Swift etal 2008 SAJB

Drought induced xylem embolism in four riparian trees from the Western
Cape Province: Insights and implications for planning
and evaluation of restoration
C.C. Swift a
, S.M. Jacobs b
, K.J. Esler b,c,⁎
a
Department of Biology, Whittier College, Whittier, CA, United States
b
Department of Conservation Ecology and Entomology, Stellenbosch University, Private Bag X1, Matieland 7602, Stellenbosch, South Africa
c
Centre for Invasion Biology, Stellenbosch University, Private Bag X1, Matieland 7602, Stellenbosch, South Africa
Received 23 August 2007; accepted 22 January 2008
Abstract
In Mediterranean climates, rainfall is restricted to periods of low evaporative demand, leaving plants to survive the summer drought. The
purpose of this study was to determine the contribution of drought tolerance to the distribution of riparian species. These physiological insights
will assist in developing target species lists for restoration of riparian ecosystems in the Western Cape, currently heavily degraded due to land
clearing and invasive aliens. We estimated P50 and P100 from vulnerability curves and Ψx in four species across a range of summer water
availability estimated by streamflow. As expected, decreasing streamflow resulted in lower predawn and mid-day xylem water potential, and
species identified in previous studies as having broad distributions, such as Brabejum stellatifolium and Metrosideros angustifolia, sustained
greater decreases in mid-day xylem water potential and were less vulnerable to cavitation than Rapanea melanophloeos or Brachyleana neriifolia,
species with more restricted distributions. These results provide preliminary evidence that a consideration of drought tolerance might be useful in
refining lists of target species for active restoration and evaluation of restoration success across projects in streams and rivers with different fluvial
regimes.
© 2008 SAAB. Published by Elsevier B.V. All rights reserved.
Keywords: Drought tolerance; Embolism; Fynbos-riparian ecotone; Vulnerability to cavitation
1. Introduction
The contribution of riparian habitats to ecosystem function
and ecosystem services has been well documented (Malanson,
1993; Naiman and Décamps, 1997; Naiman et al., 2005). Water
availability and disturbance are identified as important determi-
nants of riparian community structure (Hupp and Osterkamp,
1996; Van Coller et al., 2000; Ross and Swift, 2003). Riparian
habitats in Mediterranean-type ecosystems, characterized by
summer drought, are less well understood than riparian habitats
in temperate ecosystems (Gasith and Resh, 1999; Salinas et al.,
2000). Intense disturbance associated with peak flows during
rainfall events in winter months is coupled with water deficits in
summer months. As stream power diminishes and disturbance
intensity decreases, water availability decreases during the
period of greatest evaporative demand. This is in contrast to
temperate areas with summer rainfall where water availability
remains highest during periods of greatest evaporative demand.
The combination of summer drought and winter disturbance
creates a narrow recruitment niche making restoration of native
species to degraded riparian ecosystems difficult (Salinas et al.,
2000; Petit and Froend, 2001).
In riparian habitats of the Western Cape of South Africa, the
invasion of woody exotics has decreased surface flow, eliminated
native species important for wildlife, increased nutrient availability,
Available online at www.sciencedirect.com
South African Journal of Botany 74 (2008) 508–516
www.elsevier.com/locate/sajb
⁎ Corresponding author. Department of Conservation Ecology and Entomology,
Stellenbosch University, Private Bag X1, Matieland 7602, Stellenbosch, South
Africa. Tel.: +27 21 808 4005.
E-mail address: kje@sun.ac.za (K.J. Esler).
0254-6299/$ - see front matter © 2008 SAAB. Published by Elsevier B.V. All rights reserved.
doi:10.1016/j.sajb.2008.01.169

and altered seed banks available for regeneration of riparian spe-
cies following the natural disturbance regime that characterizes
riparian habitats (Rowntree, 1991; Brown, 1998; Prinsloo and
Scott, 1999; Le Maitre et al., 2002; Dye and Jarmain, 2004;
Richardson et al., 2007; Vosse et al., 2008-this issue). The near
complete elimination of native species from many watersheds is
complicating efforts to create a list of target species for restoration.
This is a particular problem since a large part of the effort to restore
riparian habitats in the Western Cape is devoted to the removal of
non-natives dominating many riparian habitats by the Working for
Water Programme (Van Wilgen et al., 1998). Species lists for
restoration of degraded Fynbos-riparian ecotones in the Western
Cape have been derived using ordination techniques (e.g.
Galatowitsch and Richardson, 2005; Prins et al., 2005). While
these studies have been successful in identifying co-occurring
groups of species, and the physical factors associated with their
presence or absence, no species-level, mechanistic understanding
exists to indicate how the relationship between disturbance and
drought might determine the distribution of riparian species across
streamchannels, between reaches, andbetween watersheds. Such a
mechanistic understanding would involve understanding how
drought, disturbance and substrate interact to determine
survivorship.
Vulnerability to cavitation, interpreted in combination with
water stress levels that plants experience in the field, is a useful
measure of drought tolerance (Kolb and Davis, 1994; Linton
et al., 1998; Pockman and Sperry, 2000; Maharali et al., 2004).
Cavitation via the air-seeding mechanism occurs when root or
stem xylem water pressure decreases to a point where capillary
forces responsible for the adhesion of the water column to the
vessel wall are overwhelmed, and air is pulled into the vessel at
the pit membrane forming an embolism that blocks water
transport (Zimmerman, 1983; Sperry et al., 1988; Jarbeau et al.,
1995; Hacke and Sperry, 2001). Xylem with vessels that have
thicker walls and less vessel lumen area are generally resistant
to cavitation, but the cost of increased wall to lumen ratio may
be a reduction in xylem conducting area (Pockman and Sperry,
2000; Hacke and Sperry, 2001; Hacke et al., 2001). Further-
more, there is a trade-off between resistance to cavitation and
water transport such that xylem conduit construction, and
therefore resistance to cavitation reflect the range of minimum
xylem water potentials experienced in situ (Pockman and
Sperry, 2000; Hacke and Sperry, 2001; Hacke et al., 2001).
We examined vulnerability to cavitation in four tree species
across streams with different hydrologic regimes. The selected
species varied in distribution extent and biogeographical
affinity. Rapanea melanophloeos (L.) Mez has afromontane
affinities, while the other three species are restricted to the
Western Cape province (Coates Palgrave, 2002). Metrosideros
angustifolia (L.) and Brabejum stellatifolium (L.) are wide-
spread in riparian habitats in the Western Cape occurring in each
of the four groups of species identified in Prins et al. (2005) and
identified as having ubiquitous distribution by Galatowitsch and
Richardson (2005). The other two species, R. melanophloeos
(L.) Mez. and Brachylaena neriifolia (L.f.) R. Br., were
positively associated with factors indicating higher stream
power in Galatowitsch and Richardson (2005). We hypothe-
sized that the four species selected, because of differences in
distribution suggested by previous studies, would show
differences in their vulnerability to cavitation across a range
of minimum water availabilities. Previous work on B. neriifolia
and B. stellatifolium revealed small differences in diurnal
patterns of stomatal conductance with B. neriifolia sustaining
higher mid-day xylem water potentials than B. stellatifolium
(Richardson and Kruger, 1990).
If drought tolerance is an important factor in determining
species composition in riparian habitats in the Western Cape,
M. angustifolia and B. stellatifolium should show greater
resistance to cavitation across the range of minimum water
availabilities sampled and should sustain lower mid-day water
potentials than B. neriifolia, since Galatowitsch and Richardson
(2005) identified both of these species as ubiquitous (Alder et al.,
1996; Davis et al., 1999; Pockman and Sperry, 2000). We
expected R. melanophloeos to be more resistant to cavitation
than B. stellatifolium, M. angustifolia or B. neriifolia because it
usually occurs in upstream portions of catchments where
minimum stream discharge is generally lower than for down-
stream portions of catchments. We expected B. neriifolia to be
the most vulnerable to cavitation based on its significant asso-
ciation with less invaded reaches, and with factors consistent
with higher stream power and therefore greater summer water
availability.
Understanding the relationship between drought tolerance and
fluvial regime would then lead to a more refined assessment of
restoration success and could contribute to a more refined species
palette for restoration projects in the Fynbos-riparian ecotone.
2. Methods
The study sites were all within the Jonkershoek valley (33°
57′47″S, 18° 55′32″E) in the Western Cape Province of South
Africa. Four study sites were identified: a high streamflow site
on the Eerste River at Swartbrug just below the Kleinplaas dam,
a medium streamflow site on the Eerste River at Witbrug, and
low streamflow sites at Biessievlei and Bosboukloof, both
tributaries of the Eerste River. Four species native to Fynbos
riparian zones were studied: Meterosideros angustifolia (Myrta-
ceae), B. stellatifolium (Proteaceae), R. melanophloeos (Myr-
sinaceae) and B. neriifolia (Asteraceae). M. angustifolia and
B. stellatifolium occurred at all sites, R. melanophloeos occurred
at all sites except Bosboukloof, and B. neriifolia occurred only
on the Eerste River at Swartbrug and Witbrug. Five individuals
of each species were flagged at each location, with the exception
of B. stellatifolium and M. angustifolia, which were not sam-
pled at Biessievlei because there were not enough individuals
represented.
Vulnerability to cavitation was determined using the bench
drying method (Davis et al., 1999; Pockman and Sperry, 2000;
Brodribb et al., 2003). Preliminary determination of vessel
length for all study species using the “air flow” method
(Zimmerman and Jeje, 1981; Ewers and Fisher, 1989) indicated
that no species had vessel lengths greater than 1.0 m. Eight to
ten branches, 0.5 to 1.0 m in length or greater were cut from all
individuals of one species at a study site before dawn. One
509C.C. Swift et al. / South African Journal of Botany 74 (2008) 508–516

branch from each individual was bagged in the field for the 0 h
measurements, and the rest were returned in separate bags to the
laboratory where they were allowed to dehydrate for approxi-
mately 1, 2, 4, 6, 8 and 10 h or until 90–100% of hydraulic
conductivity was lost. For M. angustifolia, branches were
allowed to dehydrate for an additional period of 12 and 24 h to
achieve the desired loss of conductivity. Branches were sealed
in separate bags for approximately 1.5 h to equilibrate before
measuring hydraulic conductivity and xylem water potential.
For each sample branch, a side branch was removed for
measurement of xylem water potential, and the sample branch
was immediately submerged in water and cut from each end [in
10 cm sections to a final 10 cm segment] to measure hydraulic
conductance. Branch segments used to measure hydraulic
conductance were always distal to side branches used to
measure xylem water potential (Ψx). Ψx was measured using a
PMS Model 600 Pressure Chamber Instrument (PMS Instru-
ments, Oregon, USA). The percentage loss of hydraulic
conductance was measured using the method described in
Sperry et al. (1988). Vulnerability curves for each individual
were constructed by plotting percent embolism as a function of
Ψx, and fitting the points to a curve using a third order
polynomial (Pockman and Sperry, 2000); R2
values for curves
of individuals ranged from 0.61 to 0.99 (Table 1). Vulnerability
curves for the study species differed in shape, and the third order
polynomial explained the largest amount of variation in percent
loss of hydraulic conductance (highest R2
values) for most
species at most sites. A review of the literature reveals multiple
curve-fitting equations including exponential (Pammenter and
Vander Willigen, 1998), second order polynomial (Jacobsen
et al., 2007), and third order polynomial (Pockman and Sperry,
2000). All vulnerability curves should show an increase in
percent loss of hydraulic conductance with decreasing water
potential; the equation used to fit data to a vulnerability curve is
less important than the R2
value obtained.
The fitted curves for individuals were used to calculate 50%
(P50), corresponding to the constant b in Pammenter and Vander
Willigen (1998), and 100% (P100) loss of hydraulic conductiv-
ity. P50 values for the high and medium streamflow sites were
analyzed using a factorial ANOVA, and the low streamflow
sites were analyzed using a one-way ANOVA (Statistica
Version 7.0, StatSoft Inc.). The low streamflow site was
excluded from the factorial ANOVA because B. neriifolia did
not occur at this site, resulting in an unbalanced model. The
choice to exclude the site rather than the missing species was
based on the fact that the low and medium streamflow sites had
very similar summer water availabilities, and that we were
interested in comparing species with different patterns of
distribution. Means were separated using Fisher's LSD multiple
range test (Pb0.05). Individuals were excluded from statistical
analysis when curves did not allow extrapolation to 100%
embolism. Points from all individuals of each species at each
site were combined and EXCEL was used to fit a third order
polynomial curve to create vulnerability curves for each species
at each site (Pockman and Sperry, 2000). In the field, pre-dawn
(Ψmax) and mid-day xylem water potential (Ψmin) for all
individuals at each site was measured in late November and
early December using a PMS Model 600 Pressure Chamber
Instrument (PMS Instruments, Oregon, USA). Mid-day water
potentials were measured several times between 11:30 and
14:00 in order to ensure that the lowest water potential was
captured. Pre-dawn and mid-day xylem water potential data
were analyzed separately using a factorial ANOVA (high and
medium streamflow sites) and a one-way ANOVA (low
streamflow site), as described for P50.
Mean streamflow data for Swartbrug were obtained from
Department of Water Affairs and Forestry, while those for
Biessievlei, and Bosboukloof were obtained from the Council
for Scientific and Industrial Research (CSIR; Stellenbosch
office). Witbrug mean streamflow data values were estimated
by subtracting the streamflow of the main tributary in the area,
Langrivier (obtained from the CSIR, Stellenbosch office) from
streamflow measured just upstream of Kleinplaas Dam
(obtained from Department of Water Affairs and Forestry).
Mean monthly streamflow was averaged from 1989–2006.
There are some minor tributaries contributing to flow, but these
Table 1
R2
values for fitted curves for individuals used to calculate P50 and P100 values
Species High streamflow Medium streamflow Low streamflow
Brabejum
stellatifolium
1. 0.94 (10) 1. 0.99 (5) 1. 0.90 (4)
2. 0.84 (7) 2. 0.98 (4) 2. 0.99 (4)
3. 0.96 (8) 3. 0.98 (5) 3. 0.99 (4)
4. 0.96 (7) 4. 0.98 (5) 4. 0.97 (6)
Meterosideros
angustifolia
1. 0.99 (5) 1. 0.91 (8) 1. 0.88 (7)
2. 0.86 (9) 2. 0.89 (10) 2. 0.97 (6)
3. 0.97 (6) 3. 0.99 (7) 3. 0.89 (5)
4. 0.90 (8) 4. 0.90 (6) 4. 0.84 (7)
5. 0.98 (7)
Brachyleana
neriifolia
1. 0.98 (4) 1. 0.89 (7)
2. 0.98 (6) 2. 0.81 (5)
3. 0.88 (4) 3. 0.64 (7)
4. 0.66 (5)
5. 0.61 (4)
Rapanea
melanophloeos
1. 0.94 (11) 1. 0.99 (5) 1. 0.99 (4)
2. 0.91 (10) 2. 0.88 (6) 2. 0.97 (6)
3. 0.92 (9) 3. 0.99 (5) 3. 0.99 (5)
4. 0.71 (7) 4. 0.80 (4)
5. 0.89 (10)
Sample sizes for vulnerability curve points are in parentheses.
Fig. 1. Mean monthly discharge at the study sites (black bars represent
Swartbrug, high streamflow; cross hatched bars represent Witbrug, medium
streamflow; horizontal bars represent Bosboukloof, low streamflow; open bars
represent Biessievlei, low streamflow).
510 C.C. Swift et al. / South African Journal of Botany 74 (2008) 508–516

are not gauged, and so the Witbrug streamflow is only an
estimated value and is likely lower than that given.
3. Results
3.1. Comparison of sites
Mean monthly streamflow differed among the three sites,
and this difference was most pronounced in the wet season
(April–November) (Fig. 1). Differences between the sites
diminished in the summer months of December–March;
however, Swartbrug, the high streamflow site, maintained
high flow relative to the other sites over all months. Differences
between the lowest streamflow sites, Bosboukloof and
Biessievlei, and the medium streamflow site, Witbrug,
decreased substantially during the dry season, and in February
and March, the low streamflow sites actually experience higher
mean streamflow than the medium streamflow site (Fig. 1).
Values of P50, Ψmax and Ψmin differed significantly among
species (Table 2), and values of P50 and Ψmin differed significantly
among sites (Table 2). There was a significant site-by-species effect
for both Ψmax and P50 (Table 2). Although the site effect for Ψmax
was not quite significant (P=0.057, Table 2), large differences
between mean values for three of four species examined, the
significant site-by-species interaction, and a probability value just
above the arbitrary 0.05 value lead us to consider the results of this
analysis further and include post hoc comparisons.
3.2. Individual species across high and medium streamflow
sites
Values of P50 for B. stellatifolium varied between −2.5 MPa
at Bosboukloof, the lowest streamflow site, and −2.3 MPa at
Swartbrug, the highest streamflow site (Fig. 2a; Table 3), but
did not differ significantly across sites (Table 3). Curves for this
Table 2
Factorial ANOVA results for P50, Ψmin and Ψmax, excluding the low streamflow
site
Trait Factor df F P
P50 Site 1 13.69 0.001
Species 3 10.56 0.000
Site×species 3 4.64 0.010
Ψmax Site 1 3.92 0.057
Species 3 4.46 0.011
Ψmin Site 1 15.87 0.000
Species 3 7.02 0.001
Fig. 2. Loss in hydraulic conductivity as a function of xylem water potential for the four species investigated. (a) B. stellatifolium. Closed black symbols and the solid
dark line represent data points and the fitted curve (R2
=0.87, n=41) for Swartbrug, grey symbols and line represent data and fitted curve (R2
=0.92, n=23) for
Witbrug, and the open symbols and dotted line represent data points and the fitted curve (R2
=0.78, n=18) for Bosboukloof. (b) M. angustifolia. Closed black symbols
and line represent data and fitted curve (R2
=0.81, n=31) for Witbrug, and open
symbols and dotted line represent data and fitted curve (R2
=0.81, n=25) for Bosboukloof. (c) B. neriifolia. Closed black symbols and solid dark line represent data
and fitted curve (R2
=0.70, n=19) for Witbrug. (d) R. melanophloeos. Closed
black symbols and line represent data and fitted curve (R2
=0.88, n=21) for
Witbrug, and open symbols and dotted line represent data and fitted curve (R2
=0.88, n=19) for Biessievlei.

species at all sites were J-shaped. Ψmax was similar and not
significantly different among sites, but mean Ψmin decreased by
almost an order of magnitude (P=0.034, Table 4). Mean Ψmin
and Ψmax observed for B. stellatifolium were consistent
with previously published work where Ψmin and Ψmax were
about −0.2 and about −2.0 MPa respectively (Richardson and
Kruger, 1990; Table 4).
The shape of the vulnerability curve of M. angustifolia
differed from that of B. stellatifolium. Individuals from the high
streamflow site exhibited a very steep increase in percent
embolism until Ψx dropped below between −1.0 and −1.5 MPa;
at lower values of Ψx, the rate of increase in percent embolism
leveled off until Ψx dropped below −3.5 to −4.5 MPa, at which
point rates of increase in percent embolism increased again
(Fig. 2b). A less pronounced, but similarly-shaped curve
described individuals at the two drier sites. Mean P50 values
decreased significantly with decreasing streamflow (Table 3).
Ψmin for M. angustifolia individuals was significantly lower at
the high streamflow site (Table 4), and Ψmax decreased by
almost two orders of magnitude from the high to medium
streamflow site (P=0.000; Table 4).
The pattern of vulnerability to cavitation for B. neriifolia was
similar to that for M. angustifolia. Individuals growing at the high
streamflow site showed increased embolism relative to individuals
growing at the medium streamflow site, especially at more
negative xylem water potentials (Fig. 2c). The mean values for P50
differed significantly between the two sites sampled for this
species (Table 3). This species did not occur at the low streamflow
sites. Ψmax did not differ significantly and Ψmin was very similar
across the two sites (P=0.325; Table 4). As with B. stellatifolium,
mean values of Ψmax, andΨmin at the high and medium streamflow
sites, −0.66 and −0.55 MPa, and −1.95 and −1.81 MPa
respectively, were consistent with values in Richardson and
Kruger (1990).
The shape of the curves of R. melanophloeos shifted from
almost linear to J-shaped with decreasing water availability
(Fig. 2d), but there was no significant difference in mean values
of P50, Ψmax, and Ψmin at the high streamflow or medium
streamflow sites, although Ψmin did decrease by an order of
magnitude (P=0.058; Table 4).
3.3. Across-species comparisons
Both B. stellatifolium and R. melanophloeos had signifi-
cantly lower values of P50 than M. angustifolia or B. nerifolia
at the high streamflow site (Table 3). At the medium streamflow
site, B. stellatifolium values of P50 were significantly lower
than R. melanophloeos (Table 3), and were close to being
significantly lower than M. angustifolia (P=0.05) and
B. neriifolia (P=0.056). At the lowest streamflow site, the
mean P50 value for B. stellatifolium was close to being sig-
nificantly lower than the mean for M. angustifolia and
R. melanophloeos, (F=3.508, P=0.075, df=2; Table 3).
Species showed considerable differences in Ψmax at all sites.
At the high streamflow site, B. neriifolia had significantly lower
mean Ψmax, than M. angustifolia or B. stellatifolium (Table 4).
At the medium streamflow site, mean Ψmax of M. angustifolia
is significantly lower than the other three species, while the
mean Ψmax of R. melanophloeos was higher than all other
species (Table 4). Ψmin is less variable at the medium
streamflow site; mean Ψmin of R. melanophloeos is significantly
higher than Ψmin of B. stellatifolium (Table 4). However, at the
high streamflow site Ψmin of M. angustifolia is significantly
lower than all other species. At the low streamflow sites, neither
Table 3
Mean values for P50 across species and streamflow regimes
Brabejum stellatifolium −2.30ab
(0.06) n=4 −2.48a
(0.13) n=4 −2.501
(0.26) n=4
Meterosideros angustifolia −1.29e
(0.09) n=5 −2.09abc
(0.29) n=4 −1.762
(0.29) n=4
Brachyleana neriifolia −1.53d
(0.07) n=5 −2.07abc
(0.04) n=3
Rapanea melanophloeos −1.96bc
(0.11) n=5 −1.85cd
(0.12) n=3 −1.772
(0.06) n=4
Values are in MPa. Standard errors are given in parentheses. Two separate ANOVA procedures were carried out. A factorial ANOVAwas used to compare the high and
medium streamflow regimes and a one-way ANOVA to determine differences within the low streamflow regime. In both cases, means were separated using the
Fishers's LSD multiple range test. Within and between the high and medium streamflow regime, means with different alphabetical superscripts differ significantly
(Pb0.05). Within the low streamflow regime, means with different numerical superscripts differ significantly (Pb0.05).
Table 4
Mean values for Ψmax and Ψmin across species and streamflow regimes
Ψmax Ψmin Ψmax Ψmin Ψmax Ψmin
Brabejum stellatifolium −0.27cd
(0.05) n=5 −2.09bc
(0.10) n=5 −0.50b
(0.06) n=5 −1.99bc
(0.10) n=5 −0.401
(0.07) n=5 −2.0812
(0.06) n=5
Meterosideros angustifolia −0.24d
(0.03) n=5 −2.43a
(0.08) n=5 −0.79a
(0.03) n=5 −1.85bcd
(0.15) n=5 −0.381
(0.12) n=4 −2.261
(0.11) n=4
Brachyleana neriifolia −0.66ab
(0.14) n=5 −1.95bc
(0.08) n=5 −0.55b
(0.07) n=4 −1.81bcd
(0.12) n=4
Rapanea melanophloeos −0.47bc
(0.09) n=5 −1.86bcd
(0.04) n=5 −0.23cd
(0.02) n=3 −1.55d
(0.03) n=3 −0.132
(0.03) n=4 −1.732
(0.19) n=4
Values are in MPa. Standard errors are given in parentheses. Two separate ANOVA procedures were carried out for Ψmax and similarly for Ψmin. A factorial ANOVA
was used to compare the high and medium streamflow regimes and a one-way ANOVA to determine differences within the low streamflow regime. In both cases,
means were separated using the Fishers's LSD multiple range test. Within and between the high and medium streamflow regime, means with different alphabetical
superscripts differ significantly (Pb0.05). Within the low streamflow regime, means with different numerical superscripts differ significantly (Pb0.05).

Ψmax nor Ψmin of B. stellatifolium and M. angustifolia differed
significantly. Mean Ψmax of M. angustifolia was significantly
lower than R. melanophloeos (F=4.10, P=0.050, df=2;
Table 4), and both B. stellatifolium and M. angustifolia
mean Ψmin were significantly lower than R. melanophloeos
(F=10.21, P=0.004, df=2; Table 4).
When mid-day xylem water potential, Ψmin, a measure of
plant water stress, is plotted as a function of values of P50 for
individuals of all species, two groups of species emerge.
Individuals that reach mid-day water potentials below measured
P50 values, the point at which 50% of xylem conducting
capacity is lost, may risk catastrophic embolism. Individuals of
M. angustifolia at both the high and low streamflow sites had
high Ψmax indicating greater water availability and fall below
the line that indicates equal values of mid-day water potential
and values of P50, as do all individuals of B. neriifolia oc-
curring in high streamflow (Fig. 3a). But, at the medium
streamflow site where Ψmax is lower for both M. angustifolia
and B. neriifolia, individuals are either above the line equat-
ing P50 and Ψmin or on the line, as are all individuals of
B. stellatifolium and most R. melanophloeos individuals at all
streamflows. Moreover, mean Ψmin exceeds mean P50 for all
species at lower water availability as indicated by Ψmax, but
Ψmin for both M. angustifolia and B. neriifolia was below
P50 at higher water availability.
If P100, the Ψx at which a total loss of conductivity occurs, is
plotted against mid-day water potential, a different pattern
emerges. M. angustifolia, with a few exceptions maintains a
larger safety margin against 100% loss of conductivity than the
other species, but B. neriifolia sustained Ψmin at 100% loss of
conductivity (Fig. 3b).
If P50 is plotted against P100, the Ψx at which a total loss of
conductivity occurs, two groups emerge. Most species at most
sites form one group, and M. angustifolia individuals from all
sites form a second group with the exception of two individuals
(Fig. 4). M. angustifolia individuals experience much lower Ψx
at 100% loss of conductivity than other species at the same sites
even though the Ψx at 50% loss of conductivity is higher.
A single B. stellatifolium individual from the low streamflow
site belongs to neither group.
4. Discussion
The large differences between Ψmax at the high and medium
sites suggests that water availability did decrease between those
two sites, at least for two of the four species examined,
B. stellatifolium and M. angustifolia. The other two species did
not exhibit decreased Ψmax, while R. melanophloeos showed an
increased Ψmax at the medium streamflow site. Although
R. melanophloeos is considered a riparian species, unlike the
other species, it has afromontane affinities. Lower Ψmax at
higher water availability might suggest that this species may be
less tolerant of saturated soils than the other species examined
(Kozlowski, 1997).
Fig. 3. Measured Ψmin for individuals plotted against xylem water potential at
different levels of hydraulic conductivity. (a) P50. R. melanophloeos is represented
by diamonds, B. stellatifolium by circles, M. angustifolia by triangles and
B. neriifolia by squares. For all species, closed symbols are Swartbrug (high
streamflow) individuals, grey symbols are Witbrug (medium streamflow)
individuals, and open symbols are either Bosboukloof (M. angustifolia and
B. stellatifolium) or Biessievlei (R. melanophloeos). (b) P100. R. melanophloeos is
represented by diamonds, B. stellatifolium by circles, M. angustifolia by triangles
and B. neriifolia by squares. For all species, closed symbols are Swartbrug (high
B. stellatifolium) or Biessievlei (R. melanophloeos).
Fig. 4. Correlation between P50 and P100. R. melanophloeos is represented by
diamonds, B. stellatifolium by circles, M. angustifolia by triangles and
B. neriifolia by squares. For all species, closed symbols are Swartbrug (high
B. stellatifolium) or Biessievlei (R. melanophloeos).

B. stellatifolium, identified as ubiquitous and R. melanoph-
loeos, identified as sensitive to invasion and associated with less
disturbed riparian habitats (Galatowitsch and Richardson,
2005), exhibited similar responses to decreasing summer
water availability across sites, since P50 values were not statis-
tically different across sites (although, R. melanophloeos did
sustain higher Ψmin than B. stellatifolium at all sites). Mean
values of Ψmin always exceeded mean P50 values at all sites.
This is in contrast to B. neriifolia and M. angustifolia, which
both exhibited significantly lower mean P50 values, and mean
Ψmin values that were more negative than P50 values at the
medium streamflow site. Vulnerability to cavitation is related to
xylem anatomy which may be phylogenetically constrained
(Vander Willigen and Pammenter, 1998; Mahareli et al., 2004;
Jacobsen et al., 2007) and may explain the observation that
vulnerability to embolism in B. stellatifolium and R. melanoph-
loeos does not change with increasing or decreasing water
availability. These results lend some support to xylem anatomy
being constrained by phylogenetic factors. However, the finding
that B. neriifolia and M. angustifolia are able to change xylem
anatomical traits in response to decreased water availability
suggests greater phenotypic plasticity for these species or that
phylogenetic constraints may not apply to all taxa, and needs
further investigation.
Vulnerability curve shapes suggest that M. angustifolia is less
drought tolerant than B. stellatifolium (Kolb and Davis, 1994;
Kavanagh et al., 1999; Kolb and Sperry, 1999). Hacke et al.
(2006), however, make the point that, despite initial rapid losses of
hydraulic conductance, this curve shape leads to sustained
hydraulic conductance at lower Ψx and this appears to be true
for M. angustifolia. This is consistent with Vander Willigen and
Pammenter's (1998) point that remaining hydraulic conductivity
may be more important than P50 in explaining the ability to
tolerate low water availability without hydraulic failure. Mean
P50 for M. angustifolia was greater than for B. stellatifolium at all
sites, but mean Ψmin of M. angustifolia was either significantly
lower (high streamflow site) or not significantly different from
B. stellatifolium. In fact, at all sites, M. angustifolia requires much
more negative Ψx to reach 100% embolism than all other species,
again supporting the idea that it is not the hydraulic capacity lost
but the hydraulic capacity remaining that is important to drought
tolerance (Vander Willigen and Pammenter, 1998). M. angusti-
folia maintains a larger safety margin between Ψmin and P100
because of the shallow slope of the vulnerability curve at more
negative Ψx, which may result in increased hydraulic transport,
and is typical of diffuse porous species (Alder et al., 1996).
Continued conductivity under drought conditions in M. angusti-
folia may reflect a range of xylem conduits with different
susceptibility to embolism (Alder et al., 1996; Hacke et al., 2006).
Native embolism has been reported for other riparian species
(Alder et al., 1996; Rood et al., 2000), and may be adaptive when
coupled with branch or leaf shedding as a result of hydraulic
segmentation (Rood et al., 2000).
The four species studied showed marked differences in
response to water availability. B. stellatifolium and R. melanoph-
loeos appear to maintain large safety margins against embolism
by not allowing Ψmin to become more negative than P50
regardless of water availability. Alternatively, both B. neriifolia
and M. angustifolia maximize hydraulic transport at the expense
of embolism, but only when water availability is high. The two
species may therefore differ with respect to degree of vulnerability
to 100% embolism under those conditions since B. neriifolia Ψmin
hovers near 100% embolism. These results suggest that the
similar distributions of these species may be the result of different
strategies to cope with increasing drought. Under high water
availability, B. stellatifolium and R. melanophloeos may limit
stomatal conductance to maintain values of Ψmin that exceed P50.
However, Richardson and Kruger (1990) found no overall
difference in stomatal conductance between B. stellatifolium
and B. neriifolia, although B. stellatifolium maintained more
constant rates of diurnal stomatal conductance than B. neriifolia.
The fact that P50 for B. stellatifolium and R. melanophloeos did
not mirror changes in water supply in this study might suggest that
phylogenetic constraints, which dictate wood density, may be
responsible for increased safety margins at high water availability
(Vander Willigen and Pammenter, 1998; Brodribb and Holbrook,
2004; Mahareli et al, 2004). The ability of M. angustifolia and B.
neriifolia to “relax” safety margins when water availability is high
may translate into higher stomatal conductance which would
translate into greater carbon gain, growth, and enhanced ability to
compete for light. However, B. neriifolia runs a greater risk of
catastrophic embolism than M. angustifolia which retains a
greater capacity for hydraulic transport even with a greater than
50% loss of hydraulic transport.
Both M. angustifolia and B. stellatifolium are more likely to
persist with invasion (Galatowitsch and Richardson, 2005;
Pretorius et al., 2008-this issue) perhaps because they are able to
tolerate decreases in water availability with invasion (Dye and
Jarmain, 2004). In the case of B. stellatifolium, xylem anatomy
may confer a greater safety margin against catastrophic
embolism. In the case of M. angustifolia, xylem traits allow
continued competition for light by maintaining higher stomatal
conductances which translate into carbon gain and growth when
water availability is high, but adjusts xylem traits when water is
more limiting to avoid catastrophic embolism. This is in
contrast to species less likely to co-exist with invasives such as
R. melanophloeos, which exhibited significantly higher Ψmin
than most other species examined at all sites, suggesting
possible limitations on stomatal conductance, or B. neriifolia
which might be highly vulnerable to water limitation or reduced
light availability imposed by invasive species (Pockman and
Sperry, 2000). A strategy involving a high risk of catastrophic
embolism, but greater gas exchange capacity, would be useful if
seasonal flood disturbance removed old branches, and kept
competing species from establishing. Field observations
indicate that this species is largely restricted to the stream
edge and is not ubiquitous across all streams in the Western
Cape (Charlie Boucher, pers. comm.).
The results of this study suggest that B. stellatifolium and
M. angustifolia would be good candidates for inclusion in plant
palettes for post-clearing restoration areas, especially in streams
with low summer water availability since they appear to be more
drought tolerant than the other two species examined. A similar
recommendation was made by Galatowitsch and Richardson

(2005), based on the presence of B. stellatifolium and M.
angustifolia in open or cleared sites in their study. Alternatively,
B. neriifolia would not be a good candidate for inclusion in a
plant palette unless water and light availability are high. The small
safety margin of this species at both P50 and P100 is consistent with
other studies of vulnerability to cavitation in riparian species and is
related to predictable water supply typical of streams (Pockman and
Sperry, 2000). This species may be a good candidate for judging
restoration of fluvial regime as a measure of restoration success.
These are preliminary results, and additional research is
needed to describe stomatal behavior and anatomical traits of
xylem of these species in these different fluvial regimes (Kolb
and Davis, 1994; Brodribb and Holbrook, 2004). This work
does, however, begin to describe potential mechanisms under-
lying the distribution of these four species. Understanding
drought tolerance facilitates refined targets for restoration that
are realistic in terms of water availability requirements of native
species, as well as illustrates the need for continued removal of
non-native invasive species in order to restore fluvial conditions
appropriate for successful restoration (Rood et al., 2003).
Acknowledgements
Financial support through a core team member grant from
the DST-NRF Centre for Invasion Biology (Karen Esler), an
international visitor's grant from Stellenbosch University
(Cheryl Swift) and a faculty research grant from Whittier
College (Cheryl Swift). Shayne Jacobs was partially supported
by a grant from Stellenbosch University. Adrian Simmers
(CSIR, site selection), Eric Prinsloo (CSIR, streamflow data)
and the Department of Water Affairs and Forestry (streamflow
data). Casper Crous and Mariette Pretorius for their assistance
with the field work, and Jessica Schelgel for assistance with
streamflow data analysis. We acknowledge valuable comments
and assistance with earlier drafts from an anonymous reviewer,
Steve Davis and Anna Jacobsen. Finally, MTO Forestry and
CapeNature are thanked for allowing us to carry out this study
on their property.
References
Alder, N.W., Sperry, J.S., Pockman, W.T., 1996. Root and stem xylem
embolism, stomatal conductance and leaf turgor in Acer grandidentatum
populations along a soil moisture gradient. Oecologia 105, 293–301.
Brown, C., 1998. The ecological status of Western Cape rivers investigated:
shock survey findings. African Wildlife 52, 27–28.
Brodribb, T.J., Holbrook, N.M., 2004. Stomatal protection against hydraulic
failure: a comparison of co-existing ferns and angiosperms. New Phytologist
162, 663–670.
Brodribb, T.J., Holbrook, N.M., Edwards, E.J., Gutierrez, M.V., 2003. Relations
between stomatal closure, leaf turgor and xylem vulnerability in eight
tropical dry forest trees. Plant Cell and Environment 26, 443–450.
Coates Palgrave, K., 2002. Trees of Southern Africa, 3rd ed. Struik Publishers,
Cape Town.
Davis, S.D., Ewers, F.W., Wood, J.J., Reeves, K.J., Kolb, K.J., 1999. Differential
susceptibility to xylem cavitation among three pairs of Ceanothus species in the
Transverse Mountain Ranges of Southern California. Ecoscience 6, 180–186.
Dye, P., Jarmain, C., 2004. Water use by black wattle (Acacia mearnsii):
implications for the link between removal of invading trees and catchment
streamflow response. South African Journal of Science 100, 40–45.
Ewers, F.B., Fisher, J.B., 1989. Variation in vessel length and diameter in stems
of six tropical and sub-tropical lianas. American Journal of Botany 76,
1452–1459.
Galatowitsch, S.M., Richardson, D.M., 2005. Riparian scrub recovery after
clearing of invasive alien trees in headwater streams of the Western Cape.
Biological Conservation 122, 509–521.
Gasith, A., Resh, V.H., 1999. Streams in Mediterranean Climate Regions: abiotic
influences and biotic responses to predictable seasonal events. Annual
Review of Ecology and Systematics 30, 51–81.
Hacke, U.G., Sperry, J.S., 2001. Functional and ecological xylem anatomy.
Perspectives in Plant Ecology, Evolution and Systematics 4, 97–115.
Hacke, U.G., Sperry, J.S., Pockman, W.T., Davis, S.D., McCulloch, K.A., 2001.
Trends in wood density and structure are linked to prevention of xylem
implosion by negative pressure. Oecologia 126, 457–461.
Hacke, U.G., Sperry, J.S., Wheeler, J.K., Castro, L., 2006. Scaling of angio-
sperm xylem structure with safety and efficiency. Tree Physiology 26,
689–701.
Hupp, C.R., Osterkamp, W.R., 1996. Riparian vegetation and fluvial geomorphic
processes. Geomorphology 14, 277–295.
Jacobsen, A.L., Pratt, R.B., Ewers, F.W., Davis, S.D., 2007. Cavitation resistance
among twenty-six chaparral species of southern California. Ecological
Monographs 77, 99–115.
Jarbeau, J.A., Ewers, F.W., Davis, S.D., 1995. The mechanism of water stress
induced embolism in two species of chaparral shrubs. Plant Cell and
Environment 18, 189–196.
Kavanagh, K.L., Bond, B.J., Gartner, B., Knowe, S., 1999. Shoot and root
vulnerability to xylem cavitation in four populations of Douglas-fir seedlings.
Tree Physiology 19, 31–37.
Kolb, K., Davis, S.D., 1994. Drought tolerance and xylem embolism in co-
occurring species of coastal sage scrub and chaparral. Ecology 75,
648–659.
Kolb, K., Sperry, J., 1999. Differences in drought adaptation between subspecies
of sagebrush (Artemesia tridentata). Ecology 80, 2372–2384.
Kozlowski, T.T., 1997. Responses of woody plants to flooding and salinity. Tree
Physiology Monograph 1, 1–29.
LeMaitre, D.C., Van Wilgen, B.W., Gelderblom,C.M., Bailey, C., Chapman, R.A.,
Nel, J.A., 2002. Invasive alien trees and water resources in South Africa:
case studies of the costs and benefits of management. Forest Ecology and
Management 160, 143–159.
Linton, M.J., Sperry, J.S., Williams, D.G., 1998. Limits to water transport in
Juniperus osteosperma and Pinus edulis: implications for drought tolerance
and regulation of transpiration. Functional Ecology 12, 906–911.
Mahareli, H., Pockman, W., Jackson, R., 2004. Adaptive variation in the
vulnerability of woody plants to xylem cavitation. Ecology 85, 2194–2199.
Malanson, G.P., 1993. Riparian landscapes. Cambridge Study in Ecology.
Cambridge University Press, Cambridge, p. 296.
Naiman, R.J., Décamps, H., 1997. The ecology of interfaces: riparian zones.
Annual Review of Ecology and Systematics 28, 621–658.
Naiman, R.J., Décamps, H., McClain, M.E., 2005. Riparia: Ecology, Conservation
and Management of Streamside Communities. Elsevier/Academic Press, San
Diego.
Pammenter, N.W., Vander Willigen, C., 1998. A mathematical and statistical
analysis of the curves illustrating vulnerability to xylem to cavitation. Tree
Physiology 18, 589–593.
Petit, N.E., Froend, R.H., 2001. Variability in flood disturbance and the impacts
on riparian tree recruitment in two contrasting river systems. Wetland
Ecology and Management 9, 13–25.
Pockman, W.T., Sperry, J.S., 2000. Vulnerability to xylem cavitation and the
distribution of Sonoran desert vegetation. American Journal of Botany 87,
1287–1299.
Pretorius, M.R., Esler, K.J., Holmes, P.M., Prins, N., 2008. The effectiveness of
active restoration following alien clearance in Fynbos riparian zones and
resilience of treatments to fire. South African Journal of Botany 74, 517–525
(this issue). doi:10.1016/j.sajb.2008.01.180.
Prins, N., Holmes, P.M., Richardson, D.M., 2005. A reference framework for the
restoration of riparian vegetation in the Western Cape, South Africa,
degraded by invasive Australian Acacias. South African Journal of Botany
70, 767–776.

Prinsloo, F.W., Scott, D.F., 1999. Streamflow responses to the clearing of alien
invasive trees from riparian zones at three sites in the Western Cape
Province. South African Forestry Journal 185, 1–7.
Richardson, D.M., Kruger, F.J., 1990. Water relations and photosynthetic
characteristics of selected trees and shrubs of riparian and hillslope habitats
in the southwestern Cape Province, South Africa. South African Journal of
Botany 56, 214–225.
Richardson, D.M., Holmes, P.M., Esler, K.J., Galatowitsch, S.M., Kirkman, S.P.,
Pyšek, P., 2007. Riparian vegetation — degradation, alien plant invasions
and restoration prospects. Diversity and Distributions 13, 126–139.
Rood, S.B., Patino, S., Coombs, K., Tyree, M.T., 2000. Branch sacrifice:
cavitation-associated drought adaptation of riparian cottonwoods. Trees
248–257.
Rood, S.B., Braatne, J.H., Hughes, F.M.R., 2003. Ecophysiology of riparian
cottonwoods: water relations and restoration. Tree Physiology 23, 1113–1124.
Ross, K., Swift, C., 2003. Determinants of southern California riparian
communities. In: Faber, P.M. (Ed.), California Riparian Systems: Process
and Floodplain Management, Ecology, and Restoration. 2001 Riparian
Habitat and Floodplain Conference Proceedings, Riparian Habitat Joint
Venture, Sacramento, CA, pp. 396–406.
Rowntree, K., 1991. An assessment of the potential impact of alien invasive
vegetation on the geomorphology of river channels in South Africa. South
African Journal of Aquatic Science 17, 28–43.
Salinas, M.J., Blanca, G., Romero, A.T., 2000. Evaluating riparian vegetation in
semi-arid Mediterranean watercourses in the south-eastern Iberian Peninsula.
Environmental Conservation 27, 24–35.
Sperry, J.S., Donnelly, J.R., Tyree, M.T., 1988. A method for measuring
hydraulic conductivity and embolism in xylem. Plant Cell and Environment
11, 35–40.
Vosse, S., Esler, K.J.., Richardson, D.M., Holmes, P.M., 2008. Can riparian seed
banks initiate restoration after alien plant invasion? Evidence from the
Western Cape, South Africa. South African Journal of Botany 74, 432–444
(this issue). doi:10.1016/j.sajb.2008.01.170.
Van Coller, A.L., Rogers, K.H., Heritage, G.L., 2000. Riparian vegetation —
environment relationships: complimentarity of gradients versus patch
hierarchy approaches. Journal of Vegetation Science 11, 337–350.
Vander Willigen, C., Pammenter, N.W., 1998. Relationship between growth and
xylem hydraulic characteristics of clones of Eucalyptus spp at contrasting
sites. Tree Physiology 18, 595–600.
Van Wilgen, B.W., Cowling, R.M., Le Maitre, D.C., 1998. Ecosystem services,
efficiency, sustainability and equity: South Africa's Working for Water
programme. Trends in Ecology and Evolution 13, 378.
Zimmerman, M.H., 1983. Xylem Structure and the Ascent of Sap. Springer Press,
New York.
Zimmerman, M.H., Jeje, A.A., 1981. Vessel-length distribution of some
American woody plants. Canadian Journal of Botany 59, 1882–1892.
Edited by ETF Witkowski

Swift etal 2008 SAJB

More Related Content

What's hot (20)

Similar to Swift etal 2008 SAJB (20)

Swift etal 2008 SAJB