American Journal of Botany 90(1): 72-77. 2003.

LEAF BIOMECHANICS, MORPHOLOGY, AND ANATOMY

OF TLE DECIDUOUS MESOPHYTE PRUNUS SERRULATA
(ROSACEAE) AND THE EVERGREEN SCLEROPHYLLOUS
SHRUB HETEROMELES ARBUTIFOLIA (RosACEAE)'

RONALD A. BALSAMO, 2 AARON M. BAUER, STEPHEN D. DAVIS,3 AND

BENITA M. RICE

Department of Biology, Villanova University, Villanova, Pennsylvania 19085 USA

Leaf tensile properties were compared between the mesic deciduous tree Prunis serrmlata (var. "Kwanzan") and the xeric and
sclerophyllous chaparral evergreen shrub Heteromeles arbutifolia(M. Roem). All values for biomechanical parameters for H. arbutifolia
wete significantly greater than those of P. serruilata. The fracture planes also differed between the two species with P. serrnlata
fracturing along the secondary veins. while H. arbutifolia most often fractured across the leaf irrespective of the vein or mesophyll
position, thus yielding qualitative differences in the stress-strain curves of the two species. Anatomically, P. sermlata exhibits features
typical for a deciduous mesophytic leaf such as a thin cuticle, a single layer of palisade mesophyll, isodiametric spongy mesophyll,
and extensive reticulation of the laminar veins. Heteromeles arbutifolia leaves, however, are typically two- to three-fold thicker with
a 35% higher dry mass/fresh mass ratio. The vascular tissue is restricted to the interface of the palisade and spongy mesophyll near
the center of the leaf. Both epidermal layers have a thick cuticle. The palisade mesophyll is tightly packed and two to three layers
thick. The spongy mesophyll cells are ameboid in shape and tightly interlinked both to other spongy cells as well as to the overlying
palisade layer. We conclude that the qualitative and quantitative biomechanical differences between the leaves of these two species
are likely due to a complex interaction of internal architectural arrangement and the physical/chemical differences in the properties of
their respective cell walls. These studies illustrate the importance that morphological and anatomical correlates play with mechanical
behavior in plant material and ultimately reflect adaptations present in the leaves of chaparral shrubs that are conducive to surviving
in arid environments.

Key words: anatomy: biomechanics: chaparral: Heteromeles: morphology: Prunus.

Sclerophylly is often used to describe the material properties
of evergreen, leaves in certain shrubs and trees growing pri-
marily in xeric habitats. These include the chaparral of the
western United States, the macchia of Chile and the Mediter-
ranean basin, the fynbos of the western Cape in South Africa,
and the heath of western Australia. The leaves of sclerophyl-
lous species are characterized as being thick, tough, and leath-
ery, with a high fiber (sclereid) to protein ratio (Loveless,
1961; Turner, 1994). However, Seddon (1974) states that many
nonxeric species may also have sclerophyllous leaves, an ob-
servation substantiated by the compendium of Rao (1992). As
such, there are several concurrent hypotheses as to why scler-
ophylly has been a successful adaptation for plants growing in
Mediterranean climates. Turner (1994) outlines three possibil-
ities: water conservation, nutrient conservation, and as a pro-
tection against herbivory. In spite of these competing hypoth-
eses to explain the incidence of sclerophyllous leaves, it is
generally recognized that the ecosystems where sclerophylly
is common typically experience periodic drought. Also, scler-
ophyllous leaves have more in common with xeric leaves than
with mesophytic leaves. For example, both sclerophylls and
xeric leaves commonly exhibit reduced surface area-to-volume
ratios, decreased intercellular airspaces, a more strongly de-
veloped palisade mesophyll, and often the presence of a hy-
podermis (Maximov, 1929; Esau, 1977; Levitt, 1980). Xero-
Manuscript received 9 October 2001: revision accepted 12 July 2002.
2 Author for correspondence (e-mail: ronald.balsamo@villanova.edu; FAX:
610-519-7863).
3 Current address: Division of Natural Sciences, Pepperdine University,
Malibu, California 90263 USA.
72
morphic characteristics can be due to a variety of environ-
mental conditions including high light intensity (Smith et al.,
1997), salinity (Sordo, Padilla, and Romero, 1982), and
drought (Esau. 1977; Levitt, 1980), all of which involve the
limitation of moisture. Gratani et al. (1989) demonstrate that
like many mesic species (for examples. see Smith et al., 1997;
Utrillas and Alegre, 1997), xeric characteristics are also en-
hanced in sclerophyllous shrubs exposed to increasing mois-
ture limitation.
One aspect that is often mentioned and subsequently ne-
glected when discussing sclerophylly is that xeric/sclerophyl-
lous leaves are strongly correlated with tolerance to dehydra-
tion. Evergreern chaparral shrubs such as those found in the
genus Ceanothus when fully hydrated exhibit water potentials
(-1.5 to -2.5 MPa at midday) that would be lethal to most
mesophytes (Larcher, 1995). For example. Heteromeles ar-
butifolia leaves average below 3.0 MPa near the end of the
drought season (Davis and Mooney, 1986a. b) and can be ar-
tificially dried to below -5.0 MPa without apparent damage.
Indeed, some species of Ceanothus exhibit water potentials
below -10 MPa in the fall dry season with no apparent ill
effect to the leaves or stems (Schlesinger et al., 1982; Davis,
Kolb, and Barton, 1998).
Despite the generally accepted observations that sclerophyll/
xeric leaves are tougher/stiffer than mesophytic leaves and that
sclerophyllous leaves are most common in habitats where
moisture is seasonally limiting, there have been few studies
that actually attempt to quantify how "tough" sclerophylls re-
ally are or to illustrate why they are stronger/stiffer. Maximov
(1929) was probably the first to point out that xerophytic
leaves can lose 30-40% of their water content before wilting,
January 2003] Jnr23BALSAMO ET AL.-RELATIONSHIP OF ANATOMY TO LEAF BIOMECHANICS
while a loss of only 1-2% causes wilting in delicate shade
plants. He attributed this discrepancy to differences in the
physical properties of the cell waEls of leaf cells. Working with
a variety of species, Turner and associates (Choong et al.,
1992; Turner et al., 1993) concluded that the fracture "tough-
ness" (actually the energy of fracture) values of sclerophyllous
species are higher than for "softer" leaves collected from trop-
ical rain forests. Recently, Edwards, Read, and Sanson (2000).
using similar techniques on Australian heath species, attempt-
ed to define sclerophylly purely by toughness and strength.
Vincent (1982, 1991) and Greenberg et al. (1989) have con-
ducted whole-leaf biomechanical studies on several grasses,
but any extrapolations to deciduous or evergreen dicotyledon-
ous leaves must be made with caution due to differing vein
architectures, as grass venation is parallel, while the latter of-
ten exhibit complex and reticulate venation. Vogel (1989) and
Niklas (1992, 1996) have worked extensively on the biome-
chanical parameters of monocotyledonous and dicotyledonous
leaves but as of this date have not looked at sclerophyllous
leaves.
In this study we report on the biomechanical properties and
leaf anatomy of the leaves of the mesophyte deciduous tree
Prunus serrulata and compare them to the leaves of the ev-
ergreen sclerophyllous chaparral shrub Heteromeles arbutifol-
ia. We chose Prunus and Heteromeles because both species
have oblong/lanceolate leaves with pinnate venation and are
roughly the same length and width. In addition, both species
are in the family Rosaceae and thus presumably share genetic
traits relating to their gross morphology and anatomy without
being so similar that a comparison would not yield useful data.
We elucidate biomechanical properties in detail and test the
hypothesis that leaf biomechanical properties are related to the
structural organization of the various leaf tissues. Further, we
discuss the importance of biomechanical properties for leaves
that exhibit dehydration tolerance and how morphology and
anatomy necessarily correlates with mechanical behavior.
MATERIALS AND METHODS
Plant material-Leaves of Prmnus serrulata var. "Kwanzen" were col-
lected from mature specimens growing on the Villanova University campus.
All tissue processing for microscopic examination and all biomechanical ex-
periments were conducted on location at Villanova University. Heteromeles
arbutifolia (M. Roem) leaves were collected from specimens growing in the
hills surrounding Pepperdine University in Malibu. Calfomia, USA. Leaves
of H.arbutifolia used in leaf tensile measurements were processed as follows:
terminal stem segments 30-60 cm were cut under water and immediately
placed in perforated plastic bags with moistened laboratory towels on ice.
Cuttings were shipped overnight to Villanova University, where they were
placed in beakers with the cut ends immersed in water and stored at 4°C.
Leaf tensile measurements were conducted within 7 d for all samples. al-
though some samples (data not shown) processed after 14 d did not signifi-
cantiy differ from those reported in this study. Leaf sections for light and
scanning electron microscopy were collected and processed on location at
Pepperdine University.
Leaf morphological measuremnents-Lamina length, width, and thickness
were measured using a Fowler ProMax digital caliper (Newton, Massachu-
setts, USA). Laminar surface area (one sided) was calculated by making leaf
rubbings onto graph paper and adding up the area inside the generated image.
Dry mass/fresh mass ratios were determined as follows: terminal shoots were
excised from healthy plants by cutting while immersed in water. Shoot sec-
tions were then transferred to a beaker of water and covered with a polyeth-
ylene bag till fully hydrated (sample lamina measured <0.1 MPa using a
73
Precision Science Instruments ([Corvallis, Oregon, USA] plant pressure cham-
ber). Laminas were excised and immediately weighed on an analytical scale
to obtain fresh masses and numbered 1-25. Samples were then placed in a
757C oven for 72 h and reweighed to obtain the dry masses.
Leaf mechanicalproperties-Leaf thickness and maximum leaf width were
measured to the nearest 0.01 mm using a Fowler ProMax digital caliper.
Whole leaves were introduced into pneumatic grips with hard rubber faces
on an MTS Bionix 100 mechanical testing system (MTS Systems Corporation,
New Prarie, Minnesota, USA). Grips were placed with a starting interval of
10 mm separation. Leaves were positioned such that their long axis (midrib)
was parallel to the load applied and their maximum width was within the gap
between the grips. Leaves were stretched at a rate of 20 mm/m/m. Force and
displacement were recorded continuously using the MTS Testworks 4 soft-
ware package and subsequently normalized to stress = (v force per cross-
sectional area at t = 0), measured in megapascals (MPa), and strain (e =
increment in length divided by the initial length). Samples were stretched to
the break point and failure load (FL, the total force necessary to achieve
catastrophic failure), failure strain (ef). and tensile strength (cm. the maximum
stress sustained prior to failure) were determined. The tensometer was set to
a test termination sensitivity of 99.5%. thus automatically ending a test when
a drop to or.= 0.5% was detected. The decision to use 0.5% as a cutoff was
made because in some cases support tissues in the leaves did not fail at a.
and continued to elongate at very low stresses. The modulus of elasticity (E),
a measure of material stiffness, was automatically calculated from the slope
of the linear portion of the stress-strain curve. The total area under the stress-
strain curve is the energy (W) required to break a unit volume of material,
expressed in joules per cubic meters, and is a measure of toughness (Gordon,
1978).
Niklas (1992) has noted that the standard clamping of tissues under uniaxial
tension can yield misleadingly high values of E because the deformation of
distal structures is prevented or restricted. He suggests that the length of sam-
ples be at least 10 times their width to avoid such problems of end-wall
effects. This was not possible or desirable in our study. We were interested
in the comparative mechanical properties of whole leaves rather than single
tissue types. Because of the irregular and species-specific architecture of the
test samples, it was not prudent to allow a long effective sample length (=
initial distance between clamps). Longer effective specimen lengths for whole
leaves generate at least two problems: uncertainty of appropriate sample width
measurement (and hence of calculated stress) due to terminal tapering, and
tendencies toward specimen slippage due to inadequate leaf surface area (and
hence friction) within the clamps. Further, in the case of the leaves tested,
length-to-width ratios would not have permitted the recommended 10 :1 as-
pect ratio under any circumstances. This could only be achieved bytrimming
the specimens down to a uniform width. As we were measuring properties of
the entire organ we considered this unacceptable as such properties may well
be altered by the disruption of the leaf's intact edge and/or removal of one
or more support elements. This is especially crucial as our comparisons were
interspecific and similar trimming of leaves may have affected each leaf type
differently. We accept that our measurement of elastic modulus may be arti-
ficially high but maintain that our test conditions allow for the appropriate
relative comparison of mechanical properties between sample leaf popula-
tions.
Pressure volume curves-Leaf tissue water relations were determined for
both species using a PMS model 1003 plant pressure chamber (PMS Instru-
ment, Corvallis, Oregon. USA) and following the protocol established by
Davis and Mooney (1986b).
Scanning electron microscopy-All leaves for scanning electron micros-
copy (SEM) were sliced into quarter pieces and fixed in a solution of 5%
glutaraldehyde in 50 mmolVL potassium phosphate buffer (pH 7.0) for 24 h
at 4°C. The pieces were then washed in 50 mmol/L Na-cacodylate buffer (pH
7.0) followed by postfixation in 1% OSO4 in 50 mmolVL Na-cacodylate buffer
(pH 7.0) for 12 h at room temperature. After washing once in 50 mmol/L
Na-cacodylate buffer (pH 7.0) the pieces were dehydrated in an acetone series.
 74 AMERICAN JOURNAL OF BoTANY [Vol. 90
TABLE 1. Leaf morphological measurements. N = 25 leaves per species. P < 0.001 indicated as 8** surface area is for one side of leaf only.
DM/FM ratio = dry mass to fresh mass ratio.

Surface area
Taxon Length (mm) Width (mm) Thickness (mm) (mrf) DM/FM ratio
Prunus serrulata 70.62 ± 5.9 34.44 ± 3.1 0.210 ± 0.02 1345.0 ± 208 0.343 ± 0.02
Heteromeles arbutifolia 76.52 ± 9.4 30.85 ± 4.0 0.513 + 0.05*** 1595.8 ± 399 0.519 ± 0.03***

Pieces were then fractured under liquid nitrogen perpendicular to the midrib leaves near the shoot tips and only in the spring months (S.
using new razor blades revealing sample fractures in cross sections. Pieces D. Davis and R. A. Balsamo, personal observations). The epi-
were then critically point dried (Polaron Jumbo Critical Point model E3100,
dermal tissues of mature leaves of H. arbutifolia were glabrous
London, UK) and gold sputter coated (Polaron SC7640). Fractured samples
were viewed and photographed using an Hitachi S-570 SEM (Tokyo, Japan).
and devoid of trichomes or glands.

Lighit microscopy-All leaves for light microscopy were sliced with a new
razor blade into 1-3 mm 2 samples from the middle of the lamina and im-
mediately fixed in a solution of 5% glutaraldehyde in 50 mmol/L potassium
phosphate buffer (pH 7.0) for 12 h at 4°C. Samples were then washed in 50
mmollL Na-cacodylate buffer (pH 7.0) followed by postfixation in I% OsO,
in 50 mmol/L Na-cacodylate buffer (pH 7.0) for 2 h at room temperature.
After washing once in 50 mmol/L Na-cacodylate buffer (pH 7.0) samples
were then subjected to a dehydration series of acetone followed by infiltration
in Spurr's (1969) resin. Polymerization occurred at 60°C over 48 h. Sections
were cut using a LKB Bronmma 2088 Ultrotome V thermal advance ultrami-
crotome (kindly on loan from Michael Wisniewski, USDA Appalachian Fruit
Research Station, Kearneysville, West Virginia, USA). Sections were placed
on glass slides and stained with a solution of 1% crystal violet followed by
a solution of 1% safranin 0. Photographs were obtained using an Olympus
SC35 SLR 35 mm camera mounted to an Olympus BX60 microscope with
UplanApo color-corrected objective lenses. Cuticle thickness was measured
from photomicrqgraphs in cross section and corrected for enlargement. Five
leaf cross sections were measured per species.
RESULTS
Leaf morphology-Both species had oblong/lanceolate
leaves with pinnate venation and were roughly the same length
and width. The deciduous leaves of P. serrulata were thinner
than H. arbutifolia and had less dry matter (fiber), but other-
wise both species were similar in their gross morphology (Ta-
ble 1). Both adaxial and abaxial epidermal layers of P. ser-
rulata leaves had numerous hair-like trichomes while in H.
arbutifolia trichomes were only observed in young, expanding
Leaf biomechanoical properties-There are commonly four
stress-strain profiles that correspond to the behavior of mate-
rials differing in the degree of elastic and plastic deformations
when subjected to a tensile force (Niklas, 1992). When leaves
of P. serrulata and H. arbutifolia were subjected to tensile
forces the stress-strain curve produced was consistently of the
elastic-plastic type (Fig. 1), but curves give no clear relative
plastic break points. However, leaves of the two species dif-
fered markedly. in their response to tensile force and their in-
herent biomechanical properties (Fig. 1, Table 2). For all pa-
rameters measured, H. arbutifolia leaves had consistently
higher mechanical values than P. serrulata (P < 0.001). Fail-
ure load was over six-fold higher in H. arbutifolia, while ten-
sile strength (Cm) was almost four-fold higher. The toughness
(W and modulus of elasticity (E) values were approximately
five-fold higher in H. arbutifolia, which also had 25% higher
values for the failure strain (ef).
Stress-strain curves for P. serrulata leaves differed quali-
tatively from H. arbutifolialeaves (Fig. 1). Break patterns fol-
lowing catastrophic failure indicated that in H. arbutifolia the
leaf broke consistently and uniformly. In contrast to this, the
leaves of P. serrulata exhibited ragged breakage patterns in-
dicating tissues within the leaf with differing resistance to ten-
sile stress and hence mechanical properties. The most common
scenario was that the leaves initially separated along the sec-
ondary veins followed by the midvein, which often snapped
long after the mesophyll.had separated.

2.5
F A 1.0 B 4

C
0.75

0.5 2

0)~~~~~~~~~~~~~~~~~02

Strain (%)
Fig. 1. Stress-strain curves from leaf tension tests. (A) Sample curve illustrating the various physical properties. The slope of the line from B to M is used
to calculate the modulus of elasticity. The point of catastrophic failure (F) is used to determine tensile strength. The area under the curve is the work of fracture
(toughness), and the point where the stress curve falls below 10% of F is recorded as the failure straih. (B) Typical stress-strain curve resulting from a tensile
experiment on leaves of Prunus serrulata: note ragged appearance of descending curve. (C) Typical stress-strain curve resulting from a tensile experiment
on
leaves of H. arbutifolia; note smooth descending curve. Also note difference in scale of the y-axes for B and C.
January 2003] BALsAMO ET AL.-RELATIONSHIP OF ANATOMY TO LEAF BIOMECHANICS
TABLE 2. Leaf tensile properties. N = 25 leaves per species, all values differ significantly (P < 0.001) between the two species, FL = failure
load, (T. = tensile strength, W = toughness, E = modulus of elasticity, e, = failure strain. Units for each set of parameters are indicated.
Taxon FL (N) (T. (MPa)
Prunus serrulata 7.44 t 1.4 0.88 ± 0.2
Heteromeles arbutifolia 46.73 ± 7.3 3.28 ± 0.5
Pressure volume curves-Pressurevolume curves revealed
significant differences (P < 0.05) in the turgor loss point (tlp),
osmotic potential (T,J, and bulk modulus of elasticity (s0 )
between the two species. Heteromneles arbutifolia exhibited
higher values for all parameters derived from the curves (Table
3). Typically, leaves of H. arbutifolia would remain rigid well
below the turgor loss point to approximately 50% relative wa-
ter content, below which they would show structural damage
from the force exerted by the pressure chamber and no further
measurements could be obtained. In contrast, the leaves of P.
serrulata became flaccid even before reaching the turgor loss
point and typically were damaged by the pressure chamber
before reaching 75% relative water content (data not shown).
Leaf anatomy-Leaves of P. serrulata and H. arbutifolia
differed markedly in their cell and tissue anatomy. Cross sec-
tions of P. serrulata(Fig. 2) revealed a thin cuticle on a small
unilayered adaxial epidermis with numerous epidermal hairs
that were also present on the abaxial epidermis. These tri-
chomes were embedded in columns of collenchyma cells (ep-
ithem) that branched off from the secondary veins to both the
abaxial and adaxial epidemal layers (not shown). A single lay-
er of palisade mesophyll cells was observed, and spongy me-
sophyll appeared isodiametric and loosely packed. In contrast,
cross sections of H. arbutifolia (Fig. 3) revealed a prominent
cuticle on the adaxial epidermis devoid of hairs or trichomes,
but like P. serrulataleaves have columns of epithem cells that
spanned from the secondary veins to both epidermal layers.
There were two to three layers of densely packed palisade
mesophyll cells, and spongy mesophyll that were also closely
packed and appearing tightly interconnected.
Paradermal sections of the abaxial half of the leaves of P.
serrulata revealed a loose assemblage of spongy mesophyll
cells and a much higher incidence of reticulation and closer
spacing of veins, most likely due to its occurrence in more
mesic habitats (Fig. 4). In contrast, the paradermal sections of
H. arbutifolia clarified the tight, interconnected nature of the
spongy mesophyll cells (Fig. 5). In addition, spongy meso-
phyll cells were much larger in H. arbutifolia and were irreg-
ular to ameboid rather than isodiametric in shape. In cross
section, the adaxial cuticle of P. serrulata was 2.40 t 0.4 «Lm
and the abaxial cuticle was 0.88 t 0.1 p1m. In H. arbutifolia
the adaxial cuticle was 26.50 ± 2.9 lim and the abaxial cuticle
was 11.50 t 2.9 pm (N = 10 different leaf cross sections per
species).
TABLE 3. Pressure volume curve parameters. N = 6 leaves per species,
all values differ significantly (P < 0.05) between the two species,
tlp = turgor loss point, RWC = relative water content, ,, = os-
motic potential, e, = bulk modulus of elasticity. Units for each set
of parameters are indicated.
Taxon tip (%RWC) .,(MPa) E, (MPa)
Prunis serrulata 93.13 ± 0.01 1.54 ± 0.10 16.93 + 2.83
Heteromneles arbutifolia 84.83 ± 0.04 3.15 ± 0.48 21.83 ± 3.76
75
3
W (J/m ) E (MPa) E
3
(%)
0.107 ± 0.03 8.52 ± 2.5 18.59 ± 2.8
0.555 ± 0.16 40.87 ± 6.94 24.62 ± 4.0
DISCUSSION
This study demonstrates that there are both quantitative and
qualitative differences in biomechanical properties between the
mesophytic leaves of P. serrulata and the sclerophyllous
leaves of H. arbutifolia. Previous studies investigating the me-
chanical properties of leaves have either used simple systems
like grasses (Vincent, 1982, 1991; Greenberg et al., 1989) or
have restricted measurements to mesophyll tissue and minor
veins by using punch or shear type of techniques (Choong et
al., 1992; Edwards, Read, and Sanson, 2000). Our interests in
this experiment focused on the mechanical behavior of whole
leaves with complex venation and internal architecture. Tis-
sues in dicotyledonous leaves can vary widely in their me-
chanical properties (Vincent, 1990; Niklas, 1992). For exam-
ple, although properties such as leaf thickness and shape may
affect absolute values for punch, shear, or tensile strength mea-
surements, more often specific components such as vascular
bundles or the epidermis (Niklas, 1990, 1999) are responsible
for bearing the bulk of the stresses due to leaf positioning,
gravity, and wind effects and, thus, have correspondingly high-
er mechanical values than other tissues such as the mesophyll.
This can be illustrated by observing a wilting tomato leaf or
that of a deciduous tree such as maple or cherry where the
lamina remains upright solely due to the stability of the midrib
and major veins. Pressure volume curves on H. arbutifolia
leaves sampled in the fall revealed that this species exhibits
osmotic potentials below 3 MPa and turgor loss points of 85%
relative water content (Table 3) that may also depend upon the
age of the leaf and the season (Davis and Mooney, 1986a).
However, leaves of this species do not wilt even if dried to
substantially below the turgor loss point (R. A. Balsamo, per-
sonal observation). In contrast, mature leaves of P. serrulata
in the fall have osmotic potentials and turgor loss points of
about 1.5 MPa and 93%, respectively (Table 3), and show
signs of wilting even earlier (R. A. Balsamo, personal obser-
vation). This discrepancy in the mechanical properties of di-
cotyledonous leaves and the consequent complexity in inter-
pretations is probably one reason why there are so few studies
reported (Vincent, 1990; Edwards, Read, and Sanson, 2000).
However, the differences in leaf morphological behavior dur-
ing water loss is precisely the reason why the studies outlined
in this manuscript were undertaken on whole leaves and not
isolated tissue fragments.
Tensile strength experiments on whole leaves can be eval-
uated with respect to the absolute values obtained as well as
to the interaction of various components by analysis of the
resultant stress-strain curves. Although the sclerophyllous
leaves of H. arbutifolia were roughly 2.5-fold thicker and had
a 35% higher dry mass/fresh mass ratio (Table 1), with the
exception of the failure strain (for which there was a 25%
difference), the sclerophyllous leaves had biomechanical val-
ues that were four- to six-fold higher than the mesophyte
leaves (Table 2). This suggests that more than just leaf mor-
phology may be involved in the determination of structural
76 AMERICAN JOURNAL OF BOTANY
.11 . ! 'r . 11
f__ , I" -
,-- li,-t-
': , i::_14:5 f .14
..
Fis.23 ScnIng -eai
Figs. 2-3. Scanning electron micrograph of PHeteros serreslataleaf cross section. Top picture photographed at 500x, bottom picture photographed at 130X.
Bar = 50Bam. 3. Scanning electon micrograph of Heteromeles arbutifolia leaf cross section, Top picture photographed at 40.X, bottom picture photograpbed
at 80x. Bar = 50 l.Lm.
integrity. Leaves of H. arbutifolia are tougher and stiffer, but
also have a higher failure strain (hence they are more stretchy)
than the leaves of P. serrulata. These results may suggest one
reason why leaves of sclerophylls are more tolerant of dehy-
dration than their more mesic counterparts and why they do
not wilt nor. are they visibly damaged even when severely
water stressed.
Analysis df the stress-strain curves of individual leaf pulls
reveal qualitative differences between the leaves of the two
species studied. particularly the descending curve following
catastrophic failure of the lamina. In P. serrulata,these curves
are typically ragged with several smaller descending peaks
corresponding to snags of stiffer and stronger material that first
resist then ultimately fail (Fig. 1). Observation during exper-
imentation reveals these "holdout" structures to be the pri-
mary and secondary vascular bundles. In contrast, descending
curves for H, arbutifolia were linear and smooth, suggesting
a more isotropic material that snaps rather than tears (Fig. 1).
It was the difference in the stress-strain curves of these species
that prompted the examination of the leaf intemal anatomy.
Examination of the leaf architecture of both species revealed
significant differences in mesophyll cell size, shape, and po-
sitions within the lamina. Mesophyll cells of P. serrulatawere
much smaller and both palisade and spongy layers appeared
loosely packed when compared to H. arbutifolia (Figs. 2, 3).
In P. serrulata, the spongy mesophyll cells were isodiametric
in shape and in both SEM (Fig. 2) and light microscopy sec-
tions (Fig. 4) showed only small regions of adjacent cells that
were in direct physical contact. Conversely, cross sections
(Fig. 3) and paradermal sections (Fig. 5) of H. arbutifolia re-
[Vol. 90
b picture p h
vealed ameboid cells with large areas of direct cell contact and
an overall interconnected or "netlike" appearance. These cells
were also much larger than the spongy mesophyll of P. ser-
rulata. These results suggest that the qualitative differences
seen in the stress-strain curves of these two species may have
been influenced by the differences in internal anatomy of the
two leaf types. It is also possible that there may be differences
in the chemical compositions of the cell walls that may also
help to explain these qualitative differences (Vicre et al.,
11999). Further studies in this area are warranted.
In conclusion, these studies have demonstrated that despite
a similar gross morphology, the sclerophyllous leaves of H.
arbictifolia differ in their mechanical properties compared to
mesic leaves of P. serrulata and that the differences are likely
due to a complex interaction of leaf architecture, internal anat-
omy, and possibly cell wall chemistry. The wilting of mesic
leaves such as those of P. serrulatamay be due primarily to
their internal design and the anisotropic nature of their vas-
cular and ground tissues. The fact that sclerophyllous leaves
like H. arbutifolia do not wilt even under severe water stress
may be an important factor in their ability to survive episodes
of dehydration without mechanical damage. These studies sug-
gest that leaf anatomy, i.e., the specific interaction of meso-
phyll cells with each other may play a larger role in tolerance
to dehydration than has generally been acknowledged. A me-
chanically tougher and stiffer leaf could translate into an abil-
ity to survive water potentials and relative water contents that
fall below the turgor loss point of the tissues. Further, bio-
mechanical techniques may prove to be a useful predictive tool
in determining the behavior of leaves during drought. These
January 2003] BALSAMO ET AL.-RELATIONSHIP OF ANATOMY TO LEAF BIOMECHANICS
Figs. 4-5. Light microscope photograph of paradermal section of the
spongy mesophyll layer of Prunus serrulata. Photographed at 250X. Bar =
50 Aim. 5. Light microscope photograph of paradermal section of the spongy
mesophyll layer of Heteromeles arbutifolia. Photographed at 250X. Bar
50 l.m.
studies indicate that subsequent investigations into the anato-
my and mechanical behavior of sclerophyll leaves during de-
hydration may provide further clues for explaining differences
in drought tolerance.
LITERATURE CITED
CHOONG, M. E. P. W. LUCAS, J. S. Y. ONG, B. PEREIRA, H. T W. TAN, AND
1. M. TURNER. 1992. Leaf fracture toughness and sclerophylly: their
correlations and ecological implications. New Phzytologist 121: 597-610.
DAvIS, S. D., K. J. KOLB, AND K. P. BARTON. 1998. Ecophysiological pro-
cesses and demographic patterns in the structuring of Califomia chap-
arral. Ecological Studies 136: 297-310.
DAVIS, S. D., AND H. A. MOONEY. 1986a. Water use pattems of four co-
occurring chaparral shrubs. Oecologia (Berlin) 70: 172-177.
77
DAVIS, S. D., AND H. A. MOONEY. 1986b. Tissue water relations of four co-
occurring chaparral shrubs. Oecologia (Berlin) 70: 527-535.
EDWARDS, C., J. READ, AND G. SANSON. 2000. Characterising sclerophylly:
some mechanical properties of leaves from heath and forest. Oecologia
(Berlin) 123: 158-167.
ESAU, K. 1977. Anatomy of seed plants, 2nd ed. John Wiley & Sons, New
York, New York, USA.
GORDON, J. E. 1978. Structures, or why things don't fall down. Penguin
Books Ltd. Harmondsworth, UK.
GRATANI, L., E. FIORENTINO, A. KUBOVA, AND P. MARZI. 1989. Effect of
microclimate on ecophysiological features of some sclerophyllous spe-
cies. Photosynthetica 23: 230-233.
GREENBERG, A. R., A. MEHLING, M. LEE, AND J. H. BOCK. 1989. Tensile
behavior of grass. Jounial of Materials Science 24: 2549-2554.
LARCHER, W. 1995. Physiological plant ecology, 3rd ed. Springer-Verlag,
Berlin, Germany.
LEVITT, J. 1980. Responses of plants to environmental stresses. Academic
Press, New York, New York, USA.
LOVELESS, A. R. 1961. A nutritional interpretation of sclerophylly based on
differences in the chemical composition of sclerophyllous and meso-
phytic leaves. Annals of Botany 25: 168-184.
MAxIMOV, N. A. 1929. The plant in relation to water. Walter George Allen
& Unwin, London, UK.
NIKLAS, K. 1990. The mechanical significance of clasping leaf sheaths in
grasses: evidence from two cultivars of Avena sativa. Annals of Botany
65: 505-512.
NIKLAS, K. 1992. Plant biomechanics: an engineering approach to plant form
and function. University of Chicago Press, Chicago, Illinois, USA.
NIKLAS, K. 1996. Differences between Acer saccharinim leaves from open
and wind-protected sites. Annals of Botany 78: 61-66.
NIKLAS, K. 1999. A mechanical perspective on foliage leaf form and function.
New Phytologist 143: 19-31.
RAo, T. A. 1992. Compendium of foliar sclereids in angiosperms. Wiley
Eastern Limited, New Debli, India.
SCHLESINGER, W. H., 1. T. GRAY, D. S. GILL, AND B. E. MAHALL. 1982.
Ceanothus megacarpus chaparral: a synthesis of ecosystem processes
during development and annual growth. Botanical Review 48: 71-117.
SEDDON, G. 1974. Xerophytes, xeromorphs and sclerophylls: the history of
some concepts in ecology. Biological Journal of the Linnean Society 6:
65-87.
SMITH, W. K., T. C. VOGELMANN, E. H. DELUCfA, D. T. BELL, AND K. A.
SHEPARD. 1997. Leaf form and photosynthesis: do leaf structure and
orientation interact to regulate internal light and carbon dioxide? Bio-
Science 47: 785-793.
SoRDo, C. M., G. PADILLA, AND L. A. ROMERO. 1982. Salt excretion in the
mangrove Avicennia germinans. Environmental Science Research 57:
383-392.
SPuRR, A. R. 1969. A low-viscosity epoxy resin embedding medium for
electron microscopy. Journal of UltrastructuralResearch 26: 31-43.
TURNER, I. M. 1994. Sclerophylly: primarily protective? FunctionalEcology
8: 669-675.
TURNER, I. M., M. E CHOONG, H. T. W. TAN, AND P. W. LUCAS. 1993. How
tough are sclerophylls? Annals of Botany 71: 343-345.
UTRILLAS, M. J., AND L. ALEGRE. 1997. Impact of water stress on leaf anat-
omy and ultrastructure in Cynodon dactylon (L.) Pers. under natural con-
ditions. InternationalJournal of Plant Science 158: 313-324.
VICRE, M., H. W. SHERwIN, A. DRIOUICH, M. A. JAFFER, AND J. M. FARRANT.
1999. Cell wall characteristics of hydrated and dry leaves of the resur-
rection plant Craterostigma wilmsii, a microscopical study. Journal of
Plant Physiology 155: 719-726.
VINcENT, J. E V. 1982. The mechanical design of grass. Journal of Materials
Science 17: 856-860.
VINCENT, J. F V. 1990. Fracture properties of plants. In Advances in botanical
research, vol. 17, 235-287. Academic Press. New York, New York, USA.
VINCENT, J. E V. 1991. Strength and fracture of grasses. Journalof Materials
Science 26: 1947-1950.
VOGEL, S. 1989. Drag and reconfiguration of broad leaves in high winds.
Journal of Experimental Botany 40: 941-948.
COPYRIGHT INFORMATION

TITLE: Leaf Biomechanics, Morphology, and Anatomy of the

Deciduous Mesophyte Prunus serrulata (Rosaceae) and
the Evergreen Sclerophyllous Shrub Heteromeles
arbutifolia (Rosaceae)
SOURCE: Am J Bot 90 no1 Ja 2003
WN: 0300104652008

The magazine publisher is the copyright holder of this article and it

is reproduced with permission. Further reproduction of this article in
violation of the copyright is prohibited.

Copyright 1982-2003 The H.W. Wilson Company. All rights reserved.