Tree Physiology 22, 747–761
© 2002 Heron Publishing—Victoria, Canada
Dependence of needle architecture and chemical composition on
canopy light availability in three North American Pinus species with
contrasting needle length
ÜLO NIINEMETS,1,2 DAVID S. ELLSWORTH,3 ALJONA LUKJANOVA4 and MARI TOBIAS4
1
Department of Plant Physiology, Institute of Molecular and Cell Biology, University of Tartu, Riia 23, Tartu 51011, Estonia
2
Author to whom correspondence should be addressed ([email protected])
3
School of Natural Resources and Environment, 430 E. University Ave., University of Michigan, Ann Arbor, MI 48109-1115, USA
4
Department of Ecophysiology, Institute of Ecology, Tallinn University of Educational Sciences, Kevade 2, Tallinn 10137, Estonia
Received August 5, 2001; accepted January 5, 2002; published online July 2, 2002
Summary Morphology and chemical composition of needles of shade-intolerant southern conifers (Pinus palustris Mill.
(mean needle length ± SD = 29.1 ± 4.1 cm), P. taeda L. (12.3 ±
2.9 cm) and P. virginiana Mill. (5.1 ± 0.8 cm)) were studied to
test the hypothesis that foliage acclimation potential to canopy
light gradients is generally low for shade-intolerant species,
and in particular, because of mechanical limitations, in species
with longer needles. Plasticity for each needle variable was defined as the slope of the foliar characteristic versus irradiance
relationship. A novel geometrical model for needle area and
volume calculation was employed for the three-needled species P. palustris and P. taeda. Needle thickness (T ) strongly
increased, but width (W ) was less variable with increasing
daily integrated quantum flux density averaged over the season
(Qint), resulting in changes in cross-sectional needle shape that
were manifested in a positive relationship between the total to
projected needle area ratio (AT /AP) and Qint in the three-needled
species. In contrast, cross-sectional needle geometry was only
slightly modified by irradiance in the two-needled conifer
P. virginiana. Needle dry mass per unit total needle area (MT)
was positively related to Qint in all species, leading to greater
foliar nitrogen contents per unit area at higher irradiances. Separate examination of the components of MT (density (D) and
the volume (V) to AT ratio; MT = DV/AT) indicated that the positive effect of light on MT resulted solely from increases in V/AT,
i.e., from increases in the thickness of foliage elements. Foliar
chlorophyll content per unit mass increased with increasing
Qint, allowing an improvement in light-harvesting efficiency in
low light. The variables characterizing needle material properties (D, the dry to fresh mass ratio, and needle carbon content
per unit mass) were generally independent of Qint, suggesting
that needles were less stiff and had greater tip deflections under
their own weight at lower irradiances because of smaller W and
T. Comparisons with the literature revealed that plasticity in foliar characteristics tended to be lower in the studied shadeintolerant species than in shade-tolerant conifers, but plasticity
among the investigated species was unaffected by needle
length. However, we argue that, because of mechanical limitations, plastic changes in needle cross section in response to low
irradiance may decrease rather than increase light-interception
efficiency in long-needled species.
Keywords: carbon content, chlorophyll content, leaf area estimation, leaf density, leaf dry mass per unit area, nitrogen content, Pinus palustris, Pinus taeda, Pinus virginiana.
Introduction
Structural characteristics of conifer needles are often strongly
related to long-term light availability gradients through the
canopy as well as across the forest. In particular, a positive
scaling of needle dry mass per unit needle area (MA ) with increasing long-term daily integrated quantum flux density
(Qint) has frequently been observed in conifers (Niinemets and
Kull 1995, Sprugel et al. 1996, Niinemets 1997b, Stenberg et
al. 1998, 1999, Niinemets et al. 2001). Such highly determined
spatial variation patterns in MA are manifestations of an important canopy acclimation response that enhances the foliage
photosynthetic capacity in high light and the light-harvesting
efficiency in low light, and thereby improves the whole-canopy carbon gain for a given biomass investment in leaves
(Gutschick and Wiegel 1988, Niinemets et al. 1998, Bond et
al. 1999). With optimality arguments it is possible to demonstrate that MA should be distributed proportionally to Qint for
optimal leaf biomass allocation in the canopy, which maximizes the carbon gain of the entire canopy (Badeck 1995,
Sands 1995). The condition of proportionality between MA
and foliage light environment for optimal allocation of resources suggests that the relationships between MA and Qint
should be conservative. However, empirical studies show important intraspecific (Niinemets et al. 2001) and interspecific
(Chen et al. 1996, Chen 1997) differences in MA versus Qint relationships, indicating large variability in foliar plasticity, the
limits to and controls of which are not yet entirely understood.
748
NIINEMETS, ELLSWORTH, LUKJANOVA AND TOBIAS
Chen et al. (1996) and Chen (1997) proposed that needle
morphological plasticity is lower in shade-intolerant species.
This suggestion arose from the observation that the slope of
the MA versus light relationship was lower in the shade-intolerant conifer Pinus contorta Dougl. ex Loud. var. latifolia
Engelm. than in the more tolerant species Pseudotsuga menziesii (Mirb.) Franco (Chen et al. 1996), and that MA of the
shade-intolerant species Pinus ponderosa Dougl. was not significantly related to irradiance, contrary to the light relations
of MA in two more shade-tolerant species studied (Chen 1997).
Given that shade-tolerant trees thrive under the shade of the
canopy formed by intolerant species in the early stages of
succession, and dominate the upper canopy in the final stages
of succession, they generally encounter a larger gradient in
light availability during their life span than intolerant trees.
Thus, a greater potential for foliage modification in shade-tolerant species may be an evolutionary consequence of the wider
light gradients these species’ experience.
Alternatively, inherent differences in needle structure and
associated constraints may explain the lower plasticity in the
shade-intolerant species examined. Among Pinaceae, shadetolerant species from the genera Abies, Tsuga and Picea have,
almost without exception, shorter needles than shade-intolerant species from the genus Pinus. Moreover, shade-tolerance
within the genus Pinus is lower for species with longer needles
(e.g., P. taeda, P. palustris) than for species with shorter needles (e.g., P. strobus L., see Baker 1949). Given that mechanical
structures become less efficient with increasing cantilever
length, because the mass is located farther from the axis of rotation, biomass costs for efficient light harvesting vary
strongly with needle length. Needle bending under the load of
a needle’s own weight potentially increases with increasing
needle length (Niklas 1991), and accordingly, greater needle
cross-sectional areas or greater investments in less mechanically elastic support structures are required in long-needled
species to keep the needles in positions that allow efficient
light harvesting. Thus, in long-needled species, maximization
of canopy light-interception efficiency depends on MA in a
complex manner. A decrease in MA results in greater total
needle area for a given biomass investment in leaves, thereby
increasing potential plant light interception. However, lower
MA is also equated with a smaller needle cross-sectional area,
which leads to lower needle inclination angles and light-harvesting efficiency per unit needle area. We suggest that a
trade-off between low MA, which is compatible with large foliar area, and high MA, which allows efficient exposure of this
area, results in low plasticity in MA in long-needled species.
Contrary to the hypothesis that foliage plasticity is lower in
long-needled, shade-intolerant species, Bond et al. (1999) observed no significant differences in the slopes of the MA versus
irradiance relationships between the relatively shade-tolerant
short-needled species Pseudotsuga menziesii and Tsuga
heterophylla (Raf.) Sarg. and the shade-intolerant long-needled species P. ponderosa. Given that Chen (Chen et al. 1996,
Chen 1997) investigated trees of the same age in the same
stand, but Bond et al. (1999) studied each species in a different
stand, and tree ages and heights varied between the selected
stands, material heterogeneity may partly explain the discrepancy between the studies. For example, it has recently been
demonstrated that site fertility may affect foliar plasticity in
conifers (Niinemets et al. 2001). In addition, the aforementioned studies (Chen et al. 1996, Chen 1997, Bond et al. 1999)
investigated only needle dry mass per unit projected area (MP),
which is the product of dry mass per unit total area (MT) and
the total to projected area ratio (AT /AP ; Niinemets and Kull
1995, Niinemets 1997b). Although the needle cross-sectional
shape has generally been considered to be constant throughout
the canopy, there is evidence of light-related changes in the
AT /AP ratio in several conifers (Niinemets and Kull 1995,
Niinemets 1997b, Niinemets et al. 2001). Needle dry mass per
unit area is also a product of needle density (D) and the needle
volume to total area ratio (V/AT), both of which may vary independently along environmental gradients (Niinemets and Kull
1995, Niinemets 1999, 2001). It is functionally important to
distinguish between these variables, because V/AT scales positively and D scales negatively with needle photosynthetic potentials (Niinemets 1999), but D is also positively associated
with needle mechanical stability (Niklas 1991) and foliage capacity for water extraction from drying soil (Niinemets 2001).
Because the same MP value may be achieved with different
combinations of D, V/AT and AT/AP, but the functional relevance of a specific MP value depends on the magnitude of each
of its components, discrimination among various hypotheses
of needle-level plasticity requires detailed investigation of
needle cross-sectional geometry.
We studied needle structure in three southern Pinus species
of contrasting needle length and light requirement to test the
hypothesis that foliage structure responds less plastically to
light environment in shade-intolerant than in shade-tolerant
species, and that among shade-intolerant species, plasticity
decreases with increasing needle length. In addition, leaf carbon contents were measured to provide an estimate of support
biomass, and foliage nitrogen and chlorophyll contents were
estimated to characterize modifications in foliage photosynthetic function. Pinus palustris, with the longest needles (20–
45 cm), is very intolerant of shade, whereas P. virginiana
(needles 3–8 cm long) and P. taeda (needles 8–18 cm long)
are intermediate to shade-intolerant species (Baker 1949).
Pinus virginiana and P. taeda co-exist in a number of forest
types; P. virginiana outcompetes P. taeda on drier sites, but is
less competitive on sites with poor drainage (Baker and
Langdon 1990, Carter and Snow 1990). Although Pinus palustris is occasionally a minor associate of P. taeda forests, this
species grows best in the complete absence of competition
(Boyer 1990) and can survive naturally only on nutrient-poor
soils with regular ground fires that control understory vegetation and soil organic matter content (Platt et al. 1988, Harcombe et al. 1993). The relative irradiance on the forest floor
(fraction of above-canopy light) of fully stocked stands varies
from about 0.1 to 0.2 in forests of P. taeda and P. virginiana
(Perry et al. 1969, Sinclair and Knoerr 1982, Pataki et al. 1998,
Sampson and Allen 1998), but is generally more than 0.5 in savanna-like stands of P. palustris (Brockway and Outcalt 1998,
Shelton and Cain 2000).
TREE PHYSIOLOGY VOLUME 22, 2002
NEEDLE ARCHITECTURE AND CHEMISTRY OF NORTH AMERICAN CONIFERS
Materials and methods
Study sites
The research was conducted in two stands in North Carolina at
the beginning of December 1998. The P. palustris and P. taeda
site (Southeast Tree Research and Education Site, SETRES)
was located in the Carolina Sandhills, Scotland County
(34°55′ N, 79°30′ W). The soil at the SETRES site is a deep
sand with low water-holding capacity and poor nutrient availability (Dougherty et al. 1998) that is classified taxonomically
as a siliceous, thermic Psammentic Hapludult from the
Wakulla series (Soil Survey Division 2001). The stand was
planted in 1985 with a ten-family mix of NC Piedmont
P. taeda seedlings after clear-cutting a 65-year-old P. palustris
forest. In 1992, the site was thinned to a density of 1260 stems
ha –1. In addition to P. taeda, a few naturally seeded individuals
of P. palustris were also present at the site during our study. A
long-term fertilization and irrigation experiment was established at the site in 1992 (Murthy and Dougherty 1997,
Dougherty et al. 1998), but foliage was collected only from the
control plots in our investigation. The P. palustris and P. taeda
trees sampled at the control plots were 7–9 m tall and 16 years
old.
Pinus virginiana was studied in the Blackwood Division of
the Duke Forest in Orange County (35°58′ N, 79°5′ W). The
naturally established trees, with a height of 8–9 m, were sampled at the edge of a 16-year-old planted P. taeda forest. The
soil is a clay loam belonging to the Enon series (thermic Ultic
Hapludalfs; Soil Survey Division 2001), which is common to
the uplands of the southeastern USA. Although these soils are
relatively infertile, tree growth data and foliar nutrient concentrations in P. taeda indicate that this site was more fertile than
the Carolina Sandhills site (cf. Murthy et al. 1996, Dougherty
et al. 1998 and Myers et al. 1999). This research area has been
thoroughly described elsewhere (Ellsworth et al. 1995, Ellsworth 1999, 2000).
The climate at both sites is characterized by humid, warm
summers and mild winters. Mean annual air temperature is
17 °C at the SETRES site and 15.5 °C at the Duke Forest site.
In the Carolina Sandhills, mean annual rainfall is 1210 mm,
and periods of drought may occur in late summer. In the Duke
Forest, precipitation (annual mean of 1150 mm) is evenly distributed throughout the year (Ellsworth et al. 1995, Ellsworth
2000).
Foliage sampling
In all species, shoots in terminal branch positions were harvested along the canopy light gradient, immediately sealed in
plastic bags and stored overnight at 4 °C before analysis of foliage structural characteristics.
Under the climatic conditions of the study sites, the studied
species form two to five growth flushes during the growing
season, depending on tree age (Baker and Langdon 1990, Harrington 1991). We generally observed three growth flushes in
all species, independent of shoot position in the canopy. To account for possible effects of needle age on foliage structural
and chemical variables, all flushes of current-year needles
749
were analyzed separately in P. palustris and P. virginiana. In
P. taeda, the last cohort of previous-year needles was also analyzed in addition to the current-year needles. Given that at both
sites bud break for the first flush generally occurs in March
(Higginbotham 1974, Dougherty et al. 1998) and the last current-year flush starts growth in July or at the beginning of August (authors’ unpublished observations), needles were 4–
9 months old in P. palustris and P. virginiana and 4–16 months
old in P. taeda at the time of sampling in December 1998. According to phenological studies, needle length and mass
growth for all current-year cohorts of P. taeda are completed
by October in the study area (Wells and Metz 1963, Kinerson
et al. 1974, Ellsworth 2000), and thus, fully mature foliage was
sampled in all instances.
Quantification of long-term light availability
For estimation of seasonal (April 1 to October 31) mean daily
integrated photosynthetically active quantum flux densities
(Qint; mol m –2 day –1) in the canopy, several hemispherical photographs were taken above each shoot with a Nikon camera
equipped with a Nikkor 8-mm fish-eye f/8 lens and a high resolution slide film. The hemispherical photographs were taken
immediately before shoot sampling. From these photographs,
we computed cosine-corrected estimates of fractional penetration of diffuse (Idif) and direct (Idir) solar radiation according to
Anderson (1964), whereas the calculation of sun paths in the
hemisphere for various dates was carried out according to
Campbell and Norman (1998). Specifically, Idif is the ratio of
diffuse solar radiation above the shoot to diffuse radiation
above the canopy for uniformly overcast sky conditions, and
Idir is the ratio of mean potential direct radiation above the
shoot to direct radiation above the canopy between the summer solstice and 1 month from the summer solstice. Values of
Idir and Idif estimated from replicate fish-eye photographs were
averaged for each shoot. The “global site factor” (the fractional penetration of global solar radiation, Isum) is given as:
I sum = pdif Idif + (1 – pdif)Idir, where pdif is the ratio of diffuse to
global radiation above the canopy (Niinemets and Kull 1998,
Niinemets et al. 2001). The values of Qint for each sampled
shoot were determined as the product of Isum at the sample location and the seasonal mean daily integrated quantum flux
0
density above the canopy (Qint
). Direct measurements with
0
estimate of 43.0 mol m –2
quantum sensors provided a Qint
–1
day for the period April 1 to October 31 (Ellsworth 2000).
The maximum, above-canopy, integrated daily quantum flux
densities were about 70 mol m –2 day –1 in June and May, and
about 30 mol m –2 day –1 at the end of October (Ellsworth
2000).
Determination of needle area and volume
Needle width (W) and thickness (T ) in all species, and needle
side length (the length of the internal needle face, L s; Figure 1)
in P. palustris and P. taeda, were measured in six to 10 needles
from each shoot flush with precision calipers (± 0.05 mm) in
the central part of the needle, and these estimates were employed to compute the needle cross-section circumference (C)
and area (AC). Needle length (L n) was measured with a ruler,
TREE PHYSIOLOGY ONLINE at http://heronpublishing.com
750
NIINEMETS, ELLSWORTH, LUKJANOVA AND TOBIAS
Figure 1. Needle geometrical models
of Pinus palustris and Pinus taeda. (A)
The shape of the needle cross section
was approximated by a sector of a circle with a central angle (2α) of 120°
(2/3πrad, where rad is the approximation of radius). Thus, both needle
thickness (T ) and side length (L s) are
equal to circle radius (r), and only
measurements of either needle width
(W ) or T are necessary to compute the
circumference (C) and area (AC) of the
needle cross section (Equations 1–4).
(B) The needle cross section was
described as a pie-shaped piece of a
circle, the L s and T of which are not
necessarily equal to r. Determination
of C and AC in (B) (Equations 5–10)
requires measurement of all three variables: T, W and L s.
and needle projected area (AP) was computed as WL n, total
area (AT) as CL n and needle volume (V) as AC L n. The product
WL n provides a rough approximation of needle capacity for interception of direct irradiance. For a more advanced estimate
of needle light-harvesting efficiency, the dependence of AP on
needle specific orientation in the shoot should also be taken
into account (Grace 1987).
In the two-needled species P. virginiana, the needle crosssectional geometry was approximated by a half-ellipse, and
AT, AP and V were calculated as described in Niinemets et al.
(2001). The fascicles of P. palustris and P. taeda consist of
three needles, and the needle cross-sectional shape in these
species has often been approximated by a circle sector with a
central angle of 120° (Kozlowski and Schumacher 1943,
Harms 1971, Wood 1971, Johnson 1984, Shelton and Switzer
1984, Johnson et al. 1985, Samuelson et al. 1992, Hultine and
Marshall 2001, Figure 1A). For a circle sector, the radius (r) is
equal to L s as well as to T (Figure 1). Thus, both C and AC
can be determined only from measurements of either T or W
(Equations 1–4 in Figure 1A). Assuming that the needle
cross-sectional shape can be approximated by a half-circle,
this model can also be employed for two-needled species (L s =
W/2 and α = 90°, Figure 1).
The suitability of a sectorial needle model for the studied
pines can be tested by simple calculations. For a circle sector
with a central angle (2α) of 120°, the ratio of needle thickness
to needle width is a constant: (2sin(60°)) –1 ≅ 0.577 (Equation 1
in Figure 1A). However, the T/W ratio varied from 0.38 to 0.69
(mean ± SD = 0.54 ± 0.06) in P. palustris and from 0.32 to 0.72
(mean ± SD = 0.46 ± 0.08) in P. taeda, suggesting that the assumption of constant needle cross-sectional shape was inappropriate. Moreover, plots of T versus L s (Figure 2A) indicated
that T was not equivalent to L s in either species. Estimates of L s
calculated from W (Equation 1 in Figure 1A) also deviated
systematically from measured values, further demonstrating
that needle cross section cannot be described as a sector of a
fascicle circle in these species. In fact, examination of the data
revealed that the needle cross-sectional shape may be more accurately characterized as a pie-shaped piece of a circle, the
thickness and side length of which are not necessarily equal to
the radius of the fascicle circle (Figure 1B). Given that the assumption of constant needle shape was disproved, Equations 5–10 (Figure 1B) were always used for needle area and
volume calculations in P. palustris and P. taeda. Although
Equations 5–10 (Figure 1B) can be employed only if T < W,
this condition was always fulfilled for both species.
TREE PHYSIOLOGY VOLUME 22, 2002
NEEDLE ARCHITECTURE AND CHEMISTRY OF NORTH AMERICAN CONIFERS
751
(M) was determined, and needle dry mass per unit total area
(MT) was calculated. Because MT is a product of the needle
volume to total area ratio (V/AT) and needle density (D;
Niinemets 1999), which may vary independently of each other
(Witkowski and Lamont 1991, Niinemets 2001), both V/AT
and D were also computed to gain detailed insight into the canopy controls on MT.
Total needle nitrogen (NM) and carbon (CM) contents were
determined with an elemental analyzer (CHN-O-Rapid, Foss
Heraeus, Hanau, Germany). Given that proteins contain more
carbon (53.5%; see Vertregt and Penning de Vries 1987) than
needles on average (means ± SD observed in the current study
were 47.9 ± 0.7% for P. palustris, 49.8 ± 0.6% for P. taeda and
49.3 ± 0.6% for P. virginiana), and that needle nitrogen content provides an estimate of foliar protein content, changes in
foliar nitrogen content may obscure variations in structural
needle carbon content. To test this possibility, protein-free carbon content (“structural” carbon, CS) was computed as: CM –
6.25(53.5NM/100), where the constant 6.25 converts nitrogen
content to protein content.
Fresh needle samples were ground in 80% (v/v) buffered
(50 mM sodium phosphate buffer, pH 7.8) aqueous acetone,
and foliage chlorophyll concentrations were determined with
a Cadas 100 spectrophotometer (Bruno Lange, Düsseldorf,
Germany) from the equations given in Porra et al. (1989).
Data analyses
Figure 2. Correlations between measurements of needle thickness (T )
and needle side length (L s) (A), and between measured and estimated
values of L s (B and C) in P. palustris and P. taeda. In (B), L s was calculated from needle width assuming that L s is equal to the radius of
the fascicle circle (Equation 1 in Figure 1A). In (C), L s was determined from needle width and thickness, assuming H = aT (Figure
1B). The mean value of a of 0.255 was determined from nonlinear regression of L s = ( T (1 − a ))2 + W 2 / 4 (cf. Equation 5 in Figure 1B). In
(A), individual linear regressions were computed for both P. palustris
and P. taeda. In (B) and (C), species were pooled, and the linear regressions between measured and estimated values of L s were forced
through zero. The dotted lines in (A) and (B) demonstrate the y = x relationships.
Estimation of needle dry mass per unit area, and needle
carbon, nitrogen and chlorophyll contents
Needles were dried at 70 °C for at least 48 h, their dry mass
Relationships of foliar chemical and structural characteristics
with Qint, as well as between foliage structural and chemical
variables, were fitted by linear regression analyses, and all relationships were considered significant at P < 0.05. The species-specific regressions were compared by co-variation analyses when Qint was significant, or by analyses of variance
when Qint was insignificant (P > 0.05). Whenever the interaction term was insignificant in the separate slope model, the
data were further explored with a common slope model.
The maximum estimates of Qint corresponding to an abovecanopy environment were similar for all species, but the minimum Qint values differed among species. The lowest value of
incident Qint was 11.1 mol m –2 day –1 (Isum = 0.26) for P. palustris, 15.4 mol m –2 day –1 (Isum = 0.36) for P. taeda and
4.4 mol m –2 day –1 (Isum = 0.10) for P. virginiana. Because of
species differences in the range of within-canopy Qint, the regressions were compared over a finite range (Niinemets and
Kull 1998) by removing all data for which Qint < 10 mol m –2
day –1 from the comparisons. Although there was evidence of a
nonlinear response of foliage structural and chemical variables in P. virginiana at low irradiance, all relationships were
essentially linear over the truncated range.
Initially, all needle cohorts were analyzed separately. However, analyses of variance demonstrated that needle age significantly affected only needle density and carbon content (P >
0.05 for all other comparisons). Therefore, all data for a given
species were pooled and single regressions were fitted to the
lumped data, except where noted.
TREE PHYSIOLOGY ONLINE at http://heronpublishing.com
752
NIINEMETS, ELLSWORTH, LUKJANOVA AND TOBIAS
Results
Various estimates of needle area and volume in P. palustris
and P. taeda
Different needle architectural models (cf. Figures 1A and 1B)
provided similar mean (± SD) estimates of total needle area
(AT). In P. palustris, the mean estimate of AT according to the
model in Figure 1A was 10.8 ± 1.9 cm2, whereas AT = 10.7 ±
1.8 cm2 according to the model in Figure 1B. In P. taeda, the
corresponding means were 3.9 ± 1.1 cm2 (Figure 1A) and 4.0 ±
1.2 cm2 (Figure 1B). Despite similar species-specific means,
there was a systematic under- or overestimation of needle area
by Equations 1–4 that was strongly related to the T/W ratio
(Figure 3A). Moreover, the discrepancies were considerably
larger for needle volume estimates (Figure 3B), especially for
P. taeda (0.053 ± 0.019 cm3 versus 0.066 ± 0.024 cm3). These
systematic differences further indicate that derivation of needle volume and area estimates from T or W alone (Figure 1A) is
not necessarily adequate.
Although the needle architectural model depicted in Figure 1B is more germane for needles of P. palustris and P. taeda
than the model of fixed needle shape (Figure 1A), application
of the more complex needle architectural model requires the
measurement of three needle characteristics. This may reduce
the precision of the area and volume estimates because of error
propagation. In particular, detection of the boundary between
the two needle sides (Figure 1) is difficult, and it is likely that
estimates of L s are less precise than measurements of T and W.
Therefore, we also explored the possibility of determining L s
from measurements of T and W. As a geometrical idealization,
L s is a side length, and W and T are the diagonals of a kite (i.e.,
a quadrilateral with two pairs of distinct adjacent sides equal in
length (Figure 1B)), and L s is given either as W/(2sinβ), where
β is the angle between L s and T, or as:
Ls = ( T − H ) 2 + W 2 / 4 ,
(1)
where H is the distance from the needle surface to the intersection of kite diagonals T and W (Figure 1B). Assuming that H is
a fixed fraction of T, i.e., H = aT, nonlinear regression equations of the form:
L s = ( T (1 − a )) 2 + W 2 / 4 ,
(2)
were fitted to the data. The estimates of a of 0.267 for P. palustris and 0.243 for P. taeda were similar, and a single fit to
pooled data yielded a = 0.255 (r 2 = 0.60, P < 0.001). This estimate of a resulted in a close correspondence between measured and modeled values of L s (Figure 2C). Thus, after determination of an appropriate a value, needle surface area and
volume may be accurately calculated from measurements of
W and T alone.
Light effects on needle size and needle cross-sectional
geometry
Figure 3. Correlations between the T/W ratio and (A) the ratio of total
needle area or (B) volume computed for flexible needle geometry
(Figure 1B) to the estimates calculated assuming a fixed needle shape
(Figure 1A). Both species were pooled, and the data were fitted by a
linear regression in A, and by a power function in B (r 2 = 0.95, P <
0.001 for both regressions). The arrows denote the value of T/W of
0.577, which is the constant estimate obtained when both T and L s are
equal to the radius of the circle sector (Figure 1A).
In the long-needled species P. palustris and P. taeda, needle
length (L n) was independent (r 2 = 0.03, P > 0.4 for both) of
Qint, but increased with increasing Qint in the short-needled
species P. virginiana (r 2 = 0.46, P < 0.001). Needle thickness
was positively related to Qint in all species (Figure 4A), and
width in P. taeda and P. virginiana (Figure 4B), leading to a
greater needle circumference at higher irradiance in all species
(r 2 = 0.26, P < 0.05 for P. palustris; r 2 = 0.30, P < 0.005 for
P. taeda; and r 2 = 0.70, P < 0.001 for P. virginiana). However,
AT scaled with Qint only in P. virginiana (r 2 = 0.65, P < 0.001).
The linear needle measures (L n, T and W) and AT were
greatest in P. palustris (Table 1). Needle length and AT were intermediate in P. taeda, whereas P. virginiana had needle width
intermediate between that of the other two species (Table 1).
The differences in needle thickness between P. taeda and
P. virginiana were dependent on Qint (Figure 4A).
The T/W ratio increased with increasing Qint in species with
a pie-shaped needle geometry (Figure 1) (r 2 = 0.63 for P. palustris and r 2 = 0.43 for P. taeda; P < 0.001 for both), indicating
that the needle cross-sectional shape was modified by irradi-
TREE PHYSIOLOGY VOLUME 22, 2002
NEEDLE ARCHITECTURE AND CHEMISTRY OF NORTH AMERICAN CONIFERS
753
Figure 4. Effects of seasonal
mean integrated daily quantum
flux density (Qint) on (A)
needle thickness, (B) width,
(C) central angle of a circle
sector (2α, see Figure 1 for
definition of needle characteristics) and (D) total to projected needle area ratio (AT/AP)
in three Pinus species. The
data were fitted by linear regression, and only significant
regressions (P < 0.05) are depicted. Symbols: 䊉,— = Pinus
palustris; 䊊, · · · = P. taeda;
and 䉭, – – – = P. virginiana.
Values of Qint were calculated
as seasonal means between
April 1 and October 31.
ance. In P. virginiana, the needle geometry of which can be approximated by a circle segment or by a half-ellipse, the T/W
ratio was negatively related to Qint (r 2 = 0.18, P < 0.01). Qualitative differences in the Qint versus T/W relationship between
P. virginiana and the other species were attributable to lower
responsiveness in needle thickness in P. virginiana than in
P. taeda (lower slope of T versus Qint relation according to separate slope ANCOVA analysis, P < 0.05), and to greater responsiveness in needle width in P. virginiana than in P. palustris (Figure 4B). Needle central angle (α, Figure 1B) scaled
positively with irradiance in both P. palustris and P. taeda
(Figure 4C), further indicating that important changes in the
Table 1. Interspecific variability in needle structural and chemical characteristics: results of one-way analyses of co-variance (common slope
model with species as main effect and irradiance as the covariate).1 Means followed by the same letter within the same row are not significantly
different (Bonferroni test, P > 0.05). Abbreviation: nd = not measured.
Variable2
Mean ± SD3
3
Central angle (2α; °)
Density (D; g cm –3)
Dry to fresh mass ratio (PD; g g –1)
Length (L n; cm)
Nitrogen content (NM; %)
Side length3 (L s; mm)
Total area (AT; cm2)
Total to projected area ratio (AT /AP)
Volume to total area ratio (V/AT; mm)
Width3 (W; mm)
1
2
3
Pinus palustris
Pinus taeda
Pinus virginiana
73.8 ± 4.6 a
0.485 ± 0.024 b
0.406 ± 0.016 b
29.1 ± 4.3 a
0.86 ± 0.10 c
1.00 ± 0.06 a
10.7 ± 1.18 a
2.344 ± 0.027 b
0.207 ± 0.014 a
1.57 ± 0.09 a
65.2 ± 4.2 b
0.56 ± 0.05 a
0.450 ± 0.022 a
12.3 ± 2.9 b
1.08 ± 0.11 b
0.84 ± 0.09 b
3.9 ± 1.1 b
2.279 ± 0.038 c
0.166 ± 0.020 b
1.38 ± 0.15 c
nd
0.467 ± 0.031 c
0.430 ± 0.016 a
5.3 ± 1.0 c
1.27 ± 0.11 a
nd
1.94 ± 0.51 c
2.477 ± 0.029 a
0.201 ± 0.020 a
1.46 ± 0.18 b
The interaction term (Qint × species) was insignificant (P > 0.05) for all variables except NM (P < 0.05) and AT /AP (P < 0.001). For these variables, P. virginiana had a lower slope than the other species (Figures 4D and 5C). Nevertheless, P. virginiana had larger values of NM and AT /AP
over the entire canopy light range.
Characteristics of needle cross section are defined in Figure 1.
The data were compared over a finite range of Qint of 10.0 to 42.2 mol m –2 day –1 (Materials and methods), and only data points with Qint ≥
10 mol m –2 day –1 were included in the comparisons and used for calculation of means.
TREE PHYSIOLOGY ONLINE at http://heronpublishing.com
754
NIINEMETS, ELLSWORTH, LUKJANOVA AND TOBIAS
needle cross-sectional shape occurred along the irradiance
gradient.
As a result of the modifications in needle cross-sectional geometry, the total needle area to projected needle area ratio
(AT /AP) was positively correlated with irradiance in P. palustris and P. taeda and negatively correlated with irradiance in
P. virginiana (Figure 4D).
larger in P. palustris than in P. virginiana. However, MP did not
differ between these species (P > 0.08). Although the slopes
were not significantly different among species, common-slope
ANCOVA demonstrated important species differences in the
V/AT ratio (Figure 5B), D and PD (Table 1).
Needle dry mass per unit area, its components, and the dry
to fresh mass ratio in relation to irradiance
Needle nitrogen content per unit dry mass (NM) was positively
related to irradiance in P. palustris and P. taeda, but not in
P. virginiana (Figure 5C). Because of strong effects of Qint on
needle dry mass per unit area, needle nitrogen contents per
unit area (NA = MT × NM) scaled positively with irradiance in
all species (Figure 5D). According to a separate slope
ANCOVA analysis, the slope of the NA versus Qint relationship
was larger in P. taeda than in P. virginiana (P < 0.001), but the
interaction term was insignificant for other species combinations.
Carbon content per unit dry mass (CM) was independent of
irradiance in P. taeda and P. virginiana, but a strong correlation between CM and Qint was observed in P. palustris (Figure 6). The protein-free carbon content (structural carbon, CS)
was also positively correlated with Qint in P. palustris (r 2 =
0.75, P < 0.001), indicating that the increase in CM with Qint
did not result from increases in needle protein content with increasing irradiance (Figure 5C), and that CS truly scaled with
irradiance in this species. Like total carbon content, CS was independent of irradiance in P. taeda and P. virginiana (P > 0.4).
In both P. taeda and P. virginiana, needle chlorophyll contents (Figure 7A) and the chlorophyll to N ratio (Chl/N; r 2 =
Needle dry mass per unit total area (MT, Figure 5A) and dry
mass per unit projected area (MP = MT (AT /AP)) were strongly
related to Qint in all species. Because of positive effects of Qint
on the AT /AP ratio in P. palustris and P. taeda (Figure 4D), the
explained variances were slightly larger for MP versus Qint
(r 2 = 0.54 for P. palustris and r 2 = 0.68 for P. taeda; P < 0.001
for both) than for MT versus Qint (Figure 5A) relations.
Light affected the components of MT (needle density (D)
and volume to total area ratio (V/AT; MT = D(V/AT)) differently.
Needle density was independent of Qint in all species (greatest
r 2 value = 0.05; P > 0.2 for P. virginiana). Thus, the increase in
MT with irradiance resulted solely from the scaling of V/AT
with Qint in all species (Figure 5B). Similarly, the needle dry
mass to fresh mass ratio (PD) was independent of irradiance in
all species (greatest r 2 value = 0.09, P > 0.2 for P. palustris).
The slopes of the MT (Figure 5A) and MP versus Qint relationships were significantly larger for P. taeda than for the
other species (P < 0.005 for MT versus Qint and P < 0.05 for MP
versus Qint), whereas the slopes did not differ between P. palustris and P. virginiana (P > 0.2). At common Qint, MT was
Foliage nitrogen, chlorophyll and carbon contents in relation
to irradiance
Figure 5. Dependencies of (A)
needle dry mass per unit total
needle area (MT), (B) needle
volume to total area ratio
(V/AT), (C) nitrogen content
per unit dry mass (NM) and (D)
nitrogen content per unit total
area (NA ) on seasonal mean integrated daily quantum flux
density in three Pinus species.
Symbols: 䊉 = Pinus palustris;
䊊 = P. taeda; and 䉭 =
P. virginiana.
TREE PHYSIOLOGY VOLUME 22, 2002
NEEDLE ARCHITECTURE AND CHEMISTRY OF NORTH AMERICAN CONIFERS
755
Figure 6. Needle carbon content in relation to seasonal mean integrated daily quantum flux density in three Pinus species. Symbols:
䊉 = Pinus palustris; 䊊 = P. taeda; and 䉭 = P. virginiana.
0.59, P < 0.05 for P. taeda and r 2 = 0.56, P < 0.001 for P. virginiana) were negatively related to Qint. Thus, the efficiency of
light interception, and the investment of foliar nitrogen in light
harvesting, were greater in low light. The chlorophyll a/b ratio
increased with increasing irradiance (Figure 7B), suggesting
that the relative amount of centers of Photosystems I and II increased at the expense of light-harvesting chlorophyll protein
complexes with increasing Qint (e.g., Hikosaka and Terashima
1995). Total chlorophyll content (P > 0.8), chlorophyll a/b
ratio (P > 0.07) and Chl/N (P > 0.07) were not significantly
different between P. taeda and P. virginiana at a common
irradiance.
Effects of needle age on needle structural and chemical
variables
Needle age, which varied from 3 to 8 months (current-year
needle cohorts) in P. palustris and P. virginiana and from 3 to
16 months in P. taeda, generally did not alter foliage structural
and chemical characteristics according to one-way ANOVA
analyses (shoot flush number as the main effect). Needle density increased with increasing needle age in P. taeda and P. virginiana (Figure 8A), and total (Figure 8B) and protein-free
carbon content (Figure 8C) in P. taeda. However, CM and CS
tended to be negatively related to needle age in P. virginiana
(Figures 8B and 8C).
Correlations between needle density, dry to fresh mass ratio,
and needle carbon and nitrogen contents
For all species pooled, D was positively correlated with PD
(Figure 9A) and total (r 2 = 0.10, P < 0.01) and structural (Figure 9B) needle carbon contents. Total (r 2 = 0.37, P < 0.001)
and structural (r 2 = 0.27, P < 0.001) needle carbon contents
were also positively related to PD, suggesting that changes in
needle structural carbon contents may provide an explanation
for species differences in PD and D. Moreover, correlations be-
Figure 7. Needle chlorophyll content per unit dry mass (A) and chlorophyll a/b ratio (B) in relation to seasonal mean integrated daily
quantum flux density in Pinus taeda and P. virginiana. Symbols: 䊉 =
Pinus palustris; 䊊 = P. taeda; and 䉭 = P. virginiana.
tween PD, D, CM and CS were also significant for P. taeda
alone, with r 2 values ranging from 0.19 to 0.33 (P < 0.05), and
the relationship between PD and D (r 2 = 0.23, P < 0.05) was
significant for P. palustris.
There was evidence of a negative relationship between NM
and CS (r 2 = 0.10, P < 0.001 for all species pooled and r 2 =
0.42, P < 0.001 for P. taeda). There were no general associations between NM and either D or PD (P > 0.1), but NM was negatively correlated with D in P. palustris (r 2 = 0.39, P < 0.05).
Discussion
Needle cross-sectional geometry in P. palustris and P. taeda
In Pinus species with three- and five-needled fascicles, including P. ponderosa (Hultine and Marshall 2001), P. radiata D.
Don (Wood 1971, 1974, Beets 1977, Grace 1987), P. rigida
Mill. (Johnson 1984, Samuelson et al. 1992), P. taeda (Kozlowski and Schumacher 1943, Harms 1971, Johnson 1984,
Shelton and Switzer 1984, Johnson et al. 1985, Samuelson et
al. 1992, Hultine and Marshall 2001) and P. strobus (Kozlowski and Schumacher 1943, Johnson 1984, Brand 1987),
TREE PHYSIOLOGY ONLINE at http://heronpublishing.com
756
NIINEMETS, ELLSWORTH, LUKJANOVA AND TOBIAS
Figure 9. Correlations of (A) needle dry mass to fresh mass ratio and
(B) protein-free needle carbon content with needle density. Symbols:
䊉 = Pinus palustris; 䊊 = P. taeda; and 䉭 = P. virginiana. All data
were pooled in the linear regressions. Protein-free needle carbon content was calculated from total carbon content and foliar nitrogen content as described in the text.
Figure 8. Modification of (A) mean (error bars = 1 SE) needle density,
(B) total carbon content (CM) and (C) protein-free carbon content (CS)
with needle age. Means within species were compared by one-way
ANOVA, and bars with the same letter are not significantly different
(P > 0.05). Needle cohorts were counted from the shoot tip to the bottom of the shoot. Shoot flushes 1–3 were formed in the current year
(3–8 months from the time of sampling), and the fourth flush in
P. taeda was the last flush formed in the previous year (16 months
old). Protein-free needle carbon content was calculated from CM and
foliar nitrogen content as described in the text.
needle cross-sectional geometry has invariably been approximated by a sector of a circle, with the needle internal faces assumed to constitute the radii of a fascicle circle. According to
this model, needle thickness (T) is equal to the length of the internal needle face (L s; Figure 1), and needle width (W ) is functionally related to T by W = 2Tsin(π/n), where n is the number
of needles in the fascicle (Figure 1A). The possibility of calculating total needle area (AT) according to the sectorial needle
model is alluring, because only measurements of T and needle
length are necessary. However, the appropriateness of the sec-
torial model for estimation of AT has not been tested by individual measurements of T, L s and W. As our study demonstrates, L s is neither equivalent to T nor is its determination
from W accurate in P. palustris and P. taeda (Figures 2A
and 2B). In fact, the central angle of the sector (2α in Figure 1B) was smaller (Figure 4C) than 120°, which is the value
expected for the circle sector of three-needled species. The radii of the fascicle circles (Equation 6 in Figure 1) were also
longer than either T or L s (data not shown). Thus, our measurements indicate that if the contiguous faces of the needles in the
fascicle were brought into contact with each other, the needles
with their specific geometry would form a fascicle cylinder,
the inner part of which would be empty (see Figure 1B). Nevertheless, because all needle linear characteristics were
strongly correlated, relative differences in estimates of AT by
various models (Figure 1A versus Figure 1B) differed by no
more than 10%, suggesting that the sectorial model may still
serve as a rough approximation of AT (Figure 3A), but not
needle volume (Figure 3B).
For three-needled conifers, the sectorial model (Figure 1A)
yields a constant ratio of total to projected needle area (AT /AP)
of 2.36 that has often been used for all leaves in the canopy to
TREE PHYSIOLOGY VOLUME 22, 2002
NEEDLE ARCHITECTURE AND CHEMISTRY OF NORTH AMERICAN CONIFERS
convert the measurements of projected needle area by optical
planimeters to total area (Raison et al. 1992, Law et al. 2001).
However, there were light-related changes in needle central
angle (Figure 4C) and needle T/W ratio in both species. Consequently, the needle cross-sectional shape was not constant
along the canopy light gradient, resulting in strong modifications in the AT /AP ratio with seasonal mean integrated daily
quantum flux density (Qint) in both P. taeda and P. palustris
(Figure 4D). Similar alteration of AT/AP with irradiance has
been observed previously in the shade-tolerant conifer Picea
abies (L.) Karst. (Niinemets and Kull 1995, Niinemets
1997b). Because decreases in the AT /AP ratio with decreasing
irradiance increase the light-intercepting needle surface area
at a common projected area, we suggest that this modification
in AT /AP is an important acclimation response to low light environments.
Needle cross-sectional geometry in P. virginiana
We have previously confirmed that the needle cross section
may be approximated by a half-ellipse or by a circle segment
in the temperate conifer Pinus sylvestris L., which possesses
two-needled fascicles (Niinemets et al. 2001). The characteristics of the needle cross section in P. virginiana also suggest
that the half-ellipse rather than the half-circle model provides
a more reasonable approximation of the needle shape in this
species. Although the cross-sectional needle geometry of
two-needled Pinus species is often approximated by a half-circle (e.g., Luoma 1997 for P. sylvestris and Johnson 1984 for
P. virginiana), i.e., W = 2T, and AT /AP is a constant of 2.57, our
investigations indicate that this simplification is generally inappropriate. Like P. sylvestris at the fertile site (Niinemets et
al. 2001), but in a marked contrast to P. palustris and P. taeda,
AT /AP was negatively related to irradiance in P. virginiana
(Figure 4D). However, AT /AP varied from 2.45 to 2.88 at the
nutrient-rich P. sylvestris site (Niinemets et al. 2001), but the
range of 2.43 to 2.54 was narrower for P. virginiana (Figure 4D).
Although the AT/AP ratio varied with irradiance in all species
(Figure 4D), the maximum changes were about 10%. This
contrasts with the canopy range of 2.2 to 3.5 in P. abies (Niinemets and Kull 1995, Niinemets 1997b) and thus partly supports the hypothesis that shade-intolerant conifers may be less
plastic than tolerant species. Nevertheless, the absolute values
observed in the three Pinus species throughout the light gradient (2.2 to 2.55) were close to the values in P. abies in low
light. Thus, we conclude that light-interception efficiency of
single needles in low light and for a hypothetical horizontal
needle orientation is not significantly lower in the studied
Pinus species than in P. abies.
Scaling of needle dry mass per unit area with Qint
In agreement with the hypothesis of lower plasticity in shadeintolerant, long-needled species, some studies (Johnson et al.
1985, Pataki et al. 1998) have found no significant vertical
variation in needle dry mass per total needle area (MT) along
canopy profiles in the long-needled species P. taeda. However,
in other studies, vertical variation in MT in P. taeda (McLaugh-
757
lin and Madgwick 1968), as well as in other long-needled
shade-intolerant species (P. radiata (Wood 1974, Benecke
1979, Ohmart and Thomas 1986) and P. ponderosa (Bond et
al. 1999)), has been observed. Given that the variability of MT
along the canopy depends on stand density (McLaughlin and
Madgwick 1968), inconsistent patterns in MT versus canopy
height relations may be the outcome of low variability in
irradiance within the crowns of open stands, rather than
limited species potentials for light acclimation of foliage morphology. We confirmed that MT is strongly linked to canopy
light climate in all studied conifers, but also that MT of the
most shade-intolerant species (P. palustris) was no less plastic
than that of the other two Pinus species (Figure 5A).
Comparisons with observations of short-needled, shade-tolerant conifers in the literature support the view that the plasticity of MT and dry mass per unit projected area (MP) is greater in
shade-tolerant than in shade-intolerant species. In the studied
Pinus species, MT varied only 1.3- to 1.7-fold, and because of
relatively low variation in the AT /AP ratio (Figure 4D), the
magnitude of variation was similar for MP (MP = MT × AT /AP;
data not shown). Likewise, MP varied 1.5-fold in another
long-needled, shade-intolerant conifer P. ponderosa (Bond et
al. 1999), and depending on site fertility, MT and MP varied
1.3- to 2-fold in the relatively short-needled, shade-intolerant
conifer P. sylvestris (Niinemets et al. 2001). In the short-needled, shade-tolerant species P. abies, MT varied 2.5-fold, and
because of a strong effect of light on AT/AP, MP varied 4-fold
(Niinemets and Kull 1995, Niinemets 1997b). In a like manner, MP changed 4-fold in the canopy of Abies amabilis Dougl.
ex Forb. (Stenberg et al. 1998). However, contrary to the hypothesis, MP differed only about 1.5-fold between the canopy
top and bottom in the shade-tolerant conifers Pseudotsuga
menziesii and Tsuga heterophylla (Bond et al. 1999). Given
that a greater variation in MP has been found in P. menziesii in
another study (Chen et al. 1996), and that there may be strong
interactions between site fertility and needle-level plasticity
(Niinemets et al. 2001), we suggest that more species from
contrasting stands should be investigated to determine the role
of foliage plasticity in shade-tolerance.
Components of MT in relation to irradiance
Increases in needle T and W (Figures 4A and 4B) were responsible for a strong effect of light on needle volume to total area
ratio (V/AT; Figure 5B). Because needle density (D) was independent of irradiance, canopy variation in MT resulted solely
from changes in V/AT with irradiance (Figure 5B). This is in
agreement with earlier findings in Picea abies (Niinemets and
Kull 1995, Niinemets 1997b), Pinus sylvestris (Niinemets et
al. 2001) and P. radiata (Beets 1977, Whitehead et al. 1994).
However, in other studies, D was moderately larger in the upper than in the lower canopy of P. taeda (Shelton and Switzer
1984) and P. radiata (Benecke 1979). Given that D is a foliar
characteristic that generally increases in response to water
limitations (Roderick et al. 2000, Niinemets 2001), and that
foliar water stress tends to increase from the bottom to the top
of the canopy (Niinemets et al. 1999b), differences in water
stress gradients between studies may provide an explanation
TREE PHYSIOLOGY ONLINE at http://heronpublishing.com
758
NIINEMETS, ELLSWORTH, LUKJANOVA AND TOBIAS
for these contrasting observations. Large errors associated
with needle volume estimation by the circle sector model (Figures 1 and 3B) may also have lead to systematic variation in
needle density in the earlier studies.
Dependence of needle nitrogen and chlorophyll contents on
irradiance
Although foliar nitrogen contents per unit dry mass (NM) are
relatively constant in tree canopies (Ellsworth and Reich 1993,
Niinemets 1997b, Niinemets and Kull 1998), there are several
examples of a positive scaling of NM with canopy height in
Pinus species (Comerford 1981, Helmisaari 1992, Niinemets
et al. 2001), as we observed in P. palustris and P. taeda (Figure
5C). Such a preferential allocation of needle nitrogen to upper
canopy leaves is possibly associated with marked apical dominance and growth in more illuminated canopy positions in this
genus (Kellomäki and Strandman 1995).
Increases in NM result in a greater increase in foliar nitrogen
per unit area (NA = NM × MT) than in MT with increasing Qint
(cf. Figures 5A and 5D). Given that foliage photosynthetic potentials are linearly correlated with needle nitrogen contents in
conifers (Dougherty et al. 1998, Bond et al. 1999, Niinemets et
al. 2001), the difference between the MT versus irradiance and
NA versus irradiance relationships suggests that, in P. palustris
and P. taeda, foliage photosynthetic capacity changes more
plastically along light gradients than foliage morphological
characteristics (see Zhang et al. 1997 and Tang et al. 1999 for
confirmation of canopy variability in foliage photosynthetic
capacity in P. taeda).
As observed in broad-leaved (Niinemets et al. 1998) and conifer species (Bond et al. 1999), foliage chlorophyll contents
per unit dry mass were negatively related to irradiance (Figure 7A). Moreover, the negative relationship between the foliar chlorophyll to N ratio and Qint further demonstrated a
greater fractional investment of foliar nitrogen in light harvesting at lower irradiance. Thus, we conclude that needle chemical composition acclimated to increase the light-interception
efficiency per unit foliar biomass with decreasing irradiance.
Light effects on foliage carbon contents and mechanical
properties: implications for species shade-tolerance
In P. palustris, needle carbon contents per unit dry mass (CM)
were strongly correlated with irradiance (Figure 6). This correlation was essentially the same for protein-free needle carbon content, indicating that the effect of Qint on CM did not
result from positive influences of light on needle protein content (Figure 5C). Scaling of CM with Qint agrees with similar
relationships in broad-leaved trees, where enhanced foliar carbon contents in high light were attributable to increases in
lignin contents (Niinemets and Kull 1998, Niinemets et al.
1999a). Because lignin is carbon-rich (65.4% carbon, calculated for Picea lignin from Freudenberg 1965), and conifer
needles contain large amounts of lignin (23% for P. taeda
(Chung and Barnes 1977) and 30% for Picea abies (Niinemets
1997c)), variability in needle lignin content with irradiance
may explain light effects on CM in P. palustris.
Given that lignin plays an important role in plant mechanical strength (Sibly and Vincent 1997), decreases in lignin content with irradiance would imply that needles in the lower canopy were more elastic than needles in the upper canopy and,
consequently, bent more at common needle cross-sectional dimensions and geometry. In fact, increases in needle bending
under the load of a needle’s own weight with decreasing
irradiance are expected in all species, because of shading-related decreases in needle thickness and width (Figures 4A
and 4B) and consequently in the needle second moment of
area (I; m4), which is the needle cross-sectional property that
characterizes the resistance to the elastic bending of needles
with the same material properties (Niklas 1991). Needle vertical deflection scales with the cube of needle length, and thus,
observed tip deflections are especially large in long-needled
species (Niklas 1991). It follows that light-related decreases in
I should result in especially large needle bendings and low
needle inclination angles in long-needled species. In fact, the
needles of P. palustris were increasingly downward-oriented
with decreasing irradiance, becoming essentially vertical (inclination angle with respect to horizontal of –90°) at the bottom of the canopy (Ü. Niinemets et al., unpublished observations). Because vertical needles intercept light inefficiently
under low light conditions (Hikosaka and Hirose 1997), we
suggest that mechanical constraints limit light-interception
potentials of needles and that changes in needle cross-sectional geometry and needle dimensions in response to low
irradiance (Figures 4A–C) adversely affected needle light-interception capacity in P. palustris. Thus, the decline in needle
thickness under low light conditions in long-needled species is
not necessarily advantageous. As a result of associated decreases in needle inclination angles with needle cross-sectional dimensions, the acclimation of needle morphology to
low irradiance may significantly curtail light-interception efficiency and photosynthesis in long-needled conifers.
Age effects on needle characteristics and relationships
between needle density, dry to fresh mass ratio and needle
carbon contents
Most foliar characteristics were independent of needle age in
our study. In general, the age-related decline in foliar nitrogen
contents and foliage photosynthetic rates is slow until the final
stages of leaf aging in P. taeda (Murthy et al. 1996, Dougherty
et al. 1998). For our study area, mean needle longevity for
P. taeda is 19 months for the first flush needles (Ellsworth
2000), and senescence had apparently not started in sampled
needles.
In contrast to nitrogen contents, there were significant
changes in needle density (Figure 8A) and total (Figure 8B)
and protein-free (Figure 8C) carbon contents with needle age.
Increases in D with increasing age is a common response
among conifers including Picea abies (Niinemets 1997a),
P. radiata (Beets 1977, Beets and Lane 1987, Whitehead et al.
1994) and P. taeda (Shelton and Switzer 1984, Johnson et al.
1985). Increases in density may result from decreases in the
fraction of air space in the needle volume because of thickening of cell walls, but possibly also because of compression of
TREE PHYSIOLOGY VOLUME 22, 2002
NEEDLE ARCHITECTURE AND CHEMISTRY OF NORTH AMERICAN CONIFERS
needle tissues as a result of secondary needle growth (Ewers
1982, Gilmore et al. 1995). The change in D was accompanied
by increases in needle dry to fresh mass ratio (Figure 9A) and
in structural needle carbon content (Figure 9B). This may indicate that increases in D resulted from thickening and enhanced
lignification of cell walls, and accordingly, that this modification improved the mechanical strength of older needles.
Although all species fitted the same relationships (Figure 9),
the ranges of variation in D, foliage carbon contents, and the
dry to fresh mass ratio differed among species (Figures 8–9).
Given that P. virginiana had shorter needles, but similar needle
width and thickness as P. taeda (Figures 4A and 4B), lower
needle densities and structural carbon contents in P. virginiana
(Figures 8–9) may be a manifestation of lower carbon requirements for structural support in this species. Except under high
light conditions, carbon contents were generally lower in
P. palustris than in the other species (Figure 6). Progressively
lower photosynthetic rates as a result of decreasing light-interception efficiencies with decreasing irradiance, and consequently the shortage of carbon, may explain low carbon
content and needle density in this species.
Conclusion
Although there were few differences in plasticity among species of contrasting needle length, we suggest that the value of
plastic modification of needle cross-sectional shape and dimensions to low irradiance is significantly altered by needle
length. Decreases in needle cross-sectional dimensions with
decreasing light availability are compatible with lower needle
inclination angles and, in turn, with reduced foliage light-harvesting efficiency. Such a decrease in light-interception efficiency may offset the benefits of greater total plant needle area
achieved by decreasing the thickness of foliage elements. Because the mechanical design of shorter needles is more robust
(Niklas 1991), changes in needle cross-sectional diameter are
expected to alter needle light-harvesting efficiency primarily
in long-needled species. Thus, needle absolute dimensions,
rather than plasticity, seem to control the light-harvesting efficiency and competitive potential of the studied Pinus species.
Inherent trade-offs between light-interception efficiency,
needle acclimation to low light and needle length may explain
the inferior performance of long-needled species in low
irradiance.
Acknowledgments
We thank the Estonian Science Foundation (Grants 3235 and 4584),
the Estonian Minister of Education (Grants 0180517s98 and
0281770Bs01), the National Research Council, the US National
Academy of Sciences (Twinning Program, COBASE) and the Bayreuther Institut für Terrestrische Ökosystemforschung (BITÖK), University of Bayreuth, Germany (BBWFT Grant 0339476C) for funding
this work. We also appreciate the apt technical assistance of Maarika
Mäesalu.
759
References
Anderson, M.C. 1964. Studies of the woodland light climate. I. The
photographic computation of light conditions. J. Ecol. 52:27–41.
Badeck, F.-W. 1995. Intra-leaf gradient of assimilation rate and optimal allocation of canopy nitrogen: a model on the implications of
the use of homogeneous assimilation functions. Aust. J. Plant
Physiol. 22:425–439.
Baker, F.S. 1949. A revised tolerance table. J. For. 47:179–181.
Baker, J.B. and O.G. Langdon. 1990. Pinus taeda L. loblolly pine. In
Silvics of North America. Vol. 1. Conifers. Eds. R.M. Burns and
B.H. Honkala. Agriculture Handbook. 654. Forest Service, United
States Department of Agriculture, Washington, DC, pp 497–512.
Beets, P. 1977. Determination of the fascicle surface area for Pinus
radiata. N.Z. J. For. Sci. 7:397–407.
Beets, P.N. and P.M. Lane. 1987. Specific leaf area of Pinus radiata as
influenced by stand age, leaf age, and thinning. N.Z. J. For. Sci. 17:
283–291.
Benecke, U. 1979. Surface area of needles in Pinus radiata—variation with respect to age and crown position. N.Z. J. For. Sci. 9:
267–271.
Bond, B.J., B.T. Farnsworth, R.A. Coulombe and W.E. Winner. 1999.
Foliage physiology and biochemistry in response to light gradients
in conifers with varying shade tolerance. Oecologia 120:183–192.
Boyer, W.D. 1990. Pinus palustris Mill. longleaf pine. In Silvics of
North America. Vol. 1. Conifers. Eds. R.M. Burns and B.H.
Honkala. Agriculture Handbook. 654. Forest Service, United
States Department of Agriculture, Washington, DC, pp 405–412.
Brand, D.G. 1987. Estimating the surface area of spruce and pine foliage from displaced volume and length. Can. J. For. Res. 17:
1305–1308.
Brockway, D.G. and K.W. Outcalt. 1998. Gap-phase regeneration in
longleaf pine wiregrass ecosystems. For. Ecol. Manage. 106:
125–139.
Campbell, G.S. and J.M. Norman. 1998. An introduction to environmental biophysics. 2nd Edn. Springer-Verlag, New York, 286 p.
Carter, K.K. and A.G. Snow, Jr. 1990. Pinus virginiana Mill. Virginia
pine. In Silvics of North America. Vol. 1. Conifers. Eds. R.M.
Burns and B.H. Honkala. Agriculture Handbook. 654. Forest Service, United States Department of Agriculture, Washington, DC,
pp 513–519.
Chen, H.Y.H. 1997. Interspecific responses of planted seedlings to
light availability in interior British Columbia: survival, growth,
allometric patterns, and specific leaf area. Can. J. For. Res. 27:
1383–1393.
Chen, H.Y.H., K. Klinka and G.J. Kayahara. 1996. Effects of light on
growth, crown architecture, and specific leaf area for naturally established Pinus contorta var. latifolia and Pseudotsuga menziesii
var. glauca saplings. Can. J. For. Res. 26:1149–1157.
Chung, H.-H. and R.L. Barnes. 1977. Photosynthate allocation in
Pinus taeda. I. Substrate requirements for synthesis of shoot biomass. Can. J. For. Res. 7:106–111.
Comerford, N.B. 1981. Distributional gradients and variability of
macroelement concentrations in the crowns of plantation grown
Pinus resinosa (Ait.). Plant Soil 63:345–353.
Dougherty, P.M., L.H. Allen, L.W. Kress, R. Murthy, C.A. Maier, T.J.
Albaugh and D.A. Sampson. 1998. An investigation of the impacts
of elevated carbon dioxide, irrigation, and fertilization on the physiology and growth of loblolly pine. In The Productivity and
Sustainability of Southern Forest Ecosystems in a Changing Environment. Eds. R.A. Mickler and S. Fox. Springer-Verlag, Berlin,
pp 149–168.
TREE PHYSIOLOGY ONLINE at http://heronpublishing.com
760
NIINEMETS, ELLSWORTH, LUKJANOVA AND TOBIAS
Ellsworth, D.S. 1999. CO2 Enrichment in a maturing pine forest: are
CO2 exchange and water status in the canopy affected? Plant Cell
Environ. 22:461–472.
Ellsworth, D.S. 2000. Seasonal CO2 assimilation and stomatal limitations in a Pinus taeda canopy. Tree Physiol. 20:435–445.
Ellsworth, D.S. and P.B. Reich. 1993. Canopy structure and vertical
patterns of photosynthesis and related leaf traits in a deciduous forest. Oecologia 96:169–178.
Ellsworth, D.S., R. Oren, C. Huang, N. Phillips and G.R. Hendrey.
1995. Leaf and canopy responses to elevated CO2 in a pine forest
under free-air CO2 enrichment. Oecologia 104:139–146.
Ewers, F.W. 1982. Secondary growth in needle leaves of Pinus longaeva (bristlecone pine) and other conifers: quantitative data. Am. J.
Bot. 69:1552–1559.
Freudenberg, K. 1965. Lignin: its constitution and formation from
p-hydroxycinnamyl alcohols. Science 148:595–600.
Gilmore, D.W., R.S. Seymour, W.A. Halteman and M.S. Greenwood.
1995. Canopy dynamics and the morphological development of
Abies balsamea: effects of foliage age on specific leaf area and secondary vascular development. Tree Physiol. 15:47–55.
Grace, J.C. 1987. Theoretical ratio between “one-sided” and total surface area for pine needles. N.Z. J. For. Sci. 17:292–296.
Gutschick, V.P. and F.W. Wiegel. 1988. Optimizing the canopy photosynthetic rate by patterns of investment in specific leaf mass. Am.
Nat. 132:67–86.
Harcombe, P.A., J.S. Glitzenstein, R.G. Knox, S.L. Orzell and E.L.
Bridges. 1993. Vegetation of the longleaf pine region of the West
Gulf coastal pine. In Proceedings of the Tall Timbers Fire Ecology
Conference. Vol. 18. The Longleaf Pine Ecosystem: Ecology,
Restoration and Management. Ed. S.M. Hermann. Tall Timbers
Research Station, Tallahassee, FL, pp 83–104.
Harms, W.R. 1971. Estimating leaf-area growth in pine. Ecology 52:
931–934.
Harrington, C.A. 1991. Retrospective shoot growth analysis for three
seed sources of loblolly pine. Can. J. For. Res. 21:306–317.
Helmisaari, H.-S. 1992. Spatial and age-related variation in nutrient
concentrations of Pinus sylvestris needles. Silva Fenn. 26:
145–153.
Higginbotham, K.O. 1974. The influence of canopy position and age
of leaf tissue on growth and photosynthesis in loblolly pine. Ph.D.
Diss., Botany Dept., Duke Univ., Durham, NC.
Hikosaka, K. and T. Hirose. 1997. Leaf angle as a strategy for light
competition: optimal and evolutionary stable light extinction coefficient within a leaf canopy. Ecoscience 4:501–507.
Hikosaka, K. and I. Terashima. 1995. A model of the acclimation of
photosynthesis in the leaves of C3 plants to sun and shade with respect to nitrogen use. Plant Cell Environ. 18:605–618.
Hultine, K.R. and J.D. Marshall. 2001. A comparison of three methods for determining the stomatal density of pine needles. J. Exp.
Bot. 52:369–373.
Johnson, J.D. 1984. A rapid technique for estimating total surface
area of pine needles. For. Sci. 30:913–921.
Johnson, J.D., S.M. Zedaker and A.B. Hairston. 1985. Foliage, stem,
and root interrelations in young loblolly pine. For. Sci. 31:
891–898.
Kellomäki, S. and H. Strandman. 1995. A model for the structural
growth of young Scots pine crowns based on light interception by
shoots. Ecol. Model. 80:237–250.
Kinerson, R.S., K.O. Higginbotham and R.C. Chapman. 1974. The
dynamics of foliage distribution within a forest canopy. J. Appl.
Ecol. 11:347–353.
Kozlowski, T.T. and F.X. Schumacher. 1943. Estimation of stomated
foliar surface of pines. Plant Physiol. 18:122–127.
Law, B.E., S. Van Tuyl, A. Cescatti and D.D. Baldocchi. 2001. Estimation of leaf area index in open-canopy ponderosa pine forests at
different successional stages and management treatments in Oregon. Agric. For. Meteorol. 108:1–14.
Luoma, S. 1997. Geographical pattern in photosynthetic light response of Pinus sylvestris in Europe. Funct. Ecol. 11:273–281.
McLaughlin, S.B. and H.A.I. Madgwick. 1968. The effects of position in crown on the morphology of needles of loblolly pine (Pinus
taeda L.). Am. Midl. Nat. 80:547–550.
Murthy, R. and P.M. Dougherty. 1997. Effect of carbon dioxide, fertilization and irrigation on loblolly pine branch morphology. Trees
11:485–493.
Murthy, R., P.M. Dougherty, S.J. Zarnoch and H.L. Allen. 1996. Effects of carbon dioxide, fertilization, and irrigation on photosynthetic capacity of loblolly pine trees. Tree Physiol. 16:537–546.
Myers, D.A., R.B. Thomas and E.H. DeLucia. 1999. Photosynthetic
capacity of loblolly pine (Pinus taeda L.) trees during the first year
of carbon dioxide enrichment in a forest ecosystem. Plant Cell Environ. 22:473–481.
Niinemets, Ü. 1997a. Acclimation to low irradiance in Picea abies:
influences of past and present light climate on foliage structure and
function. Tree Physiol. 17:723–732.
Niinemets, Ü. 1997b. Distribution patterns of foliar carbon and nitrogen as affected by tree dimensions and relative light conditions in
the canopy of Picea abies. Trees 11:144–154.
Niinemets, Ü. 1997c. Energy requirement for foliage construction depends on tree size in young Picea abies trees. Trees 11:420–431.
Niinemets, Ü. 1999. Research review. Components of leaf dry mass
per area—thickness and density—alter leaf photosynthetic capacity in reverse directions in woody plants. New Phytol. 144:35–57.
Niinemets, Ü. 2001. Climatic controls of leaf dry mass per area, density, and thickness in trees and shrubs at the global scale. Ecology
82:453–469.
Niinemets, Ü. and O. Kull. 1995. Effects of light availability and tree
size on the architecture of assimilative surface in the canopy of
Picea abies: variation in needle morphology. Tree Physiol. 15:
307–315.
Niinemets, Ü. and O. Kull. 1998. Stoichiometry of foliar carbon constituents varies along light gradients in temperate woody canopies:
implications for foliage morphological plasticity. Tree Physiol.
18:467–479.
Niinemets, Ü., O. Kull and J.D. Tenhunen. 1998. An analysis of light
effects on foliar morphology, physiology, and light interception in
temperate deciduous woody species of contrasting shade tolerance.
Tree Physiol. 18:681–696.
Niinemets, Ü., O. Kull and J.D. Tenhunen. 1999a. Variability in leaf
morphology and chemical composition as a function of canopy
light environment in co-existing trees. Int. J. Plant Sci. 160:
837–848.
Niinemets, Ü., A. Sõber, O. Kull, W. Hartung and J.D. Tenhunen.
1999b. Apparent controls on leaf conductance by soil water availability via light-acclimation of foliage structural and physiological
properties in a mixed deciduous, temperate forest. Int. J. Plant Sci.
160:707–721.
Niinemets, Ü., D.S. Ellsworth, A. Lukjanova and M. Tobias. 2001.
Site fertility and the morphological and photosynthetic acclimation
of Pinus sylvestris needles to light. Tree Physiol. 21:1231–1244.
Niklas, K.J. 1991. Biomechanical attributes of the leaves of pine species. Ann. Bot. 68:253–262.
Ohmart, C.P. and J.R. Thomas. 1986. Estimating surface area of Pinus
radiata needles using needle weight. Aust. For. Res. 16:85–89.
TREE PHYSIOLOGY VOLUME 22, 2002
NEEDLE ARCHITECTURE AND CHEMISTRY OF NORTH AMERICAN CONIFERS
Pataki, D.E., R. Oren and N. Phillips. 1998. Responses of sap flux and
stomatal conductance of Pinus taeda L. trees to stepwise reductions in leaf area. J. Exp. Bot. 49:871–878.
Perry, T.O., H.E. Sellers and C.O. Blanchard. 1969. Estimation of
photosynthetically active radiation under a forest canopy with
chlorophyll extracts and from basal area measurements. Ecology
50:39–44.
Platt, W.J., G.W. Evans and S.J. Rathbun. 1988. The population dynamics of a long-lived conifer (Pinus palustris). Am. Nat. 131:
491–525.
Porra, R.J., W.A. Thompson and P.E. Kriedemann. 1989. Determination of accurate extinction coefficients and simultaneous equations
for assaying chlorophyll a and b extracted with four different solvents: verification of the concentration of chlorophyll standards by
atomic absorption spectroscopy. Biochim. Biophys. Acta 975:
384–394.
Raison, R.J., B.J. Myers and M.L. Benson. 1992. Dynamics of Pinus
radiata foliage in relation to water and nitrogen stress. I. Needle
production and properties. For. Ecol. Manage. 52:139–158.
Roderick, M.L., S.L. Berry and I.R. Noble. 2000. A framework for
understanding the relationship between environment and vegetation based on the surface area to volume ratio of leaves. Funct.
Ecol. 14:423–437.
Sampson, D.A. and H.L. Allen. 1998. Light attenuation in a 14-yearold loblolly pine stand as influenced by fertilization and irrigation.
Trees 13:80–87.
Samuelson, L.J., J.R. Seiler and P.P. Feret. 1992. Gas exchange and
canopy structure of 9-year-old loblolly pine, pitch pine and pitch ×
loblolly hybrids. Trees 6:28–31.
Sands, P.J. 1995. Modelling canopy production. I. Optimal distribution of photosynthetic resources. Aust. J. Plant Physiol. 22:
593–601.
Shelton, M.G. and M.D. Cain. 2000. Regenerating uneven-aged
stands of loblolly and shortleaf pines: the current state of knowledge. For. Ecol. Manage. 129:177–193.
Shelton, M.G. and G.L. Switzer. 1984. Variation in the surface area
relationships of loblolly pine fascicles. For. Sci. 30:355–363.
761
Sibly, R.M. and J.F.V. Vincent. 1997. Optimality approaches to resource allocation in woody tissues. In Plant Resource Allocation.
Eds. F.A. Bazzaz and J. Grace. Academic Press, San Diego, pp
143–159.
Sinclair, T.R. and K.R. Knoerr. 1982. Distribution of photosynthetically active radiation in the canopy of a loblolly pine plantation. J.
Appl. Ecol. 19:183–191.
Soil Survey Division, N.R.C.S., United States Department of Agriculture. 2001. Official soil series descriptions. USDA-NRCS Soil Survey Division, http://www.statlab.iastate.edu/soils/osd.
Sprugel, D.G., J.R. Brooks and T.M. Hinckley. 1996. Effects of light
on shoot geometry and needle morphology in Abies amabilis. Tree
Physiol. 16:91–98.
Stenberg, P., H. Smolander, D.G. Sprugel and S. Smolander. 1998.
Shoot structure, light interception, and distribution of nitrogen in
an Abies amabilis canopy. Tree Physiol. 18:759–767.
Stenberg, P., T. Kangas, H. Smolander and S. Linder. 1999. Shoot
structure, canopy openness, and light interception in Norway
spruce. Plant Cell Environ. 22:1133–1142.
Tang, Z., J.L. Chambers, S. Guddanti and J.P. Barnett. 1999.
Thinning, fertilization, and crown position interact to control physiological responses of loblolly pine. Tree Physiol. 19:87–94.
Vertregt, N. and F.W.T. Penning de Vries. 1987. A rapid method for
determining the efficiency of biosynthesis of plant biomass. J.
Theor. Biol. 128:109–119.
Wells, C.G. and L.J. Metz. 1963. Variation in nutrient content of loblolly pine needles with season, age, soil, and position on the crown.
Soil Sci. Soc. Am. Proc. 27:90–93.
Whitehead, D., F.M. Kelliher, C.M. Frampton and M.J.S. Godfrey.
1994. Seasonal development of leaf area in a young, widely spaced
Pinus radiata D. Don stand. Tree Physiol. 14:1019–1038.
Witkowski, E.T.F. and B.B. Lamont. 1991. Leaf specific mass confounds leaf density and thickness. Oecologia 88:486–493.
Wood, G.B. 1971. Shape of Pinus radiata fascicles and the implications for estimating needle surface area. Aust. For. Res. 5:31–36.
Wood, G.B. 1974. Spatial variation in leaf chlorophyll within the
crown of a radiata pine sapling. Aust. For. Res. 6:5–14.
Zhang, S., T.C. Hennessey and R.A. Heinemann. 1997. Acclimation
of loblolly pine (Pinus taeda) foliage to light intensity as related to
leaf nitrogen availability. Can. J. For. Res. 27:1032–1040.
TREE PHYSIOLOGY ONLINE at http://heronpublishing.com