141
Evidence that branches of evergreen angiosperm
and coniferous trees differ in hydraulic
conductance but not in Huber values
Christopher H. Lusk, Mylthon Jiménez-Castillo, and Nicolás Salazar-Ortega
Abstract: The hydraulic efficiency conferred by vessels is regarded as one of the key innovations explaining the historical
rise of the angiosperms at the expense of the gymnosperms. Few studies, however, have compared the structure and function of xylem and their relationships with foliage traits in evergreen representatives of both groups. We measured sapwood
cross-sectional area, conduit diameters, hydraulic conductance, and leaf area of fine branches (2.5–7.5 mm diameter) of
five conifers and eight evergreen angiosperm trees in evergreen temperate forests in south-central Chile. Conductance of
both lineages was higher at Los Lleuques, a warm temperate site with strong Mediterranean influence, than in a cool temperate rain forest at Puyehue. At a common sapwood cross-sectional area, angiosperm branches at both sites had greater
hydraulic conductance (G) than conifers, but similar leaf areas. Branch conductance normalized by subtended leaf area
(GL) at both sites was, therefore, higher in angiosperms than in conifers. Hydraulically weighted mean conduit diameters
were much larger in angiosperms than in conifers, although this difference was less marked at Puyehue, the cooler of the
two sites. Conduits of the vesselless rain forest angiosperm Drimys winteri J.R. & G. Forst were wider than those of coniferous associates, although narrower than angiosperm vessels. However, GL of D. winteri was within the range of values
measured for vesselbearing angiosperms at the same site. The observed differences in xylem structure and function correlate with evidence that evergreen angiosperms have higher average stomatal conductance and photosynthetic capacity than
their coniferous associates in southern temperate forests. Comparisons of conifers and angiosperm branches thus suggest
that the superior capacity of angiosperm conduits is attributable to the development of higher gas-exchange rates per unit
leaf area, rather than to a more extensive leaf area. Results also suggest that the tracheary elements of some vesselless angiosperms differ in width and hydraulic efficiency from conifer tracheids.
Key words: conifer–angiosperm interactions, gymnosperms, hydraulic efficiency, sapwood, tracheids, vessels, Winteraceae.
Résumé : On considère l’efficacité hydraulique conférée aux vaisseaux comme une des innovations déterminantes expliquant la montée historique des angiospermes aux dépens des gymnospermes. Cependant, peu d’études ont comparé la
structure et la fonction du xylème, ainsi que ses relations avec les caractéristiques du feuillage, chez des espèces sempervirentes des deux groupes. Les auteurs ont mesuré la superficie en section transverse de l’aubier, les diamètres des conduits,
la conductance hydraulique et la surface foliaire des fines branches (2,5–7,5 mm de diamètre), chez cinq conifères et huit
essences sempervirentes d’angiospermes, dans des forêts tempérées du centre-sud du Chili. Chez les deux groupes, la
conductance est plus importante à Los Lleuques, un site chaud tempéré avec une forte influence de type méditerranéen,
que dans une forêt ombrophile tempérée à Puyehue. Sur les deux sites, pour une même surface d’aubier en section transverse, les branches des angiospermes montrent une conductivité hydraulique (G) plus importante que les conifères, mais
des surfaces foliaires similaires. Sur les deux sites, la conductance hydraulique sous-tendue par la surface foliaire (GL) est
donc plus grande chez les angiospermes que chez les conifères. Les diamètres moyens des conduits pondérés en termes hydrauliques sont beaucoup plus grands chez les angiospermes que chez les conifères, bien que cette différence soit moins
marquée à Puyehue, le site le plus frais des deux. Les conduits d’une angiosperme ombrophile sans vaisseaux, le Drimys
winteri J.R. & G. Forst, sont plus larges que ceux des conifères associés, bien que plus étroits que les vaisseaux des angiospermes. Cependant, le GL du D. winteri se retrouve dans l’écart des valeurs mesurées chez les angiospermes munies
de vaisseaux, sur le même site. Les différences observées dans la structure et la fonction du xylème, montrent une corrélation qui prouve que les angiospermes sempervirentes possèdent une conductance stomatale et une capacité photosynthétique en moyenne plus importantes, que les conifères qui leurs sont associés, dans les forêt tempérées du sud. La
comparaison entre les branches de conifères et d’angiospermes suggère donc que la capacité supérieure des conduits des
angiospermes est associée avec le développement de taux plus élevés d’échange gazeux par unité de surface foliaire, plutôt
qu’à une surface foliaire plus étendue. Ces résultats suggèrent également que les éléments de trachée de certaines angiospermes sans vaisseau diffèrent en diamètre et en efficacité hydraulique par rapport aux trachéı̈des des conifères.
Received 11 November 2006. Published on the NRC Research Press Web site at http://canjbot.nrc.ca on 1 May 2007.
C.H. Lusk,1,2 M. Jiménez-Castillo,3 and N. Salazar-Ortega. Departamento de Botánica, Universidad de Concepción, Concepción,
Chile.
1Corresponding
author (e-mail: [email protected]).
address: Department of Biological Sciences, Macquarie University, New South Wales, NSW 2109, Australia.
3Present address: Nucleo FORECOS, Facultad de Ciencias Forestales, Universidad Austral de Chile, Casilla 567, Valdivia, Chile.
2Present
Can. J. Bot. 85: 141–147 (2007)
doi:10.1139/B07-002
2007 NRC Canada
142
Can. J. Bot. Vol. 85, 2007
Mots-clés : interaction conifères angiospermes, gymnospermes, efficacité hydraulique, aubier, trachéı̈des, vaisseaux, Winteraceae.
[Traduit par la Rédaction]
______________________________________________________________________________________
Introduction
The development of vessels is held to be one of the key
innovations underlying the historical rise of angiosperms at
the expense of gymnosperms, and the superior competitive
ability of the former in most productive habitats (Bond
1989; Brodribb et al. 2005). As water transport is proportional to the fourth power of conduit radius (Tyree and
Zimmermann 2002), the wider lumens and open ends of
vessels should make the xylem of vesselbearing angiosperms
more hydraulically efficient than that of tracheid-bearing
conifers and other gymnosperms (Tyree and Ewers 1991;
Wang et al. 1992). Because of the intimate relationship between carbon gain and water loss by leaves, this difference
is likely to enable angiosperms to develop higher maximum
assimilation and growth rates than conifers (Brodribb et al.
2005). Considering the lesser vulnerability of narrow conduits to freeze–thaw embolism (Sperry and Sullivan 1992;
Davis et al. 1999), the same differences in xylem anatomy
may also explain the continued dominance of conifers at
many high latitude and high elevation sites.
Most empirical evidence of the greater hydraulic efficiency of angiosperm xylem comes from the strongly seasonal northern temperate zone (Tyree and Ewers 1991;
Wang et al. 1992). Such comparisons confound taxonomy
with leaf habit and leaf life-span, as most northern
temperate-zone angiosperms are deciduous and most extant
conifers evergreen (Becker 2000). In strongly seasonal
continental climates, deciduous trees are able to develop
large-diameter vessels because there is minimal risk of frost
during the growing season (Davis et al. 1999). Fewer
hydraulic data are available from temperate maritime
regions where evergreen representatives of both lineages
coexist (but see Brodribb and Feild 2000). In maritime temperate climates, the evergreen habit is advantageous for
exploiting opportunities for carbon gain, as these are often
unpredictably interspersed with freezing conditions (Feild et
al. 2002). Maintaining year-round hydraulic supply to leaves
is such environments is likely to require narrower conduits,
more resistant to freeze–thaw embolism, with the result that
angiosperms’ potential hydraulic advantage may be eroded
(Brodribb and Feild 2000).
There is also uncertainty about the relationships of possible differences in hydraulic efficiency to other functional
traits relevant to carbon gain at shoot level. Brodribb and
Feild (2000) reported that greater hydraulic supply rates in
Tasmanian and New Guinean evergreen angiosperms are
linked to higher leaf-level gas-exchange rates than those of
conifers. Greater hydraulic efficiency could also conceivably
enable angiosperms to display larger leaf areas at a given
stem diameter (i.e., develop lower Huber values) than conifers (Brodribb et al. 2005). However, there are still few data
with which to test this idea.
We measured sapwood cross-sectional area, conduit diameters, hydraulic conductance, and leaf area of fine branches
of conifer and angiosperm trees at two temperate-zone forest
sites in south-central Chile. By determining how differences
in xylem anatomy and hydraulic efficiency correlate with
traits relevant to carbon gain at the shoot level, we aimed to
contribute to the understanding of mechanisms underlying
the historical rise of the angiosperms, and competitive sorting of the two groups along environmental gradients (Bond
1989; Becker 2000). We compared data from conifers and
angiosperms at two sites differing by ca. 4.5 8C in mean annual temperature, and expected that differences in conduit
diameters and hydraulic conductance would be less evident
at the colder of the two sites.
Materials and methods
Study sites
Samples were collected between December 2004 and June
2005 at two sites in south-central Chile. Los Lleuques, in
the Andean foothills above the town of Chillán, is located
at 36852’S, 71828’W, at an elevation of ca. 800 m a.s.l.
Mean annual temperature is estimated at about 13 8C, and
annual precipitation ca. 2,000 mm, with a marked summer
minimum indicative of a strong Mediterranean influence
(Almeyda and Sáez 1958). Remaining forest cover in the
area is largely second-growth, including the deciduous tree
Nothofagus obliqua (Mirb.) Blume and the evergreens
Nothofagus dombeyi (Mirb.) Oerst., Lomatia hirsuta (Lam.)
Diels ex Macbr., Gevuina avellana Mol., Maytenus boaria
Mol., Laurelia sempervirens (Ruiz. & Pavon.) Tul., Austrocedrus chilensis (D. Don) Florin & Boutelje, and Podocarpus saligna D. Don, with localized Prumnopitys andina
(Poepp. ex Endl.).
The other site was in montane rain forest in Parque Nacional Puyehue, located at 40839’S, 72811’W, at an elevation
of ca. 700 m. Mean annual temperature is estimated at about
8.5 8C, and mean annual rainfall at ca. 3800 mm, with a
summer minimum (Almeyda and Sáez 1958). The oldgrowth forests of this part of the park are dominated by the
evergreens N. dombeyi, Nothofagus nitida (Phil.) Krasser,
Laureliopsis philippiana (Looser) Schodde, Drimys winteri
J.R. & G. Forst, Weinmannia trichosperma Cav., and Saxegothaea conspicua Lindl., with fewer Podocarpus nubigena
Lindl.
Sampling
We sampled all conifer species present at both sites (three
at Los Lleuques, two at Puyehue), as well as the commonest
evergreen angiosperm trees (Table 1). A total of eight angiosperm species were sampled, included one species present at
both sites (N. dombeyi). All species are trees attaining
heights >15 m at maturity, with leaf life-spans exceeding
12 months (Lusk et al. 2003). We anticipated that the inclusion of the vesselless angiosperm D. winteri (Winteraceae)
would add an interesting comparative dimension to our
2007 NRC Canada
GL(m–9 MPa–1 s–1)
176±77
218±55
156±38
297±92
367±70
226±44
570±110
542±90
125±23
95±12
296±60
133±28
262±58
283±68
Conduit diameter
(m–6)
7.3±0.6
7.7±0.6
7.3±0.6
16.5±0.4
16.2±1.0
16.6±1.3
16.5±1.5
16.2±0.7
7.8±0.4
7.9±0.4
13.4±0.9
14.8±1.3
14.2±1.4
11.5±1.0
Branch length
(mm)
339±54
339±51
400±59
315±65
362±59
435±83
244±58
406±68
349±81
420±54
294±52
391±67
439±72
204±23
n
7
7
8
7
6
7
6
7
8
12
13
7
9
9
Podocarpaceae
Atherospermataceae
Cunionaceae
Nothofagaceae
Winteraceae
Conifers
Angiosperms
Puyehue
Atherospermataceae
Celastraceae
Nothofagaceae
Proteaceae
Angiosperms
Species
Austrocedrus chilensis
Podocarpus saligna
Prumnopitys andina
Laurelia sempervirens
Maytenus boaria
Nothofagus dombeyi
Gevuina avellana
Lomatia hirsuta
Podocarpus nubigena
Saxegothaea conspicua
Laureliopsis philippiana
Weinmannia trichosperma
Nothofagus dombeyi
Drimys winteri
Family
Cupressaceae
Podocarpaceae
Lineage
Conifers
Site
Los Lleuques
Table 1. Hydraulic traits of conifer and angiosperm species sampled at two sites in south-central Chile.
Note: Hydraulic trais: branch length; hydraulically weighted mean diameters of conduits; hydraulic conductance of stems normalized by subtended leaf area (GL); and Huber values (sapwood area divided
by subtended leaf area). Values are means ± SE.
143
Huber
values 104
2.89±0.53
3.08±0.74
2.29±0.55
2.87±0.40
3.75±0.46
1.81±0.38
2.98±0.35
3.07±0.23
3.05±0.41
2.57±0.39
3.28±0.37
2.44±0.33
2.94±0.35
3.53±0.61
Lusk et al.
study; if vessels make a decisive difference to the hydraulic
efficiency of stems, then the hydraulic traits of angiosperms
such as D. winteri might be expected to resemble those of
conifers (Brodribb and Feild 2000).
One branch (2.5–7.5 mm diameter) was cut from the outer
crown of each of six to 12 individuals of each species.
Sampled individuals were in all cases small trees (trunk
diameter <30 cm) growing at the forest margin, with easily
accessible branches. Immediately after cutting, branches
were recut under water 150–250 mm higher, to avoid embolism. Average diameter and length of branches did not differ
significantly between conifers and angiosperms (p = 0.48 and
p = 0.24, respectively). Most cut branches were ramified.
Conductance and leaf-area measurements
As we were interested in the overall capacity of ramified
branches to supply water to leaves, and in comparing scaling
of branch size with hydraulic capacity in conifers and angiosperms, we measured hydraulic conductance of the whole
branch stem (e.g., Becker et al. 1999), rather than conductivity of a section. Flow rate of 10 mmol/L KCl solution
through branches was measured within 72 h of cutting, at
about 20 8C under laboratory conditions. This solution was
prepared with deionized water and filtered to 0.2 m. This
solution, approximately isotonic with xylem sap, has been
shown to give better results than pure water (Sperry et al.
1987; Zwieniecki et al. 2001). Leaves were excised with a
scalpel, and the branch was perfused with KCl solution using a syringe connected to the base with a watertight seal,
until guttation was observed at the leaf traces. The base of
the branch was then connected with a watertight seal to a
fluid column fed by a reservoir elevated to a height of 1 m,
providing a constant pressure of 9.8 kPa. The branch was
wrapped in polythene film to trap fluid escaping from leaf
traces and avoid losses by evaporation, and placed on an
electronic balance, which registered KCl solution flux as an
increase in sample mass. Measurements were taken when an
approximately constant flow was observed for at least 3 min.
Hydraulic nomenclature follows the guidelines suggested by
Reid et al. (2005), converting mass of water to volume.
Absolute hydraulic conductance was calculated by dividing the flow rate by the pressure resulting from the height
difference between the reservoir and the branch (G =
m3MPa–1s–1). Conductance of stems can be usefully expressed in relation to sapwood cross-sectional area, as an indicator of the hydraulic efficiency of stemwood. After
completion of hydraulic measurements, safranin dye (0.1%)
was pumped through a short segment cut from the base of
the branch, which was then cross-sectioned. The section was
photographed with a digital camera mounted on a microscope,
and the image processed using the imaging software SigmaScan Pro 5 (SPSS Inc., Chicago, Ill). The cross-sectional area
of active xylem was calculated by subtracting the area of the
pith and any unstained xylem from the total area.
Normalization of hydraulic conductance by subtended leaf
area is another useful way of comparing the efficiency of
hydraulic supply in branches of differing sizes (Ewers 1985;
Becker et al. 1999):
GL = F / (PAL)
where F is the water flow mobilized through the branch ex2007 NRC Canada
144
Can. J. Bot. Vol. 85, 2007
pressed in volume (m3s–1), P is the pressure difference
along the branch (MPa–1), and AL is the leaf area (one-sided,
m2) of each branch distal to the point of measurement. The
G value of each branch was therefore divided by leaf area,
to give GL = mMPa–1s–1). Leaf area was measured using
SigmaScan1 Pro 5, after excised leaves were photographed
with a digital camera.
Conduit diameters
Conduit lumen diameters were measured on the same section cut from the base of each branch for measuring sapwood cross-sectional area. The image of each section was
divided radially into eight sectors; one or two of these sectors were randomly chosen for subsampling. All conduits
within each of these chosen sectors were measured, yielding
between about 100 to 1500 conduits per sample, depending
on species and branch diameter. The widest diameter (a) of
the conduit lumen was recorded, and the diameter perpendicular to it (b). This enabled estimation of the effective hydraulic diameter Dh of noncircular conduits according to
formulae given in Lewis and Boose (1995) for elliptical vessels:
rffiffiffiffiffiffiffiffiffiffiffiffiffi
2a2 b2
Dh ffi
a2 þ b2
and for quasi-rectangular tracheids:
Dh ¼
2ab
aþb
We then calculated hydraulically weighted mean conduit
diameters for each sample. Diameters were weighted according to the Hagen-Poiseuille Law, which assumes that conductance is proportional to the fourth power of conduit
diameter. The hydraulically weighted mean Dh was accordingly estimated thus:
Dh = (D4/N)0.25
where D is individual vessel diameter and N is number of
vessels (Tyree and Zimmermann 2002).
Statistical analysis
We used the statistical package (S)MATR1 (Falster et al.
2003) to examine major axis relationships of sapwood crosssectional area with absolute hydraulic conductance G,
conduit diameters, and leaf area of branches. This package
computes standardized major axis, which minimizes variance from the line in both dimensions, x and y (Sokal and
Rohlf 1995; Warton and Weber 2002), in contrast to the
least squares (or ‘‘model I’’) regression. This is important
when the primary concerns are the slope and intercept of a
relationship (such as when comparing bivariate functional
relationships among two or more data sets), rather than testing for a significant correlation or predicting one variable
from another. Use of least squares regression will give misleading estimates of the slope of such relationships when
correlation coefficients are low (Falster et al. 2003).
Results
Absolute hydraulic conductance (G) was approximately
isometric with sapwood area in both lineages at both Los
Lleuques and Puyehue (Fig. 1). Angiosperm stems conducted, on average, nearly twice as much fluid as conifer
stems of the same sapwood area; however, G of both lineages at Los Lleuques averaged about twice that at Puyehue
(Fig. 1).
Conduit diameters of most species were not significantly
correlated with sapwood area (i.e., branch size) at either
Los Lleuques or Puyehue (Fig. 2). Conduit diameters were
significantly larger in angiosperms than in conifers (Table 2),
although this difference was less marked at Puyehue than
at Los Lleuques. Tracheid lumens in the rain forest angiosperms at Puyehue were intermediate in width between
those of its coniferous associates and vessel lumens of the
other angiosperms (Table 1). However, the presence of
D. winteri at Puyehue was not the only cause of the significant interaction between site and lineage (Table 2), as
conduits of vesselbearing angiosperms were also narrower
at Puyehue than at Los Lleuques (Table 1).
Branch leaf area was significantly correlated with sapwood area in most species at both sites (Fig. 3). At both
sites, angiosperm and conifer species did not differ, on average, in leaf area at a common sapwood area (Fig. 3). Mean
Huber values did not differ significantly between the two
sites or between conifers and angiosperms (Tables 1 and 2).
GL values differed significantly between sites and between lineages (Table 2). There was no evidence that the
margin between the two lineages differed between the
two sites. Overall, mean GL was much higher at Los
Lleuques (0.305 gs–1MPa–1m–2) than that at Puyehue
(0.184 gs–1MPa–1m–2) (Table 1). Similar Huber values
in the two lineages, despite large differences in hydraulic
efficiency of sapwood, meant than GL was, on average,
nearly twofold higher in angiosperms than in conifers at
both sites (Table 1). The GL values of the vesselless angiosperm D. winteri fell within the range of values
yielded by vesselbearing rain forest angiosperms at Puyehue (Table 1).
Discussion
At both Los Lleuques and Puyehue, branches of evergreen
angiosperms had higher conductance than those of conifers
(Fig. 1; Table 1). This is at least partly attributable to the
associated difference in conduit diameters, which were consistently larger in angiosperms (Fig. 2; Table 1). Contrary to
expectation, a similar twofold difference in conductance between the two lineages was seen at both sites (Fig. 1). In the
only comparable study of temperate-zone evergreen forests
that we are aware of, Brodribb and Feild (2000) showed
that specific conductivity (rather than conductance) of
branch stemwood was higher on average in angiosperms
than in conifers, but that this difference was greater at a
tropical site in New Caledonia than in cool temperate rain
forest in Tasmania, where the two lineages overlapped considerably.
The high conductance of the vesselless angiosperm D.
winteri (Table 1) is an apparent challenge to the idea that
vessels make a crucial difference to the hydraulic efficiency
of stems. This result was unexpected, as Brodribb and Feild
(2000) found that specific conductivities of branch stem2007 NRC Canada
Lusk et al.
Fig. 1. Relationships of absolute hydraulic conductance of leafless
branches with sapwood cross-sectional area of evergreen angiosperms and conifers at (a) Los Lleuques and (b) Puyehue (open
symbols, angiosperms; solid symbols, conifers). Relationships were
significant for all species except L. hirsuta at Los Lleuques (p =
0.23), and W. trichosperma at Puyehue (p = 0.07). At Los Lleuques, standardized major axes showed significant interspecific variation in elevation (p < 0.0001) but not in slope (p = 0.13). G at the
grand mean value of sapwood cross-sectional area was significantly
higher, on average, in angiosperm species than in conifers (13.8 vs.
7.5 mm3 Mpa–1s–1; p = 0.007). At Puyehue, standardized major
axes of species varied significantly in elevation (p = 0.006) but not
in slope (p = 0.24). Gh at the grand mean value of sapwood crosssectional area was significantly higher in angiosperms than in conifers (6.8 vs. 3.5 mm3MPa–1s–1 ; p = 0.03).
wood of vesselless angiosperms in Tasmanian and New
Caledonian rain forests are lower than those of vesselbearing
angiosperms and similar to those of conifers. The conduits
of D. winteri, however, although narrower than most vessels,
were wider than the tracheids of its coniferous associates
(Table 1), and Carlquist and Schneider (2002) point out that
some features of the tracheary elements of Winteraceae
arguably warrant their being considered an intermediate
between tracheids and vessels. However, the high conductance measured in D. winteri is also, to some degree, a
reflection of the relatively short length of the branches cut
from this species. Although the average length of branches
did not differ overall between conifers and angiosperms,
and there was no significant interspecific variation in branch
diameters (ANOVA: p = 0.89), there were species differences in branch length (p = 0.02). Drimys winteri had the
shortest branch length of any the study species (Table 1),
and as conductance is negatively affected by path length,
145
the unusual branch allometry of this species may partly
account for the unexpectedly high conductance that we
measured. Although our sampling procedure does not permit
accurate calculation of the hydraulic conductivity (rather
than conductance) of branches, the relatively short length of
D. winteri branches implies that the hydraulic efficiency of
this species would seem less impressive if path length were
taken into account.
Site differences in angiosperm hydraulic conductance can
be explained by parallel variation in conduit diameters, but
this was not so for conifers (Tables 1 and 2). Geographic
variation in hydraulic conductance has sometimes been
linked to the positive relationship of conduit diameters
with environmental temperatures (Tyree and Ewers 1991;
Sperry and Sullivan 1992), an idea that fits well with the
intersite variation we observed in angiosperms (Table 1).
However, although conifers also showed a twofold difference in conductance between Los Lleuques and Puyehue,
tracheid dimensions were similar at the two sites (Table 1).
We can find no obvious explanation for this pattern. As
measurements were made under laboratory conditions,
temperature-dependent variation in water viscosity should
not have influenced the results.
The superior hydraulic efficiency of angiosperm sapwood
(Fig. 1) was associated with greater hydraulic conductance
per unit leaf area (Tables 1 and 2). Brodribb and Feild
(2000), who reported conductivity rather than conductance,
found an essentially similar pattern in Tasmanian and New
Caledonian evergreen forests. A recent growth experiment
with seedlings of three tropical angiosperms and three conifers also reported greater stem hydraulic conductivity per
unit leaf area in the former (Brodribb et al. 2005). This contrasts with the results of Becker et al. (1999), who found
that shoot conductances of wildling saplings, normalized by
leaf area, are broadly similar in three tropical conifers and
nine tropical angiosperms. Of the three comparative studies
of evergreen angiosperms and conifers that we are aware of,
two indicate a more abundant hydraulic supply to leaves in
the former lineage. This correlates with the earlier finding
that evergreen angiosperms at one of our sites (Los Lleuques) have higher average stomatal conductance and photosynthetic capacity than their coniferous counterparts (Lusk
et al. 2003), although differences in these parameters are
much less marked and nonsignificant at Puyehue.
The greater conductance of angiosperm branches was not
accompanied by differences in subtended leaf area at a
given sapwood area (i.e., there was no generalized difference in Huber values) (Fig. 2). Brouat et al. (1998), reanalyzing the data of White (1983), found that both slope and
elevation of the relationship between twig cross-sectional
area and subtended leaf area are similar in evergreen angiosperms and conifers. In contrast to data from branches,
Brodribb et al. (2005) reported that seedlings of three tropical angiosperms display larger leaf areas at a given diameter
than conifer seedlings, especially in well-lit conditions. We
(C. Lusk, A. Saldaña and M. Pérez-Millaqueo, unpublished
data) also showed that seedlings of temperate-zone evergreen angiosperms develop larger leaf areas under field conditions than their coniferous associates. Controls on stem
diameter – leaf area relationships at the seedling stage may
therefore differ from those operating on lateral branches of
2007 NRC Canada
146
Fig. 2. Relationships of hydraulically weighted mean conduit diameters with sapwood cross-sectional area of branches of evergreen
angiosperms and conifers at (a) Los Lleuques and (b) Puyehue.
Relationships were significant for only two species at Puyehue:
L. philippiana (r = 0.72, p = 0.01) and W. trichosperma (r = 0.85,
p = 0.007). No significant relationships found at Los Lleuques.
adult trees. As efficiency of leaf area display on orthotropic
shoots is positively correlated with leaf size (Takenaka
1994; Falster and Westoby 2003), the small size of most
conifer leaves probably constrains the amount of leaf area
that can be displayed effectively on the central axis of small
seedlings. In contrast, about half of the study species develop plagiotropic lateral branches, which permits close
spacing of small leaves with minimal self-shading
(Takenaka 1994; King 2005). Self-shading within branches
may thus be less of an issue on the branches used in the
present study, and branch diameter – leaf area relationships
may be determined as much by mechanical support requirements as by hydraulic supply (Brouat et al. 1998).
Our results provide further evidence that, irrespective of
leaf habit, conifer and angiosperm trees tend to differ in hydraulic efficiency. Although the hydraulic differences reported in our study are less pronounced than those
separating deciduous angiosperms and evergreen conifers in
northern temperate-zone forests (Becker 2000; see Fig. 4),
their linkage to differences in leaf-level physiology
(Brodribb and Feild 2000; Lusk et al. 2003) suggests they
have bearing on ecological interactions between these two
lineages in evergreen forests. Apart from the unexpectedly
large conduit diameters and conductance of the vesselless
angiosperm D. winteri, our principal finding is that, at least
at branch level on adult trees, the greater hydraulic efficiency of angiosperm stems appears to be capitalized on
Can. J. Bot. Vol. 85, 2007
Fig. 3. Relationships of sapwood cross-sectional area with subtended leaf area in branches of evergreen angiosperms and conifers
at (a) Los Lleuques and (b) Puyehue (open symbols, angiosperms;
closed symbols, conifers). Relationships were significant for all
species except A. chilensis (p = 0.30) and P. saligna (p = 0.21) at
Los Lleuques, and D. winteri at Puyehue (p = 0.11). At Los Lleuques, standardized major axes showed significant interspecific variation in slope (p < 0.002) and elevation (p < 0.001); however, leaf
area at the grand mean value of sapwood cross-sectional area did
not differ significantly between angiosperms and conifers (p =
0.86). At Puyehue, there was no significant interspecific variation
in slope (p = 0.20) or elevation (p = 0.41) of standardized major
axes, and leaf area at the grand mean value of sapwood cross-sectional area did not differ significantly between angiosperms and
conifers (p = 0.59).
Table 2. Summaries of ANOVA examining effects of site and
lineage on conduit diameters (hydraulically weighted means),
branch conductance normalized by leaf area (GL), and Huber
values (Hv), in evergreen conifer (n = 5) and angiosperm tree
species (n = 9) from south-central Chile.
Response variable
Conduit diameter
(log) GLA
Huber values
Effect
Site
Lineage
Site lineage
Site
Lineage
Site
Lineage
F-ratio
7.61
257.66
13.51
12.59
8.69
0.54
0.22
p
0.02
< 0.0001
0.004
0.005
0.022
0.48
0.65
Note: Clearly nonsignificant interactions were removed from models
for GL and Hv. GL values were log-transformed to normalize their distribution.
2007 NRC Canada
Lusk et al.
through more abundant hydraulic supply to leaves (and
hence, higher gas-exchange potential), rather than through
development of more extensive leaf areas.
Acknowledgments
We thank the Fondo Nacional de Desarrollo Cientı́fico
y Tecnológico (FONDECYT) Programme for support
through Grant 1030811, Corporación Nacional Forestal
(CONAF) for permission to work in Parque Nacional
Puyehue, Missy Holbrook for hosting the second author’s visit to her lab, and the Programa de Mejoramiento de la Calidad y la Equidad de la Educación
Superior (MECUSUP) for funding the visit through
Grant UCO 9906. Matı́as Perez provided technical assistance.
References
Almeyda, E., and Sáez, F. 1958 Recopilation de datos climáticos
de Chile y mapas sinópticos respectivos. Ministerio de Agricultura, Santiago, Chile.
Becker, P. 2000. Competition in the regeneration niche between conifers and angiosperms: Bond’s slow seedling hypothesis. Funct.
Ecol. 14: 401–412. doi:10.1046/j.1365-2435.2000.00455.x.
Becker, P., Tyree, M.T., and Tsuda, M. 1999. Hydraulic conductances of angiosperms versus conifers: similar transport sufficiency at the whole-plant level. Tree Physiol. 19: 445–452.
PMID:12651550.
Bond, W.J. 1989. The tortoise and the hare: ecology of angiosperm
dominance and gymnosperm persistence. Biol. J. Linn. Soc. 36:
227–249.
Brodribb, T.J., and Feild, T.S. 2000. Stem hydraulic supply is
linked to leaf photosynthetic capacity: evidence from New
Caledonian and Tasmanian rainforests. Plant Cell Environ. 23:
1381–1388. doi:10.1046/j.1365-3040.2000.00647.x.
Brodribb, T.J., Holbrook, N.M., and Hill, R.S. 2005. Seedling
growth in conifers and angiosperms: impacts of contrasting xylem structure. Aust. J. Bot. 53: 749–755. doi:10.1071/BT05049.
Brouat, C., Gibernau, M., Amsellem, L., and McKey, D. 1998.
Corner’s rules revisited: ontogenetic and interspecific patterns
in leaf–stem allometry. New Phytol. 139: 459–470. doi:10.1046/
j.1469-8137.1998.00209.x.
Carlquist, S., and Schneider, E.L. 2002. The tracheid–vessel element transition in angiosperms involves multiple independent
features: cladistic consequences. Am. J. Bot. 89: 185–195.
Davis, S., Sperry, J., and Hacke, U. 1999. The relationship between
xylem conduit diameter and cavitation caused by freezing. Am.
J. Bot. 86: 1367–1372. doi:10.2307/2656919.
Ewers, F.W. 1985. Xylem structure and water conduction in conifer trees, dicot trees, and lianas. IAWA (Int. Assoc. Wood
Anat.) Bull. 6: 309–317.
Falster, D.S., and Westoby, M. 2003. Leaf size and angle vary
widely across species: what consequences for light interception?
147
New Phytol. 158: 509–525. doi:10.1046/j.1469-8137.2003.
00765.x.
Falster, D.S., Warton, D.I., and Wright, I.J. 2003. (S)MATR: standardized major axis tests and routines. Version 1.0. www.bio.
mq.edu.au/ecology/SMATR [accessed October 2006].
Feild, T.S., Brodribb, T.J., and Holbrook, N.M. 2002. Hardly a relict: freezing and the evolution of vesselless wood in Winteraceae. Evolution, 56: 464–478. doi:10.1554/0014-3820(2002)
056[0464:HARFAT]2.0.CO;2. PMID:11989678.
King, D.A. 2005. Architectural differences in saplings of temperate
versus tropical angiosperms: consequences of the deciduous habit? Can. J. Bot. 83: 1391–1401. doi:10.1139/b05-109.
Lewis, A.M., and Boose, E.R. 1995. Estimating volume flow rates
through xylem conduits. Am. J. Bot. 82: 1112–1116. doi:10.
2307/2446063.
Lusk, C.H., Wright, I.J., and Reich, P.B. 2003. Photosynthetic differences contribute to competitive advantage of evergreen angiosperm trees over evergreen conifers in productive habitats.
New Phytol. 160: 329–336. doi:10.1046/j.1469-8137.2003.
00879.x.
Reid, D.E.B., Silins, U., Mendoza, C., and Lieffers, V.J. 2005. A
unified nomenclature for quantification and description of water
conducting properties of sapwood xylem based on Darcy’s law.
Tree Physiol. 25: 993–1000. PMID:15929930.
Sokal, R.R., and Rohlf, F.J. 1995. Biometry: the principles and
practice of statistics in biological research. 3rd ed..
W.H. Freeman, San Francisco, Calif.
Sperry, J., and Sullivan, J. 1992. Xylem embolism in response to
freeze–thaw cycles stress in ring-porous, diffuse-porous, and
conifer species. Plant Physiol. 100: 605–613. PMID:16653035.
Sperry, J., Holbrook, N.M., Zimmermann, M., and Tyree, M.T.
1987. Spring filling of xylem vessels in wild grapevine. Plant
Physiol. 83: 414–417. PMID:16665259.
Takenaka, A. 1994. Effects of leaf blade narrowness and petiole
length on the light capture efficiency of a shoot. Ecol. Res. 9:
109–114. doi:10.1007/BF02347485.
Tyree, M.T., and Ewers, F. 1991. The hydraulic architecture of
trees and other woody plants. New Phytol. 19: 345–360.
Tyree, M.T., and Zimmermann, M.H. 2002. Xylem structure and
the ascent of sap. 2nd ed. Springer-Verlag, New York.
Wang, J., Ives, N.E., and Lechowicz, M.J. 1992. The relation of
foliar phenology to xylem embolism in trees. Funct. Ecol. 6:
469–475. doi:10.2307/2389285.
Warton, D.I., and Weber, N.C. 2002. Common slope tests for bivariate structural relationships. Biom. J. 44: 161–174. doi:10.
1002/1521-4036(200203)44:2<161::AID-BIMJ161>3.0.CO;2-N.
White, P.S. 1983. Evidence that temperate east North American
evergreen woody plants follow Corner’s rules. New Phytol. 95:
139–145. doi:10.1111/j.1469-8137.1983.tb03477.x.
Zwieniecki, M.A., Melcher, P.J., and Holbrook, N.M. 2001. Hydrogel control of xylem hydraulic resistance in plants. Science
(Wash., D.C.), 291: 1059–1062. doi:10.1126/science.1057175.
PMID:11161220.
2007 NRC Canada