New Phytologist - 2019 - Rosas - Adjustments and Coordination of Hydraulic Leaf and Stem Traits Along A Water Availability

Research
Adjustments and coordination of hydraulic, leaf and stem traits

along a water availability gradient
Teresa Rosas1 , Maurizio Mencuccini1,2 , Josep Barba3 , Herve Cochard4 , Sandra Saura-Mas1,5 and
Jordi Martınez-Vilalta1,5
1
CREAF, E08193 Bellaterra (Cerdanyola del Valles), Catalonia, Spain; 2ICREA, 08010, Barcelona, Spain; 3Plant and Soil Sciences Department, University of Delaware, Newark, DE 19716,
USA; 4INRA, PIAF, Universite Clermont-Auvergne, Site de Crou€el 5, chemin de Beaulieu, 63000, Clermont-Ferrand, France; 5Universitat Autonoma de Barcelona, Cerdanyola del Valles,
08193, Barcelona, Spain
Summary
Author for correspondence: Trait variability in space and time allows plants to adjust to changing environmental condi-
Teresa Rosas tions. However, we know little about how this variability is distributed and coordinated at dif-
Tel: +34 935812915
ferent organizational levels.
Email: teresa.rosas.torrent@gmail.com
For six dominant tree species in northeastern Spain (three Fagaceae and three Pinaceae) we
Received: 27 July 2018 quantified the inter- and intraspecific variability of a set of traits along a water availability
Accepted: 8 January 2019
gradient. We measured leaf mass per area (LMA), leaf nitrogen (N) concentration, carbon iso-
tope composition in leaves (d13C), stem wood density, the Huber value (Hv, the ratio of cross-
New Phytologist (2019) 223: 632–646 sectional sapwood area to leaf area), sapwood-specific and leaf-specific stem hydraulic
doi: 10.1111/nph.15684 conductivity, vulnerability to xylem embolism (P50) and the turgor loss point (Ptlp).
Differences between families explained the largest amount of variability for most traits,
Key words: drought, Huber value, hydraulic

although intraspecific variability was also relevant. Species occupying wetter sites showed
traits, interspecific variation, intraspecific higher N, P50 and Ptlp, and lower LMA, d13C and Hv. However, when trait relationships with
variation, leaf economics spectrum, turgor water availability were assessed within species they held only for Hv and Ptlp.
loss point, water availability. Overall, our results indicate that intraspecific adjustments along the water availability gradi-
ent relied primarily on changes in resource allocation between sapwood and leaf area and in
leaf water relations.
investments (Reich et al., 1997; Wright et al., 2004). However,

Introduction
how to exactly describe and integrate complex community
Understanding the patterns underlying the huge diversity in dynamics and predict ecosystem-level responses to environmental
plant form and function across different organizational levels is a changes from individual-level trait measurements remains a chal-
central goal for ecologists. This diversity arises from a combina- lenge (Shipley et al., 2016; Funk et al., 2017).
tion of genetic variation and phenotypic plasticity and results in Functional variability of plants has been frequently collapsed at
adaptations to a range of environmental conditions across space the species level by using mean values, thus ignoring intraspecific
and time (Bradshaw, 1965, 2006). In recent decades, trait-based trait variability (ITV). However, much work has shown that ITV is
ecology has emerged as a renewed discipline with the potential to relevant when making predictions about plant community assembly
be applied to dynamic global vegetation models (Van Bodegom and ecosystem functioning (Violle et al., 2012). This is particularly
et al., 2012; Harper et al., 2016) and improve predictions of vege- the case when we move from global to more regional scales (Messier
tation responses to environmental changes (Lavorel & Garnier, et al., 2010; Albert et al., 2012; Violle et al., 2012; Siefert et al.,
2002; McGill et al., 2006). The use of traits emphasizes species 2015) and from organ-level traits to integrative traits involving sev-
phenotypic values over taxonomic characteristics, facilitating the eral organs, as the latter tend to be more sensitive to the environ-
comparison among species and environments (Westoby & ment and show higher ITV as a result of local genetic adaptation
Wright, 2006). Identifying tradeoffs that appear repeatedly and phenotypic plasticity (Marks, 2007; Siefert et al., 2015). Thus,
because of evolutionary constraints has become a major research incorporating the variability of traits along environmental gradients
topic because they have the potential to reflect ecological strate- among different organizational levels (family, species, population
gies (Westoby et al., 2002; Laughlin, 2014; Adler et al., 2014). and individual) may help to elucidate how traits respond to envi-
One of the dimensions that has received more attention is the leaf ronmental variation and thus improve trait-based models. For
economics spectrum (LES), which highlights the tradeoff example, Reich et al. (2014) showed that accounting for ITV in
between the dry mass and nutrient investments in leaf construc- gymnosperm needle longevity with latitude across boreal forests
tion and the time required for obtaining returns on those impacted significantly on carbon (C) cycling projections.
632 New Phytologist (2019) 223: 632–646 Ó 2019 The Authors

www.newphytologist.com New Phytologist Ó 2019 New Phytologist Trust
14698137, 2019, 2, Downloaded from https://nph.onlinelibrary.wiley.com/doi/10.1111/nph.15684 by INPA - Instituto Nacional de Pesquisas da Amazonia, Wiley Online Library on [03/04/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
New
Phytologist Research 633
A related challenge is to understand how trait covariation communities (Skelton et al., 2015) and may provide stronger
changes at different ecological levels (organizational levels and insights into the drivers of forest dynamics than the more com-
spatial scales) (Levin, 1992; Chave, 2013). Previous work has monly measured ‘soft’ traits (Brodribb, 2017).
shown that correlation patterns are not always conserved across If trait variation across scales in commonly measured ‘soft’
scales. For example, several studies have failed to find some of the traits remains poorly understood, knowledge is even more limited
central LES tradeoffs, defined across species means at the global regarding hydraulic traits. A recent meta-analysis found that 33%
scale, when working at smaller spatial or organizational scales of the variation in P50 was contributed by differences within
(Wright & Sutton-Grier, 2012; Laforest-Lapointe et al., 2014; species (Anderegg, 2015). However, part of this variability could
Niinemets, 2015; Messier et al., 2016; Anderegg et al., 2018). be a result of methodological aspects (Cochard et al., 2013) and
This is because traits that appear closely coordinated at certain several individual studies have shown low plasticity in embolism
scales may have different sensitivities to scale-dependent drivers resistance across climatically contrasted populations (Maherali &
of variation, which can effectively decouple them at finer scales DeLucia, 2000; Martınez-Vilalta et al., 2009; Lamy et al., 2011,
(Messier et al., 2016). These results have important implications 2014; Lopez et al., 2016). The degree of coordination between
for trait-based ecology: if we want to predict species responses to leaf economics traits and hydraulic traits is also a leading research
changing environmental conditions, we need to elucidate subject. A universal ‘fast–slow’ whole-plant economics spectrum
intraspecific trait covariance structures to understand the adaptive that integrates resource use strategies (for water, C and nutrients)
value of trait combinations in different environments. At the across organs has been proposed (Reich, 2014), but the evidence
same time, we should be cautious when interpreting trait rela- remains mixed (Brodribb et al., 2007; Blonder et al., 2011;
tionships across species as fundamental tradeoffs among func- Markesteijn et al., 2011; Mendez-Alonzo et al., 2012; Sack et al.,
tions and strategy dimensions. The study of trait correlation 2013; Li et al., 2015).
networks is a step forward in formalizing multiple factors shaping To address these critical issues, we studied the variability of a
an integrated plant phenotype (Poorter et al., 2014; Messier set of hydraulic, leaf and stem traits along a water availability gra-
et al., 2017) and allowing comparisons across scales. dient in six dominant tree species in Catalonia (northeastern
The complexity of trait variation has usually been condensed Spain), focusing on the following questions:
in a few easily measured (‘soft’) traits that are not necessarily good (1) How much trait variation is observed and how is it dis-
predictors of demographic rates (Poorter et al., 2008; Paine et al., tributed among levels of organization? We hypothesize that dif-
2015; Yang et al., 2018). For example, leaf mass per area (LMA), ferences between families (Pinaceae vs Fagaceae) will explain the
one of the most commonly measured traits, is usually weakly largest part of trait variability in this temperate system, although
associated with growth rate, especially in adult plants (Wright ITV will be substantial, especially for more integrative traits such
et al., 2010; Gibert et al., 2016). Moving from ‘soft’ traits to as KL and Hv.
more mechanistic (‘hard’) traits that have a clearer physiological (2) How do traits vary along the water availability gradient
basis and are likely to be stronger determinants of fitness should within and between species? We hypothesize that hydraulic traits
improve our capacity to elucidate vegetation dynamics under will be more closely linked to water availability than are other
changing environmental conditions. This is particularly the case stem and leaf traits. Most of the trait changes along the water
for drought-related impacts on forest function and dynamics availability gradient will entail species substitutions and, thus, the
(Skelton et al., 2015; Sperry & Love, 2015; Brodribb, 2017), strength of trait–environment relationships will be weaker within
which are expected to increase in most regions of the Earth under than across species, reflecting lower capacity for functional adjust-
climate change (Allen et al., 2015). ment within species.
Several studies have related hydraulic traits to plant perfor- (3) How are traits coordinated across and within species?
mance under drought in terms of growth and mortality rates Across species, we hypothesize the existence of a general ‘fast–
(Rowland et al., 2015; Anderegg et al., 2016; Choat et al., 2018). slow’ strategy at the whole-plant level that combines LES and
Hydraulic traits define the efficiency of the plant water transport hydraulic traits (e.g. low LMA will be associated with high KS
system, usually defined in terms of stem-specific hydraulic con- and high vulnerability to embolism). At the same time, we
ductivity (KS) and its safety against failure under drought stress, expect that intraspecific correlation networks may differ from
typically characterized as the water potential at which 50% stem those across species because relatively weak evolutionary or
conductivity is lost as a result of xylem embolism (P50). In addi- physiological tradeoffs can be reversed as a result of plasticity
tion, allocation to sapwood cross-sectional area relative to leaf within species.
area (the Huber value, Hv) regulates supply capacity per unit of
water demand, and it is thus a key component of plant hydraulic
Materials and Methods
architecture (Mencuccini & Bonosi, 2001). It has been shown
that plants can respond to drier conditions by increasing the resis-
Study site and sampling design
tance to xylem embolism (e.g. Blackman et al., 2014), decreasing
the leaf water potential at turgor loss in leaves (Bartlett et al., The study area included all the forested territory of Catalonia
2012) and/or increasing their sapwood-to-leaf area ratio (northeastern Spain), encompassing 1.2 million ha, c. 38% of its
(Martınez-Vilalta et al., 2009). Thus, these hydraulic traits can be total land area. Catalonia is very diverse both topographically and
used to describe the range of plant hydraulic strategies in diverse climatically: mean annual temperature ranges from 18°C (at the
Ó 2019 The Authors New Phytologist (2019) 223: 632–646

New Phytologist Ó 2019 New Phytologist Trust www.newphytologist.com
New
634 Research Phytologist
southern coast) to 3°C (in the Pyrenees) and annual rainfall varies b ¼ 3:140 0:00222 ð% clayÞ2 3:484
from 400 to > 1500 mm (CDAC, http://www.opengis.uab.cat/ 105 ð% sandÞ2 ð% clayÞ Eqn 1
acdc/). We selected six of the most dominant tree species in
Catalonia (three Pinaceae and three Fagaceae), accounting for where less negative values of b indicate sandy soils with lower soil
c. 75% of the total forest area (Gracia et al., 2004; see also Sup- water retention capacity.
porting information Table S1): Pinus sylvestris L, Pinus nigra Forest structure data for each plot were also available from the
J.F.Arnold., Pinus halepensis Mill., Fagus sylvatica L., Quercus last Spanish forest inventory (IFN4) that was conducted over the
pubescens Willd. and Quercus ilex L. For each species, 15 plots same time period as our sampling. Forest structural data included
from the Spanish forest inventory (IFN) were resampled in which total plot basal area, stand density, mean DBH and the 90th per-
the target species was dominant (minimum 50% of the total basal centile for height of all trees in the plot. Climate data were
area), maximizing the water availability gradient occupied by obtained from the Climatic Digital Atlas of Catalonia (Ninyerola
each species in the study region. Water availability was quantified et al., 2005), a collection of digital maps at 200 9 200 m resolu-
as the precipitation to potential evapotranspiration ratio tion including average annual radiation, mean annual tempera-
(P : PET), for the spring–summer period (see later). Five plots ture, minimum annual temperature and annual precipitation for
per species were sampled for each of three species-specific P : PET the period 1951–2010. PET values were calculated according to
ranges following a stratified random design (dry, corresponding the Hargreaves–Samani method (Hargreaves & Samani, 1982)
to P : PET < 33rd/66th percentile and mild for the rest) (Figs S1, and used to estimate P : PET for the spring–summer period and
S2). Plots with the two highest stoniness values and those that P : PET for the summer period.
had been managed during the last 14 years according to previous
IFN surveys were discarded. Leaf traits and wood density
Within each plot, five nonsuppressed canopy trees of the
target species with diameter at breast height (DBH) > 12.5 cm Standard protocols (Perez-Harguindeguy et al., 2013) were
were randomly selected, all within 25 m of the centre of the followed for all trait measurements (Table 1). Previous-year
plot. All samples and data were collected from May to Decem- needles (conifers) and current-year leaves (broadleaves) were
ber 2015. To minimize phenological variation in traits within selected to measure fully expanded leaves. Before measurements,
species, species were sampled sequentially (P. halepensis, mid- twigs with leaves were cut under water and placed into flasks
May to end June; Q. pubescens, end June and July; F. sylvatica, with the cut end submerged in deionized water in the dark
August; P. sylvestris; September to mid-October; Q. ilex, mid- overnight.
October to mid-November; P. nigra, mid-November to mid- Leaf mass per area is a measure of biomass investment in leaves
December). From each tree, two branches (one for leaf per unit light interception and gas exchange (Poorter et al.,
measurements and the other for hydraulic measurements) were 2009). For LMA determinations, 20 leaves were randomly
sampled from the exposed part of the canopy in the top half selected, scanned and their areas were measured with IMAGEJ soft-
of the crown. Sampled branches were at least 70 cm long for ware (Wayne Rasband-National Institute of Health, Bethesda,
Pinus spp., 150 cm for Quercus spp. and 80 cm for Fagus, to MD, USA). Afterwards, samples were oven-dried at 60°C and
account for differences in the maximum length of xylem con- weighed, and LMA was calculated as leaf dry mass/fresh area.
duits (see later). Branches were transported to the laboratory The Hv is the ratio of cross-sectional sapwood area to sub-
inside plastic bags under cool and dark conditions and mea- tended leaf area, and it can therefore be viewed as the ratio of
surements were taken within 24 h. hydraulic and mechanical investment costs over the expected
gains obtained by leaf display. Leaves from terminal branches
(65 cm long from the tip) were oven-dried and weighed, and
Environmental variables LMA was used to convert the total dry weight of the distal leaves
At each plot, four soil samples (20 cm deep) were taken using a of each branch into total branch leaf area. In order to calculate
soil core at the four cardinal points at 5 m distance from the cen- branch level Hv, and to make values comparable across species,
tre of the plot. The topsoil (O horizon) was removed to exclude maximum leaf area was estimated, taking into account species
the organic deposit and litterfall, and the four samples were phenology and the time of sampling. Sapwood area was obtained
merged. The following variables were measured on each pooled through measuring total xylem area on digital images of stained
sample: N-NO3 concentration (colorimetric method; Keeney & (safranin-astra blue) 15–20 lm thin sections in IMAGEJ (Wayne
Nelson, 1982), phosphorus (P) content (available P-Olsen P; Rasband-National Institute of Health).
Olsen & Sommers, 1982), soil humidity (gravimetric soil water We used leaf C isotope composition (d13C) and leaf nitrogen
content; Gardner, 1986), organic matter fraction (organic C con- (N) concentrations to further characterize leaf functioning. Less
tent estimated with acid dichromate oxidation method; Nelson negative d13C values suggesting lower discrimination against the
& Sommers, 1982) and soil texture classes defined by the USDA heavier 13C are indicative of greater stomatal control and water-
system (sedimentation-Robinson pipette; Gee & Or, 2002). To use efficiency (Farquhar et al., 1989), whereas higher leaf N con-
integrate the different components of soil texture into one single centrations are usually associated with higher photosynthetic
variable, the exponent of the Saxton equation (Saxton et al., capacity because of the high N content of photosynthetic
1986) was calculated as follows: machinery (Evans, 1989). Leaf d13C and N were determined
New Phytologist (2019) 223: 632–646 Ó 2019 The Authors

New
using a PDZ Europa ANCA-GSL elemental analyser interfaced shoots (Martin-StPaul et al., 2014). An initial cut was applied to
to a PDZ Europa 20–20 isotope ratio mass spectrometer (Sercon allow xylem tension in the branch segment to relax before mea-
Ltd, Crewe, UK) at the UC-Davis Stable Isotopes Facility (Davis, surement, avoiding artefacts associated with the cutting under
CA, USA). Samples were previously oven-dried at 60° C for 4 d, tension (Wheeler et al., 2013). After the segments were cut again
grounded with a Retsch MM400 ball mill (Verder Group, Haan, to their final size (c. 2 cm in length), their proximal ends were
Germany) and placed in tin capsules for analysis. C stable isotope connected to the tubing system of the XYL’EM, which was filled
concentrations were expressed in relation to the Pee-Dee Belem- with deionized filtered water with 10 mM KCl and 1 mM CaCl2
nite standard. that had been previously degassed using a membrane contactor
Leaf osmotic potential (W0) was measured with a VAPRO (Liqui-Cell Mini-Module membrane 1.7 9 5.5; Charlotte, NC,
5500 vapor pressure osmometer (Wescor, Logan, UT, USA). USA). After measuring the initial conductivity, the segments
Leaves were wrapped in foil to limit condensation and evapora- were flushed once at 0.15 MPa for 10 min (for Quercus spp. and
tion, submerged in liquid nitrogen for 2 min and sealed in a plas- F. sylvatica) or held in the solution under partial vacuum for 48 h
tic zip bag under ambient conditions. After letting them defrost, (for Pinus spp., as flushing conifer segments often results in the
they were put inside a syringe and squeezed until 10 ll of sap pit membranes being permanently pushed against tracheid cells
were obtained. Finally, W0 was used to predict the leaf water walls) in order to measure their maximal conductivity (Kmax) as
potential at which leaf cells lose turgor, closing their stomata and described earlier. The values of Ki and Kmax were used to com-
ceasing gas exchange and growth (Ptlp) (Brodribb et al., 2003), pute the percentage loss of hydraulic conductivity (PLC). The
according to the equation described in Bartlett et al. (2012): previous measurements were repeated a second time on a differ-
ent set of stem segments after branches had been dehydrated on
Ptlp ¼ 0:832W0 0:631 Eqn 2 the bench to obtain PLC estimates at lower water potentials. The
timing of this second measurement was adjusted for different
Wood density (WD) is considered a central trait shaping the species and branches (between 2 and 8 d) to cover a wide range of
wood economics spectrum (Chave et al., 2009). We measured PLC values. The tubing system was regularly cleaned using 10%
WD on one stem core per individual extracted using a hand bleach solution for at least 20 min to prevent microorganism
increment borer (5 mm diameter; Suunto, Vantaa, Finland). The growth and, afterwards, flushed with a degassed solution. Addi-
core was sealed in plastic tubes upon collection and taken to the tionally, we used the apical part of each measured twig segment
laboratory under cold conditions. Fresh core volume of all wood to measure water potential (Ψ) with a Scholander pressure cham-
was calculated after removing the bark by the dimensional ber (Solfranc Tecnologias, Tarragona, Spain).
method, measuring its total length and its diameter using a cal- To fit vulnerability curves to each set of PLC and water poten-
liper. Cores were then oven-dried at 100°C for 48 h and weighed. tial measurements, the following sigmoid function was used
WD was calculated as the oven-dry mass divided by fresh (Pammenter & Willigen, 1998):
volume.
PLC ¼ 100=ð1 þ expðaðW P50 ÞÞÞ Eqn 3
Hydraulic traits
where Ψ is the water potential, a is the slope of the curve (and
Before hydraulic measurements, maximum vessel length was esti- thus determines the rate at which conductivity is lost as water
mated using the air infiltration technique (Ewers & Fisher, 1989) potential declines), and P50 determines the position of the curve
on eight 2 m branches per species. We pumped compressed air on the abscissa and gives the pressure causing 50% loss of con-
(c. 0.15 MPa) through the branches with their basal end ductivity. Parameters were estimated by fitting a separate nonlin-
immersed in water and successively shortened the stem until bub- ear mixed model for each species, using the NLME R package
bling was observed. Because compressed air at low pressures can- (Pinheiro et al., 2018). The model accounted for individual
not pass through vessel end walls, the bubbling indicated the nested in plot as a random effect on coefficient P50. Preliminary
presence of open xylem conduits. The resulting estimates of max- analyses confirmed that this model structure provided the best fit
imum conduit length were used to decide the minimum length to the data.
of the sampled branches (see the Study site and sampling design In addition, all distal leaves of each segment were removed to
section earlier). determinate their area as explained above. Leaf-specific hydraulic
Vulnerability curves were established by measuring the conductivity (KL) was calculated as Kmax divided by the distal leaf
hydraulic conductivity of stem segments at different water poten- area supported. Similarly, stem-specific hydraulic conductivity
tials, using a variation of the bench dehydration method (Sperry (KS) was calculated as Kmax divided by cross-sectional sapwood
& Tyree, 1988; Cochard et al., 2013; Choat et al., 2015). area.
Hydraulic conductivity was measured using a commercial
XYL’EM apparatus (Bronkhorst, Montigny-les-Cormeilles,
Statistical analyses
France) as the ratio between the flow through the stem segment
and the pressure gradient (5 kPa). The initial hydraulic conduc- To assess trait variability, the quartile coefficient of dispersion
tivity (Ki) was measured in three subsamples (segments) per (QCD) was calculated for each trait as the ratio between half the
branch that were excised underwater at the terminal part of the interquartile range ((Q3 – Q1)/2) and the average of the quartiles

New
((Q1 + Q3)/2). QCD was used as a more robust measure of dis- represented as nodes and their correlation as the edges linking
persion than the coefficient of variation (CV), as the latter is not them. Two indicators of network centrality were calculated for
appropriate for datasets including isotopic measurements (Bren- each trait: the degree (D), defined as the number of edges of a
del, 2014) or log-transformed data (Canchola et al., 2017) (see node and the weighted degree (Dw), defined as the sum of all sig-
also the Table S2). To understand the distribution of variability nificant coefficients of correlation of a node (Table S5). In these
for each trait, we used different sets of linear mixed models, latter analyses, all traits were loge-transformed to improve the lin-
always fitting separate models for each trait. In the first ones, fam- earity of relationships. All analyses were carried out using R statis-
ily, species and population were introduced as nested random fac- tical software v.3.3.2 (R Core Team, 2017).
tors to assess how trait variability was distributed among these
different levels of organization. In the second ones, models were
Results
fitted separately for each family, and included only species and
population (nested) to assess trait variability among and within
Magnitude and distribution of trait variability
species (within each family). All variables were checked for nor-
mality and loge-transformed whenever required to ensure nor- Most trait variation occurred between families (Pinaceae vs
mality. Fagaceae), with the exception of KL and Ptlp for which the contri-
Before exploring the effect of environmental factors on trait vari- bution of family was close to zero (Fig. 1b). Pinaceae tended to
ation, three separate principal component analyses (PCAs) were have higher LMA, Hv and d13C than Fagaceae, whereas the oppo-
performed to summarize soil, forest structure and climate data site was true for leaf N, WD, KS and P50 (Fig. S4). Overall, the
(Fig. S3). As before, all variables were checked for normality and proportion of variance explained at the intraspecific level (among
loge-transformed if required. For further analyses, the two most and within populations) was, on average, 23.11% (Fig. 1b).
orthogonal variables showing the highest axes loading in each PCA Within Pinaceae, KS, KL, Hv, WD and d13C showed a higher vari-
were selected as integrated measures of environmental predictors. ability within than among species, whereas in Fagaceae this was
Coefficient b from the Saxton equation (Eqn 1) and soil P were only the case for Hv (Fig. S5). Other traits, such as P50, showed
selected to describe soil characteristics, mean tree DBH and total substantial variability within families (4.51 MPa range within
plot basal area to describe forest structure, and spring–summer Pinaceae and 3.84 MPa range within Fagaceae) but most of this
P : PET and annual radiation to describe the climate. A first mixed variance occurred across species (Table S2; Fig. S5). KS, KL, LMA
model for each trait was fitted starting with the ‘saturated’ model, and Hv were the most variable traits, whereas d13C, Ptlp and WD
including all six environmental variables as fixed explanatory vari- showed the least variation (Fig. 1a).
ables (without interactions). We included plot nested in species as
random effects on the intercept of the model. Preliminary analyses
Trait responses along a water availability gradient
showed that including a random species effect on the slopes did not
improve model fit. This model was simplified stepwise, removing Traits responded differently to environmental factors (Table 2).
the least significant term until a minimal adequate model with the Regarding soil properties, only soil P concentration showed a sig-
lowest Akaike information criterion (AIC) was obtained. Models nificant effect (positive) on LMA. As for stand structure, mean
within two AIC units were considered equivalent in terms of fit and DBH had a strongest predictive effect across all models. Plots
the simplest one was selected (Zuur et al., 2009). with larger trees on average were associated with lower LMA,
To explore specifically the variability of each trait along the lower WD, lower Hv and lower KL. Stand basal area did not have
P : PET gradient imposed by our sampling design, a second significant effects on any trait. Finally, regarding climatic vari-
mixed model was fitted for each trait. To separate the intraspeci- ables, high annual radiation was associated with leaves with high
fic from the interspecific component of trait responses to P : PET, LMA and high (less negative) d13C. Plots with higher P : PET
we split P : PET into two additive variables which were included values had trees with more negative d13C, lower Hv and less neg-
as separate fixed factors in the model: mean P : PET at the species ative Ptlp. Overall, environmental variables at the plot level were
level and centred P : PET. The latter variable was calculated as not strong predictors of trait variation, as shown by relatively low
the difference between plot P : PET and the average P : PET for values of the marginal R2 (variation explained by the fixed effects)
the corresponding species. We also included plot nested within (Table 2). The fact that conditional R2 values (Table 2) were nor-
species as a random effect on the intercept. As before, preliminary mally much higher indicates that a large proportion of the vari-
analyses showed that including a random species effect on the ance in all traits is explained by differences among species and
slope did not improve model fit. Model selection was carried out plots not captured by the environmental variables included in
as described earlier. In all cases, the residuals of the selected mod- our analysis. Similar results were obtained if we used PCA axes as
els showed no obvious pattern and were approximately normally fixed factors describing environmental variation in models
distributed. Linear mixed effects models were fitted using the instead of individual variables (Table S3).
LME4 package (Bates et al., 2015). When we specifically explored the variability of each trait
Finally, to characterize trait coordination within and between along the water availability (P : PET) gradient, considering both
species, statistically significant correlations among traits were species means and plot-scale deviations from the means (centred
graphically represented using trait covariation networks with the values), higher marginal R2 values and generally stronger effects
IGRAPH package (Csardi & Nepusz, 2006). Traits were were obtained (cf. Table S4). Significant relationships between

New
(a) (b)
Fig. 1 (a) Quartile coefficient of dispersion of

the studied traits across all measurements of
the study; (b) variance partitioning across
different ecological levels of organization.
‘Within’ denotes variance between
individuals of the same population. Traits are
ordered (left to right) from higher to lower
total variation in (a), and from higher to
lower percentage variation within species in
(b). LMA, leaf mass per area; N, leaf nitrogen
concentration; d13C, leaf carbon isotope
composition; WD, wood density; Hv, Huber
value, sapwood to leaf area ratio; KL,
leaf-specific xylem hydraulic conductivity;
KS, stem-specific xylem hydraulic
conductivity; P50, pressure causing 50%
xylem embolism; Ptlp, leaf water potential at
turgor loss point.
Table 1 Traits measured in this study. Trait correlation networks

Trait Symbol Units Trait coordination differed within- and among- species
(Fig. 3). When species means were considered, LMA and Hv
Leaf mass per area LMA kg m2 were the traits showing highest values of centrality across
Leaf nitrogen concentration N mg g1
Leaf carbon isotope composition d13C & species (Table S5). These two traits were positively related to
Wood density (stem) WD g cm3 each other and tightly linked to leaf N, d13C and P50, so that
Huber value, sapwood to leaf area ratio Hv cm2 m2 higher allocation to sapwood area relative to leaf area was cor-
(branch) related with a greater construction cost per unit leaf area, lower
Leaf-specific xylem hydraulic KL kg m1 s1 MPa1 N, higher water-use efficiency (less negative d13C values) and
conductivity (branch)
Stem-specific xylem hydraulic conductivity KS kg m1 s1 MPa1
higher cavitation resistance (more negative P50). Ptlp and KL
(branch) also showed a positive relationship. Surprisingly, KL and KS
Pressure causing 50% xylem embolism P50 MPa were unrelated across species, although a consistent, positive
(branch) relationship appeared when species were analysed separately
Leaf water potential at turgor loss point Ptlp MPa (Fig. 3).
[Correction added after online publication 8 February 2019: the units for When analysing trait coordination within species, the strong
Leaf mass per area in Table 1 have been corrected.] LMA–Hv relationship observed across species was only signifi-
cant in one species (Q. ilex). At the intraspecific level, the negative
P : PET and traits across species were consistent with the results correlation between LMA and –d13C and the positive correlation
reported in the previous paragraph, but we also found a positive between KS and KL were the only relationships present in all cases
relationship between P : PET and P50 (which was only marginally (Fig. 3). KL showed the highest centrality in two out of the three
significant in the previous analysis) and a positive relationship measured gymnosperms, whereas it was never central in
with leaf N concentrations (Fig. 2). Importantly, trait–environ- angiosperms. On the other hand, LMA was the trait with the
ment relationships were scale-dependent and when these patterns highest centrality in two out of three studied angiosperm species.
were analysed within species, we only found significant relation- However, caution is needed when considering these results,
ships between centred P : PET and Hv and Ptlp. In these two owing to the limited number of species sampled within each fam-
cases, the relationships had the same (negative) sign but shallower ily. When centrality was expressed as a simple count of the num-
slopes than the corresponding relationships across species (Fig. 2; ber of significant correlations (D), d13C and Ptlp also appeared
Table S4). Similar results were obtained when the mean DBH, particularly important, especially in Fagaceae (Table S5). Finally,
the strongest explanatory variable in the initial mixed model (cf. taking into account the overall network, P. sylvestris, F. sylvatica
previous paragraph), was included as a fixed factor in this latter and Q. ilex were the species showing more correlations among
model (not shown). traits and the highest Dw (Table S5).

638 Research
www.newphytologist.com
Table 2 Results of the linear mixed models examining the relationships between traits and environmental variables characterizing the soil, the climate and the stand structure.
New Phytologist (2019) 223: 632–646

LMA Log N Log(d13C) Log WD Hv KL KS P50 Ptlp
Fixed parts
Soil P 0.08 0.08 ni 0.08 ni ni ni ni ni
(0.02 – 0.13)* (0.17–0.01) (0.00–0.17)
Log (b Saxton) ni ni ni 0.05 ni ni ni ni 0.03
(0.01–0.12) (0.13–0.07)
Log (DBHmean) 0.07 0.09 0.12 0.11 0.14 0.14 0.08 ni ni
(0.12 (0.00–0.17) (0.00–0.24) (0.19 to 0.03)** (0.26 to 0.01)* (0.24 to 0.03)* (0.16 to 0.00)
to 0.01)*
Basal area 0.05 ni ni 0.01 ni ni ni 0.08 ni
(0.10–0.01) (0.06–0.09) (0.03–0.20)
Annual 0.06 ni 0.23 ni 0.09 0.07 ni ni ni
radiation (0.01–0.10)* (0.33 (0.01–0.20) (0.02–0.17)
to 0.13)***
Log (P : PET) ni ni 0.25 ni 0.27 ni ni 0.18 0.36
(0.09–0.41)** (0.44 to 0.10)** (0.35 to 0.02) (0.53 to 0.20)***
Random part
r2 0 0.009 0.001 0.009 0.178 0.153 0.087 0.071 0.024
s00, PLOT:SP 0 0.009 0.001 0.002 0.065 0.039 0.036 0.193 0.017
s00, SP 0.005 0.084 0.002 0.035 0.134 0.216 0.521 0.775 0.053
R2 marginal 0.02 0.01 0.22 0.01 0.14 0.03 0 0.02 0.14
R2 conditional 0.94 0.91 0.84 0.8 0.59 0.64 0.86 0.93 0.78
Observations 444 444 444 444 439 444 444 444 425
A different mixed effects model including all environmental variables in the fixed part and plot nested within species in the random part was fitted for each trait. The model’s standardized coefficients
including confidence intervals (in brackets) are shown. Significant correlations (*, P < 0.05; **, P < 0.01; ***, P < 0.001) are shown in bold. Information on the random effect variances (r2, total; s00,
PLOT:SP, within-species; s00, SP, cross-species), the proportions of explained variance by fixed effects (R2 marginal) and explained variance by fixed and random effects (R2 conditional) is also pro-
vided. DBHmean, plot mean diameter at the breast height; b Saxton, b Saxton coefficient; soil P, soil phosphorus content; P : PET, precipitation to potential evapotranspiration ratio; ni, not included in
the best model; LMA, leaf mass per area; N, leaf nitrogen concentration; d13C, leaf carbon isotope composition; WD, wood density; Hv, Huber value, sapwood to leaf area ratio; KL, leaf-specific
xylem hydraulic conductivity; KS, stem-specific xylem hydraulic conductivity; P50, pressure causing 50% xylem embolism; Ptlp, leaf water potential at turgor loss point.
New
Phytologist
New Phytologist Ó 2019 New Phytologist Trust

Ó 2019 The Authors
New
(a) (b) (c)
(d) (e) (f)
(g) (h) (i)
Fig. 2 Relationship between water availability (in terms of the precipitation to potential evapotranspiration ratio, P : PET) and (a) leaf mass per area, (b) leaf
nitrogen concentration, (c) leaf carbon isotope composition, (d) wood density, (e) Huber value, sapwood to leaf area ratio, (f) leaf-specific xylem hydraulic
conductivity, (g) stem-specific xylem hydraulic conductivity, (h) pressure causing 50% xylem embolism and (i) leaf water potential at turgor loss point. The
black regression lines give the overall cross-species relationships, and the coloured lines give the corresponding within-species relationships, when significant
(P < 0.05). Variables were loge-transformed whenever required to satisfy normality assumptions. LMA, leaf mass per area; N, leaf nitrogen concentration;
d13C, leaf carbon isotope composition; WD, wood density; Hv, Huber value, sapwood to leaf area ratio; KL, leaf-specific xylem hydraulic conductivity; KS,
stem-specific xylem hydraulic conductivity; P50, pressure causing 50% xylem embolism; Ptlp, leaf water potential at turgor loss point. [Correction added after
online publication 8 February 2019: the units for LMA on the y-axis of panel (a) have been corrected.]
especially in more integrative traits (KL, Hv). Most study traits

Discussion
responded to water availability, with increasing N, P50 and Ptlp
We found that traits varied primarily between tree families but and decreasing LMA, d13C and Hv with P : PET across species.
that ITV also accounted for a relevant amount of total variation, However, at the intraspecific level, we only found trait variation

New
(a)
(b) (c) (d)
Fig. 3 (a–g) Trait correlation networks across

species (n = 6) (a) and for each studied
species separately (b–g). Solid black and grey
dashed edges show positive and negative
correlations, respectively. Correlation
strength is represented by edge thickness.
Only significant correlations are shown
(P < 0.05). Traits identified by red circles
(e) (f) (g) show the highest centrality value in terms of
weighted degree (the sum of all the
significant coefficients of correlation of a
node). All traits were loge-transformed
before analysis. LMA, leaf mass per area; N,
leaf nitrogen concentration; d13C, leaf
carbon isotope composition; WD, wood
density; Hv, Huber value, sapwood to leaf
area ratio; KL, leaf-specific xylem hydraulic
conductivity; KS, stem-specific xylem
hydraulic conductivity; P50, pressure causing
50% xylem embolism; Ptlp, leaf water
potential at turgor loss point.
along the water availability gradient for Hv and Ptlp. Finally, trait within and among species, remains to be elucidated, particu-
coordination was scale-dependent and we did not find clear evi- larly considering that in our study KS and KL did not respond
dence of a single, dominant axis of variation reflecting a fast– consistently to water availability. On the other hand, Ptlp
slow, whole-plant economics spectrum. showed very low variability in comparison to other hydraulics
traits, also in agreement with previous findings (Mencuccini
et al., 2015; Bartlett et al., 2016).
Magnitude and distribution of trait variability
Not surprisingly, trait variability was mostly distributed across
Our results show that traits differ substantially in their variabil- families, reflecting the contrasting trait syndromes between
ity along the same environmental gradient, with an order of angiosperm and gymnosperm clades (Wright et al., 2004; Chave
magnitude difference in the quartile coefficient of dispersion et al., 2009; Carnicer et al., 2013). Our results also confirm previ-
between the most variable (KS and KL) and the least variable ous findings for hydraulic traits, with higher Hv, lower KS and
traits (d13C and Ptlp). The high variability of KS and KL agrees higher resistance to embolism in conifers relative to angiosperm
with previous studies across species (Maherali et al., 2004; trees (Becker et al., 1999; Choat et al., 2012; Gleason et al.,
Martınez-Vilalta et al., 2004; Gleason et al., 2015) and may be 2015). The high proportion of variation attributed to the family
caused by their high sensitivity to small differences in wood level for KS is explained by xylem conduit properties, as unicellu-
anatomy (particularly conduit diameter), which varies substan- lar conifer tracheids are substantially narrower and more than an
tially across and within species (Tyree et al., 1994; Sperry et al., order of magnitude shorter than angiosperm vessels (Sperry et al.,
2008). The higher variability of KS relative to KL probably 2006). Besides the direct effect of these different dimensions on
reflects that the latter is normalized by water demand in terms KS, the fact that we measured relatively short segments implies
of leaf area. More generally, however, the ecological implica- that our KS estimates corresponded mostly to lumen conductivity
tions of this high variability in xylem transport capacity, both for the Fagaceae and to total conductivity (lumen and end wall)

New
for Pinaceae species. Interestingly, the family effect disappeared higher at drier sites as a result of water stress adaptation through
when xylem conductivity was normalized by leaf area (KL) increasing wilting resistance (Schulze et al., 1998; Cunningham
because conifers also tend to have more sapwood per unit leaf et al., 1999). Regarding the relationship between N and water
area (higher Hv; Fig. S4) (see also Becker et al., 1999). availability, contrasting results have been reported. Although
Intraspecific trait variability contributed to a substantial some studies have reported that species from drier sites present
amount of the total variance (from 6% to 42% depending on the higher N leaf concentration to enhance water conservation dur-
trait). This is consistent with a growing body of evidence showing ing photosynthesis (Wright & Westoby, 2002), others have
that ITV is relevant (Albert et al., 2012; Laforest-Lapointe et al., found no general relationship (Killingbeck & Whitford, 1996;
2014), especially when we move from organ-specific traits (leaves, Vila-Cabrera et al., 2015).
stems or roots) to more integrative traits involving several organs Vulnerability to xylem embolism was lower (more negative
(e.g. KL, Hv) (Siefert et al., 2015). Studies addressing ITV in P50) in species occupying drier sites, consistent with the notion
hydraulic traits are less frequent (but see Martınez-Vilalta et al., that cavitation resistance is a key determinant of species distribu-
2009; Wortemann et al., 2011; Lamy et al., 2014; Hajek et al., tions (Maherali et al., 2004; Jacobsen et al., 2007; Martınez-
2016). In line with our results, Hv and KL have been reported to Vilalta et al., 2012; Choat et al., 2012; Blackman et al., 2014;
be among the most plastic hydraulic properties in pines (DeLucia Trueba et al., 2017; Li et al., 2018; Skelton et al., 2018). Simi-
et al., 2000; Martınez-Vilalta et al., 2009) whereas other traits larly, another key drought tolerance trait, Ptlp, also showed a sig-
such as P50 usually show low plasticity (Maherali & DeLucia, nificant relationship with P : PET across species, with lower
2000; Martınez-Vilalta et al., 2009; Lamy et al., 2014; Lopez (more negative) Ptlp associated with drier habitats, allowing the
et al., 2016). Further studies are needed to investigate whether maintenance of leaf turgor and gas exchange under drier condi-
these patterns are generalizable across other plant families. It tions (Brodribb et al., 2003; Lenz et al., 2006; Bartlett et al.,
should also be noted that we probably underestimated the magni- 2012). This did not, however, prevent an increase in water-use
tude of ITV because we did not cover the whole species distribu- efficiency (less negative d13C values) and increased allocation to
tion range, species were sampled sequentially to minimize sapwood area relative to leaf area (Hv) at drier sites, consistent
phenological variation within species, and we always selected with previous reports (Warren et al., 2001; Martınez-Vilalta
healthy-looking mature trees with sun-exposed branches accord- et al., 2004, 2009; Gebrekirstos et al., 2011). Interestingly,
ing to standard trait sampling protocols (Perez-Harguindeguy species hydraulic efficiency (KS, KL) did not vary consistently
et al., 2013). These factors, however, would also affect total trait along the water availability gradient. Overall, our results across
variation and it remains unclear what their impact would be on species suggest that increasing tolerance to hydraulic dysfunction
the percentage contribution of ITV. in drier sites implies increasing C costs per unit leaf area in terms
of leaf and sapwood construction.
Importantly, trait–environment relationships were scale-
Trait responses along a water availability gradient
dependent (Anderegg et al., 2018) and, as hypothesized, rela-
In agreement with findings reported in other studies (Vila-Cabrera tionships within species were generally weaker than across
et al., 2015; Anderegg et al., 2018), trait–environment relationships species. Hv and Ptlp, two of the three traits with higher per-
were not very tight, suggesting that unaccounted species-specific dif- centage ITV, were the only traits that responded to P : PET
ferences and/or other plot variables not included in our study were within species. These two intraspecific relationships had the
stronger drivers of trait variability. Mean DBH was the strongest same sign but shallower slopes than the corresponding relation-
determinant of trait variation. Specifically, plots with larger trees on ships with P : PET among species, which probably reflects
average tended to have lower LMA, WD, Hv and KL, in line with lower capacity for hydraulic adjustment within than among
previous findings (Laforest-Lapointe et al., 2014; Gleason et al., species as a result of relatively fixed drought response strategies
2018). The effect of P : PET, our target environmental factor, was at the species level. This result highlights the importance of
significant or marginally significant for d13C, Hv, Ptlp and P50, but Hv and Ptlp in shaping plastic responses along water availabil-
not for LMA, N or WD when controlling for the effect of other ity gradients. Lower leaf area per unit of sapwood (which
environmental factors. This indicates that hydraulic and water- reduced water demand) and osmotic adjustment may be
related traits responded more strongly to water availability than did needed to balance water and C costs under reduced water
LES or other stem traits, as hypothesized. availability in the context of relatively constant hydraulic safety
When we assessed the overall response of each trait to P : PET, thresholds within species, measured here as stem P50. This is
without accounting for the effect of other environmental vari- consistent with the view that P50 is an (evolutionarily) canal-
ables that covaried along the environmental gradient studied, a ized trait buffered against genetic and environmental variation
higher proportion of trait variance was explained, because species (Lamy et al., 2014). Overall, adjustments along the water avail-
means were explicitly included in the model (Table S4). In this ability gradient in the six species studied rely more on changes
broader assessment, LMA and N were also related to water avail- in stomata closure and resource allocation between sapwood
ability, besides the hydraulic/water relations variables identified and leaf area than changes in hydraulic safety and efficiency,
in the previous analysis. Wetter sites were associated with species consistent with previous results comparing pine populations
with leaf traits related to acquisitive resource strategies (low LMA (Mencuccini & Grace, 1995; Mencuccini & Bonosi, 2001;
and high N). Several studies have shown that LMA tends to be Poyatos et al., 2007; Martınez-Vilalta et al., 2009).

New
does not allow associations resulting from fundamental con-

Trait correlation networks
straints to be disentangled from those arising from indirect rela-
To our knowledge, our study is the first attempt to test tionships through third variables (in our case driven by changing
simultaneously the covariation between traits related to leaf water availability), which should constitute a priority for future
economics (LMA, N), xylem hydraulics in terms of safety research.
and efficiency (P50, KS), allocation (KL, HV) and traits related
to leaf gas exchange (d13C, Ptlp), at both the interspecific and
intraspecific levels. We found weak evidence for the existence Conclusion
of a unique coordination between hydraulics and more stan- Our study shows that plant adjustment along a water avail-
dard leaf and stem traits, which would be required for the ability gradient involves many different suites of traits, and
existence of universal, resource use strategies at the whole- highlights the importance of ITV for understanding the
plant level (our last hypothesis, cf. Reich (2014)). In our capacity of plants to buffer against environmental changes.
study, species with conservative leaf economic strategies (i.e. The availability of individual/plot-level trait data coupled with
higher LMA) presented a safer xylem (lower P50), possibly to environmental and site information will allow more accurate
support longer leaf life spans (Wright et al., 2004). However, model parameterization and, therefore, better predictions of
this interpretation should also consider that species occupying species’ responses to global change (Moran et al., 2016). We
drier sites are also likely to be exposed to lower water poten- show that, within species, plant adjustments along a water
tials, which would affect their hydraulic safety margins and availability gradient rely more on changes in allocation (Hv)
possibly result in higher hydraulic risk in drier locations. On and leaf tolerance to low water potentials (Ptlp) than on
the other hand, although higher LMA species also showed changes in xylem safety or efficiency. Finally, we show that
higher Hv, this pattern did not result in any relationship with the use of trait networks could accommodate the intricate,
xylem transport efficiency (either KS or KL). This lack of a multivariate relationships shaping plant strategies to a much
universal ‘fast–slow’ whole-plant economics spectrum is rein- greater degree than approaches based on bivariate relationships
forced when we assess trait covariation at the intraspecific (Poorter et al., 2014; Messier et al., 2017). Scale-dependent
level. We provide evidence that rather than a single dominant trait covariation networks can provide powerful insights when
axis of ‘fast–slow’ plant economics spectrum, multiple combi- assessing the architecture of plant plasticity and its limits
nations of traits are possible depending on the species and under changing environmental conditions.
the environment. Caution is thus needed when interpreting
the comparatively simple trait covariation structures revealed
in global studies using relatively few traits (Dıaz et al., 2016), Acknowledgements
and comprehensive assessments including wider sets of traits
may improve our ability to represent the patterns underlying We thank Carles Batlles and Ingrid Regalado for their valuable
the huge diversity in plant form and function. field work and N uria Serra, Pau Agost and Julieta V. for field and
The increase in water-use efficiency (estimated from d13C) laboratory assistance. This research was supported by the Spanish
with increasing LMA was the only correlation present in all Ministry of Economy and Competitiveness (MINECO) via com-
studied trait networks. This relationship is commonly petitive grant CGL2013-46808-R (FUN2FUNproject). TR was
reported (K€orner et al., 1991; Hultine & Marshall, 2000) and supported by a FPI scholarship from the MINECO. JMV bene-
it is probably a result of an increase in length in the internal fited from an ICREA Academia award.
diffusion pathway from the stomata to the chloroplasts reduc-
ing carbon dioxide supply at the site of carboxylation (Evans Author contributions
et al., 1986). We did not find support for a tradeoff between
hydraulic safety and efficiency across species and only in two JM-V, MM and TR planned and designed the research; TR and
cases within species, consistent with a recent global synthesis JB performed the measurements with contributions from HC,
that found that many species presented low safety and low MM and JM-V; TR analysed the data; and TR wrote the first
efficiency (Gleason et al., 2015). At the intraspecific level, of draft with contributions from JM-V, MM, SS-M, JB and HC.
the two traits that responded to water availability at the
intraspecific level, Hv was typically loosely linked to the rest ORCID
of the trait network (except in Q. ilex), whereas Ptlp retained
a more central role. Higher leaf tolerance to low water poten- Josep Barba https://orcid.org/0000-0001-9094-9516
tials (more negative Ptlp) was associated with higher water-use Herve Cochard https://orcid.org/0000-0002-2727-7072
efficiency (less negative d13C) and higher leaf construction Jordi Martınez-Vilalta https://orcid.org/0000-0002-2332-
costs (higher LMA) in most species, suggesting an adaptation 7298
to drier and hotter conditions (Wright et al., 2005). It should Maurizio Mencuccini https://orcid.org/0000-0003-0840-
be noted, however, that our results on trait coordination 1477
across species should be considered with caution, as only six Teresa Rosas https://orcid.org/0000-0002-8734-9752
species were measured. In addition, our experimental design Sandra Saura-Mas https://orcid.org/0000-0001-8539-427X

New
embolism – bench dehydration. Prometheus Wiki. [WWW document] URL

References http://prometheuswiki.publish.csiro.au [accessed 22 September 2015]
Adler PB, Salguero-Gomez R, Compagnoni A, Hsu JS, Ray-mukherjee J. 2014. Choat B, Jansen S, Brodribb TJ, Cochard H, Delzon S, Bhaskar R, Bucci SJ,
Functional traits explain variation in plant life history strategies. Proceedings of Feild TS, Gleason SM, Hacke UG, et al. 2012. Global convergence in the
the National Academy of Sciences, USA 111: 740–745. vulnerability of forests to drought. Nature 491: 752–755.
Albert CH, de Bello F, Boulangeat I, Pellet G, Lavorel S, Thuiller W. 2012. On Cochard H, Badel E, Herbette S, Delzon S, Choat B, Jansen S. 2013. Methods
the importance of intraspecific variability for the quantification of functional for measuring plant vulnerability to cavitation: a critical review. Journal of
diversity. Oikos 121: 116–126. Experimental Botany 64: 4779–4791.
Allen CD, Breshears DD, McDowell NG. 2015. On underestimation of global Csardi G, Nepusz T. 2006. The igraph software package for complex network
vulnerability to tree mortality and forest die-off from hotter drought in the research. InterJournal Complex Sy: 1695.
Anthropocene. Ecosphere 6: 1–55. Cunningham SA, Summerhayes B, Westoby M. 1999. Evolutionary divergences
Anderegg LDL, Berner LT, Badgley G, Sethi ML, Law BE, HilleRisLambers J. in leaf structure and chemistry, comparing rainfall and soil nutrient gradients.
2018. Within-species patterns challenge our understanding of the leaf Ecology 69: 569–588.
economics spectrum. Ecology Letters 21: 734–744. DeLucia EH, Maherali H, Carey EV. 2000. Climate-driven changes in biomass
Anderegg WRL. 2015. Spatial and temporal variation in plant hydraulic traits allocation in pines. Global Change Biology 6: 587–593.
and their relevance for climate change impacts on vegetation. New Phytologist Dıaz S, Kattge J, Cornelissen JHC, Wright IJ, Lavorel S, Dray S, Reu B, Kleyer
205: 1008–1014. M, Wirth C, Prentice IC, et al. 2016. The global spectrum of plant form and
Anderegg WRL, Klein T, Bartlett M, Sack L, Pellegrini AFA, Choat B, Jansen function. Nature 529: 167–171.
S. 2016. Meta-analysis reveals that hydraulic traits explain cross-species Evans JR. 1989. Photosynthesis and nitrogen relationships in leaves of C3 plants.
patterns of drought-induced tree mortality across the globe. Proceedings of the Oecologia 78: 9–19.
National Academy of Sciences, USA 113: 5024–5029. Evans J, Sharkey T, Berry J, Farquhar G. 1986. Carbon isotope discrimination
Bartlett MK, Klein T, Jansen S, Choat B, Sack L. 2016. The correlations and measured concurrently with gas exchange to investigate CO2 diffusion in leaves
sequence of plant stomatal, hydraulic, and wilting responses to drought. of higher plants. Australian Journal of Plant Physiology 13: 281–292.
Proceedings of the National Academy of Sciences, USA 113: 13098–13103. Ewers FW, Fisher JB. 1989. Techniques for measuring vessel lengths and
Bartlett MK, Scoffoni C, Sack L. 2012. The determinants of leaf turgor loss diameters in stems of woody plants. American Journal of Botany 76: 645–656.
point and prediction of drought tolerance of species and biomes: a global meta- Farquhar GD, Hubick KT, Condon AG, Richards RA. 1989. Carbon isotope
analysis. Ecology Letters 15: 393–405. fractionation and plant water-use efficiency. In: Rundel PW, Ehleringer JR,
Bates D, M€a chler M, Bolker B, Walker S. 2015. Fitting linear mixed-effects Nagy KA, eds. Stable isotopes in ecological research. New York, NY, USA:
models using {lme4}. Journal of Statistical Software 67: 1–48. Springer, 21–40.
Becker P, Tyree MT, Tsuda M. 1999. Hydraulic conductances of angiosperms Funk JL, Larson JE, Ames GM, Butterfield BJ, Cavender-Bares J, Firn J,
versus conifers: similar transport sufficiency at the whole-plant level. Tree Laughlin DC, Sutton-Grier AE, Williams L, Wright J. 2017. Revisiting the
Physiology 19: 445–452. Holy Grail: using plant functional traits to understand ecological processes.
Blackman CJ, Gleason SM, Chang Y, Cook AM, Laws C, Westoby M. 2014. Biological Reviews 92: 1156–1173.
Leaf hydraulic vulnerability to drought is linked to site water availability across Gardner WR. 1986. Water content. In: Klute A, ed. Methods of soil analysis: Part
a broad range of species and climates. Annals of Botany 114: 435–440. 1. Physical and mineralogical methods. Madison: American Society of
Blonder B, Violle C, Bentley LP, Enquist BJ. 2011. Venation networks and the Agronomy, Soil Science Society of Americ, 63–73.
origin of the leaf economics spectrum. Ecology Letters 14: 91–100. Gebrekirstos A, van Noordwijk M, Neufeldt H, Mitl€ohner R. 2011.
Bradshaw AD. 1965. Evolutionary significance of phenotypic plasticity in plants. Relationships of stable carbon isotopes, plant water potential and growth: an
Advances in Genetics 13: 115–155. approach to asses water use efficiency and growth strategies of dry land
Bradshaw AD. 2006. Unravelling phenotypic plasticity – why should we bother? agroforestry species. Trees – Structure and Function 25: 95–102.
New Phytologist 170: 644–648. Gee G, Or D. 2002. Particle size analysis. In: Dane J, Topp G, eds. Methods of
Brendel O. 2014. Is the coefficient of variation a valid measure for variability of soil analysis. Part 4. Physical methods. Madison, WI, USA: Soil Science Society
stable isotope abundances in biological materials? Rapid Communications in of America, 383–411.
Mass Spectrometry 28: 370–376. Gibert A, Gray EF, Westoby M, Wright IJ, Falster DS. 2016. On the link
Brodribb TJ. 2017. Progressing from ‘functional’ to mechanistic traits. New between functional traits and growth rate: meta-analysis shows effects change
Phytologist 215: 9–11. with plant size, as predicted. Journal of Ecology 104: 1488–1503.
Brodribb TJ, Feild TS, Jordan GJ. 2007. Leaf maximum photosynthetic rate and Gleason SM, Stephens AEA, Tozer WC, Blackman CJ, Butler DW, Chang Y,
venation are linked by hydraulics. Plant Physiology 144: 1890–1898. Cook AM, Cooke J, Laws CA, Rosell JA, et al. 2018. Shoot growth of woody
Brodribb TJ, Holbrook NM, Edwards EJ, Gutierrez MV. 2003. Relations trees and shrubs is predicted by maximum plant height and associated traits.
between stomatal closure, leaf turgor and xylem vulnerability in eight tropical Functional Ecology 32: 247–259.
dry forest trees. Plant, Cell & Environment 26: 443–450. Gleason SM, Westoby M, Jansen S, Choat B, Hacke UG, Pratt RB, Bhaskar R,
Canchola JA, Tang S, Hemyari P, Paxinos E, Marins E. 2017. Correct use of Brodribb TJ, Bucci SJ, Cao K, et al. 2015. Weak tradeoff between xylem safety
percent coefficient of variation (% CV) formula for log-transformed data. MOJ and xylem- specific hydraulic efficiency across the world’s woody plant species.
Proteomics & Bioinformatics 6: 00200. New Phytologist 209: 123–136.
Carnicer J, Barbeta A, Sperlich D, Coll M, Pe~ nuelas J. 2013. Contrasting trait Gracia C, Burriel JA, Iba n ~ez JJ, Mata T, Vayreda J. 2004. Inventari ecolo gic i
syndromes in angiosperms and conifers are associated with different responses forestal de Catalunya. Catalunya, 10 volumes. Bellaterra, Spain: CREAF.
of tree growth to temperature on a large scale. Frontiers in Plant Science 4: Hajek P, Kurjak D, von W€ uhlisch G, Delzon S, Schuldt B. 2016. Intraspecific
1–19. variation in wood anatomical, hydraulic, and foliar traits in ten european beech
Chave J. 2013. The problem of pattern and scale in ecology: what have we provenances differing in growth yield. Frontiers in Plant Science 7: 1–14.
learned in 20 years? Ecology Letters 16: 4–16. Hargreaves GH, Samani ZA. 1982. Estimating potential evapotranspiration.
Chave J, Coomes D, Jansen S, Lewis SL, Swenson NG, Zanne AE. 2009. Journal of the Irrigation & Drainage Division – ASCE 108: 225–230.
Towards a worldwide wood economics spectrum. Ecology Letters 12: 351–366. Harper AB, Cox PM, Friedlingstein P, Wiltshire AJ, Jones CD, Sitch S,
Choat B, Brodribb TJ, Brodersen CR, Duursma RA, Lopez R, Medlyn BE. Mercado LM, Groenendijk M, Robertson E, Kattge J, et al. 2016. Improved
2018. Triggers of tree mortality under drought. Nature 558: 531–539. representation of plant functional types and physiology in the Joint UK Land
Choat B, Creek AD, Lo Gullo MA, Nardini A, Oddo E, Raimondo F, Torres- Environment Simulator (JULES v4.2) using plant trait information.
Ruiz JM, Triiflo P, Vilagrosa A. 2015. Quantification of vulnerability to xylem Geoscientific Model Development 9: 2415–2440.

New
Hultine KR, Marshall JD. 2000. Altitude trends in conifer leaf morphology and McGill BJ, Enquist BJ, Weiher E, Westoby M. 2006. Rebuilding community
stable carbon isotope composition. Oecologia 123: 32–40. ecology from functional traits. Trends in Ecology and Evolution 21: 178–185.
Jacobsen AL, Pratt RB, Davis SD, Ewers FW. 2007. Cavitation resistance and Mencuccini M, Bonosi L. 2001. Leaf/sapwood area ratios in Scots pine show
seasonal hydraulics differ among three arid Californian plant communities. acclimation across Europe. Canadian Journal of Forest Research 31: 442–456.
Plant, Cell and Environment 30: 1599–1609. Mencuccini M, Grace J. 1995. Climate influences the leaf area/sapwood area
Keeney D, Nelson D. 1982. Nitrogen-inorganic forms. In: Page A, ed. ratio in Scots pine. Tree Physiology 15: 1–10.
Methods of soil analysis. Part 2: Chemical and microbiological properties. Mencuccini M, Minunno F, Salmon Y, Martınez-Vilalta J, H€oltt€a T. 2015.
Madison, WI, USA: American Society of Agronomy, Soil Science Society Coordination of physiological traits involved in drought-induced mortality of
of America, 643–698. woody plants. New Phytologist 208: 396–409.
Killingbeck KT, Whitford WG. 1996. High foliar nitrogen in desert shrubs: Mendez-Alonzo R, Paz H, Cruz RZ, Rosell JA, Olson ME. 2012. Coordinated
an important ecosystem trait or defective desert doctrine? Ecology 77: 1728– evolution of leaf and stem economics in tropical dry forest trees. Ecology 93:
1737. 2397–2406.
K€orner C, Farquhar GD, Wong SC. 1991. Carbon isotope discrimination by Messier J, Lechowicz MJ, Mcgill BJ, Violle C, Enquist BJ. 2017. Interspecific
plants follows latitudinal and altitudinal trends. Oecologia 88: 30–40. integration of trait dimensions at local scales: the plant phenotype as an
Laforest-Lapointe I, Martınez-Vilalta J, Retana J. 2014. Intraspecific variability integrated network. Journal of Ecology 105: 1775–1790.
in functional traits matters: case study of Scots pine. Oecologia 175: 1337– Messier J, Mcgill BJ, Enquist BJ, Lechowicz MJ. 2016. Trait variation and
1348. integration across scales: is the leaf economic spectrum present at local scales?
Lamy JB, Bouffier L, Burlett R, Plomion C, Cochard H, Delzon S. 2011. Ecography 40: 685–697.
Uniform selection as a primary force reducing population genetic Messier J, McGill BJ, Lechowicz MJ. 2010. How do traits vary across ecological
differentiation of cavitation resistance across a species range. PLoS ONE 6: scales? A case for trait-based ecology. Ecology Letters 13: 838–848.
e23476. Moran EV, Hartig F, Bell DM. 2016. Intraspecific trait variation across scales:
Lamy JB, Delzon S, Bouche PS, Alia R, Vendramin GG, Cochard H, Plomion implications for understanding global change responses. Global Change Biology
C. 2014. Limited genetic variability and phenotypic plasticity detected for 22: 137–150.
cavitation resistance in a Mediterranean pine. New Phytologist 201: 874–886. Nelson DW, Sommers LE. 1982. Total carbon, organic carbon, and organic
Laughlin DC. 2014. The intrinsic dimensionality of plant traits and its relevance matter. In: Page A, ed. Methods of soil analysis. Part 2. Chemical and
to community assembly. Journal of Ecology 102: 186–193. microbiological properties. Madison, WI, USA: American Society of Agronomy,
Lavorel S, Garnier E. 2002. Predicting changes in community composition and Soil Science Society of America, 961–1010.
ecosystem functioning from plant traits: revisiting the Holy Grail. Functional Niinemets U. € 2015. Is there a species spectrum within the world-wide leaf
Ecology 16: 545–556. economics spectrum? Major variations in leaf functional traits in the
Lenz TI, Wright IJ, Westoby M. 2006. Interrelations among pressure-volume Mediterranean sclerophyll Quercus ilex. New Phytologist 205: 79–96.
curve traits across species and water availability gradients. Physiologia Plantarum Ninyerola M, Pons X, Roure J. 2005. Atlas clima tico digital de la Penı nsula
127: 423–433. Ibe rica. Metodologı as y aplicaciones en bioclimatologıa y geobota nica. Metodologı a
Levin SA. 1992. The problem of pattern and scale in ecology. Ecology 73: 1943– y aplicaciones en bioclimatologı a y geobota nica. Bellaterra, Spain: Universidad
1967. Autonoma de Barcelona.
Li L, McCormack ML, Ma C, Kong D, Zhang Q, Chen X, Zeng H, Niinemets Olsen SR, Sommers LE. 1982. Phosphorous. In: Page A, ed. Methods of soil
€ Guo D. 2015. Leaf economics and hydraulic traits are decoupled in five
U, analysis. Part 2. Chemical and microbiological properties. Madison, WI, USA:
species-rich tropical-subtropical forests. Ecology Letters 18: 899–906. American Society of Agronomy, Soil Science Society of America, 403–430.
Li X, Blackman CJ, Choat B, Duursma RA, Rymer PD, Medlyn BE, Tissue Paine CET, Amissah L, Auge H, Baraloto C, Baruffol M, Bourland N,
DT. 2018. Tree hydraulic traits are co-ordinated and strongly linked to Bruelheide H, Da€ınou K, de Gouvenain RC, Doucet JL, et al. 2015. Globally,
climate-of-origin across a rainfall gradient. Plant, Cell & Environment 41: 646– functional traits are weak predictors of juvenile tree growth, and we do not
660. know why. Journal of Ecology 103: 978–989.
Lopez R, Cano FJ, Choat B, Cochard H, Gil L. 2016. Plasticity in vulnerability Pammenter NW, Willigen CV. 1998. A mathematical and statistical analysis of
to cavitation of pinus canariensis occurs only at the driest end of an aridity the curves illustrating vulnerability of xylem to cavitation. Tree Physiology 18:
gradient. Frontiers in Plant Science 7: 769. 589–593.
Maherali H, DeLucia EH. 2000. Xylem conductivity and vulnerability to Perez-Harguindeguy N, Dıaz S, Lavorel S, Poorter H, Jaureguiberry P, Bret-
cavitation of ponderosa pine growing in contrasting climates. Tree Physiology Harte MS, Cornwell WK, Craine JM, Gurvich DE, Urcelay C, et al. 2013.
20: 859–867. New Handbook for standardized measurment of plant functional traits
Maherali H, Pockman WT, Jackson RB. 2004. Adaptive variation in the worldwide. Australian Journal of Botany 23: 167–234.
vulnerability of woody plants to xylem cavitation. Ecology 85: 2184–2199. Pinheiro J, Bates D, DebRoy S, Sarkar D, R Core Team. 2018. {nlme}: Linear
Markesteijn L, Poorter L, Bongers F, Paz H, Sack L. 2011. Hydraulics and life and Nonlinear Mixed Effects Models. R package v. 3.1-137, [WWW document]
history of tropical dry forest tree species: coordination of species’ drought and URL https://CRAN.R-project.org/package=nlme. Vienna, Austria: R
shade tolerance. New Phytologist 191: 480–495. Foundation for Statistical Computing.
Marks CO. 2007. The causes of variation in tree seedling traits: the roles of Poorter H, Lambers H, Evans JR. 2014. Trait correlation networks – a whole-
environmental selection versus chance. Evolution 61: 455–469. plant perspective on the recently criticized leaf economic spectrum. New
Martınez-Vilalta J, Cochard H, Mencuccini M, Sterck F, Herrero A, Korhonen Phytologist 201: 378–382.
JFJ, Llorens P, Nikinmaa E, Nole A, Poyatos R, et al. 2009. Hydraulic Poorter H, Niinemets U, € Poorter L, Wright IJ, Villar R, Niinemets U, Poorter
adjustment of Scots pine across Europe. New Phytologist 184: 353–364. L, Wright IJ, Villar R. 2009. Causes and consequences of variation in leaf mass

Martınez-Vilalta J, Mencuccini M, Alvarez X, Camacho J, Loepfe L, Pi~ nol J, per area (LMA): a meta-analysis. New Phytologist 182: 565–588.
Cochard H, Sterck F, Herrero A, Korhonen JFJ, et al. 2012. Spatial Poorter L, Wright SJ, Paz H, Ackerly DD, Condit R, Ibarra-Manrıquez G,
distribution and packing of xylem conduits. American Journal of Botany 99: Harms KE, Licona JC, Martınez-Ramos M, Mazer SJ, et al. 2008. Are
1189–1196. functional traits good predictors of demographic rates? Evidence from five
Martınez-Vilalta J, Sala A, Pi~nol J. 2004. The hydraulic architecture of Pinaceae neotropical forests. Ecology 89: 1908–1920.
–a review. Plant Ecology 171: 3–13. Poyatos R, Martınez-Vilalta J, Cerm a k J, Ceulemans R, Granier A, Irvine J,
Martin-StPaul NK, Longepierre D, Huc R, Delzon S, Burlett R, Joffre R, K€ostner B, Lagergren F, Meiresonne L, Nadezhdina N, et al. 2007. Plasticity
Rambal S, Cochard H. 2014. How reliable are methods to assess xylem in hydraulic architecture of Scots pine across Eurasia. Oecologia 153: 245–259.
vulnerability to cavitation? The issue of ‘open vessel’ artifact in oaks. Tree R Core Team. 2017. R: a language and environment for statistical computing.
Physiology 34: 894–905. Vienna, Austria: R Foundation for Statistical Computing.

New
Reich PB. 2014. The world-wide ‘fast-slow’ plant economics spectrum: a traits Westoby M, Falster DS, Moles AT, Vesk PA, Wright IJ. 2002. Plant ecological
manifesto. Journal of Ecology 102: 275–301. strategies: some leading dimensions of variation between species. Annual
Reich PB, Rich RL, Wang Y, Lu X, Oleksyn J. 2014. Biogeographic variation in Review of Ecology and Systematics 33: 125–159.
evergreen conifer needle longevity and impacts on boreal forest carbon cycle Westoby M, Wright IJ. 2006. Land-plant ecology on the basis of functional
projections. Proceedings of the National Academy of Sciences, USA 111: 13703– traits. Trends in Ecology and Evolution 21: 261–268.
13708. Wheeler JK, Huggett BA, Tofte AN, Rockwell FE, Holbrook NM. 2013.
Reich PB, Walters MB, Ellsworth DS. 1997. From tropics to tundra: global Cutting xylem under tension or supersaturated with gas can generate PLC and
convergence in plant functioning. Proceedings of the National Academy of the appearance of rapid recovery from embolism. Plant, Cell & Environment
Sciences, USA 94: 13730–13734. 36: 1938–1949.
Rowland L, da Costa ACL, Galbraith DR, Oliveira RS, Binks OJ, Oliveira Wortemann R, Herbette S, Barigah TS, Fumanal B, Alia R, Ducousso A,
AAR, Pullen AM, Doughty CE, Metcalfe DB, Vasconcelos SS, et al. 2015. Gomory D, Roeckel-Drevet P, Cochard H. 2011. Genotypic variability and
Death from drought in tropical forests is triggered by hydraulics not carbon phenotypic plasticity of cavitation resistance in Fagus sylvatica L. across Europe.
starvation. Nature 528: 119–122. Tree Physiology 31: 1175–1182.
Sack L, Scoffoni C, John GP, Poorter H, Mason CM, Mendez-Alonzo R, Wright SJ, Kitajima K, Kraft NJB, Reich PB, Wright IJ, Bunker DE,
Donovan LA. 2013. How do leaf veins influence the worldwide leaf economic Condit R, Dalling JW, Davies SJ, Dıaz S, et al. 2010. Functional traits
spectrum? Review and synthesis. Journal of Experimental Botany 64: and the growth—mortality trade-off in tropical trees. Ecology 91:
4053–4080. 3664–3674.
Saxton KE, Rawls WJ, Romberger JS, Papendick RI. 1986. Estimating Wright IJ, Reich PB, Cornelissen JHC, Falster DS, Groom PK, Hikosaka K,
generalized soil-water characteristics from texture. Soil Science Society of America Lee W, Lusk CH, Niinemets U, € Oleksyn J, et al. 2005. Modulation of leaf
Journal 50: 1031. economic traits and trait relationships by climate. Global Ecology and
Schulze E, Williams R, Farquhar G, Schulze W, Langridge J, Miller JM, Walker Biogeography 14: 411–421.
BH. 1998. Carbon and nitrogen isotope discrimination and nitrogen nutrition Wright IJ, Reich PB, Westoby M, Ackerly DD, Baruch Z, Bongers F,
of trees along a rainfall gradient in northern Australia. Australian Journal of Cavender-Bares J, Chapin T, Cornellssen JHC, Diemer M, et al. 2004. The
Plant Physiology 25: 413–425. worldwide leaf economics spectrum. Nature 428: 821–827.
Shipley B, De Bello F, Cornelissen JHC, Laliberte E, Laughlin DC, Reich PB. Wright JP, Sutton-Grier A. 2012. Does the leaf economic spectrum hold within
2016. Reinforcing loose foundation stones in trait-based plant ecology. local species pools across varying environmental conditions? Functional Ecology
Oecologia 180: 923–931. 26: 1390–1398.
Siefert A, Violle C, Chalmandrier L, Albert CH, Taudiere A, Fajardo A, Aarssen Wright IJ, Westoby M. 2002. Leaves at low versus high rainfall: coordination of
LW, Baraloto C, Carlucci MB, Cianciaruso MV, et al. 2015. A global meta- structure, lifespan and physiology. New Phytologist 155: 403–416.
analysis of the relative extent of intraspecific trait variation in plant Yang J, Cao M, Swenson NG. 2018. Why functional traits do not predict tree
communities. Ecology Letters 18: 1406–1419. demographic rates. Trends in Ecology & Evolution 33: 1–11.
Skelton RP, Dawson TE, Thompson SE, Shen Y, Weitz AP, Ackerly D. 2018. Zuur AF, Ieno EN, Walker NJ, Saveliev AA, Smith G. 2009. Mixed effects
Low vulnerability to xylem embolism in leaves and stems of North American models and extensions in ecology with R. New York, NY, USA: Springer
oaks. Plant Physiology 177: 1066–1077. Science.
Skelton RP, West AG, Dawson TE. 2015. Predicting plant vulnerability to
drought in biodiverse regions using functional traits. Proceedings of the National
Academy of Sciences, USA 112: 5744–5749.
Sperry JS, Hacke UG, Pittermann J. 2006. Size and function in conifer tracheids
Supporting Information
and angiosperm vessels. American Journal of Botany 93: 1490–1500. Additional Supporting Information may be found online in the
Sperry JS, Love DM. 2015. What plant hydraulics can tell us about responses to
climate-change droughts. New Phytologist 207: 14–27.
Supporting Information tab for this article:
Sperry JS, Meinzer FC, McCulloh KA. 2008. Safety and efficiency conflicts in
hydraulic architecture: scaling from tissues to trees. Plant, Cell and Environment Fig. S1 Map of study area showing the distribution of sampling
31: 632–645. plots along the water availability gradient.
Sperry JS, Tyree MT. 1988. Mechanism of water stress-induced. Plant Physiology
88: 581–587.
Trueba S, Pouteau R, Lens F, Feild TS, Isnard S, Olson ME, Delzon S. 2017.
Fig. S2 Distribution of the precipitation to potential evapotran-
Vulnerability to xylem embolism as a major correlate of the environmental spiration ratio (P : PET) plot values for each sampled species.
distribution of rain forest species on a tropical island. Plant, Cell &
Environment 40: 277–289. Fig. S3 Principal component analysis (PCA) conducted to sum-
Tyree MT, Davis SD, Cochard H. 1994. Biophysical perspectives of xylem marize environmental variables at the plot level (climate, forest
evolution: is there a tradeoff of hydraulic efficiency for vulnerability to
dysfunction? IAWA Journal 15: 335–360.
structure and soil characteristics).
Van Bodegom PM, Douma JC, Witte JPM, Ordo~ nez JC, Bartholomeus
RP, Aerts R. 2012. Going beyond limitations of plant functional types Fig. S4 Distribution of each trait as a function of species and
when predicting global ecosystem-atmosphere fluxes: exploring the family.
merits of traits-based approaches. Global Ecology and Biogeography 21:
625–636.
Vila -Cabrera A, Martınez-Vilalta J, Retana J. 2015. Functional trait variation
Fig. S5 Variance partitioning within Pinaceae and Fagaceae
along environmental gradients in temperate and Mediterranean trees. Global families.
Ecology and Biogeography 24: 1377–1389.
Violle C, Enquist BJ, McGill BJ, Jiang L, Albert CH, Hulshof C, Jung V, Table S1 Characterization of the dominant species in the study
Messier J. 2012. The return of the variance: intraspecific variability in plots.
community ecology. Trends in Ecology and Evolution 27: 244–252.
Warren CR, McGrath JF, Adams MA. 2001. Water availability and carbon
isotope discrimination in conifers. Oecologia 127: 476–486. Table S2 Statistical descriptors of studied traits.

New
Table S3 Results of the linear mixed models examining the rela- Table S5 Descriptors of trait networks across species and for each
tionships between environmental variables using the PCA axis studied species separately.
and traits.
Please note: Wiley Blackwell are not responsible for the content
Table S4 Results of the linear mixed models examining the effect or functionality of any Supporting Information supplied by the
of precipitation to potential evapotranspiration ratio (P : PET) authors. Any queries (other than missing material) should be
on each trait within and among species. directed to the New Phytologist Central Office.
New Phytologist is an electronic (online-only) journal owned by the New Phytologist Trust, a not-for-profit organization dedicated
to the promotion of plant science, facilitating projects from symposia to free access for our Tansley reviews and Tansley insights.
Regular papers, Letters, Research reviews, Rapid reports and both Modelling/Theory and Methods papers are encouraged.
We are committed to rapid processing, from online submission through to publication ‘as ready’ via Early View – our average time
to decision is <26 days. There are no page or colour charges and a PDF version will be provided for each article.
The journal is available online at Wiley Online Library. Visit www.newphytologist.com to search the articles and register for table
of contents email alerts.
If you have any questions, do get in touch with Central Office (np-centraloffice@lancaster.ac.uk) or, if it is more convenient,
our USA Office (np-usaoffice@lancaster.ac.uk)
For submission instructions, subscription and all the latest information visit www.newphytologist.com
See also the Commentary on this article by Skelton et al. 223: 505–507.


New Phytologist - 2019 - Rosas - Adjustments and Coordination of Hydraulic Leaf and Stem Traits Along A Water Availability

Uploaded by

Copyright:

Available Formats

New Phytologist - 2019 - Rosas - Adjustments and Coordination of Hydraulic Leaf and Stem Traits Along A Water Availability

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

New Phytologist - 2019 - Rosas - Adjustments and Coordination of Hydraulic Leaf and Stem Traits Along A Water Availability

Uploaded by

Copyright:

Available Formats

Research

Adjustments and coordination of hydraulic, leaf and stem traits

Key words: drought, Huber value, hydraulic

investments (Reich et al., 1997; Wright et al., 2004). However,

632 New Phytologist (2019) 223: 632–646 Ó 2019 The Authors

Ó 2019 The Authors New Phytologist (2019) 223: 632–646

New Phytologist (2019) 223: 632–646 Ó 2019 The Authors

Ó 2019 The Authors New Phytologist (2019) 223: 632–646

New Phytologist (2019) 223: 632–646 Ó 2019 The Authors

Fig. 1 (a) Quartile coefficient of dispersion of

Table 1 Traits measured in this study. Trait correlation networks

Ó 2019 The Authors New Phytologist (2019) 223: 632–646

New Phytologist (2019) 223: 632–646

New Phytologist Ó 2019 New Phytologist Trust

(a) (b) (c)

(d) (e) (f)

(g) (h) (i)

especially in more integrative traits (KL, Hv). Most study traits

Ó 2019 The Authors New Phytologist (2019) 223: 632–646

(b) (c) (d)

Fig. 3 (a–g) Trait correlation networks across

New Phytologist (2019) 223: 632–646 Ó 2019 The Authors

Ó 2019 The Authors New Phytologist (2019) 223: 632–646

does not allow associations resulting from fundamental con-

New Phytologist (2019) 223: 632–646 Ó 2019 The Authors

embolism – bench dehydration. Prometheus Wiki. [WWW document] URL

Ó 2019 The Authors New Phytologist (2019) 223: 632–646

New Phytologist (2019) 223: 632–646 Ó 2019 The Authors

Ó 2019 The Authors New Phytologist (2019) 223: 632–646

New Phytologist (2019) 223: 632–646 Ó 2019 The Authors

You might also like