RESEARCH New Phytol. (2000), 146, 119–127
Effects of salinity on xylem structure and
water use in growing leaves of sorghum
STUART F. BAUM", PHONG N. TRAN#  WENDY K. SILK#*
"Section of Plant Biology, University of California, Davis, CA 95616, USA
#Department of Land, Air, and Water Resources, University of California, Davis,
CA 95616, USA
Received 12 July 1999 ; accepted 13 December 1999

Within the growth zone, salt-affected leaves of sorghum (Sorghum bicolor) had narrower protoxylem and
metaxylem cells than controls. Leaf width and cross-sectional area were also reduced, so that the salt treatment
had no effect on the area of protoxylem per area of leaf cross section. Dye uptake studies suggested that in controls
most of the veins, but in salt-affected leaves only half of the veins, are functional in water transport. Volumetric
water flow was greatly diminished in the salt-affected plants. The reduced flow rate was largely explained by the
salt-induced decrease in leaf surface area. Some decreases in flow rates per unit leaf mass or area were also
produced by salinity, particularly in late developmental stages. Not surprisingly, leaf conductance measured with
a diffusion porometer did not appear to be correlated with the diameter of the protoxylem or metaxylem elements.
By contrast, published values of water deposition rates are strongly related to the size of the protoxylem elements :
the rates of water deposition into the growing leaf tissue are proportional to the square of the protoxylem radius.
Thus environmentally produced change of the hydraulic architecture of monocot leaves may cause change in local
growth rates.
Key words : protoxylem, water use, Sorghum bicolor (sorghum), leaf, transport law, growth rate, salinity, hydraulic
architecture.

Worldwide, effects of salinity on plant development
are increasingly problematic for irrigated crops.
Sorghum, a moderately salt-tolerant species, shows
decreased growth rates and changes in spatial
patterns of nutrient deposition in NaCl-affected
leaves (Bernstein et al., 1993, 1995). The local rates
of water deposition associated with leaf growth are
reduced by salinity. And salinity reduces tissue
content and deposition rates of calcium, an element
thought to move primarily in the transpiration
stream.
In this paper we explore some aspects of hydraulic
architecture that might be related to transpiration
and also to water use during growth. Our basic
premise follows the classical work of Sharman (1942)
and Esau (1953) who concluded that protoxylem
vessels supply water to expanding monocot leaves.
The protoxylem becomes occluded during leaf
expansion, while metaxylem vessels mature after
cessation of tissue expansion and conduct large
volumes of water into the mature leaves during
transpiration.
*Author for correspondence (fax j00 1 530 752 1552 ; e-mail
wksilk!ucdavis.edu).
In recent years considerable attention has been
focussed on the development of conducting tissue in
monocot leaves. A great deal has been learned about
the ontogeny of anatomy. Dengler and colleagues
have delineated the development of bundle sheath
cells in C and C species (Dengler et al., 1994),
$
%
while Evert and colleagues have described the
development of the vascular system in barley leaves
(Trivett & Evert, 1998). Tomos and colleagues have
shown that there is considerable cell-to-cell variation
in the patterns of water and ion relations (Fricke et
al., 1995). Early reports in the literature indicate that
salinity causes the development of xylem elements
with decreased diameter (Hayward & Long, 1941).
Here, we asked whether salt-induced changes in
metaxylem diameter affect transpiration rate, and
whether changes in protoxylem diameter affect water
deposition rates into growing tissue. We measured
the effects of salinity on number and width of xylem
elements in growth zones of salt-affected sorghum
leaves, and water flow rates and stomatal conductance of salt affected and control plants. We then
compared previously published effects of salt on
deposition rates of water to the effects on the xylem
element radius shown here. Finally, we sought a
power law to describe the dependence of water
120
RESEARCH S. F. Baum et al.
deposition rates (associated with the local growth
rates) on protoxylem radius.
  
Plant material and growth conditions
Seeds of sorghum (Sorghum bicolor (L.) Moench, cv.
NK 265) were germinated and seedlings were
cultivated as previously described (Bernstein et al.,
1993). As in the earlier experiments, plants were
grown in modified half-strength Hoagland’s solution, with iron supplied as 50 µM Fe-EDTA plus 20
µM Fe(NH SO ) and the control Na level elevated to
% %
1 mM, while pH was maintained at 5n7. Humidity
was 65% during the light period ; the photon flux
was 450 µmol m−# s−" ; temperature was 28mC at the
base of the leaves (the growth zone). Eight days after
germination, sodium chloride was added to make 25
mol m−$ saline treatments. At this time leaf 4 had not
yet emerged from the whorl of enclosing older leaf
sheaths ; and leaf 6, if visible as a primordium, had
not yet begun rapid elongation. Then for three
consecutive days sodium chloride concentration was
increased until the desired concentration of 100 mol
m−$ was reached, 11 d after germination.
Flow, transpiration, and conductance rates
Water flow through plants (as g water per plant per
day) was calculated by weighing cultivation flasks
before and after 48 h of undisturbed growth. To
assess the effect of the plant architecture produced
by 10 d of treatment, on day 18 some salinized plants
were placed in control medium for 48 h ; and some
control plants were placed in saline medium for 48 h.
To obtain transpiration rates (gravimetric flow rates
per unit leaf mass or leaf surface area), a separate set
of plants was destructively harvested at 2-d intervals
to obtain average fresh weight and surface area of
leaves for the different treatments over time. Abaxial
stomatal conductance was measured with a steadystate diffusion porometer designed and built by T.
Hsiao (Puech-Suanzes et al., 1989). Measurements
were made near the centre of the lamina on unshaded
leaves held 88 cm from the overhead light source.
Anatomy
Leaf 6 was selected for anatomical analysis because it
had not yet begun to elongate visibly when the salt
treatment was initiated. Leaves were harvested when
the leaf tip had emerged at least 20 mm (salt-affected
plants) or 30 mm (controls) above the ensheathing
base of the older leaves. Leaves of the treated and
control plants were the same chronological age.
Leaves of control plants were elongating more
rapidly than those of treated plants, but all sampled
leaves were elongating linearly with time. Since salt
changes the rates of some developmental processes,
some developmental effects might be included in the
apparent differences between treatments.
We avoided harvesting those plants with leaves
that had torn during development in the saline
treatment. Older leaves were removed, and the
younger leaf bases were cut into three sections with
excision at 13, 25 and 35 mm from the base of leaf 6.
The 13-mm location was chosen because the relative
elemental growth rate is maximal at 13 mm from the
leaf base in both salt-affected and control leaves
elongating at a constant rate (Bernstein et al., 1995).
For paraffin sectioning the leaf material was fixed in
chromic acid : acetic acid : formaldehyde (Craf III)
for 48 h (Sass, 1958) and then dehydrated in a graded
tertiary-butanol series followed by infiltration and
embedding in Paraplast Plus paraffin. The embedded
material was sectioned on a rotary microtome
producing 10-µm-thick sections. The sectioned material was stained with safranin and fast green
(Johansen, 1940). Five plants were sectioned from
each treatment, and all the measurements were
repeated on a second set of plants grown at a different
time.
Functionality of xylem bundles
An aqueous solution of eosin has been used to
visualize the path of water movement through plants
(Eichhorn et al., 1986), and also to stain cellular
cytoplasm (Sass, 1958). Therefore eosin was used to
determine whether water transport occurred primarily through protoxylem elements. Sorghum
shoots were excised at the root-shoot junction, and
the base of the stem was placed in 1 ml of eosin (50
g l−") for 15 min (preliminary observations suggested
that the dye solution spread through the cut stem in
less than this time period). The plants were then
removed from the dye solution, and the first four
leaves were removed. Sections 2 mm long were
excised at 13, 25 and 35 mm from the base of leaf 6.
The tissue sections were quickly submerged in
Tissue-Tek (Miles Co., Elkhart, IN, USA) and then
frozen on dry ice. The frozen tissue was sectioned
using a Reichert-Jung freezing microtome to give 10
µm thick sections. The sections were viewed using an
Olympus Vanox-T fluorescence microscope (Olympus, Sunnyvale, CA, USA) with exciter filter BP545
(excitation wavelength maximum l 546 nm), barrier
filter 0590 and a DM580 dichroic mirror (tritc filter
set). To calculate the percentage of functional
bundles, the numbers of visible xylem bundles were
first counted using bright field microscopy and then
using fluorescence microscopy. Bundles that appeared dark under the fluorescence microscope were
assumed to be non-functional for transpiration.
Image analysis
Paraffin sections of leaves were photographed with
an Olympus Vanox-T photomicroscope using Kodak
RESEARCH Salinity and hydraulics of growing sorghum leaves
(a)
(c)
121
(b)
(d)
Fig. 1. Transverse sections of sorghum leaves 6 and 7 taken 13 mm from the base of leaf 6 (the region of highest
growth rate in both treatments). (a) Control plant. (b) Salt-treated plant. (a) and (b) are presented at the same
magnification ; bar, 0n3 mm. Large arrows point to the inside edge of leaf 6 ; small arrows point to that of leaf
7. White asterisks, location of an outside lateral bundle 4 (O4). (c, d) Magnified view of the fourth lateral bundle
on the outside edge of sorghum leaf 6 at 13 mm ; bar, 25 µm. (c) Control plant. (d) Salt-treated plant. M,
metaxylem vessel ; P, protoxylem vessel. The metaxylem vessels are immature (i.e. lignified secondary wall is
absent in all four metaxylem cells).
2415 technical pan film shot at ASA 50. Negatives
were printed at a magnification of i100–200, and
prints were digitized using an SII digitizing pad
(Seiko Instruments, San Jose, CA 95131, USA)
using the Easydij software package (Geocomp Ltd.,
Golden, CO 80401, USA) to calculate the cross-
122
RESEARCH S. F. Baum et al.
sectional area of the leaf. Protoxylem and metaxylem
diameters were measured (directly from the paraffin
sections) using the Galai Scanarray image analysis
software (Galai, Migdal Ha’Emek, Israel) with a
i40 objective. Protoxylem elements were measured
in sections from the growth zone (13 mm), while
metaxylem elements (assumed functional after cessation of growth) were measured in sections from 35
mm. The xylem elements were considered circular
in cross-section, and a single measurement was taken
for each diameter. Two protoxylem and two
metaxylem elements were measured in each bundle.
The larger protoxylem element in each bundle was
averaged per treatment for the graphical display of
protoxylem radius, while all measured r and calculated r#, r$ and r% were summed to obtain average
values per leaf for comparison to water deposition
rates.
Table 1. The effects of salinity on morphology of leaf
6 (cross sections at 13 mm)
Morphological
variable
Cross-sectional
area (mm#)
Mid-rib thickness
(mm)
Leaf width (mm)
Leaf spiral
constant ∆r\∆θ (mm)

M
Morphology
P
Some basic morphological features were unchanged
by salinity. On both control and salt affected plants,
each leaf grew in a spiral form, wrapped within the
sheaths of older leaves (Fig. 1a, b). Viewed from
below, leaf 6 could be traced from the inside edge in
a counterclockwise spiral, while leaf 7 was wrapped
clockwise in a tighter spiral sheet. Growth in width
involves propagation of the spiral form during
development. This can be seen by examining the
cross section of a particular leaf at different ages, or,
if growth is approximately steady, a cross section
close to the node can be compared to a cross section
of older tissue farther from the node. In crosssections taken at 13 mm from the leaf base, leaf 6 is
wrapped in approximately 3" spiral gyres (a ‘ gyre ’
#
represents one turn of the spiral or 2π radians around
the spiral form.) In the more mature sections taken
at 35 mm, the spiral has four complete gyres (not
shown). The cross-sections taken at 35 mm also
show some unrolling of the outermost gyre. This
unrolling continues during development, as the leaf
will become approximately flat for a period after
emergence from the sheathing base.
In a polar coordinate system r, θ the spiral form
can be characterized by a parameter, ∆r\∆θ, that
gives the change in radius for a particular angular
displacement. In the cross sections taken at 13 mm,
salt-affected leaves were more tightly wrapped than
controls, and ∆r\∆θ is smaller for the salt-affected
leaves. This effect is apparent when lines are drawn
on the leaf cross section perpendicular to the spiral
form (Table 1). The tighter spiral form is due partly
to the thinner shape of the salt-affected leaves. The
greater curvature of the leaf edge is probably also
due to a greater cross-sectional growth gradient (Silk
& Erickson, 1978), since the leaf gyres are often well
Control
Salt treated
1.86p0.48
0.825p0.26
0.525p0.093
0.335p0.052
17.09p1.25
0.118p0.04
12.36p0.60
0.078p0.010
M
P
Fig. 2. Magnified view using fluorescence microscopy of
lateral bundles on the outside edge of sorghum control leaf
6 at 13 mm. The excised shoot had been allowed to
transpire an eosin solution, as described in the Materials
and Methods section. The protoxylem walls are strongly
fluorescent, whereas the metaxylem walls are dark.
separated. Salt-treated leaves had narrower midribs,
and smaller leaf width and cross-sectional area than
controls (Fig. 1, Table 1).
Anatomy
The anatomical study was repeated on two sets of
five plants per treatment, with similar results. The
salt treatment did not affect the number of longitudinal vascular bundles in the leaf growth zone. On
most plants, leaf 6 had 13 ‘ lateral veins ’ (terminology
of Sharman (1942)). We assumed only the lateral
veins could contain functional xylem during leaf
expansion, as the smaller or intermediate veins are
not mature (Sharman, 1942). In the sections from 13
mm, in the rapidly elongating region, each vein had
RESEARCH Salinity and hydraulics of growing sorghum leaves
123
Table 2. Some developmental variables related to water use
Developmental variable
Units
Control
Salt-treated
plants
Σ (radii of protoxylem)
Area of protoxylem
Area of metaxylem
Bundles actively conducting (%)
Area of protoxylem
Leaf cross-sectional area
µm
µm#
µm#
154p20
2161p428
6183p223
96
94p13
955p221
3914p138
56
µm# µm−#
1.16i10−$
1.11i10−$
Transpiration rate
Fresh weight of leaves at 12 dap
g g−" d−"
16
11
Transpiration rate
Fresh weight of leaves at 16 dap
g g−" d−"
14
11
Transpiration rate
Fresh weight of leaves at 19 dap
g g−" d−"
15
5
dap, days after planting.
either one large protoxylem vessel or one large and
one small protoxylem vessel. Two metaxylem vessels
per vein were present as viewed in transverse section
(Fig. 1c, d). The widest xylem elements occurred in
the central vein (along the midrib), and width of
both protoxylem and metaxylem elements decreased
with distance from the midrib (Figs 1, 3). The
sections at 35 mm, in the more mature part of the
leaf, often had one protoxylem vessel and an
accompanying lacuna as well as two metaxylem
elements (not shown).
Salinity affected the width of the conducting
element. As shown in Fig. 3, protoxylem cells in the
first lateral vein (O1) had a radius of 11n9p2n3 µm in
control leaves and 7n5p1n3 µm in salinized leaves ;
protoxylem cells in the outer veins (O6) were
6n2p1n6 µm in controls and 5n5p0n8 µm in salt
affected leaves. Metaxylem cells were also affected
by salinity : Controls had cell radii of 19n2p2n1 and
8n9p1n6 µm in central (M) and peripheral (O6)
veins, respectively ; salt-affected leaves had metaxylem radii of 15n1p1n3 and 6n9p0n9 µm in central (M)
and peripheral (O6) veins, respectively.
The total cross-sectional area of protoxylem
elements in control leaves was approximately twice
that of the salt-affected leaves (Table 2). Since the
leaf cross-sectional area was also reduced, the area of
protoxylem elements per unit of leaf cross-sectional
area was unchanged by the salt treatment (Table 2).
Development of xylem in relation to function
In both salt-treated and control plants, the protoxylem was mature at 13 mm from the base of the
leaf while the metaxylem was immature (i.e. a
lignified secondary wall was absent and cytoplasm
was present in metaxylem cells in the actively
growing region ; Fig. 1c,d). At 35 mm from the base
of the control leaf, where growth had stopped, there
were both mature and immature metaxylem elements
with xylem parenchyma cells present among the
metaxylem elements (not shown). In salt-affected
plants at 35 mm, all the lateral veins had mature
metaxylem elements, and the xylem parenchyma
cells appeared sclerified. Presumably, water conduction was restricted to the protophloem and
protoxylem during leaf elongation (Fig. 2 ; Paolillo,
1995).
Eosin is used to visualize the vascular elements
associated with water movement in plants (Eichhorn
et al., 1986). In preliminary experiments conducted
to validate the use of this dye, we found that the
brightest fluorescence due to the eosin was in mature
vessel element walls, and we also observed some
slight fluorescence in parenchyma cells surrounding
the functioning vessels (Fig. 2). In some instances
there was also a small amount of fluorescence
associated with the phloem and a smaller amount of
fluorescence in parenchyma cells surrounding the
phloem. In fluorescent tissue, the eosin fluorescence
was observed in both vessels and in neighboring
parenchyma cell walls ; whereas in dark tissue,
fluorescence appeared neither in vessels nor in
parenchyma cells. This supports the interpretation
of eosin fluorescence as evidence of water-conducting
ability.
According to our dye-transport studies, most of
the lateral veins in control plants were functional in
transporting the dye (Table 2). By contrast, the dye
studies revealed that only about half the lateral veins
in the salt-affected leaves were actively conducting.
Transpiration
Water flow was diminished by salinity (Fig. 4). The
flow rate was large and increased rapidly with time in
control plants. In salt-affected plants the water flow
rate was small, increased slightly and then decreased
(a)
Transpiration rate (g H2O d–1 per plant)
RESEARCH S. F. Baum et al.
124
16
14
Radius (µm)
12
10
8
6
4
2
0
O6 O5 O4 O3 O2 O1 M I1 I2 I3 I4 I5 I6
Leaf bundle position
(b)
20
18
50
40
30
20
10
0
12
16
Days after planting
18
Fig. 4. Transpiration rate per plant, measured as gravimetric water loss from the incubation vessels. Salinity
effects increased over time and persisted for at least 2 d
after the treatments were reversed. Controls, closed bars ;
salt-treated plants, open bars ; controls transferred to salt
treatment, horizontally hatched bars ; salt-treated plants
transferred to control medium diagonally hatched bars.
Radius (µm)
16
14
12
10
8
6
4
2
0
O6 O5 O4 O3 O2 O1 M I1 I2 I3 I4 I5 I6
Leaf bundle position
Fig. 3. (a) Radii of protoxylem in transverse section of
sorghum leaves taken 13 mm from the base of leaf number
6. (b) Radii of metaxylem vessel elements taken 35 mm
from the base of leaf number 6. The bundles are arranged
in sequence as if the leaves were unfolded (i.e. I5 designates
the fifth major vein from the midvein (M) on the portion
of the leaf that wraps to the inside of the midvein while O5
designates the fifth bundle from the midvein on the
portion of the leaf that wraps to the outside of the
midvein). Controls, closed bars ; salt-treated plants, open
bars.
with time. When plants grown under saline conditions were transferred to control medium, their
water-flow rates tripled relative to plants maintained
continuously in salt solution (day 18, Fig. 4).
However, this treatment (saline medium followed by
control medium) produced a water flow only 25% of
that of controls. In contrast, when plants grown in
control solution were transferred to saline solution,
they maintained an elevated transpiration rate relative to the salt-grown plants for at least 48 h. The
treatment consisting of control medium followed by
saline medium reduced the water flow rate to 75% of
controls. In absolute rather than relative terms the
diminution of the control flow rate by the 2-d salt
treatment late in development was similar to the
augmentation of the flow rate by the 2-d control
treatment late in development, c. 9 g of water per
plant per day. These results were probably due to
pleiotropic effects of salinity on rate of leaf production, stomatal functioning, and rate of leaf
senescence.
The effect of salt on total plant water flow seems
most strongly explained by the salt-induced decrease
in plant growth (with associated decrease in fresh
weight and surface area of leaves). However, salt also
affected the leaf-specific flow rates, especially after
16 d of age (Table 2). Per unit fresh weight of leaves,
control plants transpired approx. 15 g g−" d−"
days 12–19. Per unit of leaf surface area, this
corresponded to transpiration rates in the range 0n20
g cm−# d−" at day 12 to 0n24 g cm−# d−" at day 19 (not
shown). In comparison, per unit fresh weight of
leaves, salt-treated plants transpired 11 g g−" d−"
from day 12 to day 16 corresponding, on a leaf
surface area basis, to the range 0n14–0n19 g cm−# d−".
By day 19 there was a large decrease in transpiration
rate of salt affected leaves, to 5 g g−" d−" or 0n09 g
cm−# d−". The salt-treated plants appeared senescent
(lower leaves yellow and somewhat curled) compared
with controls. Beyond the seven-leaf stage, the
difference in transpiration rates might be related to a
more rapid aging of salt-affected leaves.
Measurements with the diffusion porometer
showed salt-induced decreases in leaf conductance.
However, the effects of the hydraulic architecture
were not obvious (Fig. 5). Leaves 3 and 4 had lower
conductance in salt treatments, even though their
xylem was formed before the treatment began. This
suggests an effect of salinity on stomatal opening.
Leaf 5, measured at full expansion on day 21 (Fig. 5,
bottom), had a conductance that was not affected by
RESEARCH Salinity and hydraulics of growing sorghum leaves
6
Day 12
5
4
3
2
1
2
3
4
6
Leaf conductance (mm s–1)
Day 16
5
4
3
2
1
2
3
4
5
6
7
6
Day 21
5
4
3
2
1
0
draulic architecture might be involved at least
indirectly ; and stomatal opening could certainly be
involved, as discussed below. Since the relative effect
of salt on conductance of leaf 6 was never greater
than on that of leaf 4 at day 12 ; and leaf 4 had
developed prior to the imposition of the salt stress ; it
seems likely that the salt treatment has a large effect
on stomatal opening at several stages of leaf development.

0
0
125
5
6
Leaf number
Fig. 5. Leaf conductance, measured with a diffusion
porometer. For leaf 6, the transpiration rate of the control
on day 16 can be compared with the transpiration rate of
the salt-affected leaf on day 21 to minimize the confounding effects of developmental stage. Controls, closed bars ;
salt-treated plants, open bars ; salt treated but still
expanding leaves, horizontally hatched bars.
salinity. We observed that the older leaves in the
salt-affected plants were somewhat senescent when
leaf 5 was maturing (this is suggested also by the
transpiration rates in Table 2). Presumably a large
proportion of the volumetric water flow through the
salinized plants was occurring through leaf 5 at day
21. Leaf 6 showed a large effect of salinity on
conductance. The expanding control leaf lost water
vapor 34% more rapidly, and the mature control leaf
66% more rapidly than the salt treated leaf. Hy-
Our results show that salt greatly decreases flow of
water through the plant. This is apparent from other
studies, such as that of Munns & Passioura (1984) of
barley and Nicolas et al. (1993) of wheat. The
magnitude of the effect on flow rates increases with
time or developmental stage for at least 21 d.
Physically, the decreased volumetric water flow is
mostly explained by the decrease in plant growth
rate and associated decrease in leaf surface area. The
physiological basis for the effect of salinity on growth
rate remains to be determined. The role of hormones
such as ABA deserves to be explored (Dodd &
Davies, 1996).
Salinity also produces a decrease in transpiration
rate (water flow per unit of leaf mass or surface area)
that becomes more marked with plant development.
Not surprisingly, this decrease is less than that
predicted by Poiseuille’s Law that flow rates with
similar driving forces should vary as r%. Our
anatomical studies show that salinity causes typically
a 25% decrease in diameter of the water-conducting
elements ; if Poiseuille’s Law held, transpiration
rates in salt-affected leaves would be reduced to less
than one-third of the control values. However, the
phase change during the evaporation process in the
leaf would be expected to lead to a compensating
increase of the water potential gradient to maintain
similar flow rates through the narrower vessels.
Thus the effect of the increased hydraulic resistance
on transpiration and conductance probably occurs
indirectly, via a decrease in water potential in the
leaves. The salt treatment also produces a direct
effect on leaf water relations, as the saline medium
has 0n45 M Pa more negative water potential. A more
negative water potential in salt-affected leaves could
cause stomatal closing, either directly or via hormonal signals. The measured transpiration rates and
stomatal conductances could have been affected by
stomatal aperture changes.
Other mechanisms might also be operating to
change water flows during salt stress. The large
decrease in leaf conductance during maturation of
leaf 6 under salt treatment might be due to changes
in leaf habit. Rolling, curling, or loss of longitudinal
curvature would reduce leaf temperature by reducing
flux of incident radiation onto the leaf. The lower
leaf temperature would reduce the water-vapor
126
RESEARCH S. F. Baum et al.
concentration in the substomatal cavities and greatly
reduce the driving force for water movement out of
the leaf.
The literature contains many examples of variation
in hydraulic conductivity not explained by xylem
anatomy. Schultz & Matthews (1993) found that the
conductivity per xylem cross-sectional area increases
by an order of magnitude during the period of leaf
expansion in grape. Neumann (1987) found pulvinar
xylem conductivity decreases four-fold during senescence of bean leaves. In general, a possible
explanation for the small effect of xylem diameter on
hydraulic conductivity is that the conductivity in
mature leaves is affected less by the diameter of
xylem elements of the growth zone than by hydraulic
bottlenecks – constriction zones in roots, stems or
leaf blades. This idea has been explored by Zimmerman & Brown (1971) who found constrictions in
xylem elements leading to lateral tree branches, by
Richards & Passioura (1989) studying wheat roots,
and by Calkin et al. (1986) analysing flow in fern
tracheids. Meinzer & Grantz (1990) found convincing evidence for synchrony between stomatal
and hydraulic conductance in sugar cane, a plant
closely related to sorghum. They believe steady
water flow and water potential gradients are regulated
by hydraulic bottlenecks. Meinzer et al. (1992)
found hydraulic constrictions at nodes, where hydraulic conductivity dropped an order of magnitude
relative to internodes, and along the leaf sheath and
blade, where hydraulic conductivity dropped another
order of magnitude. Thus there is good evidence in
the literature to support the hypothesis of at least
indirect regulation of hydraulic conductivity by
occasional xylem constrictions. It would be interesting to study hydraulic conductivity in conjunction
with xylem fine structure to find the effect of salinity
on the putative hydraulic bottlenecks of sorghum.
Comparison of anatomical effects to water deposition
rates in the growth zone
Bernstein et al. (1995) used growth data with a
continuity equation (Meiri et al., 1992) to calculate
local net rates of deposition of water and calcium into
the growing tissue of salt-affected and control leaves
of sorghum. To explore the functional significance of
the salt induced effects on xylem element diameter,
we compare the published water deposition rates to
the anatomical measurements reported here.
In contrast to the lack of correlation between
transpiration rate and xylem anatomy, it is striking
that the water deposition rate in the growth zone
varies with the diameter of the functional protoxylem. From the data of Bernstein et al. (1995) we
calculate that, in the zone 10n5–22n5 mm from the
leaf base (the location of fastest growth rate and near
the anatomical section cut at 13 mm), control leaves
have approx. 3n9 times the water deposition rate of
salt-affected leaves. The sum of the radii of our
protoxylem elements in control leaves is 1n6 times
the sum of the radii of the protoxylem elements in
salt-affected leaves. However, fewer of the vascular
bundles in the stressed leaves are conducting (Table
2). Thus the vein circumference (conducting water
to the expanding cells) of control leaves is 2n8 times
that in salt-affected leaves. We see that, when the
percentage of actively conducting veins is taken into
account, it seems that the best fit power dependence
that could be proposed is to the cross-sectional area,
or r#, of the functional protoxylem. Control leaves
have 3n9 times the functional protoxylem area of the
salt-affected leaves (Table 2). The ratios for r$ and r%
are, respectively, 5n8 and 9n0. Apparently, water
deposition rate is correlated to r#, the square of the
protoxylem radius.
Canny and colleagues have modelled water flow in
plants as a tension driven flow through a set of leaky
pipes. Their model suggests that the radial transfer
rate of water per unit length of xylem should be
proportional to the circumference, i.e. the first power
of the element radius (Altus et al., 1985 ; Canny,
1991). Our ‘ water deposition rate ’ is equivalent to
their radial transfer rate. Thus we look for a
dependence of water deposition rate on the radius of
the protoxylem element. The r# dependence of the
water deposition rate is more sensitive to xylem
radius than the prediction of the leaky pipe model. A
possible explanation for the discrepancy is that the
wider interveinal spacing in the control tissue creates
a larger radial gradient in water potential. That is,
the larger control leaves may have a greater driving
force, as well as greater absorptive surface, for water
extraction from veins into growing leaf cells. Nonami
& Boyer (1993) and Silk and colleagues (Silk &
Wagner, 1980 ; Silk, 1994) have discussed the
importance of such growth-sustaining water potentials. This interpretation seems consistent with
some recent results of Martre et al. (2000) who
measured both water deposition rate and xylem
radius through the growth zone of cultivated fescue
leaves. They found that within a particular leaf the
growth-associated water deposition rate was directly
proportional to the local xylem radius throughout
the growth zone. This is consistent with the leaky
pipe model and also with the idea that the first power
dependence (of water deposition rate on xylem
radius) requires equal spacing of veins.
Our results provide strong support for the hypothesis that hydraulic architecture is correlated
with the water deposition rates associated with leaf
growth. Future work could include detailed characterization of anatomical fine structure, measurement
of cell lengths and the hydraulic properties of
interelement connections, and more physiological
studies to measure fluxes and hydraulic conductivity
of the growth zone under various environmental
conditions. Our data suggest the need for a com-
RESEARCH Salinity and hydraulics of growing sorghum leaves
prehensive hydraulic model, including driving forces
and resistances and a realistic geometry, for the
water relations of leaf expansion.
              
The authors’ research has had support from award IBN9604230 from the National Science Foundation. We thank
Dr. Nirit Bernstein for helpful consultation and George
Scheer for skilful technical assistance. T. Hsiao and F.
Meinzer provided insightful criticism.

Altus DP, Canny MJ, Blackman DR. 1985. Water pathways in
wheat leaves. II. Water-conducting capacities and vessel
diameters of different vein types, and the behaviour of the
integrated network. Australian Journal of Plant Physiology 12 :
183–189.
Bernstein N, Silk WK, La$ uchli A. 1993. Growth and development of sorghum leaves under conditions of NaCl stressspatial and temporal aspects of leaf growth inhibition. Planta
191 : 433–439.
Bernstein N, Silk WK, La$ uchli A. 1995. Growth and development of sorghum leaves under conditions of NaCl stress :
possible role of some mineral elements in growth inhibition.
Planta 196 : 699–705.
Calkin H, Gibson A, Nobel P. 1986. Biophysical model of xylem
conductance in tracheids of the fern Pteris vittata. Journal of
Experimental Botany 37 : 1054–1064.
Canny MJ. 1991. The xylem wedge as a functional unit –
speculations on the consequences of flow in leaky tubes. New
Phytologist 118 : 367–374.
Dengler NG, Dengler RE, Donnelly PM, Hattersley PW.
1994. Quantitative leaf anatomy of C and C grasses (Poaceae)
$
%
– bundle sheath and mesophyll surface area relationships.
Annals of Botany 73 : 241–255.
Dodd IC, Davies WJ. 1996. The relationship between leaf
growth and ABA accumulation in the grass leaf elongation zone.
Plant, Cell and Environment 19 : 1047–1056.
Eichhorn SE, Perry JW, Evert RF. 1986. Preparation guide for
laboratory topics in botany. New York, USA : Worth Publishers.
Esau K. 1953. Plant anatomy. New York, USA : John Wiley &
Sons.
Fricke W, Hinde PS, Leigh RA, Tomos AD. 1995. Vacuolar
solutes in the upper epidermis of barley leaves : intercellular
differences follow patterns. Planta 196 : 40–49.
Hayward H, Long E. 1941. Anatomical and physiological
responses of the tomato to varying concentrations of sodium
chloride, sodium sulfate and nutrient solutions. Botanical
Gazette 102 : 437–462.
Johansen DA. 1940. Plant microtechnique. New York, USA :
McGraw-Hill.
127
Martre P, Durand JL, Cochard H. 2000. Changes in axial
hydraulic conductivity along elongating leaf blades in relation
to xylem maturation in tall fescue. New Phytologist. (In press.)
Meinzer FC, Goldstein G, Neufeld HS, Grantz DA, Crisosto
GM. 1992. Hydraulic architecture of sugarcane in relation to
patterns of water use during plant development. Plant, Cell and
Environment 15 : 471–477.
Meinzer FC, Grantz DA. 1990. Stomatal and hydraulic
conductance in growing sugarcane : stomatal adjustment to
water transport capacity. Plant, Cell and Environment 13 :
383–388.
Meiri A, Silk WK, La$ uchli A. 1992. Growth and deposition of
inorganic nutrient elements in developing leaves of Zea mays L.
Plant Physiology 99 : 972–978.
Munns R, Passioura J. 1984. Hydraulic resistance of plants. III.
Effects of NaCl in barley and lupin. Australian Journal of Plant
Physiology 11 : 351–359.
Neumann P. 1987. Sequential leaf senescence and correlatively
controlled increases in the xylem flow resistance. Plant
Physiology 83 : 941–944.
Nicolas ME, Munns R, Samarakoon AB, Gifford RM. 1993.
Elevated CO improves the growth of wheat under salinity.
#
Australian Journal of Plant Physiology 20 : 349–360.
Nonami H, Boyer JS. 1993. Direct demonstration of a growthinduced water potential gradient. Plant Physiology 102 : 13–19.
Paolillo DJ. 1995. Protoxylem maturation in the seedling leaf of
wheat. American Journal of Botany 82 : 337–346.
Puech-Suanzes I, Hsiao TC, Fereres E, Henderson DS. 1989.
Water-stress effects on the carbon exchange rates of three
upland cotton (Gossypium hirsutum) cultivars in the field. Field
Crops Research 21 : 239–255.
Richards R, Passioura J. 1989. Breeding program to reduce the
diameter of the major xylem vessel in the seminal roots of wheat
and its effect on grain yield in rain-fed environments. Australian
Journal of Agricultural Research 40 : 943–950.
Sass JE. 1958. Elements of botanical microtechnique. New York,
USA : McGraw-Hill.
Schultz HR, Matthews MA. 1993. Xylem development and
hydraulic conductance in sun and shade shoots of grapevine
(Vitis vinifer L.) : evidence that low light uncouples water
transport capacity from leaf area. Planta 190 : 393–406.
Sharman BC. 1942. Developmental anatomy of the shoot of Zea
mays L. Annals of Botany 6 : 245–281.
Silk WK. 1994. Kinematics and dynamics of primary growth.
Biomimetics 2 : 199–213.
Silk WK, Erickson RO. 1978. Kinematics of hyocotyl curvature.
American Journal of Botany 65 : 310–319.
Silk WK, Wagner KK. 1980. Growth-sustaining water potential
distributions in the primary corn root. A noncompartmented
continuum model. Plant Physiology 66 : 859–863.
Trivett CL, Evert RF. 1998. Ontogeny of the vascular bundles
and contiguous tissues in the barley leaf blade. International
Journal of Plant Sciences 159 : 716–723.
Zimmermann MH, Brown CL. 1971. Trees, structure and
function. Berlin, Germany : Springer-Verlag.