Oecologia (2003) 136:213–219
DOI 10.1007/s00442-003-1273-9
ECOPHYSIOLOGY
Frank W. Ewers · Michael C. Lawson ·
Timothy J. Bowen · Stephen D. Davis
Freeze/thaw stress in Ceanothus of southern California chaparral
Received: 10 October 2002 / Accepted: 28 March 2003 / Published online: 10 May 2003
Springer-Verlag 2003
Abstract Freeze/thaw stress was examined in chaparral
shrubs of the genus Ceanothus to determine the interactive effects of freezing and drought and to consider which
is the more vulnerable component, the living leaves
(symplast) or the non-living water transport system
(apoplast). We hypothesized that where Ceanothus species co-occurred, the more inland species C. crassifolius
would be more tolerant of low temperatures than the
coastal species C. spinosus, both in terms of leaf survival
(LT50, or the temperature at which there is 50% loss of
function or viability) and in terms of resistance to
freezing-induced embolism (measurements of percent
loss hydraulic conductivity due to embolism following
freeze/thaw). Cooling experiments on 2 m long winteracclimated shoots resulted in LT50 values of about –10C
for C. spinosus versus –18C for C. crassifolius. Freezethaw cycles resulted in no change in embolism when the
plants were well hydrated (0.7 to –2.0 MPa). However,
when plants were dehydrated to –5.0 MPa, C. spinosus
became 96% embolized with freeze/thaw, versus only
61% embolism for C. crassifolius. Stems of C. crassifolius became 90% and 97% embolized at –6.6 and
–8.0 MPa, respectively, meaning that even in this species,
stems could be more vulnerable than leaves under
conditions of extreme water stress combined with
freeze/thaw events. The dominance of C. crassifolius at
colder sites and the restriction of C. spinosus to warmer
sites are consistent with both the relative tolerance of their
symplasts to low temperatures and the relative tolerance
of their apoplasts to freeze events in combination with
drought stress.
F. W. Ewers ())
Department of Plant Biology, Michigan State University,
East Lansing, MI 48824, USA
e-mail: [email protected]
Fax: +1-517-3531926
M. C. Lawson · T. J. Bowen · S. D. Davis
Natural Science Division, Pepperdine University,
Malibu, CA 90263–4321, USA
Keywords Xylem embolism · Freezing injury ·
Cavitation · LT50 · Drought stress
Introduction
Winter freezing tolerance is one of the factors that limit
the survival and distribution of plants in many habitats
(Shreve 1914; Sakai and Larcher 1987; Pockman and
Sperry 1997), including evergreen shrubs in chaparral
communities of southern California (Langan et al. 1997).
However, the critical factors for the survival of chaparral
plants when they are exposed to subfreezing winter
temperatures remains unclear, since both the living tissues
(symplast) of the leaves and the functioning of the nonliving water transport system (apoplast) of stems are
vulnerable to freezing injury. For Ceanothus megacarpus,
a combination of extreme water stress plus freezing can
result in catastrophic embolism in the stem apoplast
(Langan et al. 1997). In contrast, the symplast of the
leaves of that species can tolerate, with winter acclimation, temperatures down to –9C (Boorse et al. 1998a).
In the present study we chose to compare C. spinosus
to C. crassifolius, each of whose distribution within the
chaparral community might be determined in part by their
relative tolerance of low temperatures. C. spinosus is
narrowly restricted to coastal sites in southern California,
from Santa Barbara to San Diego County (Fig. 1a). At
such sites, freezing events are extremely rare, whereas, at
more inland sites, winter freezing events can occur almost
every year (Boorse et al. 1998a). In contrast, C. crassifolius is restricted to much more inland sites and it is a
dominant shrub only where winter freezing events are
common (Fig. 1c). Only rarely does it co-occur with C.
spinosus.
Inland sites in southern California experience much
lower minimum winter temperatures than coastal sites,
and topography can also greatly impact the seasonal low
temperatures that plants will experience. For instance, we
established a “warm” study site at coastal Malibu that
rarely has winter air temperatures below 0C and a “cold”
214
Fig. 1. a The general distribution of Ceanothus spinosus and C.
crassifolius in southern California. The two species overlap but C.
crassifolius has a more inland distribution. This figure was
reconstructed from a website (http://www.orst.edu/Dept/botany/
herbarium/projects/ceanothus) with permission from C.L. Schmidt.
The vertical arrow indicates the orientation for the transect used in
b. b Transect from Amarillo Beach in Malibu 6.5 km north to Cold
Creek Canyon reconstructed from the Trail Map of the Santa
Monica Mountains (Harrison 1996). The minimum air temperatures
are shown for different microsites in 1989 (see Boorse et al. 1998a).
c Transect through the cold site showing species distributions. Note
that C. crassifolius was restricted to low areas of Cold Creek
Canyon from 4 to 7 km inland, consistent with a vegetation map by
Wieslander (1930–1934; see also Wieslander 1961) and confirmed
in the present study with the use of a GPS III Plus instrument,
Garmin International, 1200 East 151st Street, Olathe, KA 66062,
USA. The arrow shows where plants were sampled
site at Cold Creek Canyon, 5–6 km inland, which can
have winter minimum air temperatures as low as –12C in
some years (Langan et al. 1997; Boorse et al. 1998a). C.
spinosus and C. crassifolius overlap in distribution at the
cold site, and C. crassifolius dominates at the colder
microsites, which are in the lower parts of the canyon, due
to the nighttime hill-and-valley effect (Fig. 1). Sampling
species at the same microsite in the present study helped
to minimize environmentally induced differences, including variance due to seasonal or long-term acclimation.
We used intact shoots for cooling experiments to
determine the differential tolerance of living and nonliving components to freeze/thaw events and the potential
interaction of drought and freezing stress. We hypothesized that where the two species co-occurred, C. crassifolius would be more resistant to low temperatures than C.
spinosus. Furthermore, since both living leaves and the
xylem transport system are vital to the survival of
evergreen shoots, we predicted that both systems of C.
crassifolius would be more resistant to low temperature
damage than those of C. spinosus. Lastly, it has been
suggested that vessel diameter is the critical factor
determining vulnerability to freeze-induced embolism,
and more specifically, that vessels greater than 44 m are
quite vulnerable to embolism under conditions of modest
water stress, based upon empirical data from a broad array
of species (Davis et al. 1999b). We hypothesized that C.
crassifolius would have narrower vessel diameters than C.
spinosus.
Materials and methods
The primary study site was the “cold site” described by Langan et
al. (1997), located on or adjacent to the Malibu Forestry Unit of Los
Angeles County, California, USA, along Malibu Canyon Road, at
an elevation of 180 m (3450 N, 118420 W). Throughout the study,
temperature data were recorded with data loggers at both the warm
and cold sites shown in Fig. 1b, as described in Langan et al.
(1997). The temperatures indicated in Fig. 1b, c are the lowest
temperatures recorded at those sites in the last 15 years (October
1988 to February 2003, based on data logger and nearby weather
station records; Langan et al. 1997; Boorse et al. 1998a, recent
unpublished data). Similarly, adjacent the cold site, early morning
temperatures of 11C were recorded on 3 days during the winter
of 1989–1990 and on 4 days in 1994–1995 (weather station records,
215
Malibu Forestry Unit, Los Angles County). At the bottom of the
valley the winter minimum temperatures were 3C lower than they
were at the microsites 70 m higher in elevation, up the hills
(Fig. 1c), and up to 12C lower than at the warm coastal exposure
site (Fig. 1b, c; Boorse et al. 1998a).
Plants of Ceanothus spinosus (Torrey & A. Gray) and C.
crassifolius (Torrey), were sampled at the “ecotone” demarcated by
the arrow at the cold site in Fig. 1c. The mixed chaparral
community at the cold site was composed of C. crassifolius, C.
spinosus, C. megacarpus (Nutt.), C. oliganthus (Nutt.), Adenostoma
fasciculatum (Hook. & Arn.), Cercocarpus betuloides (Torrey & A.
Gray), Malosma laurina (Nutt.), Rhus ovata (S. Watson), and
Quercus berberidifolia (Liemb). Nomenclature follows Hickman
(1993). The sampled plants were between 25 and 30 years old,
since the site was completely burned in the Malibu wildfire of 1970
(Los Angeles County Fire Department), and plants were sampled
between 1995 and 2000. Typical rates of cooling and warming in
the winter were determined previously at the site (Langan et al.
1997), allowing for realistic cooling experiments.
Stems of C. spinosus and C. megacarpus were also sampled at
the “warm site” shown in Fig. 1b, but C. crassifolius did not occur
there. This site contained a mature stand of mixed chaparral
adjacent Pepperdine University in Los Angeles County (see “Warm
Site A” in Boorse et al. 1998a).
At the cold site, native xylem embolism was measured in winter
and in summer using the hydraulic conductivity method to
determine the percent loss in conductivity due to embolism (Sperry
et al. 1988; Langan et al. 1997). For each of n=12 individuals, one
branch greater than 2 m in length was sampled at the ecotone. The
branches were collected soon after sunrise (0600 to 0800 hours),
placed in a moist plastic bag to avoid dehydration, and taken to an
air-conditioned laboratory. A small twig was then sampled from
each shoot for measurement of xylem water potential with a
pressure chamber (Scholander et al. 1965). The stem segments
sampled for conductivity and embolism were about 6 mm in
diameter and were cut under water to a length of 0.1 m.
Cooling experiments on shoots greater than 2 m in length were
modified from those described by Langan et al. (1997) for
measurements of xylem embolism, and included measurements of
leaf injury and LT50, that is, the temperature at which there is a 50%
loss of leaf function (Boorse et al. 1998b). For each sampled
individual, one experimental and one control shoot were collected
at the field site soon after sunrise, placed in a moist plastic bag to
avoid dehydration, and taken immediately to the laboratory.
Controls were kept in the plastic bag in the air-conditioned
laboratory while the experimental shoots were cooled in a chamber
described previously (Langan et al. 1997). Unlike in the Langan et
al. study, none of the shoots were placed with their stem in a bucket
of water during the experiment; instead the water potentials were
kept close to what the plants were experiencing in the field. As in
the Langan et al. study (1977), temperature gradients within the
chamber were minimized, such that the temperature gradients
throughout the chamber always totaled less than 1C, and no fan
was used in the chamber, to mimic the still cool night air of
freezing events in the chaparral. The rate of cooling was set at
0.08C min1 when decreasing from 15C to 0C, and at 0.02C
min1 when decreasing from 0 to –25C. The warming rate was
0.08C min1 from –25C to 15C. Thermocouples were used to
record air temperatures, stem temperatures, stem exotherms and
leaf temperatures inside the chamber. In most cases the shoot water
potentials were kept the same as when they were field collected, by
storing the shoots in large plastic bags when they were not in the
chamber, but in one series of experiments, shoots were allowed to
dehydrate in the air conditioned laboratory until they reached about
–5 MPa of xylem pressure. Then the control shoots were put into
moist bags to prevent further water loss while the experimental
shoots were put through the freeze/thaw treatment in the chamber.
It was clear that freeze/thaw occurred during the experiments,
since all experimental shoots of both species showed exotherms
during the cooling cycle and endotherms upon warming. By way of
controls, dried wooden dowels in the chamber, of the same
dimensions as the experimental stems, were monitored and they did
not show exotherms or endotherms. The mean temperatures for the
exotherms were –2.9C (SE=0.45) and –2.4C (SE=0.32) for C.
crassifolius and C. spinosus, respectively.
In some experiments, six shoots were cooled in the chamber per
experimental run, and dark-adapted, photosynthetic fluorescence
capacity (Fv/Fm) was performed on each shoot after the freeze/thaw
experiment was completed (Boorse et al. 1998b). In other
experiments, just three shoots were cooled at one time and one
leaf per shoot was sampled, at 1 increments throughout the cooling
experiment, to determine the LT50, that is, the temperature at which
there was a 50% loss of leaf function or viability. In those
experiments, before putting the plants in the chamber, small
clippers were fastened to twigs, with strings attached so that leaves
could be remotely excised and pulled out of the bottom of the
chamber, without disrupting the cooling cycle. Three different leaf
viability tests were performed on each shoot at the 1 increments in
cooling (1) photosynthetic fluorescence capacity (Fv/Fm), (2) the
percentage of palisade mesophyll cells stained by fluorescein
diacetate (vital stain), and (3) visual score of leaf color. Following
the cooling treatment, experimental and control (untreated) leaves
were stored in the dark, in a zip lock bags, in an ice cooler, until
used in the viability tests (Boorse et al. 1998b). For all three
techniques, the treated leaves quickly changed in their viability
scores as they neared the critical LT50 temperatures, going from
near 100% viability to very low viability over a range of about 3C.
Liner regression lines (with r2 values always greater than 0.90)
were used to determine the temperature at which there was a 50%
loss in viability.
For photosynthetic fluorescence capacity, a dark adaptation
cuvette was placed on each leaf for 30 min, and a chlorophyll
fluorescence meter (Model CF-1000, P.K. Morgan Instrument,
Andover, Mass.) was set at an exposure time of 15 s and an actinic
light level of 200 mmole m2 s1 PFD to measure Fv/Fm. Such
measurements were performed 1 day and 3 days after freezing
treatment to detect the possibility of fluorescence recovery, but
since there was no significant recovery detected, calculations were
based on values taken 1 day after the cooling treatment. Healthy
leaves had Fv/Fm values of about 0.80, and, following severe freeze
damage, the values were at about 0.20; they flattened out at about
0.20 and they did not drop below 0.20 even at the lowest
temperatures tested. The LT50 values occurred at Fv/Fm values of
about 0.50 in species of Ceanothus, as reported previously (Boorse
et al. 1998b).
For the vital stain technique, 1 day after the cooling treatment a
0.01% solution of fluorescein diacetate was applied to hand
sections of leaves to observe palisade parenchyma cells, which,
under blue fluorescent light (Excitation filter BG-12 on a
Microphot-FX Nikon scope with epifluorescence, Garden City,
N.Y.), fluoresced either yellow-green (alive) or exhibited red, chla
autofluorescence (dead). The sample size was n=100 cells per leaf,
with healthy leaves typically having nearly 100% of their palisade
cells living, and severely freeze-damaged leaves with close to 0%
living palisade cells (Boorse et al. 1998b).
For the visual score technique leaves were stored in the cooler
for two days after the treatment. With the use of a Munsell plant
tissue color chart (1977 edn., Forestry Supply, Jackson, Miss.)
leaves were given one of three scores, depending on how discolored
the leaves were relative to control leaves. The scores were 0,
similar to control in color; 0.5, darker green than control; and 1.0,
brownish green (Boorse et al. 1998b).
After the freeze/thaw cycle was completed, one twig per shoot
was sampled for measurement of water potential, and percent loss
in conductivity due to embolism was measured in a 6 mm diameter,
0.1 m long stem segment.
For measurements of vessel diameters in 6 mm diameter stems,
populations of C. megacarpus and C. spinosus were sampled at
both the warm and cold study sites, whereas C. crassifolius only
occurred at the cold site. Plants of C. megacarpus were sampled
further up the slope from the ecotone at the cold site, with care
taken to avoid apparent hybrids between C. megacarpus, which has
alternate leaves, and C. crassifolius, which has opposite leaves.
Vessel diameters were measured on a compound microscope from
216
transverse sections using an ocular micrometer, with total magnification of 400. All the vessels in one sector, defined by rays,
including inner and outer areas of the xylem, were sampled such
that at least 100 vessels per stem were measured.
For statistical comparisons, a one-way ANOVA was followed,
when appropriate, by a Fisher’s PLSD post-hoc test. For comparing
percentages such as percent loss in conductivity due to embolism,
results were ARCSIN-transformed prior to analysis. Differences
were considered as statistically significant when P was <0.05.
Results
At the cold site, Ceanothus crassifolius had significantly
greater native embolism than C. spinosus, both during
winter and summer. Both species showed significantly
greater embolism in winter than in the following summer
(Fig. 2). We also sampled native embolism of C. spinosus
at the warm site in the winter, where freezing temperatures had not occurred. The percent loss conductivity was
just 32.3% (SE 3.3%). The winter value at the warm site
was not significantly different from the summer value at
the cold site (data not shown).
The plants should have been fully acclimated to low
temperatures at the cold site in January of 1996, given that
minimum air temperatures of –5C and –6C occurred
during the early mornings of January 23 and 24. During
the time period that followed, freeze/thaw experiments
resulted in no significant change in embolism for either
species; C. crassifolius remained more embolized than C.
spinosus (Fig. 3). The plants were not drought stressed
during these experiments since they were performed
during the winter rainy season. However, the twigs of C.
crassifolius had significantly lower water potentials than
those of C. spinosus. The mean € SE values for the
stems following the freeze/thaw treatments were 1.41€
0.09 MPa for C. crassifolius and –1.12€0.09 MPa for C.
spinosus. For both species, the water potentials for the
control stems were not significantly different from the
freeze/thaw stems, indicating that the freeze/thaw experiment did not dehydrate the shoots.
The vulnerability of the leaves to low temperatures
was quite different for the two species. Depending upon
the type of viability test, the LT50 values ranged from
about –9 to –12C for C. spinosus versus from –17.7 to
–18.7C for C. crassifolius. For every test, the LT50
values were significantly lower for C. crassifolius than
for C. spinosus (Fig. 3).
During July 1996, when native embolism was at a low
level, shoots were dehydrated in the laboratory to about –
5 MPa before undergoing a cooling experiment. Under
those conditions, the freeze-induced embolism was much
more severe for C. spinosus, which ended up with 96%
embolism, which would likely be fatal in nature (Fig. 4).
In contrast, experimental stems of C. crassifolius, while
showing a significant increase in embolism as a result of
the freeze/thaw treatment, had only 61% embolism,
similar to normal winter embolism levels for that species.
In other experiments we dehydrated both species to –
8 MPa, which is close to the driest that these species are
Fig. 2 Winter and summer native embolism (percent loss conductivity) for C. spinosus and C. crassifolius at Cold Creek Canyon;
mean€SE, n=12 individuals. Winter embolism was measured
between 1 February and 7 March 1996 following temperatures
below –5C in the proceeding month; summer embolism was
measured between 15 and 22 July 1996. Values with a different
letter are significantly different from one another (P<0.05)
Fig. 3 Winter freezing experiments showing, on the left graph,
percent loss xylem conductivity with freeze/thaw and, on the right
graph, the temperature at which there was a 50% loss of leaf
function (LT50) based upon photosynthetic fluorescence (Fv/Fm),
visual score, and a vital stain. For each species, means€SE for n=6
experiments are shown; SE not shown when less than 0.3. Three
individuals were sampled per experiment, with all experiments
done on native plants between 17 January and 4 March 1996. The
water potentials of the experimental and control shoots ranged from
–0.7 to –2.0 MPa for these experiments. Values with a different
letter are significantly different from one another
found in nature. With the freeze/thaw treatment both
species then became 97% embolized (data not shown).
To determine whether a combination of water stress
and freezing would be more hazardous to the xylem
transport system or to the living leaves of native plants,
freeze/thaw experiments were done on plants of C.
crassifolius that were experiencing an unusually prolonged summer drought which extended into November
of 1995. For the control shoots the mean water potential
was –6.6 MPa (SE=0.10), whereas following the freeze/
thaw treatment, the water potential was –5.7€0.1 MPa,
217
Fig. 4 Effect of freeze/thaw on percent loss conductivity when
water potentials were at –5.0 MPa in two species of Ceanothus.
Shoots were dehydrated to –5 MPa during July of 1996, prior to the
freeze/thaw experiments. For each species the experiment was
repeated three times on a total of n=18 shoots. Values with a
different letter are significantly different from one another
Fig. 6 Maximum vessel diameters in 6 mm diameter stems of three
species of Ceanothus. C. megacarpus and C. spinosus were
sampled at both the warm and cold sites, whereas C. crassifolius
only occurs at the cold site. Sample size was n=6 individuals at the
warm site and n=14 at the cold site. Values with a different letter
are significantly different from one another
low temperatures, but there were no recent freeze/thaw
events to induce embolisms. The freeze/thaw treatment,
with a minimum temperature of –10C, resulted in a
significant increase in xylem embolism from 65%
(SE=3.7%) in controls to 90% (SE=1.6%) in experimental
shoots (Fig. 5). However, the effect of drought plus the
10C freeze/thaw cycle on leaf vitality was minimal.
There was no significant difference in the overall mean €
SE Fv/Fm value for the controls (0.79€0.014) versus
experimental shoots (0.78€0.016). For comparative purposes, in cooling experiments for this species, the average
Fv/Fm value at LT50, meaning 50% loss of cell function,
was 0.46 (Fig. 5).
Mean and maximum vessel diameters were significantly greater for the populations at the warm site than at
the cold site. At the cold site, maximum vessel diameters
were less than 44 m for all three species of Ceanothus
(Fig. 6) and mean vessel diameters were 24.7€1.0 m,
22.8€0.8 m and 22.6€1.0 m for C. megacarpus, C.
spinosus, and C. crassifolius, respectively. Differences
among species at the cold site were not statistically
significant.
Fig. 5 Freezing experiments on plants of C. crassifolius that were
collected at Cold Creek Canyon under severe natural drought
conditions, from 13 to 30 November 1995, with shoots at xylem
water potentials of from –5.0 to –6.4 MPa. Freezing to –10C had
little effect on leaf vitality (Fv/Fm) but a significant effect on xylem
embolism (Percent loss conductivity). The dotted line in the top
graph shows the average Fv/Fm value at LT50 that we determined
for this species (0.46). For each of the six trials, an asterisk
indicates a statistically significant difference between the experimental and control, with n=6 individuals per trial
suggesting the freeze-induced embolism may have released some water to the leaves. Data loggers indicated
the plants had experienced nighttime temperatures as low
as +1C during the autumn prior to the experiments.
Therefore the symplasts should have been acclimated to
Discussion
The higher amounts of embolism in the winter than in the
summer, in both species at the cold site, suggest that
freezing-induced embolism occurs there. Freezing experiments done during the winter had no effect on the levels
of embolism, probably because those plants had already
recently experienced freeze-induced embolism events,
and one additional freeze/thaw cycle, when the plants
were fairly well hydrated, had no additional impact. In the
spring, chaparral plants can produce new xylem, which
lowers the measured embolism values. The summer is a
period of drought stress in the chaparral, and drought-
218
induced embolism can then be a limiting factor for some
species of chaparral plants (Kolb and Davis 1994;
Williams et al. 1997; Davis et al. 2002).
Under well-hydrated conditions, the leaves of plants of
C. spinosus were vulnerable to temperatures lower than
–10 C at the cold site. The LT50 values for this species
were similar to those reported earlier for C. megacarpus
during the winter (Boorse et al. 1998a). However, plants
of C. crassifolius were much more tolerant of low
temperatures, with LT50 values of about –18C. Therefore
the leaves of C. crassifolius are well adapted to the
coldest temperatures that have been recorded at the cold
site (12C). In contrast, leaf survival in the face of
winter low temperatures could help limit the distribution
of both C. spinosus and C. megacarpus. Unlike reports for
plants of Larrea tridentata (Pockman and Sperry 1997;
Martnez-Vilalta and Pockman 2002), in C. crassifolius
and C. spinosus it made no difference to xylem embolism
as to whether the minimum temperature was –10C or
–20C, as long as freezing had occurred in the xylem sap
(data not shown). This is consistent with a previous study
for C. megacarpus (Langan et al. 1997).
C. spinosus resprouts from the base following destruction of the aerial shoots by wildfire. In contrast, C.
megacarpus and C. crassifolius are non-sprouters following wildfire (Thomas and Davis 1989; Davis et al. 1998,
1999a). The ability to resprout following leaf and shoot
death could extend the range of a species. Perhaps this is
why C. spinosus is found at slightly colder sites than is C.
megacarpus, which appears to have nearly identical
tolerance of low temperatures. Another perspective is
that leaves of C. crassifolius, which occurs at the coldest
sites, can tolerate temperatures about 8C colder than the
other two species. That safety margin may be essential for
long-term survival at those sites since C. crassifolius is
not able to resprout following shoot death.
When plants of Ceanothus are dehydrated, as in the
summer and early fall in the chaparral, then a freeze/thaw
event could seriously impair the hydraulic transport
system. This is particularly the case for C. spinosus;
when plants were dehydrated to –5 MPa, a freeze/thaw
treatment resulted in 97% embolism. In nature, such
elevated embolism levels would, in all likelihood, result
in shoot death (Davis et al. 2002). Unlike for C. spinosus
and unlike a previous report for C. megacarpus (Langan
et al. 1997), when plants of C. crassifolius were
dehydrated to –5 MPa, the freeze/thaw treatment resulted
in only 61% embolism, a value that is typical for winter
embolism in this species. Therefore, as hypothesized, C.
crassifolius is more tolerant of low temperatures than C.
spinosus and C. megacarpus, both in terms of tolerance of
the living leaves, and tolerance of the xylem to freeze/
thaw events under drought stress conditions.
However, for both the freeze tolerant C. crassifolius,
and for the less tolerant species C. spinosus, which system
is more vulnerable to low temperatures, the xylem
transport system, or the living leaves? Under well
hydrated conditions, certainly the leaves are the more
vulnerable; they lose viability at about –18C (or at just
–10C for C. spinosus), whereas the xylem transport
systems appear to be tolerant of freeze/thaw events at
least to –25C, the minimum temperature that our
chamber achieved, and much lower than experienced at
our cold study site. However, under special conditions of
a freeze/thaw event under very severe drought conditions,
the xylem transport system of both species could be more
vulnerable than the leaves. As noted above, at –8 MPa, a
freeze/thaw treatment resulted in 97% embolism for C.
crassifolius (with a similar result at just –5 MPa for C.
spinosus), but do such conditions occur in nature for these
species? Perhaps, but only in a rare year, when a
prolonged summer drought lasted into the winter, with a
freeze event occurring before the first rains of the winter
season. There is a weather station very near to the cold
site (Malibu Forestry Unit, Los Angles County), but
unfortunately most of the older records have been lost and
so we are dependant on fairly recent data. In November
1995 we did freeze/thaw treatments on shoots of C.
crassifolius with water potentials of –6.6 MPa, before the
first rains of the winter, and the result was 90% embolism
of the shoots. In 1997, a prolonged summer drought that
lasted into November resulted in field water potentials of
as low as –11.2 MPa in some branches of C. crassifolius
at the cold site (Davis et al. 2002). However, in both of
those years, the lowest temperature reached before the
first winter rains came was +1C, and so the plants
narrowly averted a catastrophic freeze event. Catastrophic
freeze-induced embolism in this species might be rare, but
possible.
The vessels of the three Ceanothus species were
remarkably narrow at the cold site. In fact they were all
narrower than the 44 m freeze/thaw vulnerability
threshold noted by Davis et al. (1999b). That threshold
was for water potential conditions of –0.5 MPa, and the
diameter threshold would probably be lower when the
freezing occurred under extreme drought conditions. The
fact that the three species were not significantly different
in vessel diameter at the cold site, and the remarkable
resilience of C. crassifolius relative to the other species,
suggests that factors other than vessel diameter help to
determine vulnerability to freeze/thaw embolism. Those
factors might include vessel length, biophysical properties
of the vessel walls and/or activities of living parenchyma
cells of the xylem.
Since C. crassifolius is clearly more cold tolerant than
the other two species of Ceanothus, it is understandable
that it would dominate at the colder microsites. However,
what prevents it from occurring closer to the coast?
Apparently it is not quite as drought tolerant as C.
megacarpus, and not as deep rooted as C. spinosus. At
sites closer to the coast in southern California, the
summer drought tends to be more prolonged, and the soil
water potentials become more severely negative (Miller
and Poole 1979). At the cold site, C. crassifolius shows
considerable drought-induced shoot dieback (Davis et al.
2002), which is consistent with drought stress limiting the
coastal distribution of this species. As noted above,
freezing stress should not limit the distribution of C.
219
crassifolius in the Santa Monica Mountains unless the
freezing events co-occurred with severe drought conditions.
Acknowledgements Sincere thanks to Mike Takeshita and Bradley
Yocum of the Malibu Forestry Unit of Los Angeles County for
assistance with field work and weather data, and to Brandon Pratt
for critical review of the manuscript. This work was funded by NSF
Grants BIR-9225034 and IBN-950753.
References
Boorse GC, Ewers FW, Davis SD (1998a) Response of chaparral
shrubs to below-freezing temperatures: acclimation, ecotypes,
seedlings vs adults. Am J Bot 85:1224–1230
Boorse GC, Bosma TL, Meyer A-C, Ewers FW, Davis SD (1998b)
Comparative methods of estimating freezing temperatures and
freezing injury in leaves of chaparral shrubs. Int J Plant Sci
159:513–521
Davis SD, Kolb KJ, Barton KP (1998) Ecophysiological processes
and demographic patterns in the structuring of California
chaparral. In: Rundel PW, Montenegro G, Jaksic F (eds)
Landscape disturbance and biodiversity in Mediterranean-type
ecosystems, vol 136. Springer, Berlin Heidelberg New York, pp
297–310
Davis SD, Ewers FW, Wood J, Reeves JJ, Kolb KJ (1999a)
Differential susceptibility to xylem cavitation among three
pairs of Ceanothus species in the transverse mountain ranges of
southern California. Ecoscience 6:180–186
Davis SD, Sperry JS, Hacke UG (1999b) The relationship between
xylem conduit diameter and cavitation caused by freezing. Am
J Bot 86:1367–1372
Davis SD, Ewers FW, Sperry JS, Portwood KA, Crocker MC,
Adams GC (2002) Shoot dieback during prolonged drought in
Ceanothus (Rhamnaceae) chaparral of California: a possible
case of hydraulic failure. Am J Bot 89:820–828
Harrison T (1996) Trail map of the Santa Monica Mountains
Central. Tom Harrison Cartography, San Rafael, Calif.
Hickman JC (1993) The Jepson manual. University of California
Press, Berkeley
Kolb KJ, Davis SD (1994) Drought-induced xylem embolism in cooccurring species of coastal sage and chaparral of California.
Ecology 75:648–659
Langan S J, Ewers FW, Davis SD (1997) Xylem dysfunction
caused by water stress and freezing in two species of cooccurring chaparral shrubs. Plant Cell Environ 20:425–437
Martnez-Vilalta J, Pockman WT (2002) The vulnerability to
freezing-induced xylem cavitation of Larrea tridentate (Zygophllaceae) in the Chihuahuan desert. Am J Bot 89:1916–
1924
Miller PC, Poole DK (1979) Patterns of water use in shrubs in
southern California. For Sci 25:84–98
Pockman WT, Sperry JS (1997) Freezing-induced xylem cavitation
and the northern limit of Larrea tridentata. Oecologia 109:19–
27
Sakai A, Larcher W (1987) Frost survival of plants. Ecological
Studies 62. Springer, Berlin Heidelberg New York
Scholander PF, Hammel HT, Bradstreet ED, Hemmingsen EA
(1965) Sap pressure in vascular plants. Science 148:339–346
Shreve F (1914) The role of winter temperatures in determining the
distribution of plants. Am J Bot 1:194–202
Sperry JS, Donnelly JR, Tyree MT (1988) A method for measuring
hydraulic conductivity and embolism in xylem. Plant Cell
Environ 11:35–40
Thomas CM, Davis SD (1989) Recovery patterns of three chaparral
shrub species after wildfire. Oecologia 80:309–320
Wieslander AE (1930–1934) A vegetation type map of California,
Santa Monica Mountains, Malibu. U.S. Forest Service, California
Wieslander AE (1961) California’s vegetation maps. Recent
Advances in Botany, University of Toronto Press, Toronto,
1402–1405
Williams JE, Davis SD, Portwood KA (1997) Xylem embolism in
seedlings and resprouts of Adenostoma fasciculatum after fire.
Aust J Bot 45:291–300