater xylem tensions in the green variant.
Although the percentage of embolism was large in both
variants, and higher in the green variant (81 versus 65%), it did
not lead to runaway embolism (Tyree and Sperry 1989). In
E. largiflorens, cavitation may be an important way of releasing water in the stem under conditions of extreme water stress
(Grace 1993). High degrees of embolism (83%) were also
measured for chaparral shrub of southern California where
water potentials as low as –10 MPa were measured (Williams
et al. 1997). For E. camaldulensis Dehnh. seedlings, Franks et
al. (1995) showed that water potentials of about –4 MPa are
required to induce a high degree (80%) of embolism. Water
potentials of –4 MPa were frequent in both variants of E. largiflorens on the Chowilla floodplain.
After a major rainfall, stomatal conductance and water potential increased in both variants. This could only occur as a
result of water uptake by surface roots, suggesting an opportunistic water-use strategy, as has also been observed for E. camaldulensis growing on the floodplain (Mensforth et al. 1994)
and for Eucalyptus species with dimorphic root systems
(Dawson and Pate 1996). The more variable g in the grey variant trees throughout the year may indicate that the grey variant
is more reliant on surface roots than the green variant.
Although it was apparent that changes in g throughout the
year correlated partially with environmental conditions, g did
not correlate with tree health. A similar result was observed by
Crombie and Milburn (1988) in a comparative study of
healthy eucalypts and eucalypts showing dieback. They concluded that the physiological effects of dieback were too small
for g to be useful as an indicator of stress. Similarly, several
studies have documented suppression of growth and dieback
of limbs before an effect on photosynthesis was evident (Papp
et al. 1983, Seeman and Critchley 1985, Munns 1993). We
found that predawn water potential showed a significant correlation with substrate severity, as measured by salinity of the
groundwater. Based on our predawn water potential data, we
conclude that, at Coppermine North, the most degraded site at
which grey variants are still surviving, the trees may not be
able to extract sufficient water to support long-term growth if
there is further salinization of the soil.
The ci /ca ratios in the green variants calculated from the discrimination values were higher than in the grey variants, indicating a greater biochemical limitation to photosynthesis in
the green variant, perhaps resulting from the effects of operating at low water potentials. However, direct effects of salinity
on the photosynthetic apparatus may result in non-stomatal inhibition of photosynthesis (Brugnoli and Lauteri 1991). Turgor-dependent processes are also likely to be affected by
salinity if osmotic adjustment is insufficient to counteract the
decline in xylem water potential (Walker et al. 1983).
The ∆ values for the two variants are similar when compared with the range reported for Eucalyptus species from diverse habitats growing in a common garden (Anderson et al.
1996). However, when comparing sites and variants, both ∆
and g values appear to lie along the same trend line when plotted against predawn Ψl, suggesting that lower water potential
is the sole cause for the differences between the green and grey
variant trees. Thus, the green variant may have the same characteristics in terms of photosynthesis and stomatal conductance as the grey variant at any given time, provided that it has
the same water potential.
The putative hybrid origin of the green variant appears pertinent to its ability to tolerate highly saline groundwater conditions. Small differences in anatomy and root growth patterns
may account for the green variant’s ability to lower its water
potential further than the grey variant. Nevertheless, as the upwardly moving front of salt nears the surface of the soil, the
green variant will ultimately suffer water and salt stress, as indicated by dieback of green variants at the Tareena site. Thus,
although green variant trees are outperforming grey variant
trees under the arid conditions of the Chowilla Anabranch region, the long-term survival of the green variant is dependent
on periodic flooding of the Chowilla floodplain to leach salt
and recharge the soil horizons above the saline water table
(Akeroyd et al. 1998).
Acknowledgments
This research was supported financially by the Flinders University of
South Australia and the Murray Darling Basin Commission. We
gratefully acknowledge the help of Drs. Tony Condon and Owen
Atkin for feedback on the manuscript and Professor George Stewart
13
for kindly providing the C-isotope analysis.
References
Akeroyd, M.D., S.D. Tyerman, G.R. Walker and I.D. Jolly. 1998. Impact of flooding on the water use of semi-arid riparian eucalypts.
J. Hydrol. 206:104–117.
Anderson, J.E., J. Williams, P.E. Kriedemann, M.P. Austin and G.D.
Farquhar. 1996. Correlations between carbon isotope discrimination and climate of native habitats for diverse eucalypt taxa growing in a common garden. Aust. J. Plant Physiol. 23:311–320.
TREE PHYSIOLOGY VOLUME 20, 2000
SALT TOLERANCE IN A EUCALYPTUS LARGIFLORENS HYBRID
Brugnoli, E. and M. Lauteri. 1991. Effects of salinity on stomatal conductance, photosynthetic capacity and carbon isotope discrimination of salt tolerant (Gossypium hirsutum L.) and salt-sensitive
(Phaseolus vulgarus L.) C3 non-halophytes. Plant Physiol. 95:
628–635.
Cochard, H. and M.T. Tyree. 1990. Xylem dysfunction in Quercus:
vessel sizes, tyloses, cavitation and seasonal changes in embolism.
Tree Physiol. 6:393–407.
Crombie, D.S. and J.A. Milburn. 1988. Water relation of rural dieback. Aust. J. Bot. 36:233–237.
Davidson, N.J. and J.B. Reid. 1989. Response of eucalypt species to
drought. Aust. J. Ecol. 14:139–156.
Dawson, T.E. and J.S. Pate. 1996. Seasonal water uptake and movement in root systems of Australian phreatophytic plants of dimorphic root morphology: a stable isotope investigation. Oecologia
107:13–20.
Ehleringer, J.R., S.L. Phillips and J.P. Comstock. 1992. Seasonal
variation in the carbon isotopic composition of desert plants. Funct.
Ecol. 6:396–404.
Farquhar, G.D., J.R. Ehleringer and K.T. Hubick. 1989. Carbon isotope discrimination and photosynthesis. Annu. Rev. Plant Physiol.
Plant Mol. Biol. 40:503–537.
February, E.C., W.D. Stock, W.J. Bond and D.J. Le Roux. 1995. Relationships between water availability and selected vessel characteristics in Eucalyptus grandis and two hybrids. IAWA J.
16:269–276.
Franks, P.J., A. Gibson and E.P. Bachelard. 1995. Xylem permeability and embolism susceptibility in seedlings of Eucalyptus
camaldulensis Dehnh. from two different climatic zones. Aust. J.
Plant Physiol. 22:15–21.
Greacen, E.L., G.R. Walker and P.G. Cook. 1989. Procedure for filter
paper method of measuring soil water suction. CSIRO Division of
Soils Divisional Report No. 108. CSIRO, Canberra, Australia, 7 p.
Grace, J. 1993. Consequences of xylem cavitation for plant water deficits. In Water Deficits, Plant Responses from Cell to Community.
Eds. J.A.C. Smith and H. Griffiths. Bios Scientific Publishers, Oxford, U.K., pp 109–128.
Jolly, I.D., G.R. Walker and P.J. Thorburn. 1993. Salt accumulation
in semi-arid floodplain soils with implications for forest health.
J. Hydrol. 150:589–614.
Klob, K.J., J.S. Sperry and B.B. Lamont. 1996. A method for measuring xylem hydraulic conductance and embolism in entire root and
shoot systems. J. Exp. Bot. 47:1805–1810.
Lo Gullo, M.A., S. Salleo, E.C. Piaceri and R. Rosso. 1995. Relations
between vulnerability to xylem embolism and xylem conduit dimensions in young trees of Quercus cerris. Plant Cell Environ.
18:661–669.
Mensforth, L.J., P.J. Thorburn, S.D. Tyerman and G.R. Walker. 1994.
Sources of water used by riparian Eucalyptus camaldulensis overlying highly saline groundwater. Oecologia 100:21–28.
Munns, R. 1993. Physiological processes limiting plant growth in saline soils: some dogmas and hypotheses. Plant Cell Environ.
16:15–24.
963
Myers, B.A. 1995. The influence of the lignotuber on hydraulic conductance and leaf conductance in Eucalyptus behriana seedlings.
Aust. J. Plant Physiol. 22:857–863.
Myers, B.A. and T.F. Neales. 1984. Seasonal changes in the water relations of Eucalyptus behriana F. Muell. and Eucalyptus microcarpa (Maiden) Maiden in the field. Aust. J. Bot. 32:495–510.
Noyce, T. and K. Nicholson. 1992. Chowilla flood and groundwater
modelling. Information Systems Branch, South Australian Department of Environment and Planning, Adelaide, Australia.
Papp, J.C., M.C. Ball and N. Terry. 1983. A comparative study of the
effects of NaCl salinity on respiration, photosynthesis and leaf extension growth in Beta vulgaris L. (sugar beet). Plant Cell Environ.
6:687–693.
Revesz, K. and P.H. Woods. 1990. A method to extract soil water for
stable isotope analysis. J. Hydrol. 115:397–406.
Seeman, J.R. and C. Critchley. 1985. Effects of salt stress on the
growth, ion content stomatal behaviour and photosynthetic capacity of a salt sensitive species, Phaseolus vulgaris L. Planta 164:
151–162.
Simpson, H.J. and A.L. Herczeg. 1991. Salinity and evaporation in
the River Murray evaporation basin, Australia. J. Hydrol.
124:1–27.
Sinclair, R. 1980. Water potential and stomatal conductance of three
Eucalyptus species in the Mount Lofty Ranges, South Australia:
responses to summer drought. Aust. J. Bot. 28:499–510.
Sperry, J.S., M.T. Tyree and J.R. Donnelly. 1988. Vulnerability to
embolism in mangrove vs. an inland species of Rhizophraceae.
Physiol. Plant. 74:276–283.
Thorburn, P.J., T.J. Hatton and G.R. Walker. 1993. Combining measurements of transpiration and stable isotopes of water to determine groundwater discharge from forests. J. Hydrol. 150:563–587.
Thorburn, P.J., G.R. Walker and I.D. Jolly. 1995. Uptake of saline
groundwater by plants: an analytical model for semi-arid and arid
areas. Plant Soil. 175:1–11.
Tyree, M.T. and F.W. Ewers. 1991. The hydraulic architecture of
trees and other woody plants. New Phytol. 119:345–360.
Tyree, M.T. and J.S. Sperry. 1989. Vulnerability of xylem to cavitation and embolism. Annu. Rev. Plant Physiol. Plant Mol. Biol.
40:19–38.
Walker, R.R., E. Torokfalvy, A.M. Grieve and L.D. Prior. 1983. Water relations and ion concentrations of leaves on salt-stressed citrus
plants. Aust. J. Plant Physiol. 10:265–277.
Williams, J.E., S.D. Davis and K. Portwood. 1997. Xylem embolism
in seedlings and resprouts of Adenostoma fasciculatum after fire.
Aust. J. Bot. 45:291–300.
Zubrinich, T.M. 1996. An investigation of the ecophysiological, morphological and genetic characteristics of Eucalyptus largiflorens
F. Muell and Eucalyptus gracilis F. Muell in relation to soil salinity
and groundwater conditions throughout the Chowilla Anabranch.
Ph.D. Thesis, Flinders Univ. of South Australia, Australia, 338 p.
TREE PHYSIOLOGY ON-LINE at http://www.heronpublishing.com