Directory UMM :Data Elmu:jurnal:T:Tree Physiology:Vol15.1995:
Tree Physiology 15, 649--656
© 1995 Heron Publishing----Victoria, Canada
Systematic variation in xylem hydraulic capacity within the crown of
white ash (Fraxinus americana)
BRIAN J. JOYCE1 and KIM C. STEINER2
1
School of Forest Resources, 206 Forest Resources Laboratory, The Pennsylvania State University, University Park, PA 16802, USA
2
School of Forest Resources, 213 Ferguson Building, The Pennsylvania State University, University Park, PA 16802, USA
Received September 19, 1994
Summary A 7-m tall white ash tree (Fraxinus americana
Marsh.) was dissected, and hydraulic parameters of the xylem
were determined by inducing a steady-state flow of water
through the stem segments and monitoring volume and velocity
flow rates. Leaf-specific conductivity (LSC) was highest in the
main stem and lowest in some of the leaf-bearing lateral shoots.
The LSC was higher in the main stem than in branches and
higher in primary than in secondary branch axes. Terminal
leaf-bearing shoots were larger and had a significantly greater
mean LSC than subjacent lateral shoots. A significant reduction
in LSC was associated with the transition between 1- and
2-year-old growth. In branches of the same age, there was a
close correspondence among LSC, branch position and branch
size. The average LSC of leaf-bearing shoots from south-facing
branches was 43% greater than that of shoots from north-facing
branches.
Within-crown variation in LSC was associated with variation in velocity flow rate (V). By contrast, the ratio of potentially functional xylem area to supported leaf area (Apf/Al) was
relatively stable throughout the crown. Stratification of stems
by Strahler order accounted for approximately 70% of the total
variation in LSC. These results suggest that (1) there exists a
systematic pattern of variation in LSC distribution within the
crown of white ash, (2) within-crown variability in LSC is
primarily the result of variation in mean vessel diameter, and
(3) there is a physiological linkage between LSC and crown
morphology that is maintained through a positive feedback
mechanism during branch ontogeny.
Keywords: hydraulic architecture, hydraulic dominance, hydraulic segmentation, leaf-specific conductivity.
Introduction
Xylem tissue in trees serves several biological functions including mechanical support and vascular transport. Because
the relative importance of these functions varies according to
branch location, there may exist some measure of order in the
structure of the internal water transport system and its relationship to the morphology of a tree’s crown.
Long distance transport of water and mineral nutrients is a
vital function of the xylem. A practical measure of the xylem’s
hydraulic supply capacity is leaf-specific conductivity (LSC)
(Zimmermann 1978). For example, the LSC of a stem segment
can be used to relate the average transpirational water loss (E)
from leaves supported by the segment to the decline in water
potential (dΨ) per unit path length (dx) within the segment
(dΨ/dx = E/LSC). Thus, the distribution of LSC within a tree
influences patterns of Ψ throughout the crown and can impose
constraints on such physiological processes as transpiration
and photosynthesis (Tyree and Sperry 1988, Shumway et al.
1993, Yang and Tyree 1993).
Within-tree patterns of LSC distribution have been quantified for several angiosperm and conifer tree species (Zimmermann 1978, 1983, Tyree et al. 1983, Ewers and Zimmermann
1984a, 1984b, Tyree et al. 1991). In most species examined,
LSC increases with increasing stem diameter and is higher in
the main stem than in branches. Another consistently observed
feature of tree hydraulic architecture is the increased hydraulic
resistance associated with leaf and branch junctions (Isebrands
and Larson 1977, Zimmermann 1978, Tyree et al. 1983, Ewers
and Zimmermann 1984a, 1984b).
Zimmermann (1983) considered that LSC distribution patterns in tree crowns were evidence of hydraulic segmentation,
defined as a structural feature that confines cavitation events to
relatively expendable plant parts in favor of parts more important to plant survival. Hydraulic segmentation is considered an
architectural advantage in dicotyledonous tree species because
leaves and small diameter stems are more readily replaced than
older, supporting stem structures.
Xylem also provides mechanical support to the crown. The
distribution of xylem anatomical features within a tree’s crown
likely reflects a compromise between hydraulic and mechanical support functions. Because the multiple roles of xylem
depend directly on its anatomical characteristics, we hypothesized that crown development will result in an organized pattern of xylem anatomical features. The objective of this study
was to describe this pattern of organization in white ash (Fraxinus americana Marsh.) by characterizing variations in several
xylem hydraulic parameters throughout an entire tree crown.
We also examined the relationship between functional and
morphological attributes.
650
JOYCE AND STEINER
Materials and methods
Plant material and branch segmentation
An 18-year-old white ash tree that was approximately 7 m in
height was selected for study from within a plantation population in Center County, PA. The entire crown and main stem of
the tree was destructively sampled. Starting in early September, all branches in the crown were labeled, and leaf areas were
measured. Stem segments were coded starting with the distal
end of the lowest branch. The first internode of at least 0.075 m
in length was identified as segment one, and the leaves supported by that segment were collected and placed in a high
humidity bag. This procedure was continued basipetally for all
lateral and terminal branch segments. Segment codes were
used to identify parent and descendent segments associated
with branch trifurcations (ash has opposite leaves). Leaf area
(Al) supported by each current-year shoot was measured to the
nearest 0.01 cm2 with a model LI-3000 surface area meter
(Li-Cor Inc., Lincoln, NE). After all stem segments were
identified on the tree (390 total), detailed drawings were made
of each branch system noting the relative position of all labeled
segments and the compass quadrant orientation of each branch
from the main stem: NE (1--90°), SE (91--180°), SW (181-270°), and NW (271--360°). The Strahler order (McDonald
1983) of each segment was also noted.
To facilitate within-crown comparisons, stem segments
were assigned to a hierarchy of crown components (Figure 1):
the main stem with its primary branches, and primary branch
systems comprising a central axis and secondary branches.
Measurement of xylem hydraulic parameters
Xylem hydraulic parameters were measured during the dormant season from November 3 to February 11. On each sampling day, an entire branch system was cut from the tree with
hand pruners and transported to the laboratory. Segments approximately 0.15 m long were excised from labeled internodes
and cut to 0.075 m in length with a miter trimmer, to assure a
smooth stem surface without crushing the vessels. Segments
were immersed in a water bath at 20 °C until measurements
were made. A 1.0-cm-wide band of bark and phloem tissue was
removed from the ends of the segments, and the segments were
fitted with plastic tubing with a neoprene seal.
To remove air emboli, stem segments were then attached to
an apparatus similar in design to that described by Kelso et al.
(1963), and pressurized at their proximal ends with filtered
(0.2 µm retention) and de-aerated, distilled water at a gradient
of 0.75 MPa m −1 until a steady-state flow condition was
achieved. The pressure gradient was then reduced to 0.14 MPa
m −1, and the system was allowed to reach steady-state flow.
Because there was no decrease in flow with time, a perfusing
agent was not used (see Zimmermann 1978). Steady-state
mass flow rate (M, kg s −1) was then measured gravimetrically
by collection of induced flow in beakers over a 15-min interval
and converted to volume flow rate (Q, m3 s −1). At the end of
the 15-min interval, 0.5 M KCl was injected with a syringe into
the water stream as it entered the segment. Electrical resistance
was monitored at the segment output. The average velocity
flow rate (V, m s −1) was estimated from segment length and the
elapsed time between injection of KCl and the average change
in electrical resistance of the output stream. The potentially
functional transverse xylem area (Apf, m2) (Shumway et al.
1991) was determined as Apf = Q/V (Preston 1952, Heine
1971). Because air emboli were removed under positive pressure, the derived Apf parameter represents a maximum conductive xylem area and may include vessels that were
dysfunctional in situ. Leaf specific conductivity (LSC, m2) was
calculated as:
Apf L η
LSC = V
,
Al P
where L is the segment length (m), η is the absolute viscosity
of water at 20 °C (10 −9 MPa s), Al is the leaf area supported by
the stem segment (m2), and P is the pressure difference across
the segment (MPa). Most published values of LSC do not
include units of viscosity and must be multiplied by 10 −12 m3
kg −1 MPa s (at 20 °C) to obtain the same units for LSC.
For each stem segment, total xylem (less pith) cross-sectional area (Ax) (m2) was measured on a thin, transverse section
from the segment’s midpoint by means of a digital caliper and
stereomicroscope.
Statistical methods
Figure 1. Diagrammatic representation of the descriptive terminology
used in the text. Diagram is representative of the branching pattern of
the white ash crown examined.
Paired comparison analysis (see Sokal and Rohlf 1969) was
used to test differences in both hydraulic and morphological
variables between branches with opposing compass directions
and between terminal and lateral leaf-bearing shoots. Each
HYDRAULIC ARCHITECTURE OF WHITE ASH
individual branch pair (originating from the same node) and
each terminal-lateral shoot pair was treated as a within-crown
replicate. Comparisons of xylem characteristics between parent and descendant segments of branch trifurcations were
made by paired t-tests (α = 0.05). Trifurcation comparisons
were made using the mean of the descendant segment values
and the parent segment value. Paired t-tests were also used to
test differences in hydraulic parameters among branches arising from a common annual extension increment of the main
stem, between segments immediately above and below the
transition between the 1-year-old and 2-year-old extension
growth of branch axes and between different year extension
increments of primary and secondary branch axes. Analysis of
variance (ANOVA) and mean separation procedures were conducted with Statistical Analysis Systems software (SAS Inc.,
Cary, NC). Variance component analysis (Steel and Torie
1980) was used to calculate the proportion of total variance in
the measured variables explained by Strahler order.
651
der (ultimate) segments and increased significantly (P < 0.05)
with increasing branch order (Table 1). The Apf/Al ratio showed
little variation (P > 0.05) among branch orders. Strahler order
accounted for 72.53, 0.68 and 84.87% of the total variability
in LSC, Apf /Al and V, respectively. Values of LSC for the main
stem were higher than those for primary branch axes as a result
of substantially greater V and Apf/Al values in the main stem
(Table 2). The mean LSC of the main stem (mean = 20.91) was
more than 2.9 times that of the proximal segments of adjacent
branches (mean = 7.80). Values of LSC were similar among
segments along the main stem until the last extension increment (1990), where there was a sharp decrease from 22.60 to
Table 1. Mean (± 1 SE) leaf specific conductivity (LSC) × 1016 in m2,
potentially functional xylem area to leaf area ratio (Apf/Al) × 106
(unitless), and velocity flow rate (V) × 102 in m s − 1 for all segments
from white ash crown and for segments stratified by Strahler order.
Means within a column appearing with the same letter are not significantly different (P > 0.05) from one another using Fisher’s LSD test.
Results
Strahler
order
n
LSC
Apf /Al
V
General distribution of xylem hydraulic and morphological
parameter values
All orders
Order 1
Order 2
Order 3
Order 4
390
257
102
26
5
4.58 (0.27)
1.72 (0.07) a
9.06 (0.52) b
12.60 (1.18) c
18.62 (3.10) d
2.21 (0.59)
2.21 (0.78) a
2.21 (1.10) a
2.10 (1.10) a
2.30 (2.90) a
2.94 (1.51)
1.18 (0.47) a
5.63 (2.10) b
8.00 (4.71) c
12.75 (13.78) d
Values of LSC and V varied considerably (CV = 1.16 and 1.01,
respectively) within the crown, but the Apf/Al ratio was less
variable (CV = 0.53). Both LSC and V were lowest in first-or-
Table 2. A comparison of mean leaf specific conductivity (LSC) × 1016 in m2, potentially functional xylem area to leaf area ratio (Apf/Al) × 106
(unitless), and velocity flow rate (V) × 102 in m s − 1 between the sequence of annual extension increments that comprise the main stem and the
central axes of primary branches originating from the same annual increment of the main stem (n = number of segments representing each year of
growth). Primary branch values within a column represent successive annual segments along those branch axes. All values within a row represent
segments of the same age but at different crown positions.
Year of growth
LSC (n)
1990
1989
1988
1987
1986
1985
Main stem
5.27 (4)
22.60 (2)
25.65 (2)
25.60 (2)
23.27 (3)
23.10 (1)
Primary branches originating from main stem increment of
1984
1985
1986
1987
1988
1989
1.88 (12)
7.40 (3)
11.06 (11)
11.19 (8)
10.91 (10)
7.60 (3)
0.75 (4)
9.50 (1)
5.65 (4)
5.70 (5)
6.38 (4)
1.98 (18)
7.89 (11)
10.30 (10)
7.53 (9)
2.01 (9)
13.75 (6)
12.08 (5)
2.04 (15)
10.61 (9)
2.61 (20)
Apf/Al
1990
1989
1988
1987
1986
1985
4.24
4.96
3.06
2.82
2.57
1.41
2.37
2.53
2.13
2.08
2.02
1.26
1.15
2.63
1.78
1.26
1.28
2.57
2.32
2.30
1.51
2.13
3.32
2.63
2.10
2.60
1.35
V
1990
1989
1988
1987
1986
1985
1.80
6.84
11.60
12.50
13.13
15.00
1.19
4.55
6.38
7.56
7.67
8.53
0.94
5.00
5.22
5.66
6.81
1.10
4.68
6.24
7.59
1.31
5.51
6.31
1.65
5.59
2.64
652
JOYCE AND STEINER
5.27. Values of V declined steadily along the main stem with a
pronounced decrease (from 6.84 to 1.80) at the 1990 extension
increment. The Apf /Al ratio increased acropetally along the
main stem, with only a slight decline (4.96 to 4.24) at the 1990
extension increment. Similar trends for LSC, V and Apf/Al
occurred along the axes of primary branches that originated in
1985 from the 1984 increment of the main stem.
Differences in xylem hydraulic parameters between the central and lateral axes of primary branch systems were similar to
the observed differences between the main stem and the central
axes of primary branches from the main stem (Table 3). Values
of LSC, V and Apf/Al of primary branch axes were consistently
higher than those of secondary branch axes.
In all branches sampled, the transition between 1- and
2-year-old extension growth represented a point of reduction
in hydraulic supply capacity, i.e., there were significant reductions (P < 0.05) in mean LSC and V across the transition
(Table 4). The Apf/Al ratio did not differ significantly in segments from above and below the transition.
Xylem hydraulic parameters in terminal versus lateral
leaf-bearing shoots
The values of LSC varied in a consistent manner between
terminal and lateral leaf-bearing shoots. The mean LSC of
shoots arising from terminal buds was significantly greater
(P < 0.01) than that of shoots arising from axillary buds immediately below the terminal (Table 5). The difference is attributable to a difference in V, because there was no difference (P >
0.05) in Apf /Al. Terminal shoots were longer than lateral shoots
and had a greater (P < 0.05) mean Ax.
Variation in LSC among branches arising from a common
annual extension increment
Within an annual extension increment of the main stem, there
was a pattern of hydraulic dominance (higher LSC) expressed
by branches in distal (superior) positions compared to their
subordinates. Because of the small number of comparisons
(n = 5 pairs) and the large amount of variation that existed
among the various extension increments, a significant differ-
Table 3. A comparison of mean leaf specific conductivity (LSC) × 1016 in m2, potentially functional xylem area to leaf area ratio (Apf/Al) × 106
(unitless), and velocity flow rate (V) × 102 in m s −1 between the central axis and lateral branches of 2-, 4- and 6-year-old primary branch systems.
Each branch age group is represented by three branches (n = number of segments in sample). For a given variable within a branch age group, means
within a row appearing with the same letter are not significantly different from one another (P > 0.05) using paired t-tests.
Year of growth
n (axis)
n (branch)
Apf/Al
LSC
V
Axis
Branch
Axis
Branch
Axis
Branch
6-Year-old primary branches
1990
10
1989
2
1988
9
1987
6
1986
5
1985
4
39
9
12
7
1
1.90 a
6.30 a
10.14 a
11.00 a
10.32
7.08
1.12 b
2.97 b
2.27 b
3.64 b
6.30
2.37 a
2.53 a
2.25 a
2.13 a
2.15
1.22
1.77 b
1.89 a
1.67 b
1.39 b
1.44
1.23 a
4.55 a
6.41 a
7.58 a
6.82
8.10
0.93 a
2.89 a
2.85 b
3.66 b
5.80
4-Year-old primary branches
1990
10
1989
7
1988
6
1987
4
38
7
1
2.61 a
9.43 a
11.42
9.67
1.21 b
4.12 b
1.00
2.98 a
2.42 a
2.30
1.98
2.13 a
1.99 b
1.04
1.24 a
5.35 a
6.96
6.56
0.83 b
2.85 b
1.27
2-Year-old primary branches
1990
12
1989
7
17
2.07 a
11.15
0.98 b
2.10 a
2.73
2.42 a
1.74 a
5.14
0.68 b
Table 4. Mean leaf specific conductivity (LSC) × 1016 in m2, velocity
flow rate (V) × 102 in m s −1, and potentially functional xylem area to
leaf area ratio (Apf/Al) × 106 (unitless) for segments immediately above
and below the transition between the 1-year-old and 2-year-old growth
of primary branch axes and results of paired t-tests (n = 19).
LSC
Above transition
(1-year-old growth)
Below transition
(2-year-old growth)
P-value
V
Apf/Al
Table 5. A comparison of hydraulic parameters for terminal and lateral
leaf-bearing shoots and results of paired comparison F-tests (n = 18
pairs). Table values are leaf specific conductivity (LSC) × 1016 in m2,
velocity flow rate (V) × 102 in m s −1, and potentially functional xylem
area to leaf area ratio (Apf/Al) × 106 (unitless). Lateral means were
computed using values from the two laterals below each terminal.
1.44
1.37
2.13
Shoot position
LSC
V
Apf/Al
8.17
4.78
2.34
< 0.001
< 0.001
> 0.300
Terminal
Lateral
P-value
2.13
1.26
< 0.005
1.49
0.90
< 0.005
2.04
2.05
> 0.100
HYDRAULIC ARCHITECTURE OF WHITE ASH
ence in LSC between distal and proximal branches within an
annual increment was not detected (P > 0.05); however, whenever a branch system in a more distal branch position was more
vigorous (i.e., greater length, larger basal diameter, more extensive branching) than a branch in a proximal position, it
consistently had a higher LSC (Figure 2).
LSC in branches with opposing orientations
Values of LSC at the base of terminal shoots of branches from
the southwest (SW) quadrant (mean = 3.02) were greater (P <
0.05) than those of branches from the northeast (NE) quadrant
(mean = 2.14), and LSC values at the base of shoots from the
southeast (SE) quadrant (mean = 2.35) were greater (P < 0.05)
than those of shoots from the northwest (NW) quadrant (mean
= 1.61). The mean LSC at the base of terminal shoots with a
southern exposure (SW and SE quadrants combined) was 43%
greater (P < 0.05) than that of shoots with a northern exposure
(NW and NE quadrants combined). The average LSC for all
primary and secondary branch segments was also greater (P <
0.05) for branches with a southern exposure (mean = 4.32) than
for branches with a northern exposure (mean = 3.10). Although
not significantly different, both V and Apf/Al values were consistently higher in south-facing branches than in north-facing
branches.
Comparison of parent and descendant segments
Measured values of LSC, V and Apf/Al for parent and descendant segments for 17 trifurcations of third-order branches are
shown in Table 6. There were no significant differences (P >
0.05) between the parent and terminal descendant segments for
any of the parameters. In contrast, significant differences were
detected between the parent and lateral descendant segments
for all three parameters. Mean lateral segment LSC was
63.22% lower than the mean parent LSC as the result of a
Figure 2. Leaf specific conductivity (LSC in m2 × 1016) and transverse
xylem area (Ax in m2 × 106) (in parentheses) along the central axes of
three primary branches arising from the proximal, middle and distal
positions along a common annual extension increment (1986) of the
main stem.
653
Table 6. A comparison of hydraulic parameters between parent (P) and
terminal (DT) and lateral descendant (DL) stem segments of 17 trifurcations of third-order branches and results of paired t-tests. Table
values are leaf specific conductivity (LSC) × 1016 in m2, velocity flow
rate (V) × 102 in m s − 1, and potentially functional xylem area to leaf
area ratio (Apf/Al) × 106 (unitless).
Variable
P
DT
P-value
DL
P-value
LSC
V
Apf /Al
10.06
6.54
2.16
9.22
6.30
2.10
0.455
0.772
0.692
3.64
2.46
1.52
< 0.001
< 0.001
0.001
reductions in both V and Apf/Al. The average V for lateral
segments was 62.38% lower than for parent segments, and the
average Apf/Al ratio was 29.63% lower.
Discussion
Variations in LSC values within the tree crown and within its
component parts resulted primarily from variation in V; by
comparison, the Apf/Al ratio was much less variable. Variations
in both LSC and V were systematically related to branch order.
Stratification by Strahler order accounted for slightly more
than 70% of the total variation in both LSC and V.
The highest LSC values occurred in the main stem. The
hydraulic dominance of the main stem over the primary
branches appeared to arise in the year that shoots were formed
by the elongation of terminal (stem axis) and axillary (branch)
buds on the main stem and was attributable to differences in
both V and Apf/Al (Table 2). Because velocity flow rate is
proportional to the square of mean vessel diameter (Zimmermann and Brown 1971), the observed pattern of xylem hydraulic capacity must be associated with a greater mean vessel
diameter in the main stem relative to that in the branches as has
been previously observed in both conifers (Ewers and Zimmermann 1984a, 1984b) and angiosperms (Zimmermann 1978).
The pattern of LSC distribution within the main stem and its
branches was repeated within the primary branch axes and
their secondary branches. As with the main stem, the pattern of
LSC dominance in the primary branch axis compared to
branches from that axis appeared to arise in the first year of
shoot elongation and persisted over subsequent years (Table 3).
A substantial reduction in the hydraulic capacity of the
xylem was associated with the lateral segments of trifurcations
of primary branches. Although differences in LSC between
parent and lateral descendent segments were associated with
reductions in both components of LSC (Apf/Al and V), they
resulted primarily from differences in V (Table 6), suggesting
that mean vessel diameter is smaller in lateral segments than
in parent segments.
The reduction in LSC at the transition between the 1- and
2-year-old growth resulted from a decrease in V, indicating a
reduction in mean vessel diameter across the transition. Similar reductions have been observed in the xylem of Quercus
alba L. and Quercus rubra L. (Cochard and Tyree 1990),
654
JOYCE AND STEINER
where vessel diameters were significantly smaller in 1-yearold shoots than in 2-year-old shoots. Substantial reductions in
LSC at this transition have also been reported for several
diffuse-porous species (Zimmermann 1978) and may result
from the developmental lag between cambial reactivation and
secondary xylem development observed in elongating shoots
(Larson 1976). The slight acropetal increase in Apf/Al observed
at the transition provided some compensation for the relatively
large decrease in V across this growth node.
Because the terminal growth increment of a main axis consistently had a lower LSC than older portions of the same axis,
the terminal increment itself was analogous to a branch from
the main axis. However, the analogy is imperfect because in
every instance the terminal extension increment had higher
LSC and Ax values and greater extension growth than lateral
shoots of the same age located immediately below the terminal. We have found a similar pattern of hydraulic dominance
within the crowns of eight other white ash trees (unpublished
data). A similar pattern of hydraulic dominance of the main
axis was also observed in Abies balsamea (L.) Mill. (Ewers
and Zimmermann 1984b), where LSC was higher in the terminal leader than in lateral segments of equal diameter, and was
associated with a strong expression of apical control (as defined by Zimmermann and Brown 1971). In contrast, the
plagiotropic leader of Tsuga canadensis (L.) Carr. did not
exhibit hydraulic dominance or a strong expression of apical
control (Ewers and Zimmermann 1984a).
Sellin (1987) suggested that hydraulic architecture plays a
significant role in apical control and overall crown form. He
observed that, in Norway spruce (Picea abies (L.) Karst.), the
terminal shoot grows faster and for a longer period than the
lateral shoots, and that this growth habit is repeated in the
branches, resulting in a conical crown. In white ash, we found
that the terminal shoots exhibited greater elongation and diameter growth than lateral shoots, and the terminal shoots also
had lower resistance to water transport, suggesting a mechanistic linkage between xylem hydraulic capacity and shoot development.
The relationship between hydraulic dominance and morphology observed in the leaf-bearing shoots was also evident
among primary branches arising from a common annual extension increment of the main stem. Typically, distal (upper)
branches arising from a given increment had a higher LSC and
were better developed (greater diameter and length, and more
extensive branching) than their subordinates. Because LSC
values are scaled to leaf area, LSC and branch size are potentially independent. Thus, an association between high LSC and
large size suggests the existence of a physiological connection
between LSC and stem size during branch ontogeny.
Initial differences in LSC and diameter between terminal
and lateral shoots and among lateral shoots originating from
the same extension increment may be attributable to differences in the size and phenology of the buds from which the
shoots developed. Gill (1971) found that current-year increments of white ash have large terminal buds and decreasingly
smaller lateral buds at successively lower nodes. We found a
similar order of dominance in the size and LSC of shoots
arising from buds in these respective positions, indicating that
there is a strong linkage between the size of the bud and
subsequent LSC development (cf. Larson 1976).
We suggest that a high LSC and large stem size in dominant
branches are maintained by a positive feedback mechanism,
whereby initial patterns of LSC dominance that develop in the
year of shoot formation are sustained over subsequent cycles
of growth and dormancy. Presumably, a branch with an initial
superiority in LSC will be better able to acquire water and
nutrients and to sustain higher rates of transpiration than a
branch having a low LSC. Leaves supported by such branches
may also photosynthesize at higher rates (Ewers and Zimmermann 1984a) with an accompanying increase in net assimilation rate. A portion of this photosynthate will be reinvested in
the stem and its developing buds, thus favoring LSC dominance in subsequent years. Thus, the observed patterns of
hydraulic dominance suggest a physiological connection between the LSC of a shoot, the leaves supported by that shoot,
the buds that develop on that shoot, and the LSC of shoots
elongating from those buds.
A similar type of physiological connection between leaf and
shoot is suggested by the observed differences in LSC between
shoots of north- and south-facing branches. In ash in central
Pennsylvania, south-facing shoots supported leaves having a
high evaporative demand relative to leaves supported by northfacing shoots (Joyce and Steiner, unpublished data). The finding that these shoots developed a high LSC suggests that a
functional balance is maintained between transpirational demand and xylem hydraulic supply capacity. Similarly, Shumway et al. (1993) found that LSC in seedlings of Q. rubra and
Liriodendron tulipifera L. developed in close correspondence
with the water regime and evaporative potential of the environment.
The observation of relatively stable Apf /Al values tends to
support the notion of a pipe model of tree architecture (Shinozaki et al. 1964), although the concept of a unit pipe supplying each unit of leaf area is anatomically inaccurate. The
similar Apf /Al values between large and small diameter stems
were the result of large stems having fewer but wider vessels
(as suggested by V values) than small stems. Therefore, the
number of vessels supplying a unit of leaf area must decrease
basipetally. This anatomical plasticity presumably helps to
maintain a functional balance between xylem hydraulic supply
capacity and transpirational demand during crown development. Aloni (1991) suggested that such changes in xylem
anatomy were necessary to compensate for increased transport
distances and supported leaf areas associated with large
branches.
The distribution of flow resistance within the white ash
crown supports Zimmermann’s (1978) theory of hydraulic
segmentation. Several points of hydraulic segmentation were
observed: between segments from above and below the junctions of the main stem and primary branch axes, between
segments from above and below the junctions of primary and
secondary branch axes, and at the transition between the 2year-old and 1-year-old growth of both the main stem and
branch axes. The greatest degree of hydraulic segmentation
HYDRAULIC ARCHITECTURE OF WHITE ASH
was associated with the 1-year-old leaf-bearing shoots. Within
the population of leaf-bearing shoots, a finer level of segmentation was evident between terminal and subordinate, lateral
axes. This pattern of hydraulic architecture insures that water
will flow more readily to the terminal shoots and the main axes
(main stem and primary branches), increasing the likelihood
that the crown’s apices will survive a drought. As a result, these
portions of the crown are preferentially maintained during
drought, thereby contributing to the tree’s long-term survival
under stress. During two dry summers, we observed that
drought-induced leaf death on ash occurred first on lateral and
terminal shoots of branches in lower portions of the crown and
on lateral shoots only of branches in the upper crown. On a few
trees, all leaves had dried except those on the current year’s
extension of the main stem.
Zimmermann’s segmentation hypothesis should be considered together with vulnerability segmentation (Tyree and Ewers 1991). The large diameter vessels associated with the
ring-porous xylem anatomy of white ash conduct water efficiently, but may be prone to drought-induced cavitation. Because embolized vessels were refilled under high pressure, the
LSC values calculated in this study represent a potential maximum and may be an overestimation of the in situ values.
Variability in LSC within the sampled crown was due primarily to variation in V, indicating differences in vessel diameter distribution throughout the crown. The distribution of V
values suggested a basipetal increase in mean vessel diameter
along the main axes and a larger mean vessel diameter in the
main stem relative to branches. Basipetal increases in the mean
diameter of conducting elements have been documented in
angiosperms (Zimmermann 1978, Zimmermann and Potter
1982) and conifers (Ewers and Zimmermann 1984a, 1984b),
and support the hormonal gradient hypothesis of xylem differentiation proposed by Aloni and Zimmermann (1983).
The observed systematic patterns of LSC distribution, corresponding to similar patterns of V distribution, reveal that
regulated changes in xylem anatomy (i.e., vessel diameter
distribution) accompany crown development. Because Strahler order describes relative crown position and presumably is
associated with variation in hydraulic and mechanical demands within the crown, the observed trend of increasing V
with increasing branch order reflects an anatomical plasticity
that enables the xylem to compensate for changes in hydraulic
and mechanical requirements during crown development. In
addition, large decreases in mean vessel diameter (as suggested by velocity values) at certain points within the crown
lead to a segmented pattern of hydraulic architecture, influencing the growth habit and potential survival of the tree.
References
Aloni, R. 1991. Wood formation in deciduous hardwood trees. In
Physiology of Trees. Ed. A.S. Raghavendra. Wiley, New York,
pp 175--197.
Aloni, R. and M.H. Zimmermann. 1983. The control of vessel size and
density along the plant axis: a new hypothesis. Differentiation
24:203--208.
655
Cochard, H. and M.T. Tyree. 1990. Xylem dysfunction in Quercus:
vessel size tyloses, cavitation and seasonal changes in embolism.
Tree Physiol. 6:393--407.
Ewers, F.W. and M.H. Zimmermann. 1984a. The hydraulic architecture of eastern hemlock (Tsuga canadensis). Can. J. Bot. 62:940-946.
Ewers, F.W. and M.H. Zimmermann. 1984b. The hydraulic architecture of balsam fir (Abies balsamea). Physiol. Plant. 60:453--458.
Gill, A.M. 1971. The formation, growth and fate of buds of Fraxinus
americana L. in central Massachusetts. Harvard Forest Paper No.
20, Harvard University, Petersham, MA, 16 p.
Heine, R.W. 1971. Hydraulic conductivity in trees. J. Exp. Bot.
22:503--511.
Isebrands, J.G. and P.R. Larson. 1977. Vascular anatomy of the nodal
region in eastern cottonwood. Am. J. Bot. 64:1066--1077.
Kelso, W.C., C.O. Gertjejanson and R.L. Hossfield. 1963. The effect
of air blockage upon the permeability of wood to liquids. Univ.
Minn. Agric. Res. Stn. Tech. Bull., 242 p.
Larson, P.R. 1976. The leaf--cambium relation and some prospects for
genetic improvement. In Tree Physiology and Yield Improvement.
Eds. M.G.R. Cannell and F.T. Last. Academic Press, New York,
pp 261--282.
McDonald, N. 1983. Trees and networks in biological models. John
Wiley and Sons, New York, pp 160--167.
Preston, R.D. 1952. Movement of water in higher plants. In Deformation and Flow in Biological Systems. Ed. A. Frey-Wyssling. North
Holland Pub. Co., pp 257--321.
Sellin, A. 1987. Hydraulic conductivity of the water transport system
in Norway spruce. Fiziologiya Rastenii 34:545--553 (Translated in
English).
Shinozaki, K., K. Yoda, K. Hozumi and T. Kira. 1964. A quantitative
analysis of plant form: the pipe model theory. I. Basic analyses. Jpn.
J. Ecol. 14:97--105.
Shumway, D.L., K.C. Steiner and M.D. Abrams. 1991. Effects of
drought stress on the hydraulic architecture of seedlings from five
populations of green ash. Can. J. Bot. 69:2158--2164.
Shumway, D.L., K.C. Steiner and T.E. Kolb. 1993. Variation in seedling hydraulic architecture as a function of species and environment. Tree Physiol. 12:41--54.
Sokal, R.R. and F.J. Rohlf. 1969. Biometry. W.H. Freeman and Co.,
San Francisco, 776 p.
Steel, R.G.D. and J.H. Torrie. 1980. Principles and procedures of
statistics. McGraw-Hill, New York, 633 p.
Tyree, M.T. and F.W. Ewers. 1991. The hydraulic architecture of trees
and other woody plants. New Phytol. 119:345--360.
Tyree, M.T. and J.S. Sperry. 1988. Do plants operate near the point of
catastrophic xylem dysfunction caused by dynamic water stress.
Plant Physiol. 88:574--580.
Tyree, M.T., M.E.D. Graham, K.E. Cooper and L.J. Bazos. 1983. The
hydraulic architecture of Thuja occidentalis. Can. J. Bot. 61:2105-2111.
Tyree, M.T., D.A. Snyderman, T.R. Wilmot and J.-L. Machado. 1991.
Water relations and hydraulic architecture of a tropical tree (Scheffelera morotoni): data, models and a comparison of two temperate
species. Plant Physiol. 96:1105--1113.
Yang, S. and M.T. Tyree. 1993. Hydraulic resistance in Acer saccharum shoots and its influence on leaf water potential and transpiration. Tree Physiol. 12:231--242.
Zimmermann, M.H. 1978. Hydraulic architecture of some diffuse-porous trees. Can J. Bot. 56:2286--2295.
Zimmermann, M.H. 1983. Xylem structure and the ascent of sap.
Springer-Verlag, New York, 139 p.
656
JOYCE AND STEINER
Zimmermann, M.H. and C.L. Brown. 1971. Trees: structure and function. Springer-Verlag, New York, 336 p.
Zimmermann, M.H. and D. Potter. 1982. Vessel length distribution in
branches, stem and roots of Acer rubrum L. IAWA Bull. 3:103--109.
© 1995 Heron Publishing----Victoria, Canada
Systematic variation in xylem hydraulic capacity within the crown of
white ash (Fraxinus americana)
BRIAN J. JOYCE1 and KIM C. STEINER2
1
School of Forest Resources, 206 Forest Resources Laboratory, The Pennsylvania State University, University Park, PA 16802, USA
2
School of Forest Resources, 213 Ferguson Building, The Pennsylvania State University, University Park, PA 16802, USA
Received September 19, 1994
Summary A 7-m tall white ash tree (Fraxinus americana
Marsh.) was dissected, and hydraulic parameters of the xylem
were determined by inducing a steady-state flow of water
through the stem segments and monitoring volume and velocity
flow rates. Leaf-specific conductivity (LSC) was highest in the
main stem and lowest in some of the leaf-bearing lateral shoots.
The LSC was higher in the main stem than in branches and
higher in primary than in secondary branch axes. Terminal
leaf-bearing shoots were larger and had a significantly greater
mean LSC than subjacent lateral shoots. A significant reduction
in LSC was associated with the transition between 1- and
2-year-old growth. In branches of the same age, there was a
close correspondence among LSC, branch position and branch
size. The average LSC of leaf-bearing shoots from south-facing
branches was 43% greater than that of shoots from north-facing
branches.
Within-crown variation in LSC was associated with variation in velocity flow rate (V). By contrast, the ratio of potentially functional xylem area to supported leaf area (Apf/Al) was
relatively stable throughout the crown. Stratification of stems
by Strahler order accounted for approximately 70% of the total
variation in LSC. These results suggest that (1) there exists a
systematic pattern of variation in LSC distribution within the
crown of white ash, (2) within-crown variability in LSC is
primarily the result of variation in mean vessel diameter, and
(3) there is a physiological linkage between LSC and crown
morphology that is maintained through a positive feedback
mechanism during branch ontogeny.
Keywords: hydraulic architecture, hydraulic dominance, hydraulic segmentation, leaf-specific conductivity.
Introduction
Xylem tissue in trees serves several biological functions including mechanical support and vascular transport. Because
the relative importance of these functions varies according to
branch location, there may exist some measure of order in the
structure of the internal water transport system and its relationship to the morphology of a tree’s crown.
Long distance transport of water and mineral nutrients is a
vital function of the xylem. A practical measure of the xylem’s
hydraulic supply capacity is leaf-specific conductivity (LSC)
(Zimmermann 1978). For example, the LSC of a stem segment
can be used to relate the average transpirational water loss (E)
from leaves supported by the segment to the decline in water
potential (dΨ) per unit path length (dx) within the segment
(dΨ/dx = E/LSC). Thus, the distribution of LSC within a tree
influences patterns of Ψ throughout the crown and can impose
constraints on such physiological processes as transpiration
and photosynthesis (Tyree and Sperry 1988, Shumway et al.
1993, Yang and Tyree 1993).
Within-tree patterns of LSC distribution have been quantified for several angiosperm and conifer tree species (Zimmermann 1978, 1983, Tyree et al. 1983, Ewers and Zimmermann
1984a, 1984b, Tyree et al. 1991). In most species examined,
LSC increases with increasing stem diameter and is higher in
the main stem than in branches. Another consistently observed
feature of tree hydraulic architecture is the increased hydraulic
resistance associated with leaf and branch junctions (Isebrands
and Larson 1977, Zimmermann 1978, Tyree et al. 1983, Ewers
and Zimmermann 1984a, 1984b).
Zimmermann (1983) considered that LSC distribution patterns in tree crowns were evidence of hydraulic segmentation,
defined as a structural feature that confines cavitation events to
relatively expendable plant parts in favor of parts more important to plant survival. Hydraulic segmentation is considered an
architectural advantage in dicotyledonous tree species because
leaves and small diameter stems are more readily replaced than
older, supporting stem structures.
Xylem also provides mechanical support to the crown. The
distribution of xylem anatomical features within a tree’s crown
likely reflects a compromise between hydraulic and mechanical support functions. Because the multiple roles of xylem
depend directly on its anatomical characteristics, we hypothesized that crown development will result in an organized pattern of xylem anatomical features. The objective of this study
was to describe this pattern of organization in white ash (Fraxinus americana Marsh.) by characterizing variations in several
xylem hydraulic parameters throughout an entire tree crown.
We also examined the relationship between functional and
morphological attributes.
650
JOYCE AND STEINER
Materials and methods
Plant material and branch segmentation
An 18-year-old white ash tree that was approximately 7 m in
height was selected for study from within a plantation population in Center County, PA. The entire crown and main stem of
the tree was destructively sampled. Starting in early September, all branches in the crown were labeled, and leaf areas were
measured. Stem segments were coded starting with the distal
end of the lowest branch. The first internode of at least 0.075 m
in length was identified as segment one, and the leaves supported by that segment were collected and placed in a high
humidity bag. This procedure was continued basipetally for all
lateral and terminal branch segments. Segment codes were
used to identify parent and descendent segments associated
with branch trifurcations (ash has opposite leaves). Leaf area
(Al) supported by each current-year shoot was measured to the
nearest 0.01 cm2 with a model LI-3000 surface area meter
(Li-Cor Inc., Lincoln, NE). After all stem segments were
identified on the tree (390 total), detailed drawings were made
of each branch system noting the relative position of all labeled
segments and the compass quadrant orientation of each branch
from the main stem: NE (1--90°), SE (91--180°), SW (181-270°), and NW (271--360°). The Strahler order (McDonald
1983) of each segment was also noted.
To facilitate within-crown comparisons, stem segments
were assigned to a hierarchy of crown components (Figure 1):
the main stem with its primary branches, and primary branch
systems comprising a central axis and secondary branches.
Measurement of xylem hydraulic parameters
Xylem hydraulic parameters were measured during the dormant season from November 3 to February 11. On each sampling day, an entire branch system was cut from the tree with
hand pruners and transported to the laboratory. Segments approximately 0.15 m long were excised from labeled internodes
and cut to 0.075 m in length with a miter trimmer, to assure a
smooth stem surface without crushing the vessels. Segments
were immersed in a water bath at 20 °C until measurements
were made. A 1.0-cm-wide band of bark and phloem tissue was
removed from the ends of the segments, and the segments were
fitted with plastic tubing with a neoprene seal.
To remove air emboli, stem segments were then attached to
an apparatus similar in design to that described by Kelso et al.
(1963), and pressurized at their proximal ends with filtered
(0.2 µm retention) and de-aerated, distilled water at a gradient
of 0.75 MPa m −1 until a steady-state flow condition was
achieved. The pressure gradient was then reduced to 0.14 MPa
m −1, and the system was allowed to reach steady-state flow.
Because there was no decrease in flow with time, a perfusing
agent was not used (see Zimmermann 1978). Steady-state
mass flow rate (M, kg s −1) was then measured gravimetrically
by collection of induced flow in beakers over a 15-min interval
and converted to volume flow rate (Q, m3 s −1). At the end of
the 15-min interval, 0.5 M KCl was injected with a syringe into
the water stream as it entered the segment. Electrical resistance
was monitored at the segment output. The average velocity
flow rate (V, m s −1) was estimated from segment length and the
elapsed time between injection of KCl and the average change
in electrical resistance of the output stream. The potentially
functional transverse xylem area (Apf, m2) (Shumway et al.
1991) was determined as Apf = Q/V (Preston 1952, Heine
1971). Because air emboli were removed under positive pressure, the derived Apf parameter represents a maximum conductive xylem area and may include vessels that were
dysfunctional in situ. Leaf specific conductivity (LSC, m2) was
calculated as:
Apf L η
LSC = V
,
Al P
where L is the segment length (m), η is the absolute viscosity
of water at 20 °C (10 −9 MPa s), Al is the leaf area supported by
the stem segment (m2), and P is the pressure difference across
the segment (MPa). Most published values of LSC do not
include units of viscosity and must be multiplied by 10 −12 m3
kg −1 MPa s (at 20 °C) to obtain the same units for LSC.
For each stem segment, total xylem (less pith) cross-sectional area (Ax) (m2) was measured on a thin, transverse section
from the segment’s midpoint by means of a digital caliper and
stereomicroscope.
Statistical methods
Figure 1. Diagrammatic representation of the descriptive terminology
used in the text. Diagram is representative of the branching pattern of
the white ash crown examined.
Paired comparison analysis (see Sokal and Rohlf 1969) was
used to test differences in both hydraulic and morphological
variables between branches with opposing compass directions
and between terminal and lateral leaf-bearing shoots. Each
HYDRAULIC ARCHITECTURE OF WHITE ASH
individual branch pair (originating from the same node) and
each terminal-lateral shoot pair was treated as a within-crown
replicate. Comparisons of xylem characteristics between parent and descendant segments of branch trifurcations were
made by paired t-tests (α = 0.05). Trifurcation comparisons
were made using the mean of the descendant segment values
and the parent segment value. Paired t-tests were also used to
test differences in hydraulic parameters among branches arising from a common annual extension increment of the main
stem, between segments immediately above and below the
transition between the 1-year-old and 2-year-old extension
growth of branch axes and between different year extension
increments of primary and secondary branch axes. Analysis of
variance (ANOVA) and mean separation procedures were conducted with Statistical Analysis Systems software (SAS Inc.,
Cary, NC). Variance component analysis (Steel and Torie
1980) was used to calculate the proportion of total variance in
the measured variables explained by Strahler order.
651
der (ultimate) segments and increased significantly (P < 0.05)
with increasing branch order (Table 1). The Apf/Al ratio showed
little variation (P > 0.05) among branch orders. Strahler order
accounted for 72.53, 0.68 and 84.87% of the total variability
in LSC, Apf /Al and V, respectively. Values of LSC for the main
stem were higher than those for primary branch axes as a result
of substantially greater V and Apf/Al values in the main stem
(Table 2). The mean LSC of the main stem (mean = 20.91) was
more than 2.9 times that of the proximal segments of adjacent
branches (mean = 7.80). Values of LSC were similar among
segments along the main stem until the last extension increment (1990), where there was a sharp decrease from 22.60 to
Table 1. Mean (± 1 SE) leaf specific conductivity (LSC) × 1016 in m2,
potentially functional xylem area to leaf area ratio (Apf/Al) × 106
(unitless), and velocity flow rate (V) × 102 in m s − 1 for all segments
from white ash crown and for segments stratified by Strahler order.
Means within a column appearing with the same letter are not significantly different (P > 0.05) from one another using Fisher’s LSD test.
Results
Strahler
order
n
LSC
Apf /Al
V
General distribution of xylem hydraulic and morphological
parameter values
All orders
Order 1
Order 2
Order 3
Order 4
390
257
102
26
5
4.58 (0.27)
1.72 (0.07) a
9.06 (0.52) b
12.60 (1.18) c
18.62 (3.10) d
2.21 (0.59)
2.21 (0.78) a
2.21 (1.10) a
2.10 (1.10) a
2.30 (2.90) a
2.94 (1.51)
1.18 (0.47) a
5.63 (2.10) b
8.00 (4.71) c
12.75 (13.78) d
Values of LSC and V varied considerably (CV = 1.16 and 1.01,
respectively) within the crown, but the Apf/Al ratio was less
variable (CV = 0.53). Both LSC and V were lowest in first-or-
Table 2. A comparison of mean leaf specific conductivity (LSC) × 1016 in m2, potentially functional xylem area to leaf area ratio (Apf/Al) × 106
(unitless), and velocity flow rate (V) × 102 in m s − 1 between the sequence of annual extension increments that comprise the main stem and the
central axes of primary branches originating from the same annual increment of the main stem (n = number of segments representing each year of
growth). Primary branch values within a column represent successive annual segments along those branch axes. All values within a row represent
segments of the same age but at different crown positions.
Year of growth
LSC (n)
1990
1989
1988
1987
1986
1985
Main stem
5.27 (4)
22.60 (2)
25.65 (2)
25.60 (2)
23.27 (3)
23.10 (1)
Primary branches originating from main stem increment of
1984
1985
1986
1987
1988
1989
1.88 (12)
7.40 (3)
11.06 (11)
11.19 (8)
10.91 (10)
7.60 (3)
0.75 (4)
9.50 (1)
5.65 (4)
5.70 (5)
6.38 (4)
1.98 (18)
7.89 (11)
10.30 (10)
7.53 (9)
2.01 (9)
13.75 (6)
12.08 (5)
2.04 (15)
10.61 (9)
2.61 (20)
Apf/Al
1990
1989
1988
1987
1986
1985
4.24
4.96
3.06
2.82
2.57
1.41
2.37
2.53
2.13
2.08
2.02
1.26
1.15
2.63
1.78
1.26
1.28
2.57
2.32
2.30
1.51
2.13
3.32
2.63
2.10
2.60
1.35
V
1990
1989
1988
1987
1986
1985
1.80
6.84
11.60
12.50
13.13
15.00
1.19
4.55
6.38
7.56
7.67
8.53
0.94
5.00
5.22
5.66
6.81
1.10
4.68
6.24
7.59
1.31
5.51
6.31
1.65
5.59
2.64
652
JOYCE AND STEINER
5.27. Values of V declined steadily along the main stem with a
pronounced decrease (from 6.84 to 1.80) at the 1990 extension
increment. The Apf /Al ratio increased acropetally along the
main stem, with only a slight decline (4.96 to 4.24) at the 1990
extension increment. Similar trends for LSC, V and Apf/Al
occurred along the axes of primary branches that originated in
1985 from the 1984 increment of the main stem.
Differences in xylem hydraulic parameters between the central and lateral axes of primary branch systems were similar to
the observed differences between the main stem and the central
axes of primary branches from the main stem (Table 3). Values
of LSC, V and Apf/Al of primary branch axes were consistently
higher than those of secondary branch axes.
In all branches sampled, the transition between 1- and
2-year-old extension growth represented a point of reduction
in hydraulic supply capacity, i.e., there were significant reductions (P < 0.05) in mean LSC and V across the transition
(Table 4). The Apf/Al ratio did not differ significantly in segments from above and below the transition.
Xylem hydraulic parameters in terminal versus lateral
leaf-bearing shoots
The values of LSC varied in a consistent manner between
terminal and lateral leaf-bearing shoots. The mean LSC of
shoots arising from terminal buds was significantly greater
(P < 0.01) than that of shoots arising from axillary buds immediately below the terminal (Table 5). The difference is attributable to a difference in V, because there was no difference (P >
0.05) in Apf /Al. Terminal shoots were longer than lateral shoots
and had a greater (P < 0.05) mean Ax.
Variation in LSC among branches arising from a common
annual extension increment
Within an annual extension increment of the main stem, there
was a pattern of hydraulic dominance (higher LSC) expressed
by branches in distal (superior) positions compared to their
subordinates. Because of the small number of comparisons
(n = 5 pairs) and the large amount of variation that existed
among the various extension increments, a significant differ-
Table 3. A comparison of mean leaf specific conductivity (LSC) × 1016 in m2, potentially functional xylem area to leaf area ratio (Apf/Al) × 106
(unitless), and velocity flow rate (V) × 102 in m s −1 between the central axis and lateral branches of 2-, 4- and 6-year-old primary branch systems.
Each branch age group is represented by three branches (n = number of segments in sample). For a given variable within a branch age group, means
within a row appearing with the same letter are not significantly different from one another (P > 0.05) using paired t-tests.
Year of growth
n (axis)
n (branch)
Apf/Al
LSC
V
Axis
Branch
Axis
Branch
Axis
Branch
6-Year-old primary branches
1990
10
1989
2
1988
9
1987
6
1986
5
1985
4
39
9
12
7
1
1.90 a
6.30 a
10.14 a
11.00 a
10.32
7.08
1.12 b
2.97 b
2.27 b
3.64 b
6.30
2.37 a
2.53 a
2.25 a
2.13 a
2.15
1.22
1.77 b
1.89 a
1.67 b
1.39 b
1.44
1.23 a
4.55 a
6.41 a
7.58 a
6.82
8.10
0.93 a
2.89 a
2.85 b
3.66 b
5.80
4-Year-old primary branches
1990
10
1989
7
1988
6
1987
4
38
7
1
2.61 a
9.43 a
11.42
9.67
1.21 b
4.12 b
1.00
2.98 a
2.42 a
2.30
1.98
2.13 a
1.99 b
1.04
1.24 a
5.35 a
6.96
6.56
0.83 b
2.85 b
1.27
2-Year-old primary branches
1990
12
1989
7
17
2.07 a
11.15
0.98 b
2.10 a
2.73
2.42 a
1.74 a
5.14
0.68 b
Table 4. Mean leaf specific conductivity (LSC) × 1016 in m2, velocity
flow rate (V) × 102 in m s −1, and potentially functional xylem area to
leaf area ratio (Apf/Al) × 106 (unitless) for segments immediately above
and below the transition between the 1-year-old and 2-year-old growth
of primary branch axes and results of paired t-tests (n = 19).
LSC
Above transition
(1-year-old growth)
Below transition
(2-year-old growth)
P-value
V
Apf/Al
Table 5. A comparison of hydraulic parameters for terminal and lateral
leaf-bearing shoots and results of paired comparison F-tests (n = 18
pairs). Table values are leaf specific conductivity (LSC) × 1016 in m2,
velocity flow rate (V) × 102 in m s −1, and potentially functional xylem
area to leaf area ratio (Apf/Al) × 106 (unitless). Lateral means were
computed using values from the two laterals below each terminal.
1.44
1.37
2.13
Shoot position
LSC
V
Apf/Al
8.17
4.78
2.34
< 0.001
< 0.001
> 0.300
Terminal
Lateral
P-value
2.13
1.26
< 0.005
1.49
0.90
< 0.005
2.04
2.05
> 0.100
HYDRAULIC ARCHITECTURE OF WHITE ASH
ence in LSC between distal and proximal branches within an
annual increment was not detected (P > 0.05); however, whenever a branch system in a more distal branch position was more
vigorous (i.e., greater length, larger basal diameter, more extensive branching) than a branch in a proximal position, it
consistently had a higher LSC (Figure 2).
LSC in branches with opposing orientations
Values of LSC at the base of terminal shoots of branches from
the southwest (SW) quadrant (mean = 3.02) were greater (P <
0.05) than those of branches from the northeast (NE) quadrant
(mean = 2.14), and LSC values at the base of shoots from the
southeast (SE) quadrant (mean = 2.35) were greater (P < 0.05)
than those of shoots from the northwest (NW) quadrant (mean
= 1.61). The mean LSC at the base of terminal shoots with a
southern exposure (SW and SE quadrants combined) was 43%
greater (P < 0.05) than that of shoots with a northern exposure
(NW and NE quadrants combined). The average LSC for all
primary and secondary branch segments was also greater (P <
0.05) for branches with a southern exposure (mean = 4.32) than
for branches with a northern exposure (mean = 3.10). Although
not significantly different, both V and Apf/Al values were consistently higher in south-facing branches than in north-facing
branches.
Comparison of parent and descendant segments
Measured values of LSC, V and Apf/Al for parent and descendant segments for 17 trifurcations of third-order branches are
shown in Table 6. There were no significant differences (P >
0.05) between the parent and terminal descendant segments for
any of the parameters. In contrast, significant differences were
detected between the parent and lateral descendant segments
for all three parameters. Mean lateral segment LSC was
63.22% lower than the mean parent LSC as the result of a
Figure 2. Leaf specific conductivity (LSC in m2 × 1016) and transverse
xylem area (Ax in m2 × 106) (in parentheses) along the central axes of
three primary branches arising from the proximal, middle and distal
positions along a common annual extension increment (1986) of the
main stem.
653
Table 6. A comparison of hydraulic parameters between parent (P) and
terminal (DT) and lateral descendant (DL) stem segments of 17 trifurcations of third-order branches and results of paired t-tests. Table
values are leaf specific conductivity (LSC) × 1016 in m2, velocity flow
rate (V) × 102 in m s − 1, and potentially functional xylem area to leaf
area ratio (Apf/Al) × 106 (unitless).
Variable
P
DT
P-value
DL
P-value
LSC
V
Apf /Al
10.06
6.54
2.16
9.22
6.30
2.10
0.455
0.772
0.692
3.64
2.46
1.52
< 0.001
< 0.001
0.001
reductions in both V and Apf/Al. The average V for lateral
segments was 62.38% lower than for parent segments, and the
average Apf/Al ratio was 29.63% lower.
Discussion
Variations in LSC values within the tree crown and within its
component parts resulted primarily from variation in V; by
comparison, the Apf/Al ratio was much less variable. Variations
in both LSC and V were systematically related to branch order.
Stratification by Strahler order accounted for slightly more
than 70% of the total variation in both LSC and V.
The highest LSC values occurred in the main stem. The
hydraulic dominance of the main stem over the primary
branches appeared to arise in the year that shoots were formed
by the elongation of terminal (stem axis) and axillary (branch)
buds on the main stem and was attributable to differences in
both V and Apf/Al (Table 2). Because velocity flow rate is
proportional to the square of mean vessel diameter (Zimmermann and Brown 1971), the observed pattern of xylem hydraulic capacity must be associated with a greater mean vessel
diameter in the main stem relative to that in the branches as has
been previously observed in both conifers (Ewers and Zimmermann 1984a, 1984b) and angiosperms (Zimmermann 1978).
The pattern of LSC distribution within the main stem and its
branches was repeated within the primary branch axes and
their secondary branches. As with the main stem, the pattern of
LSC dominance in the primary branch axis compared to
branches from that axis appeared to arise in the first year of
shoot elongation and persisted over subsequent years (Table 3).
A substantial reduction in the hydraulic capacity of the
xylem was associated with the lateral segments of trifurcations
of primary branches. Although differences in LSC between
parent and lateral descendent segments were associated with
reductions in both components of LSC (Apf/Al and V), they
resulted primarily from differences in V (Table 6), suggesting
that mean vessel diameter is smaller in lateral segments than
in parent segments.
The reduction in LSC at the transition between the 1- and
2-year-old growth resulted from a decrease in V, indicating a
reduction in mean vessel diameter across the transition. Similar reductions have been observed in the xylem of Quercus
alba L. and Quercus rubra L. (Cochard and Tyree 1990),
654
JOYCE AND STEINER
where vessel diameters were significantly smaller in 1-yearold shoots than in 2-year-old shoots. Substantial reductions in
LSC at this transition have also been reported for several
diffuse-porous species (Zimmermann 1978) and may result
from the developmental lag between cambial reactivation and
secondary xylem development observed in elongating shoots
(Larson 1976). The slight acropetal increase in Apf/Al observed
at the transition provided some compensation for the relatively
large decrease in V across this growth node.
Because the terminal growth increment of a main axis consistently had a lower LSC than older portions of the same axis,
the terminal increment itself was analogous to a branch from
the main axis. However, the analogy is imperfect because in
every instance the terminal extension increment had higher
LSC and Ax values and greater extension growth than lateral
shoots of the same age located immediately below the terminal. We have found a similar pattern of hydraulic dominance
within the crowns of eight other white ash trees (unpublished
data). A similar pattern of hydraulic dominance of the main
axis was also observed in Abies balsamea (L.) Mill. (Ewers
and Zimmermann 1984b), where LSC was higher in the terminal leader than in lateral segments of equal diameter, and was
associated with a strong expression of apical control (as defined by Zimmermann and Brown 1971). In contrast, the
plagiotropic leader of Tsuga canadensis (L.) Carr. did not
exhibit hydraulic dominance or a strong expression of apical
control (Ewers and Zimmermann 1984a).
Sellin (1987) suggested that hydraulic architecture plays a
significant role in apical control and overall crown form. He
observed that, in Norway spruce (Picea abies (L.) Karst.), the
terminal shoot grows faster and for a longer period than the
lateral shoots, and that this growth habit is repeated in the
branches, resulting in a conical crown. In white ash, we found
that the terminal shoots exhibited greater elongation and diameter growth than lateral shoots, and the terminal shoots also
had lower resistance to water transport, suggesting a mechanistic linkage between xylem hydraulic capacity and shoot development.
The relationship between hydraulic dominance and morphology observed in the leaf-bearing shoots was also evident
among primary branches arising from a common annual extension increment of the main stem. Typically, distal (upper)
branches arising from a given increment had a higher LSC and
were better developed (greater diameter and length, and more
extensive branching) than their subordinates. Because LSC
values are scaled to leaf area, LSC and branch size are potentially independent. Thus, an association between high LSC and
large size suggests the existence of a physiological connection
between LSC and stem size during branch ontogeny.
Initial differences in LSC and diameter between terminal
and lateral shoots and among lateral shoots originating from
the same extension increment may be attributable to differences in the size and phenology of the buds from which the
shoots developed. Gill (1971) found that current-year increments of white ash have large terminal buds and decreasingly
smaller lateral buds at successively lower nodes. We found a
similar order of dominance in the size and LSC of shoots
arising from buds in these respective positions, indicating that
there is a strong linkage between the size of the bud and
subsequent LSC development (cf. Larson 1976).
We suggest that a high LSC and large stem size in dominant
branches are maintained by a positive feedback mechanism,
whereby initial patterns of LSC dominance that develop in the
year of shoot formation are sustained over subsequent cycles
of growth and dormancy. Presumably, a branch with an initial
superiority in LSC will be better able to acquire water and
nutrients and to sustain higher rates of transpiration than a
branch having a low LSC. Leaves supported by such branches
may also photosynthesize at higher rates (Ewers and Zimmermann 1984a) with an accompanying increase in net assimilation rate. A portion of this photosynthate will be reinvested in
the stem and its developing buds, thus favoring LSC dominance in subsequent years. Thus, the observed patterns of
hydraulic dominance suggest a physiological connection between the LSC of a shoot, the leaves supported by that shoot,
the buds that develop on that shoot, and the LSC of shoots
elongating from those buds.
A similar type of physiological connection between leaf and
shoot is suggested by the observed differences in LSC between
shoots of north- and south-facing branches. In ash in central
Pennsylvania, south-facing shoots supported leaves having a
high evaporative demand relative to leaves supported by northfacing shoots (Joyce and Steiner, unpublished data). The finding that these shoots developed a high LSC suggests that a
functional balance is maintained between transpirational demand and xylem hydraulic supply capacity. Similarly, Shumway et al. (1993) found that LSC in seedlings of Q. rubra and
Liriodendron tulipifera L. developed in close correspondence
with the water regime and evaporative potential of the environment.
The observation of relatively stable Apf /Al values tends to
support the notion of a pipe model of tree architecture (Shinozaki et al. 1964), although the concept of a unit pipe supplying each unit of leaf area is anatomically inaccurate. The
similar Apf /Al values between large and small diameter stems
were the result of large stems having fewer but wider vessels
(as suggested by V values) than small stems. Therefore, the
number of vessels supplying a unit of leaf area must decrease
basipetally. This anatomical plasticity presumably helps to
maintain a functional balance between xylem hydraulic supply
capacity and transpirational demand during crown development. Aloni (1991) suggested that such changes in xylem
anatomy were necessary to compensate for increased transport
distances and supported leaf areas associated with large
branches.
The distribution of flow resistance within the white ash
crown supports Zimmermann’s (1978) theory of hydraulic
segmentation. Several points of hydraulic segmentation were
observed: between segments from above and below the junctions of the main stem and primary branch axes, between
segments from above and below the junctions of primary and
secondary branch axes, and at the transition between the 2year-old and 1-year-old growth of both the main stem and
branch axes. The greatest degree of hydraulic segmentation
HYDRAULIC ARCHITECTURE OF WHITE ASH
was associated with the 1-year-old leaf-bearing shoots. Within
the population of leaf-bearing shoots, a finer level of segmentation was evident between terminal and subordinate, lateral
axes. This pattern of hydraulic architecture insures that water
will flow more readily to the terminal shoots and the main axes
(main stem and primary branches), increasing the likelihood
that the crown’s apices will survive a drought. As a result, these
portions of the crown are preferentially maintained during
drought, thereby contributing to the tree’s long-term survival
under stress. During two dry summers, we observed that
drought-induced leaf death on ash occurred first on lateral and
terminal shoots of branches in lower portions of the crown and
on lateral shoots only of branches in the upper crown. On a few
trees, all leaves had dried except those on the current year’s
extension of the main stem.
Zimmermann’s segmentation hypothesis should be considered together with vulnerability segmentation (Tyree and Ewers 1991). The large diameter vessels associated with the
ring-porous xylem anatomy of white ash conduct water efficiently, but may be prone to drought-induced cavitation. Because embolized vessels were refilled under high pressure, the
LSC values calculated in this study represent a potential maximum and may be an overestimation of the in situ values.
Variability in LSC within the sampled crown was due primarily to variation in V, indicating differences in vessel diameter distribution throughout the crown. The distribution of V
values suggested a basipetal increase in mean vessel diameter
along the main axes and a larger mean vessel diameter in the
main stem relative to branches. Basipetal increases in the mean
diameter of conducting elements have been documented in
angiosperms (Zimmermann 1978, Zimmermann and Potter
1982) and conifers (Ewers and Zimmermann 1984a, 1984b),
and support the hormonal gradient hypothesis of xylem differentiation proposed by Aloni and Zimmermann (1983).
The observed systematic patterns of LSC distribution, corresponding to similar patterns of V distribution, reveal that
regulated changes in xylem anatomy (i.e., vessel diameter
distribution) accompany crown development. Because Strahler order describes relative crown position and presumably is
associated with variation in hydraulic and mechanical demands within the crown, the observed trend of increasing V
with increasing branch order reflects an anatomical plasticity
that enables the xylem to compensate for changes in hydraulic
and mechanical requirements during crown development. In
addition, large decreases in mean vessel diameter (as suggested by velocity values) at certain points within the crown
lead to a segmented pattern of hydraulic architecture, influencing the growth habit and potential survival of the tree.
References
Aloni, R. 1991. Wood formation in deciduous hardwood trees. In
Physiology of Trees. Ed. A.S. Raghavendra. Wiley, New York,
pp 175--197.
Aloni, R. and M.H. Zimmermann. 1983. The control of vessel size and
density along the plant axis: a new hypothesis. Differentiation
24:203--208.
655
Cochard, H. and M.T. Tyree. 1990. Xylem dysfunction in Quercus:
vessel size tyloses, cavitation and seasonal changes in embolism.
Tree Physiol. 6:393--407.
Ewers, F.W. and M.H. Zimmermann. 1984a. The hydraulic architecture of eastern hemlock (Tsuga canadensis). Can. J. Bot. 62:940-946.
Ewers, F.W. and M.H. Zimmermann. 1984b. The hydraulic architecture of balsam fir (Abies balsamea). Physiol. Plant. 60:453--458.
Gill, A.M. 1971. The formation, growth and fate of buds of Fraxinus
americana L. in central Massachusetts. Harvard Forest Paper No.
20, Harvard University, Petersham, MA, 16 p.
Heine, R.W. 1971. Hydraulic conductivity in trees. J. Exp. Bot.
22:503--511.
Isebrands, J.G. and P.R. Larson. 1977. Vascular anatomy of the nodal
region in eastern cottonwood. Am. J. Bot. 64:1066--1077.
Kelso, W.C., C.O. Gertjejanson and R.L. Hossfield. 1963. The effect
of air blockage upon the permeability of wood to liquids. Univ.
Minn. Agric. Res. Stn. Tech. Bull., 242 p.
Larson, P.R. 1976. The leaf--cambium relation and some prospects for
genetic improvement. In Tree Physiology and Yield Improvement.
Eds. M.G.R. Cannell and F.T. Last. Academic Press, New York,
pp 261--282.
McDonald, N. 1983. Trees and networks in biological models. John
Wiley and Sons, New York, pp 160--167.
Preston, R.D. 1952. Movement of water in higher plants. In Deformation and Flow in Biological Systems. Ed. A. Frey-Wyssling. North
Holland Pub. Co., pp 257--321.
Sellin, A. 1987. Hydraulic conductivity of the water transport system
in Norway spruce. Fiziologiya Rastenii 34:545--553 (Translated in
English).
Shinozaki, K., K. Yoda, K. Hozumi and T. Kira. 1964. A quantitative
analysis of plant form: the pipe model theory. I. Basic analyses. Jpn.
J. Ecol. 14:97--105.
Shumway, D.L., K.C. Steiner and M.D. Abrams. 1991. Effects of
drought stress on the hydraulic architecture of seedlings from five
populations of green ash. Can. J. Bot. 69:2158--2164.
Shumway, D.L., K.C. Steiner and T.E. Kolb. 1993. Variation in seedling hydraulic architecture as a function of species and environment. Tree Physiol. 12:41--54.
Sokal, R.R. and F.J. Rohlf. 1969. Biometry. W.H. Freeman and Co.,
San Francisco, 776 p.
Steel, R.G.D. and J.H. Torrie. 1980. Principles and procedures of
statistics. McGraw-Hill, New York, 633 p.
Tyree, M.T. and F.W. Ewers. 1991. The hydraulic architecture of trees
and other woody plants. New Phytol. 119:345--360.
Tyree, M.T. and J.S. Sperry. 1988. Do plants operate near the point of
catastrophic xylem dysfunction caused by dynamic water stress.
Plant Physiol. 88:574--580.
Tyree, M.T., M.E.D. Graham, K.E. Cooper and L.J. Bazos. 1983. The
hydraulic architecture of Thuja occidentalis. Can. J. Bot. 61:2105-2111.
Tyree, M.T., D.A. Snyderman, T.R. Wilmot and J.-L. Machado. 1991.
Water relations and hydraulic architecture of a tropical tree (Scheffelera morotoni): data, models and a comparison of two temperate
species. Plant Physiol. 96:1105--1113.
Yang, S. and M.T. Tyree. 1993. Hydraulic resistance in Acer saccharum shoots and its influence on leaf water potential and transpiration. Tree Physiol. 12:231--242.
Zimmermann, M.H. 1978. Hydraulic architecture of some diffuse-porous trees. Can J. Bot. 56:2286--2295.
Zimmermann, M.H. 1983. Xylem structure and the ascent of sap.
Springer-Verlag, New York, 139 p.
656
JOYCE AND STEINER
Zimmermann, M.H. and C.L. Brown. 1971. Trees: structure and function. Springer-Verlag, New York, 336 p.
Zimmermann, M.H. and D. Potter. 1982. Vessel length distribution in
branches, stem and roots of Acer rubrum L. IAWA Bull. 3:103--109.