Forest Ecology and Management 255 (2008) 3659–3666

Contents lists available at ScienceDirect

Forest Ecology and Management

journal homepage: www.elsevier.com/locate/foreco

Beech regeneration of seed and root sucker origin: A comparison

of morphology, growth, survival, and response to defoliation
Marilou Beaudet *, Christian Messier
Centre d’étude de la Forêt (CEF), Département des Sciences Biologiques, Université du Québec à Montréal, C.P. 8888, Succ. Centre-Ville,
Montréal, Québec, Canada H3C 3P8

A R T I C L E I N F O A B S T R A C T

Article history: American beech (Fagus grandifolia Ehrh.) reproduces sexually, and vegetatively by root suckers. Although
Received 30 November 2007 many studies have investigated its regeneration response, most did not account for differences that may
Received in revised form 28 January 2008 exist between its two modes of reproduction. This study was performed in an old-growth Acer - Fagus
Accepted 6 March 2008
forest in southern Quebec, where beech bark disease had only a minor effect at the time of the study. We
compared the density and frequency of occurrence of beech seedlings and root suckers (height < 30 cm),
Keywords: as well as their morphology, growth, survival, and response to experimental defoliation. Root suckers
Fagus grandifolia
accounted for 13% of beech regeneration at our site. Density and frequency of occurrence were greater
Root sprouts
for seedlings than suckers, but did not vary with light availability, which was low at our study site (mean:
Vegetative reproduction
Leaf removal 2.9%). Seedlings and suckers did not differ in leaf characteristics, but several differences were observed in
Leaf display terms of plant morphology, growth, and survival. Root suckers showed more lateral growth than height
growth, and had a lower leaf area index than seedlings. Root suckers had both a greater growth in height
and diameter, and a higher survivorship than seedlings (height and diameter growth were, respectively,
five and two times greater for suckers than seedlings, and 74% of suckers survived more than 1 year,
compared to 52% for seedlings). Defoliation treatments, which included levels of defoliation of 50% and
100% (1) did not affect current-year extension growth of seedlings and suckers; (2) did not affect seedling
diameter growth, but had a negative impact on sucker diameter growth; and (3) affected survivorship for
both origins, but had a much greater negative impact on seedling survivorship (none of the completely
defoliated seedlings survived over one year, while 55% of the suckers did). This study showed that several
differences exist between small beech seedlings and root suckers in traits that are important
determinants of a species’ competitive ability. We therefore expect that variation in the relative
importance of root suckering among sites might have several community-level implications.
ß 2008 Elsevier B.V. All rights reserved.

1. Introduction species (e.g., Populus tremuloides Michx., Frey et al., 2003), this
mode of vegetative reproduction is less well understood in shade-
Vegetative reproduction occurs in many tree species, through tolerant tree species (Jones and Raynal, 1988), partly because few
layering, stump sprouting and root suckering (Del Tredici, 2001). of those species produce root suckers (Peterson and Jones, 1997).
Potential advantages associated with vegetative reproduction American beech (Fagus grandifolia Ehrh., hereafter ‘‘beech’’) is a
include increased competitivity and greater survival under adverse very shade-tolerant deciduous tree species found in diverse forest
environmental conditions (Bond and Midgley, 2001). In general, types throughout eastern North America (Tubbs and Houston,
the relative importance of vegetative reproduction is greater near 1990). Beech reproduces both sexually and vegetatively, the latter
the limits of a species’ altitudinal or latitudinal range where primarily by root suckers (Jones and Raynal, 1987). Root suckering
environmental conditions become harsher (Peterson and Jones, in beech has been observed throughout the species’ distributional
1997). Although vegetative reproduction through root suckering range, with great variation among sites in the relative importance
has been extensively studied in many early successional tree of this reproductive mode (e.g., Ward, 1961; Held, 1983; Kitamura
and Kawano, 2001; Morris et al., 2004).
In beech, root suckers arise from adventitious buds formed on
* Corresponding author. callus tissues that develop following root injury (Jones and Raynal,
E-mail address: marilou.beaudet@sympatico.ca (M. Beaudet). 1986, 1988). Potential sources of injury to shallow roots include

0378-1127/$ – see front matter ß 2008 Elsevier B.V. All rights reserved.
doi:10.1016/j.foreco.2008.03.010
3660 M. Beaudet, C. Messier / Forest Ecology and Management 255 (2008) 3659–3666

soil freezing and thawing, abrasion by rocks, animal browsing, 2. Methods

human activity, and logging disturbance (Jones and Raynal, 1986;
Houston, 2001). Apical control of root sucker production is 2.1. Study site
relatively weak in beech; thus, adventitious buds may expand
under a closed canopy (Jones and Raynal, 1988). Moreover, the The study was conducted at the Boisé-des-Muir Ecological
subsequent survival of understory suckers is relatively high (Jones Reserve, which is an 11 ha old-growth forest located in southern
and Raynal, 1987). Beech root suckers can remain connected to the Quebec (Canada), about 70 km southwest of Montréal. The reserve
parent tree for several years (Jones and Raynal, 1986). Evidence of is in the sugar maple–bitternut hickory (Carya cordiformis
physiological connection is limited, but translocation of herbicide [Wangenh.] K. Koch.) bioclimatic domain (Robitaille and Saucier,
from parent stems to adjacent suckers has been reported 1998). This forest has not been subjected to major anthropogenic
(Abrahamson, 1983; Kochenderfer et al., 2004, 2006). Such results disturbance for the last 300 years (Brisson et al., 1992). The forest
suggest that resource translocation might occur from parent trees comprises two areas that differ in their drainage and vegetation
to beech root suckers. If it does, it could provide a critical advantage type. An American elm–black ash (Ulmus americana L.–Fraxinus
for root sucker survival and growth. nigra Marsh.) community occupies the more hydric portion of the
Most of the studies characterising the regeneration response forest (approximately 15% of the forest area), while the more mesic
of beech have not taken into account its two different modes of portion is dominated by a sugar maple–beech community (Brisson
reproduction. For practical reasons (e.g., greenhouse or trans- et al., 1992). This study was performed in the mesic part of the
plant experiments, or because there were too few root suckers at Boisé-des-Muir forest where sugar maple dominates (67% of basal
a site), only beech of seed origin was included in some studies area), followed by beech (12%), basswood (Tilia americana L.) (10%),
(Loach, 1970; Latham, 1992; Reid and Strain, 1994; Kobe et al., and hemlock (Tsuga canadensis [L.] Carr.) (5%) (Brisson et al., 1992).
2002; Caspersen and Saprunoff, 2005). In other studies, the origin The stand basal area and density of trees (DBH > 15 cm) were
of beech individuals was not specified (Amthor et al., 1990; 29 m2/ha and 277 stems/ha, respectively (Brisson et al., 1992). At
Beaudet and Messier, 1998; Gill et al., 1998; Wilder et al., 1999; the time of the study (1995–1996), canopy gaps were generally
Beaudet et al., 2000; Finzi and Canham, 2000; McClure et al., small, resulting from branch- and single tree-falls. The first signs of
2000; Messier and Nikinmaa, 2000; Lin et al., 2001; Ricard et al., beech bark disease were observed in 1990, but the disease had only
2003). Nevertheless, differences between seedlings and root a minor effect on the forest at the time of the study (Brisson et al.,
suckers were reported in a few studies that have addressed the 1996). In the mesic part of the forest, drainage varies from
question of beech origin. For instance, (1) beech seedlings, on moderate to good, with slopes less than 5%. The humus is a Mull,
average, were dispersed slightly further away from parent trees and the soil is a brown stony loam underlain with surface deposits
than beech root suckers (Ribbens et al., 1994); (2) small beech of morainal origin (Beaudet et al., 1999). The region has a humid
seedlings suffered less parasite-induced leaf damage than did continental climate. The mean annual precipitation is 1102 mm,
root suckers (Burt and Bell, 1991); and (3) higher growth rates and the mean monthly temperature ranges from 9.1 8C in January
were reported for beech root suckers compared to seedlings in to 21.3 8C in July (Huntingdon Meteorological Station, Environ-
Ward (1961), Houston (2001), and Beaudet et al. (2007), but no ment Canada, 2004).
difference was found in Canham (1988) and Takahashi and
Lechowicz (2008). 2.2. Sampling of beech seedlings and root suckers
The question thus can be raised regarding a need to distinguish
beech root suckers from seedlings when investigating the Sample plots (radius = 3 m) were established every 10 m along
community dynamics of forests with a beech component. If beech four 100-m-long parallel transects, which in turn were established
originating from seeds and root suckers differs in traits that are 20 m apart in the mesic part of the forest; there were a total of 44
important determinants of the species’ competitive ability, then systematically distributed plots. In early May 1995, we recorded
variation in the relative importance of root suckering among sites the number of beech seedlings and root suckers that were 5–30 cm
and regions might have several community-level implications. For in height (excluding 1995 extension growth) in each of the 44
example, Beaudet et al. (2007) suggested that the effect of canopy plots. This height range was chosen to minimise variation among
gaps might differ among communities depending on the pre- sampled individuals, while allowing us to reach an adequate
valence of root suckering in beech, with beech possibly being sample size (larger individuals were much less abundant). Root
favoured over sugar maple (Acer saccharum Marsh.) where suckers suckers were generally easy to distinguish from seedlings, since
are abundant. Obtaining a better knowledge of the differences that they often originated from exposed roots. Seedlings were
may exist between individuals of seed and root sucker origin is characterised by a tap root. When there was doubt regarding
therefore essential to better understand the community dynamics the origin of an individual, its root system was partially and
of forests with a beech component. carefully excavated. The number of beech of each origin that were
Our study had three objectives. The first was to characterise recorded in the plots were used to calculate the density (stems/ha)
and compare the density, frequency of occurrence and growing and frequency of occurrence (% of plots with 1 individual) of
conditions of beech seedlings and root suckers in the understory seedlings and suckers.
of an old-growth sugar maple–beech forest located in southern Since only 20 beech root suckers 5–30 cm in height were
Quebec, near the northern range limit of the species. The second found in the 44 systematic plots, we decided to increase our
objective was to determine how beech individuals originating sample size for the study of the morphology, growth, survival
from seed versus root suckering differed in terms of leaf and response to defoliation (see below) by establishing three
and plant morphology, growth in height and diameter, additional plots (radius = 3 m) in nearby areas of the stand where
and survival. The third objective was to determine whether root suckers were more abundant, possibly because there had
growth and survival of beech seedlings and suckers would be been some soil disturbance (e.g., old forest trails). One of the
similarly affected by different levels of experimental defoliation. three plots was located less than 20 m away from the end of one
The defoliation was meant to induce stress (e.g., Canham et al., of the transects, while the other two plots were less than 50 m
1999; Myers and Kitajima, 2007), allowing us to further from the other end of our sampling transects. All of the beech
differentiate responses of beech root suckers from those of root suckers (i.e., n = 20) found in the 44 plots were tagged, as
seedlings. were an equal number of seedlings randomly selected among
 M. Beaudet, C. Messier / Forest Ecology and Management 255 (2008) 3659–3666 3661

those found in the 44 plots (i.e., n = 20 out of 129). We 2.6. Measurement of crown morphology, growth, and survival
supplemented this sample by randomly selecting 70 seedlings
and 70 root suckers in the three additional plots described above, In May 1995 (i.e., prior to applying the defoliation treatments),
for a total of 90 seedlings and 90 suckers. we determined the height, age (based on bud scale scars), number
of leaves, and number of lateral branches of all individuals. Stem
2.3. Light measurements diameter was measured at the stem base using calipers. The exact
location where the calipers were positioned was marked on the
Light availability was determined at 30 cm above the ground at stem using a permanent marker. Moreover, calipers were
the center of each of the 44 systematic plots and three additional positioned so that stem diameter would always be measured in
plots by measuring the percent transmission of above-canopy the same direction. At the end of August 1995, stem diameter was
photosynthetic photon flux density (400–700 nm, %PPFD) under re-measured on all individuals at the same height along the stem,
completely overcast sky conditions. Measurements were taken on and in the same direction as for measurements made in May.
June 2, 1995. Two light sensors (LI-190SA point quantum sensor, Diameter increment was calculated as the difference between the
LI-COR, Lincoln, NE, USA) were used simultaneously; one sensor two measurements. Current-year height growth was determined,
was installed in an adjacent open area and linked to a datalogger based on bud scale scars on the shoot reaching the highest point in
(LI-1000, LICOR, Lincoln, NE, USA) that recorded every minute the an individual’s crown. Height growth was also assessed for all
average of readings taken at 5 s intervals. Those readings were previous years. The current-year extension growth of each lateral
used as estimates of above-canopy light (PPFD0). The second light branch was also assessed, based on bud scale scars. The latter
sensor was used to take instantaneous measurements of light at measurements were summed for each individual to yield total
each sampling point (PPFDi), and the light transmission (%PPFD) lateral growth.
was calculated as (PPFDi/PPFD0)  100, where PPFDi and PPFD0 In addition, the following measurements were performed only
were values recorded at the same time (1 min). on control individuals in June 1995. Leaf width was measured on
each leaf in situ and individual leaf area was estimated using
2.4. Defoliation treatments Eq. (1). These values were summed to yield the total leaf area in the
crown of each individual. The crown projection area was
Three defoliation treatments were randomly assigned to measured. The distance between each beech individual and the
seedlings and suckers: complete defoliation, partial defoliation, nearest beech tree with a DBH  10 cm was determined, and the
and control. The defoliation was performed on June 2 and 5, 1995. DBH of that tree was measured. The following ratios were
Complete defoliation was performed on 20 seedlings and 20 calculated: leaf area index (LAI: total leaf area/crown projection
suckers by manually removing all leaves (including petioles) from area, cm2/cm2), the height/stem diameter ratio (based on
each individual. Partial defoliation was performed on 20 seedlings measurements taken in May 1995, cm/mm), and the fraction of
and 20 suckers by removing every second leaf on each individual, the total extension growth that was lateral branch growth (i.e.,
while retaining the leaf located closest to the end of the leader. lateral growth/[lateral growth + height growth], cm/cm).
Finally, 50 seedlings and 50 suckers were kept intact as controls. Since each individual was tagged, we were able to monitor the
Leaves that were removed from completely defoliated individuals survivorship of seedlings and suckers from May to August 1995,
were brought back to the laboratory and kept in the freezer for and again in the following spring (early June 1996). Individuals
subsequent analysis. that could not be located in June 1996 were considered dead
(n = 17). This assumption seemed justified, since our plots were
2.5. Leaf measurements relatively small and live individuals could be easily located.

One-sided leaf surface area, together with the maximum width 2.7. Statistical analysis
and length of the leaf blade, were determined for each of the
harvested leaves using MacFolia software (Régent Instruments The density of beech seedlings and root suckers was compared
Inc., Quebec, Canada). The leaves were then oven-dried (at 65 8C for using a non-parametric Wilcoxon signed-rank test with density
3 days) and weighed. Specific leaf area (SLA) was calculated for values of seedlings and suckers paired per plot. The coefficient of
each individual beech as the ratio of total leaf area/total leaf dry variation (CV) for the density estimates was calculated among
mass. Leaf area, leaf width, and leaf length data were used to plots for beech of each origin. The presence of a significant
develop a predictive equation of leaf area as a function of leaf width correlation between the density of seedlings and suckers was
or leaf length so we could subsequently estimate non-destructively tested using a Pearson product-moment correlation coefficient (r).
the leaf area of beech individuals in the field, based on in situ The frequency of occurrence was compared between seedlings and
measurements of leaf width or length. An equation of the form suckers using a McNemar 2  2 contingency table test using plots
S = b(x) + c(x2) was used, where S was the leaf surface area (cm2), x as the pairing criteria. The distance to, and the DBH of, the nearest
was either leaf width (cm) or length (cm), and b and c were the beech tree was compared between seedlings and suckers using,
estimated parameters. Data from seedlings and suckers were first respectively, a t-test and a non-parametric Mann–Whitney U-test
analysed separately. Since b and c estimates did not differ (t-test assumptions could not be met, despite data transformation).
significantly between the two groups (based on overlap of their We compared light availability among plots with (1) no beech, (2)
95% confidence intervals), we pooled all data and obtained an seedlings only, (3) root suckers only, and (4) both seedlings and
allometric relation that was subsequently used for individuals of suckers present, using one-way ANOVA, after verifying normality
either origin. Leaf width was a slightly better predictor of leaf area and homoscedasticity assumptions.
than leaf length, based the r2 value of its equation (r2 = 0.945 and For each variable describing leaf- and plant-level traits,
0.935 for leaf width and leaf length, respectively). This difference comparisons were made between seedlings and suckers using a
was likely due to the fact that leaf length was more variable due to t-test. Logarithmic (log10[x + 1]) or square root ([x + 0.5]^0.5)
leaf tip damage in some leaves. The relationship between leaf area transformations were used when needed to meet test assump-
(S, cm2) and leaf width (W, cm) was: tions. A non-parametric Mann–Whitney U-test was used when t-
test assumptions could not be met. These comparisons were aimed
S ¼ 0:827W þ 1:028W 2 ðr 2 ¼ 0:945; n ¼ 346 leavesÞ (1) at describing differences between seedlings and suckers not
3662 M. Beaudet, C. Messier / Forest Ecology and Management 255 (2008) 3659–3666

Table 1
Comparison of the density (mean  1S.E., range in parentheses), frequency of occurrence (% of plots including  1 individual), and location with respect to beech trees of beech
seedlings and root suckers in the Boisé-des-Muir old-growth forest

Seedlings Root suckers P

Density (stems/ha) 1037  273 (0–9197) 161  47 (0–1415) <0.001¥

Coefficient of variation in density 1.75 1.92
Frequency of occurrence (%) 54 27 0.005§
Distance (m) to the nearest beech tree with DBH  10 cm 3.30  0.38 (1.00–6.35) 3.27  0.31 (1.38–5.75) 0.946£
DBH (cm) of the nearest beech tree with DBH  10 cm 15.7  1.8 (10.0–44.1) 16.6  1.7 (10.0–42.0) 0.568*

The density and frequency of occurrence were evaluated from a census of 44 systematic plots, while the distance to, and DBH of, the nearest beech tree was evaluated for 20
juveniles of each origin, sampled in the 44 plots. Note: ¥Wilcoxon signed-rank test; §McNemar 2  2 contingency table test; £t-test; *Mann–Whitney U-test.

affected by defoliation, and therefore, were made among all plots, Table 1). Density of beech seedlings and root suckers was
individuals (for those traits that had been measured at the onset of uncorrelated among the plots (r = 0.173, P = 0.261, n = 44). Root
the study and that had not been affected by the defoliation suckers were not located closer to a beech tree than seedlings, with
treatments), among control individuals only, or among completely an average distance for both of about 3.3 m (Table 1). The nearest
defoliated individuals for leaf characteristics measured on beech tree had an average DBH of 16 cm (Table 1).
harvested leaves.
Two-way analyses of variance (ANOVA) were used to test the 3.2. Light conditions
effects of beech origin and defoliation on height growth, total
lateral growth, and stem diameter growth. Height growth and Light availability at 30 cm above ground level was 2.9  0.2%
lateral growth required logarithmic transformation to meet (mean  1S.E.) and ranged from 1.1% to 6.2%PPFD among plots. Among
normality and homoscedasticity assumptions. Tukey multiple the 44 systematic plots, the presence/absence of beech seedlings and
comparison tests were used to identify significantly different suckers was not related to light availability (ANOVA, F3,40 = 1.346,
groups. P = 0.273; data not shown). There was no correlation between root
A linear categorical model (analogous to logistic regression, but sucker density and light availability (r = 0.139, P = 0.668, n = 12) in
the independent factors are categorical variables) was used to test plots where root suckers were present, and no correlation between
for beech origin and defoliation treatments as potential predictors seedling density and light availability in plots where seedlings were
of the status (dead or alive) of beech individuals at the end of the present (r = 0.288, P = 0.173, n = 24; data not shown).
first growing season (in August 1995), or one whole year after the
3.3. Leaf- and plant-level characteristics
defoliation was performed (in June 1996). The complete model was
first tested. When it was significant (P < 0.05), we also report the
Seedlings and root suckers did not differ in terms of SLA and
significance of each factor and their interaction based on
mean surface area of individual leaves (Table 2). However, large
likelihood-ratio x2 statistics. The latter analyses were performed
and significant differences were observed between seedlings and
using JMP statistical software (v. 4.0.2, SAS Inc., Cary, NC, USA),
suckers for many plant-level characteristics. Suckers had more
while other analyses were performed with Systat (v. 10, SSI, San
leaves than seedlings, and thus, a larger total leaf area (Table 2).
Jose, CA, USA).
Suckers also had a larger crown projection area, but a lower LAI
3. Results (Table 2). Root suckers had more branches than seedlings (in fact,
most seedlings were unbranched) (Table 2). The height/stem
3.1. Density and distribution diameter ratio was greater for seedlings than for suckers (Table 2).
Lateral branch growth represented a greater proportion of the total
The mean density of beech seedlings was more than six times extension growth in suckers than in seedlings (Table 2).
greater than that of root suckers, while the density of root suckers
3.4. Age
was slightly more heterogeneous among plots than that of
seedlings (Table 1). Beech regeneration, regardless of origin, was
Beech seedlings and root suckers were about same age, i.e., 2
present in 61% of the plots. The frequency of occurrence of beech
years old at the onset of the study in May 1995 (mean age  1S.E.,
seedlings was greater than that of root suckers (i.e., 54% vs. 27% of

Table 2
Leaf- and plant-level characteristics of beech seedlings and root suckers from the Boisé-des-Muir old-growth forest

Variable Time of measurement n/origin Seedlings Root suckers P

Mean  1S.E. Range Mean  1S.E. Range

Specific leaf area (SLA, cm2/g) Early June 1995 20 426  13 313–544 430  11 333–537 0.844¥
Surface area of individual leaves (cm2) Early June 1995 50 16.0  0.8 5.6–29.0 16.7  1.0 3.6–33.2 0.567¥
Number of leaves Early June 1995 90 4.0  0.2 2–10 12.7  0.8 3–42 <0.001§
Total leaf area in the crown (cm2) Early June 1995 50 63.8  4.0 16.7–145.1 195.8  19.1 7.1–767.0 <0.001£
Crown projection area (cm2) Early June 1995 50 74.5  6.9 19–213 291.8  29.0 10–1040 <0.001£
Number of lateral branches May 1995 90 0.06  0.03 0–1 0.67  0.11 0–5 <0.001*
Leaf area index (LAI, cm2/cm2) Early June 1995 50 0.97  0.03 0.50–1.46 0.77  0.04 0.37–1.40 <0.001¥
Height/stem diameter ratio (cm/mm) May 1995 90 6.9  0.1 3.0–9.9 5.8  0.2 2.4–11.9 <0.001¥
Lateral growth/(total height growth + 1995 season 50 0.17  0.03 0.00–0.68 0.42  0.05 0.00–0.88 <0.001¥
lateral growth) (cm/cm)
The data presented here are for control individuals (i.e., no defoliation). Sample size varies among variables because some measurements were performed on all individuals
prior to defoliation treatments (n = 90 individuals/origin), while other measurements are for control individuals only (n = 50 individuals/origin). For SLA, leaves were
collected from completely defoliated individuals (n = 20 individuals/origin). Note: ¥t-test; §t-test (log-transformed); £t-test (square root-transformed); *Mann–Whitney U-
test.
 M. Beaudet, C. Messier / Forest Ecology and Management 255 (2008) 3659–3666 3663

Fig. 1. Annual height growth recorded on control (i.e., no defoliation treatment)

beech seedlings (A) and root suckers (B). Each line corresponds to an individual.
Note that for some root suckers, the line ends in 1994. This is due to individuals
selected for study in early 1995 for which the shoot that had been identified as the
leader died during the 1995 growing season. Note the different scales for seedlings
and suckers.

1.99  0.01 yrs for seedlings vs. 2.12  0.08 yrs for suckers, Mann– Fig. 2. Effect of defoliation treatment (performed in early June 1995) and beech
Whitney U-test P = 0.247). On average, 1995 was their third growing origin on (A) height growth; (B) total lateral branch growth; and C) stem diameter
growth recorded at the end of August 1995, on beech of seed and root sucker origin,
season. Almost all seedlings originated in 1993, while root suckers
5–30 cm in height at the onset of the study, growing in the shaded understory of the
originated at various times from 1992 to 1994 (Fig. 1). Boisé-des-Muir old-growth forest. Error bars are standard errors. In each panel,
different letters indicate significantly different growth (P < 0.05, Tukey post hoc
3.5. Growth multiple comparisons test). See text and Table 3 for more details on ANOVA results.

Beech seedlings and root suckers had very different patterns of Stem diameter growth in root suckers was twice that of
height growth over time (Fig. 1). Seedlings reached a height of seedlings (0.73  0.08 mm/yr for control root suckers vs.
approximately 10–15 cm during their first growing season, but had 0.29  0.03 mm/yr for control seedlings; Fig. 2C). Diameter growth
a much lower height growth rate in subsequent years (ranging was affected by both beech origin and defoliation, with a significant
from 1 to 3 cm/yr), and this pattern was quite consistent among interaction between the two factors indicating that defoliation did
individuals (Fig. 1A). In contrast, root suckers showed much more not significantly affect the diameter growth of seedlings, but had a
variation in height growth between years and among individuals negative impact on the diameter growth of suckers (Table 3(C);
(Fig. 1B). Fig. 2C). Among root suckers, the diameter growth was significantly
Annual height growth rate of control suckers was more than lower in defoliated individuals compared to controls, but did not
five times greater than for control seedlings (13.9  1.6 cm/yr for differ between the two defoliation levels (Fig. 2C).
suckers vs. 2.6  0.3 cm/yr for seedlings; Fig. 2A). Height growth over
the 1995 growing season was not affected by the defoliation 3.6. Survival
treatments applied in early June 1995, and this was true for both
seedlings and root suckers (Fig. 2A, Table 3(A)). Overall, only beech Survivorship of beech individuals was relatively high over the
origin affected height growth (Fig. 2A; Table 3(A)). first growing season, ranging from 85% to 100%, depending on
The total growth of the lateral branches was much greater in beech origin and defoliation level (Fig. 3). Among control
root suckers than in seedlings (Fig. 2B), partly due to the numerous individuals, 92% of the suckers and 94% of the seedlings survived
branches that the root suckers had compared to seedlings from May to August 1995 (Fig. 3). Survivorship from May to August
(Table 2). Like height growth, lateral growth was not affected by 1995 did not differ as a function of beech origin or defoliation level
defoliation (Fig. 2B, Table 3(B)). (Categorical model, whole model x2 = 4.90, df = 5, P = 0.428). Even
3664 M. Beaudet, C. Messier / Forest Ecology and Management 255 (2008) 3659–3666

Table 3
ANOVA results testing the effect of beech origin (seedlings vs. root suckers) and defoliation treatment (control, partial, and complete; performed in early June 1995) on (A)
height growth (log-transformed), (B) total growth of lateral branches (log-transformed), and (C) stem basal diameter growth (untransformed), as recorded at the end of
August 1995

Variable Beech origin Sample size Source df Mean square F ratio P

(A) Height growth Both 149 Origin 1 12.860 220.219 <0.001

Defoliation treatment 2 0.073 1.254 0.288
Origin  defoliation 2 0.119 2.047 0.133
Error 143 0.058

(B) Total lateral growth Both 166 Origin 1 19.506 111.997 <0.001
Defoliation treatment 2 0.179 1.027 0.361
Origin  defoliation 2 0.023 0.133 0.875
Error 160 0.174

(C) Diameter growth Both 154 Origin 1 1.537 12.687 <0.001

Defoliation treatment 2 1.139 9.404 <0.001
Origin  defoliation 2 0.509 4.208 0.017
Error 148 0.121

Root suckers 80 Defoliation treatment 2 1.713 9.204 <0.001

Error 77 0.186

Seedlings 74 Defoliation treatment 2 0.061 1.211 0.304

Error 71 0.051

Since the interaction between origin and defoliation treatment was significant for diameter growth, separate analyses were performed for beech of each origin. Sample sizes
are lower than the number of individuals selected at the onset of the study (i.e., 180) because some individuals died during the growing season. They differ among variables
because for height growth, the branch identified as the leader died during the growing season on some individuals, while for diameter growth (obtained from re-
measurement of stem diameter) some negative increment values were obtained and excluded from analysis.

complete defoliation did not greatly affect beech survival in the seedlings (Fig. 3). The survivorship of suckers subjected to
year of treatment, since 100% of the root suckers and 90% of the complete defoliation was 55%, while none of the seedlings
seedlings subjected to that treatment in early June survived until subjected to that treatment were still alive after one year
the end of August 1995 (Fig. 3). The survivorship recorded in early (Fig. 3). Effects of partial defoliation were intermediate relative
June 1996 was lower and much more variable (Fig. 3), both as a to the two other treatments, with a 30% and 80% survivorship for
function of beech origin and defoliation levels (Categorical model, seedlings and suckers, respectively (Fig. 3).
whole model x2 = 50.22, df = 5, P < 0.001). However, a significant
interaction between the two factors (origin x defoliation x2 = 9.66, 4. Discussion
df = 2, P = 0.008) indicated that the effect of defoliation varied as a
function of beech origin, having a much greater negative impact on 4.1. Density and distribution
the survivorship of seedlings (Fig. 3). In root suckers, only the
complete defoliation treatment noticeably impacted their survi- Relative densities of beech seedlings and root suckers can vary
vorship, while survivorship of seedlings decreased more or less considerably among sites, from complete absence of suckers to
linearly with increasing defoliation level (Fig. 3). Overall, survivor- almost complete dominance of suckers over seedlings (Ward,
ship of root suckers remained much higher than that of seedlings 1961; Held, 1983; Houston, 2001; Kochenderfer et al., 2004;
(Fig. 3). Seventy-four percent (74%) of control suckers survived Morris et al., 2004). Variation among sites in proportion of beech
from May 1995 to June 1996, compared to 52% of the control root suckers can be influenced by many factors, including genetic
differentiation of beech populations (Kitamura and Kawano, 2001),
harvesting, and beech bark disease (Jones et al., 1989; Houston,
2001; Farrar and Ostrofsky, 2006). At our study site, an old-growth
stand little affected by beech bark disease at the time of sampling
(in 1995), root suckers accounted for 13% of beech regeneration
5–30 cm in height. Beaudet et al. (1999) reported for the same site
somewhat higher proportions of root suckers among individuals
10–30 cm in height. Differences in year of sampling, size of the
plots, and minimum size of the individuals sampled might explain
some of the discrepancy. Beaudet et al. (1999) also observed that
the proportion of root suckers increased with increased beech size.
For instance, they reported that more than 70% of beech individuals
1–4 m in height were of root sucker origin. The latter would be
consistent with a greater survival rate in root suckers compared to
seedlings, as observed in our study. Therefore, we suggest that at
our study site, a relatively undisturbed forest with low-light
conditions, the elevated proportion of suckers reported previously
by Beaudet et al. (1999) for sapling-sized individuals might be
more a result of a lower survival rate among beech seedlings than
of a particularly great production of root suckers.
Fig. 3. Percentage of survival over time for beech seedlings and root suckers
At our study site, beech originating from seed was present in a
submitted to various defoliation treatments (control, partial, and complete) in early
June 1995. The initial cohorts comprised 50 individuals of each origin for the control greater proportion of the plots than root suckers. This is consistent
treatment, and 20 individuals of each origin for each of the two defoliation levels. with trends observed at most of the study sites in Ward (1961) and
 M. Beaudet, C. Messier / Forest Ecology and Management 255 (2008) 3659–3666 3665

Held (1983) for individuals less than one inch in diameter and for 1 m in height. However, Takahashi and Lechowicz (2008), and
‘‘juvenile beech stems’’, respectively. Although the mean distance Canham (1988) found no difference in the growth rate of beech
to beech trees did not differ between seedlings and suckers, the originating from either seeds or root suckers based on samples that
greater frequency of occurrence observed for seedlings might comprised individuals larger than in our study (i.e., with height up
indicate that they are less limited in their dispersal from parent to 1.9 and 4 m, respectively). Although marked differences in
trees than root suckers. The latter is consistent with Ribbens et al. growth exist between small beech seedlings and root suckers, such
(1994), who reported a greater mean dispersal distance for beech results suggest that they might tend to decrease with increasing
seedlings than root suckers. size. Differences in methodology used for growth measurement
Even if there are reasons to think that the low-light conditions might also be involved; both Canham (1988) and Takahashi and
might limit the survival of beech regeneration at our study site Lechowicz (2008) assessed height growth as a vertical increment in
(especially for seedlings - see below), we did not detect any height, whereas in our study and in Beaudet et al. (2007), values of
relationship between the density or frequency of occurrence of extension growth were reported.
beech seedlings and root suckers, and light availability. Light Survivorship over one summer, i.e., from May to the end of
conditions were low and relatively uniform in the understory, with August 1995, was relatively high and did not differ between beech
the highest recorded value equal to 6%PPFD. Possibly, the narrow seedlings and root suckers. Most of the mortality occurred during
range of light conditions was not sufficient to create substantial the subsequent fall/winter period. Survivorship from May 1995 to
variation among plots with respect to survivorship of this highly June 1996 was greater among root suckers than seedlings (74% vs.
shade-tolerant species. 52%). The annual survival rate that we observed for root suckers
agrees with rates reported by Jones and Raynal (1987). Although
4.2. Age distribution root suckers might survive better than seedlings under low light
conditions, that might not be true under higher light levels. For
Most of the beech seedlings found in the study plots in 1995 example, Houston (2001) observed a slightly greater survival rate
originated in 1993, while root suckers originated at various times in beech root suckers than seedlings in his control plots, but the
from 1992 to 1994. The fact that only one 1994 seedling (and no reverse was true in harvested plots. Similarly, Peterson (unpub-
1995 germinants) was found in the plots is in agreement with the lished results cited in Peterson and Jones, 1997, p. 277) found a
notion that seed production, and hence germinant abundance, can slightly greater survivorship among beech root suckers than
vary considerably from year to year. Good seed crops in beech are seedlings in a shady understory, while no difference in survivor-
generally produced every 2–8 years (Tubbs and Houston, 1990). ship was observed between origins in a forest blowdown.
How root sucker production varies from year to year is much less
known, but Jones and Raynal (1987) observed large variations in 4.5. Response to the defoliation treatments
the number of suckers produced per tree between years, and noted
that this variation was not synchronised among trees within a The lack of an effect for the early springtime defoliation
stand. treatment on the current-year height growth and branch growth
was not surprising since in species with determinate growth, such
4.3. Leaf- and plant-level morphological differences as beech (Marks, 1975; Bicknell, 1982), shoot elongation observed
in a given year is mostly a function of the conditions prevailing in
At the leaf-level, we did not observe any differences between the previous year, when buds were formed. The defoliation
beech seedlings and root suckers in terms of SLA and individual leaf treatment, however, had a negative effect on diameter growth
area. At the plant-level, our results clearly showed several (significant only in root suckers). The presence of a negative effect
differences in morphology and growth patterns between beech was consistent with expectation, since stem diameter growth
seedlings and root suckers. Beech seedlings had mostly occurs later in the growing season and is more dependent on the
unbranched stems with a few leaves, whereas root suckers had carbon gained during the current growing season (Kozlowski,
numerous branches and leaves. Growth of lateral branches 1992). What was not expected, however, was that defoliation
accounted for a greater proportion of total extension growth in would have a greater negative impact on the diameter growth of
suckers than in seedlings (42% vs. 17%). Such growth patterns root suckers compared to seedlings.
resulted in suckers having a rather ‘‘lateral-growth’’ type and Defoliation had a more pronounced negative impact on the
seedlings a ‘‘vertical-growth’’ type (sensu Takahashi et al., 2001). A annual survival rate of seedlings than root suckers. This response
crown architecture based on greater lateral crown expansion than was consistent with our expectation that root suckers, because of
height growth is considered more efficient for light interception their possible access to reserves from parental root system, would
under a closed canopy, as long as leaf overlap is minimal (Canham, be less affected by defoliation than seedlings. Nevertheless, it was
1988; Takahashi et al., 2001). This was the case for root suckers. surprising that root suckers had a higher survival rate than
They had a lower LAI than seedlings, and a low LAI is generally seedlings over the winter while showing greater diameter growth
related to a decrease in within-crown self-shading (Henry and reduction after defoliation. Diameter growth is generally con-
Aarssen, 1997). sidered to be related to vigour and carbohydrate reserves, and
therefore, is considered a good predictor of the probability of
4.4. Growth and survivorship of control seedlings and root suckers survival for seedlings and saplings (Kobe et al., 1995). However, for
suckers, which may have access to parental subsidies, diameter
Our results showed that root suckers had greater growth in growth might not be related to the amount of carbohydrate
height and diameter than seedlings (height and diameter growth reserves to which they have access, and thus, that measure might
were, respectively, five and two times greater for suckers than not be a good predictor of their probability of survival (Kobe et al.,
seedlings in individuals less than 30 cm in height). Similarly, Ward 1995).
(1961) reported greater diameter growth for root suckers
compared to seedlings, in individuals less than 20 year old. Our 5. Conclusions
results also agree with those reported by Houston (2001) and
Beaudet et al. (2007), where beech root suckers had higher height This study showed that although small-sized beech seedlings
and diameter growth rates than seedlings, in individuals less than and root suckers were similar in terms of leaf-level characteristics,
3666 M. Beaudet, C. Messier / Forest Ecology and Management 255 (2008) 3659–3666

they differed for several traits that are important determinants of a Gill, D.S., Amthor, J.S., Bormann, F.H., 1998. Leaf phenology, photosynthesis, and the
persistence of saplings and shrubs in a mature northern hardwood forest. Tree
species’ competitive ability. In terms of plant-level morphology, Physiol. 18, 281–289.
root suckers had more traits considered characteristic of a shade- Held, M.E., 1983. Pattern of beech regeneration in the east-central United States.
tolerant species. For instance, they showed more lateral growth Bull. Torrey Bot. Club 110, 55–62.
Henry, H.A.L., Aarssen, L.W., 1997. On the relationship between shade tolerance and
than height growth, and had a lower leaf area index than seedlings. shade avoidance strategies in woodland plants. Oikos 80, 575–582.
In terms of growth and survival, root suckers had both a greater Houston, D.R., 2001. Effect of harvesting regime on beech root sprouts and seedlings
growth in height and diameter, and a higher survivorship than in a North-Central Maine forest long affected by beech bark disease. Research
paper NE-717. USDA Forest Service, Newtown Square, PA, USA.
seedlings. Based on such results, small-sized beech root suckers Jones, R.H., Raynal, D.J., 1986. Spatial distribution and development of root sprouts
appear to be better competitors than beech seedlings; therefore, in Fagus grandifolia (Fagaceae). Am. J. Bot. 73, 1723–1731.
we expect that variation among sites in proportion of root suckers Jones, R.H., Raynal, D.J., 1987. Root sprouting in American beech: production,
survival, and the effect of parent tree vigor. Can. J. For. Res. 17, 539–544.
will likely affect understory tree dynamics.
Jones, R.H., Raynal, D.J., 1988. Root sprouting in American beech (Fagus grand-
ifolia): effects of root injury, root exposure, and season. For. Ecol. Manage. 25,
Acknowledgements 79–90.
Jones, R.H., Nyland, R.D., Raynal, D.J., 1989. Response of American beech regenera-
tion to selection cutting of northern hardwoods in New York. North. J. Appl. For.
We would like to thank Pierre Gagné for assistance in the field, 6, 34–36.
Dominique Gravel, Juan Posada and two anonymous reviewers for Kitamura, K., Kawano, S., 2001. Regional differentiation in genetic components for
the American beech, Fagus grandifolia Ehrh., in relation to geological history and
comments on earlier versions of the manuscript, and William F.J. mode of reproduction. J. Plant Res. 114, 353–368.
Parsons for linguistic revision. Financial support was provided by Kobe, R.K., Pacala, S.W., Silander Jr., J.A., Canham, C.D., 1995. Juvenile tree survivor-
the Ministère de l’Environnement du Québec (Direction du ship as a component of shade tolerance. Ecol. Appl. 5, 517–532.
Kobe, R.K., Likens, G.E., Eagar, C., 2002. Tree seedling growth and mortality
patrimoine écologique), the Fonds québécois de la recherche sur
responses to manipulations of calcium and aluminium in a northern hardwood
la nature et les technologies (FQRNT, Programme des Actions forest. Can. J. For. Res. 32, 954–966.
concertées), and the Natural Sciences and Engineering Research Kochenderfer, J.D., Kochenderfer, J.N., Warner, D.A., Miller, G.W., 2004. Preharvest
Council of Canada (NSERC). manual herbicide treatments for controlling American beech in Central West
Virginia. North. J. Appl. For. 21, 40–49.
Kochenderfer, J.D., Kochenderfer, J.N., Miller, G.W., 2006. Controlling beech root and
stump sprout using the cut-stump treatment. North. J. Appl. For. 23, 155–165.
References Kozlowski, T.T., 1992. Carbohydrate sources and sinks in woody plants. Bot. Rev. 58,
107–221.
Abrahamson, L.P., 1983. Control of beech root and stump sprouts by herbicide Latham, R.E., 1992. Co-occurring tree species change rank in seedling performance
injection of parent trees. For. Res. Note RN-SOF-83-001, SUNY Coll. Environ. Sci. with resources varied experimentally. Ecology 73, 2129–2144.
For., Syracuse, NY. Lin, J., Harcombe, P.A., Fulton, M.R., 2001. Characterizing shade tolerance by the
Amthor, J.S., Gill, D.S., Bormann, F.H., 1990. Autumnal leaf conductance and appar- relationship between mortality and growth in tree saplings in a southeastern
ent photosynthesis by saplings and sprouts in a recently disturbed northern Texas forest. Can. J. For. Res. 31, 345–349.
hardwood forest. Oecologia 84, 93–98. Loach, K., 1970. Shade tolerance in tree seedlings. II. Growth analysis of plants raised
Beaudet, M., Messier, C., 1998. Growth and morphological responses of yellow under artificial shade. New Phytol. 69, 273–286.
birch, sugar maple, and beech seedlings growing under a natural light gradient. Marks, P.L., 1975. On the relation between extension growth and successional
Can. J. For. Res. 28, 1007–1015. status of deciduous trees of the northeastern United States. Bull. Torrey Bot.
Beaudet, M., Messier, C., Paré, D., Brisson, J., Bergeron, Y., 1999. Possible mechanisms Club 102, 172–177.
of sugar maple regeneration failure and replacement by beech in the Boisé-des- McClure, J.W., Lee, T.D., Leak, W.B., 2000. Gap capture in northern hardwoods:
Muir old-growth forest, Quebec. Écoscience 6, 264–271. patterns of establishment and height growth in four species. For. Ecol. Manage.
Beaudet, M., Messier, C., Hilbert, D.W., Lo, E., Wang, Z.M., Lechowicz, M.J., 2000. Leaf- 127, 181–189.
and plant-level carbon gain in yellow birch, sugar maple and beech seedlings Messier, C., Nikinmaa, E., 2000. Effects of light availability and sapling size on the
from contrasting forest light environments. Can. J. For. Res. 30, 390–404. growth, biomass allocation, and crown morphology of understory sugar maple,
Beaudet, M., Brisson, J., Gravel, D., Messier, C., 2007. Effect of a major canopy yellow birch, and beech. Écoscience 7, 345–356.
disturbance on the coexistence of Acer saccharum and Fagus grandifolia in the Morris, A.B., Small, R.L., Cruzan, M.B., 2004. Variation in frequency of clonal
understorey of an old-growth forest. J. Ecol. 95, 458–467. reproduction among populations of Fagus grandifolia Ehrh. in response to
Bicknell, S.H., 1982. Development of canopy stratification during early succession in disturbance. Castanea 69, 38–51.
northern hardwoods. For. Ecol. Manage. 4, 41–51. Myers, J.A., Kitajima, K., 2007. Carbohydrate storage enhances seedling shade and
Bond, W.J., Midgley, J.J., 2001. Ecology of sprouting in woody plants: the persistence stress tolerance in a neotropical forest. J. Ecol. 95, 383–395.
niche. Trends Ecol. Evol. 16, 45–51. Peterson, C.J., Jones, R.H., 1997. Clonality in woody plants: a review and comparison
Brisson, J., Bergeron, Y., Bouchard, A., 1992. The history and tree stratum of an old- with clonal herbs. In: de Kroon, H., van Groenendael, J. (Eds.), The Ecology and
growth forest of Haut-Saint-Laurent Region, Quebec. Nat. Areas J. 12, 3–9. Evolution of Clonal Plants. Backuys Publishers, Leiden, The Netherlands, pp.
Brisson, J., Hauser, S., Beauchamps, É., Bouchard, A., 1996. État de la maladie 263–289.
corticale du hêtre dans la réserve écologique du Boisé-des-Muir. Institut de Reid, C.D., Strain, B.R., 1994. Effects of CO2 enrichment on whole-plant carbon
recherche en biologie végétale (IRBV) Montréal, Québec. budget of seedlings of Fagus grandifolia and Acer saccharum in low irradiance.
Burt, A., Bell, G., 1991. Seed production is associated with a transient escape from Oecologia 98, 31–39.
parasite damage in American beech. Oikos 61, 145–148. Ribbens, E., Silander Jr., J.A., Pacala, S.W., 1994. Seedling recruitment in forests:
Canham, C.D., 1988. Growth and canopy architecture of shade-tolerant trees: calibrating models to predict patterns of tree seedling dispersion. Ecology 75,
response to canopy gaps. Ecology 69, 786–795. 1794–1806.
Canham, C.D., Kobe, R.K., Latty, E.F., Chazdon, R.L., 1999. Interspecific and intras- Ricard, J.-P., Messier, C., Delagrange, S., Beaudet, M., 2003. Do understory saplings
pecific variation in tree seedling survival: effects of allocation to roots versus respond to both light and below-ground competition?: A field experiment in a
carbohydrate reserves. Oecologia 121, 1–11. north-eastern American hardwood forest and a literature review. Ann. Sci. For.
Caspersen, J.P., Saprunoff, M., 2005. Seedling recruitment in a northern temperate 60, 749–756.
forest: the relative importance of supply and establishment limitation. Can. J. Robitaille, A., Saucier, J.-P., 1998. Paysages régionaux du Québec méridional. Les
For. Res. 35, 978–989. Publications du Québec, Sainte-Foy, Québec.
Del Tredici, P., 2001. Sprouting in temperate trees: a morphological and ecological Takahashi, K., Lechowicz, M.J., 2008. Do interspecific differences in sapling growth
review. Bot. Rev. 67, 121–140. traits contribute to the co-dominance of Acer saccharum and Fagus grandifolia?
Environment Canada, 2004. Canadian climate normals 1971–2000. National Cli- Ann. Bot. 101, 103–109.
mate Data and Information Archive, http://www.climate.weatheroffi- Takahashi, K., Seino, T., Kohyama, T., 2001. Responses to canopy openings in
ce.ec.gc.ca/climate_normals/, data last reviewed on 2004/02/25, web site architectural development of saplings in eight deciduous broad-leaved tree
accessed on 2006/05/10. species. Can. J. For. Res. 31, 1336–1347.
Farrar, A., Ostrofsky, W.D., 2006. Dynamics of American beech regeneration 10 years Tubbs, C.H., Houston, D.R., 1990. Fagus grandifolia Ehrh.—American beech. In: Burns,
following harvesting in a beech bark disease-affected stand in Maine. North. J. R.M., Honkala, B.H. (Eds.), Silvics of North America. U. S. Department of
Appl. For. 23, 192–196. Agriculture, Washington, DC, pp. 325–332.
Finzi, A.C., Canham, C.D., 2000. Sapling growth in response to light and nitrogen Ward, R.T., 1961. Some aspects of the regeneration habits of the American beech.
availability in a southern New England forest. For. Ecol. Manage. 131, 153–165. Ecology 42, 828–832.
Frey, B.R., Lieffers, V.J., Landhausser, S.M., Comeau, P.G., Greenway, K.J., 2003. An Wilder, C.M., Holtzclaw, F.W., Clebsch, E.E.C., 1999. Succession, sapling density and
analysis of sucker regeneration of trembling aspen. Can. J. For. Res. 33, 1169– growth in canopy gaps along a topographic gradient in a second growth East
1179. Tennessee forest. Am. Midl. Nat. 142, 201–212.