Research
Bark thickness across the angiosperms: more than just fire
Julieta A. Rosell
Departamento de Ecologıa de la Biodiversidad, Instituto de Ecologıa, Universidad Nacional Autonoma de Mexico, CP 04510, Mexico, DF, Mexico
Summary
Author for correspondence:
Julieta A. Rosell
Tel: +52 55 5623 7718
Email: [email protected]
Received: 21 September 2015
Accepted: 6 January 2016
New Phytologist (2016)
doi: 10.1111/nph.13889
Key words: adaptation, allometry, bark
thickness, fire ecology, inner bark, outer
bark, phloem, water storage.
Global variation in total bark thickness (TBT) is traditionally attributed to fire. However, bark
is multifunctional, as reflected by its inner living and outer dead regions, meaning that, in
addition to fire protection, other factors probably contribute to TBT variation.
To address how fire, climate, and plant size contribute to variation in TBT, inner bark thickness (IBT) and outer bark thickness (OBT), I sampled 640 species spanning all major
angiosperm clades and 18 sites with contrasting precipitation, temperature, and fire regime.
Stem size was by far the main driver of variation in thickness, with environment being less
important. IBT was closely correlated with stem diameter, probably for metabolic reasons,
and, controlling for size, was thicker in drier and hotter environments, even fire-free ones,
probably reflecting its water and photosynthate storage role. OBT was less closely correlated
with size, and was thicker in drier, seasonal sites experiencing frequent fires. IBT and OBT
covaried loosely and both contributed to overall TBT variation. Thickness variation was higher
within than across sites and was evolutionarily labile.
Given high within-site diversity and the multiple selective factors acting on TBT, continued
study of the different drivers of variation in bark thickness is crucial to understand bark ecology.
Introduction
As their outermost covering, bark plays a crucial role in protecting plant stems (Romero & Bolker, 2008; Midgley et al., 2010;
Ferrenberg & Mitton, 2014). The thickness of bark is routinely
cited as the main trait providing this protection, especially from
fire (Hoffmann et al., 2012; Lawes et al., 2013; Pausas, 2015).
However, in addition to protection, variation in thickness probably reflects selection on other bark functions as well. For example,
thickness is crucial in the contribution of bark to stem mechanics
(Niklas, 1999; Rosell & Olson, 2014b), and in allowing stem
photosynthesis (Pfanz et al., 2002; Cernusak & Hutley, 2011;
Rosell et al., 2015). In drier areas, not necessarily prone to fire,
bark water and nutrient storage also affects thickness, an issue
that has received little attention (Rosell & Olson, 2014b). Further complicating understanding of the role of bark thickness is
the fact that bark is composed of two contrasting regions that can
be termed inner and outer bark (Fig. 1). It is unclear how variation in the thicknesses of these two regions produces total bark
thickness (TBT) variation across habitats and species. Moreover,
bark thickness variation across species is overlain on a positive
ontogenetic relationship between stem size and bark thickness
(Paine et al., 2010; Poorter et al., 2014). As a result, ecologists
still lack a global understanding of the patterns and causes of bark
thickness variation across angiosperm lineages, habitats (both
fire-prone and non-fire-prone), and climates. Here, I examined
the contributions of fire and climate, along with plant size, to
variation in TBT, inner bark thickness (IBT), and outer bark
thickness (OBT), using a global data set of 640 species in an
Ó 2016 The Author
New Phytologist Ó 2016 New Phytologist Trust
attempt to understand some of the multiple selective forces acting
on bark.
The different functions of bark are performed by its two major
regions, inner and outer bark (Romero, 2014). Inner bark is
mostly made up of living cells and includes, from the inside out:
the secondary phloem, the tissue specialized in photosynthate
translocation also including fibers and parenchyma (Ryan &
Asao, 2014); the cortex, a parenchymatous tissue of primary
origin; and the phelloderm, a usually thin layer of parenchyma
(Fig. 1). Outer bark includes dead cells of homogeneous structure, as in the case of phellem, or of a more complex structure, as
in the case of the rhytidome (Evert & Eichhorn, 2006; Fig. 1).
Although anatomical differences suggest divergent functions carried out by inner and outer bark, a global understanding of their
relative variation in thickness and the selective factors behind this
variation is still lacking. As a consequence of its abundance of
parenchyma, inner bark probably has a key role in the storage of
water and photosynthates (Srivastava, 1964; Scholz et al., 2007;
Romero, 2014). If selection favors thicker inner bark in areas
where water storage would bring a selective advantage, then inner
bark thickness should correlate negatively with water availability
and positively with temperature (Srivastava, 1964; Rosell &
Olson, 2014b). In turn, it has been suggested that fire protection
is mainly provided by outer bark (Graves et al., 2014), although
other studies suggest that fire protection is mainly the result of
TBT, with bark structure being less important (Vines, 1968;
Brando et al., 2012). If outer bark is mainly involved in fire protection, its thickness should correlate positively and more
strongly with fire regime than the thickness of total or inner bark.
New Phytologist (2016) 1
www.newphytologist.com
New
Phytologist
2 Research
phellogen
vascular
cambium
Dead ‘outer’ bark
successive periderms (phellem
+ phelloderm) = rhytidome
phellem
phelloderm
2˚ cortex w/ sclereids
Living ‘inner’ bark
crushed 2˚ phloem
phloem fibers
phloem ray
active 2˚ phloem
Wood
Fig. 1 Bark structure in cross-section. Bark includes all tissues outside the vascular cambium and is made up of a mostly living inner region (inner bark) and
a mostly dead outer region (outer bark). In turn, inner and outer bark are made up of tissues with different ontogenetic origins and functions. Bark is
produced by two meristems, the vascular cambium, which produces secondary phloem, and the phellogen, which produces phellem to the outside and
phelloderm to the inside. In species with a single phellogen, the phelloderm is underlain by a secondary cortex, cells produced by the apical meristem that
subsequently divide radially as the stem grows. In some species, phellogens form one after another, successively cutting off layer after layer of secondary
phloem and producing thick, fiber-bearing outer bark. This dead outer bark formed by successive phellogens is known as rhytidome.
Here, I carried out the first broad-scale tests of environmental
explanations for variation in total, inner, and outer bark across
the angiosperms.
To examine the effects of climate and fire on TBT, IBT, and
OBT, I analyzed variation in the bark of 1947 samples from 640
species spanning all major clades of angiosperms from a very wide
range of environments (Table 1) and bark morphologies (Fig. 2).
Because the samples were all collected specifically for this study,
this data set had the advantage, as compared with data sets in the
literature, of including samples from wild populations using the
same sampling criteria and methods. The sampling of habitats
emphasized variability in fire regime, with sampled habitats ranging from fire-free areas such as rainforests to frequently burned
savannas, variability in precipitation, with habitats ranging from
very dry areas such as deserts to very wet rainforests, and variability in temperature, with habitats ranging from alpine vegetation
to lowland tropical forests. This wide range of habitats made it
possible to examine the effects of these environmental conditions
on, and their relative importance for, bark thickness traits.
Such cross-species comparisons of bark thickness need to take
into account stem size (Hempson et al., 2014), given that bark
becomes thicker as stems grow wider (Schwilk et al., 2013). This
observation raises the question of how much of the interspecific
variation in thickness can be explained by plant size, and how
much variation is left to be explained by environmental factors.
Moreover, it is unclear whether the strong bark thickness–plant
size association applies equally to both inner and outer bark.
Inner bark is mainly the product of the vascular cambium, the
meristem producing wood to the inside and secondary phloem to
the outside, whereas outer bark production is regulated by the
phellogen (Roth, 1981; Fig. 1). Given that inner bark includes
the photosynthate-translocating secondary phloem, inner bark
would be expected to reflect metabolic needs more strongly and
thus scale more clearly with plant size (Jensen et al., 2012). Associated mainly with protection and produced by a distinct meristem, outer bark could have more flexibility to change
evolutionarily with environmental conditions. This greater lability would support the general view that outer bark is the main
driver of TBT variation (Paine et al., 2010), a hypothesis that I
New Phytologist (2016)
www.newphytologist.com
tested here. Although some bark anatomical traits seem to be
highly conserved (e.g. the family-characteristic configurations of
phloem fibers in Malvaceae or Bignoniaceae; Roth, 1981;
Junikka & Koek-Noorman, 2007), bark functional traits generally seem evolutionarily labile (Romero et al., 2009; Rosell et al.,
2014). Here, I tested this lability in thickness traits using the
largest comparative data set on bark thickness published to date.
Understanding the lability and degree of independence in the
production, as well as in the function, of inner and outer bark is
essential to understand the ecology and evolution of bark as a
multifunctional structure.
Based on a data set of wild samples from extremely diverse
environments, this study represents the first attempt to examine
variation in bark thickness traits in a global context considering
stem size, precipitation, temperature, seasonality, and fire regime.
I aimed to answer the following questions. How are TBT, OBT,
and IBT correlated with plant size, and how much variation is
left over for environmental explanations, including fire, once size
is taken into account? Which environmental factors are most
important in the explanation of thickness variation and how does
this information aid our understanding of inner and outer bark
ecology? Is outer bark the main driver of TBT, and how strongly
correlated is variation in OBT and IBT? How evolutionarily
labile are thickness traits?
Materials and Methods
Sampling and measurements
I collected 1947 samples from 640 species in 153 families of
angiosperms covering most major clades (Supporting Information Fig. S1) and bark morphologies (Fig. 2). Samples included
species ranging from very tall trees to subshrubs, succulents, mangroves, and parasites. Habitats spanned freezing-prone alpine
vegetation, very dry to very wet tropical and temperate forests,
savannas, and deserts (Table 1).
For trees, a wedge of bark was sampled from the base, above
buttresses or roots, whereas whole stem segments were collected
for small plants. Stem diameter (SD) was measured at the
Ó 2016 The Author
New Phytologist Ó 2016 New Phytologist Trust
Ó 2016 The Author
New Phytologist Ó 2016 New Phytologist Trust
1
14.7
16.5
33.8°S, 151.1°E
46.1°N, 12.5°E
22.9°S, 48.5°W
17.7°S, 145.5°E
Pordenone, Italy
Botucatu, Brazil
Atherton
Tablelands, Australia
Mount Field National
Park, Tasmania
Howard Springs
Nature Park, Australia
New South Wales
rainforests, Australia
Daintree National
Park, Australia
Los Tuxtlas
Reserve, Mexico
18.6°N, 95.1°W
16.1°S, 145.45°E
34.1°S, 151.0°E
12.5°S, 131.1°E
42.7°S, 146.6°E
24.0
25.2
15.6
27.3
4.6
18.9
19.0
11.7
15.6
19.31°N, 99.19°W
43.8°N, 10.3°E
26.2
19.5°N, 105.04°W
16.3
3356.5
2081.0
1697.4
1570.0
1512.3
1331
1382.0
1284
1162.2
905
847.0
795.7
792.3
733.0
580
575.0
324.0
213.0
MAP
(mm)
18
> 100
35
> 100
52
18
> 100
> 100
24
21
> 100
2–5
40
11
22
> 100
3–10
> 100
62
5–20
7
57
> 100
20
23
5–20
12
41
> 100
5–20
24
27
> 100
70–100
43
n
> 100
FRI
(yr)1
34.9 (1.2, 130.2)
15.1 (0.7, 72.6)
17.4 (1.3, 63.2)
11.7 (0.3, 35.5)
1.3 (0.4, 33.0)
8.6 (0.7, 39.8)
8.9 (1.9, 22.9)
5.6 (0.4, 16.9)
1.3 (0.2, 119.4)
6.0 (0.2, 11.9)
10.8 (0.9, 32.5)
15.8 (0.1, 59.4)
2.9 (0.3, 46.0)
2.6 (0.4, 68.0)
7.5 (0.8, 27.4)
2.6 (0.4, 15.0)
1.4 (0.4, 14.4)
4.2 (0.3, 24.7)
SD (cm)
15.8 (1.3, 32.5)
10.0 (0.3, 46.8)
10.2 (1.2, 22.0)
6.3 (0.1, 15.6)
0.8 (0.3, 13.7)
3.5 (0.7, 6.2)
5.0 (2.0, 12.0)
6.0 (0.6, 11.2)
1.6 (0.2, 31.3)
2.9 (0.2, 4.8)
4.5 (1.0, 19.0)
8.0 (0.2, 25.0)
2.5 (0.3, 35.0)
2.0 (1.1, 21.0)
5.0 (0.9, 12.6)
2.1 (0.6, 7.1)
1.2 (0.3, 4.9)
2.0 (0.1, 7.7)
H (m)
7.0 (0.9, 21.1)
4.6 (0.6, 26.8)
4.4 (0.4, 13.0)
11.7 (0.3, 26.6)
1.0 (0.4, 10.6)
9.7 (0.4, 53.8)
2.0 (0.7, 9.9)
1.8 (0.3, 11.0)
1.0 (0.2, 35.2)
2.3 (0.2, 6.2)
6.7 (0.5, 22.1)
6.2 (0.3, 48.5)
2.9 (0.4, 52.9)
2.0 (0.4, 22.1)
3.5 (0.3, 19.0)
1.5 (0.3, 4.0)
1.2 (0.5, 7.8)
2.6 (0.2, 31.4)
TBT (mm)
References for FRI can be found in Supporting Information Table S1. Medians and ranges (in parentheses) are shown for SD, H, TBT, IBT and OBT.
Tropical rainforest 3
Tropical rainforest 2
Temperate rainforest
Savanna 2
Cool sclerophyll forest
Coastal Mediterranean
woodland
Temperate
sclerophyll forest
Temperate
broad-leaved
deciduous forest
Savanna 1
Tropical rainforest 1
Seasonally dry
tropical forest 2
Xerophytic shrubland
32.8°S, 150.9°E
7.1
22.3
Yengo National Park,
Australia
Chamela-Cuixmala
Reserve, Mexico
Pedregal de San Angel
Reserve, Mexico
Migliarino San
Rossore Park, Italy
Sydney area, Australia
42.4°S, 147°E
7.5°S, 36.9°W
Bothwell, Tasmania
Cariri, Brazil
14.9
34.1°N, 116.6°W
34.1°N, 118.7°W
16.3
23.03°N, 109.72°W
San Jose del
Cabo, Mexico
Morongo Valley,
California
Stunt Ranch, California
Sarcocaulescent
shrubland
Desert
Mediterranean
shrubland
Seasonally dry
tropical forest 1
Cool temperate
woodland
Temperate woodland
23.8
Lat/Lon
Locality
Vegetation
MAT
(°C)
5.6 (0.7–17.1)
3.5 (0.4–24.3)
3.2 (0.3–8.6)
6.8 (0.5–11.7)
0.7 (0.1–9.8)
4.5 (0.2–17.0)
1.8 (0.3–9.7)
1.4 (0.1–5.2)
0.8 (0.2–17.2)
1.2 (0.1–3.6)
4.4 (0.3–15.5)
4.4 (0.6–27.7)
1.8 (0.1–29.9)
1.3 (0.3–18.7)
2.8 (0.1–14.2)
0.5 (0.1–3.8)
0.9 (0.2–3.2)
2.7 (0.3–30.6)
IBT (mm)
0.7 (0.1–8.6)
0.5 (0.1–3.0)
1.1 (0.06–6.7)
6.2 (0.3–19.8)
0.2 (0.1–1.5)
2.5 (0.04–38.9)
0.2 (0.05–1.7)
0.3 (0.05–5.8)
0.3 (0.05–22.3)
0.5 (0.2–2.6)
1.0 (0.07–9.3)
0.8 (0.07–20.8)
0.3 (0.1–41.3)
0.2 (0.03–4.0)
0.4 (0.02–10.2)
0.4 (0.05–2.1)
0.3 (0.03–5.5)
0.5 (0.04–4.8)
OBT (mm)
Table 1 Vegetation, locality, latitude (Lat), longitude (Lon), mean annual temperature (MAT), mean annual precipitation (MAP), fire return interval (FRI), number of species (n), stem diameter (SD),
height (H), total bark thickness (TBT), inner bark thickness (IBT) and outer bark thickness (OBT) for the 18 sampled sites (sites ordered by precipitation)
New
Phytologist
Research 3
New Phytologist (2016)
www.newphytologist.com
New
Phytologist
4 Research
(a)
(b)
(c)
(d)
(e)
Fig. 2 Bark diversity within a fire-free seasonally dry tropical forest, showing the relative contributions of different bark regions to the often large total bark
thickness. (a) Thick bark of Jacaratia mexicana with very thick inner bark, mostly secondary phloem, and very thin outer bark; (b) thick bark of
Cochlospermum vitifolium with abundant phloem fibers, widely dilated phloem rays, and thin phellem; (c) thick bark of Heliocarpus pallidus with similar
amounts of inner and outer bark; (d) thick bark of Aralia mexicana with very thick phellem and even thicker secondary phloem; (e) thick bark of
Pachycereus pecten-aboriginum with abundant secondary phloem, very abundant cortex, and thin phellem. Gray bars indicate inner bark and black bars
indicate outer bark. Scale bars, 5 mm.
sampling point. I measured TBT as the maximum distance from
the stem surface to the cambium, in fresh samples or in samples
fixed in 70% aqueous ethanol. I measured TBT with digital
calipers and a hand lens, or on thin sections using a light microscope when the bark was very thin (< 4 mm). I measured IBT at
the same point where TBT was measured. I identified inner bark
based on the presence of living tissue, using criteria such as color,
texture, and cell types (Figs 1, 2) with a hand lens or on thin sections when necessary. Data on IBT were available for 93% (592)
of the species. Plant height was measured with a Tru-Pulse 200B
laser rangefinder (Laser Technology Inc., Centennial, CO, USA)
or a tape measure, or extracted from the literature. Two to five
adults of similar size per species were collected, but only one
adult was available for 11% of the species.
I averaged TBT, SD, and height of samples to calculate species
means. For the few species collected from more than one site, I
calculated per site means to reflect potential differences in TBT
between sites. For each sample, I calculated the proportion of
TBT represented by inner bark. This proportion had a lower
within-species coefficient of variation than direct inner bark measurements. Therefore, I used the within-species average of this
proportion and multiplied it by species mean TBT to calculate a
mean IBT per species. OBT was calculated as the difference
between mean TBT and IBT. As defined here, inner bark and
outer bark do not correspond to Roth’s (1981) terminology,
which was based on structural criteria rather than delimiting the
clearly functionally different inner living and outer dead regions.
These definitional differences preclude comparisons with her
work and other work based on her terminology of bark regions.
The data set was uploaded to the TRY Plant Trait Database
(Kattge et al., 2011).
Associations between bark thickness traits and plant size
I examined how closely thickness traits were associated with plant
size through univariate linear regression models. I predicted log10
TBT, IBT or OBT based on log10 SD and examined the goodness of fit and significance of the models and their coefficients.
New Phytologist (2016)
www.newphytologist.com
Associations between bark thickness traits and
environment
I used the residuals of log–log thickness–SD regressions to
examine the relationships between thickness and environment
taking into account stem size. I examined whether thicker or
thinner than expected bark (given plant size) was associated
with precipitation, temperature, and seasonality. To this end,
I used the geographic coordinates of the 624 species in the
data set that were native to sampling sites to extract 19 climate variables from WORLDCLIM v.1.4 (Hijmans et al., 2005)
using ARCGIS 9.2 (ESRI, 2006). Climate variables were
closely correlated with one another, forming groups. Based on
these groups, I used principal component analysis to build
indices reflecting (1) precipitation of the wet season, (2) precipitation of the dry season, (3) mean temperature, and (4)
temperature seasonality. After scaling, I carried out a principal
component analysis for each group of variables. I used the
first principal component of each analysis as an index. I calculated Spearman correlations between residual TBT, OBT
and IBT and the four environmental indices.
I also examined the effect of fire on thickness traits based on a
subset of species from 18 communities differing strongly in fire
regime (Table 1). Collections included seven to 62 species per
community, emphasizing the most common species while ensuring a wide phylogenetic diversity (Fig. S1). Sampling included
communities in which fire is a major selective factor, as in very
frequently burned savannas and eucalypt woodlands (Murphy
et al., 2013). To represent the opposite extreme, I also included
vegetation such as deserts, rainforests, and seasonally dry tropical
forests where fire has not been an important selective force, given
that there is no evidence of natural fires, or fire return intervals
far surpass plant longevity (Janzen, 2002; Rodrıguez Trejo, 2008;
Pennington et al., 2009; Cantarello et al., 2011). Intermediate
communities included Mediterranean vegetations (Keeley et al.,
2012) and temperate forests.
To examine the effect of fire, I calculated Spearman correlations between TBT, IBT and OBT residuals and fire return
Ó 2016 The Author
New Phytologist Ó 2016 New Phytologist Trust
New
Phytologist
interval using the 537 species of these 18 sites. This interval,
expressed in years, was derived from the literature and transformed into a variable taking values from 1 (fire return interval
> 100 yr) to 5 (fire return interval 2–10 yr; Table S1). Although
the effect of fire on vegetation depends also on fire intensity, this
information was not available for the great majority of the sites.
Nevertheless, the span of fire interval represented by my sampling, from sites where fires dependably occur every 2 or 3 yr to
areas in which fire has never been known, is so marked that even
in the absence of intensity data there is every reason to expect that
it should reveal the effects of the vastly different fire selective
pressures experienced by bark thickness in these communities.
I examined the association with fire return interval also taking
into account size-associated fire survival strategies. Small plants
tend to lose their stems to fire and to reseed or resprout from
underground organs after fires, whereas larger plants tend to
resprout from persistent aerial stems, though there is some variation; for example, some large Mediterranean plants are also
reseeders (Keeley et al., 2012; Clarke et al., 2013; Dantas &
Pausas, 2013; Charles-Dominique et al., 2015). If fire does in fact
select for thicker bark, then it should do so most markedly in the
bark of persistent stems of larger plants. For this reason, I recalculated thickness–fire interval correlations based on larger and
smaller plants separately. To separate the two size groups, I used
a cut-off of 2 m plant height, a threshold that has been associated
with flame height dividing low- from high-intensity fires (Hoffmann et al., 2012). Low-intensity fires, with flame lengths ≤ 2 m,
would mainly be expected to affect short plants. Fires with flames
> 2 m would affect larger plants as well.
Roles of plant size, precipitation, temperature, and fire in
explaining bark thickness traits
I used multiple regression to compare the relative importance
of size, climate indices, and fire return interval in explaining
thickness traits. Based on species with information on fire
return interval (Table 1), I fitted models predicting log10 TBT,
IBT, or OBT based on log10 SD, the dry and wet season precipitation indices, the mean temperature and temperature seasonality indices, and the fire return interval. To select the best
models from all possible combinations of predictors, I used
stepwise model selection as well as the dredge function of the
R package MUMIN (Barton, 2015). Model selection procedures
always converged on the same model. In many cases, some
terms in the model were dropped after finding that their associated coefficient was not statistically significant. For the final
model, I examined the significance of two-way interactions,
aiming to examine whether thickness–SD scaling was affected
by environmental variables. Model assumptions were also
checked and the effect of collinearity was ruled out using variance inflation factors. Finally, I calculated squared standardized
coefficients using the R package RELAIMPO (Gr€omping, 2006)
to compare the relative importance of predictors in the model
without interactions. I recalculated this metric using only
species > 2 m, for which fire would be expected to have a
stronger effect than for all species combined.
Ó 2016 The Author
New Phytologist Ó 2016 New Phytologist Trust
Research 5
Variation of thickness traits within and across plant
communities
I compared variation within and across sites following the procedure of Messier et al. (2010). Briefly, I fitted a nested ANOVA
with random effects in which species were nested within sites. I
then calculated the variance components with 95% confidence
intervals (CIs) using a bootstrap procedure run 1000 times
(Manly, 1997). I fitted this model using the R packages APE (Paradis et al., 2004) and NLME (Pinheiro et al., 2015). I then compared TBT, IBT, and OBT across the 18 sites using parametric
or nonparametric ANOVAs followed by post hoc comparisons
using the R package PGIRMESS (Giraudoux, 2015).
Inner and outer bark thickness variation and covariation
I examined the levels of variation in IBT and OBT across species
using standard deviations and coefficients of variation. I also
examined whether a species with thick outer bark tended also to
produce thick inner bark by calculating the correlation between
the two traits.
Evolutionary lability of bark thickness traits
I built a phylogeny using the Angiosperm Phylogeny Group backbone and specialized literature to resolve relationships within
groups (Fig. S1), setting branch lengths to 1. I assessed the evolutionary lability of TBT, IBT and OBT through a randomization
procedure based on phylogenetically independent contrasts and the
K statistic of Blomberg et al. (2003). This procedure was implemented in the R package PHYTOOLS (Revell, 2012). To account for
the uncertainty caused by several polytomies of the phylogenetic
tree, I repeated these calculations for 1000 randomly and fully
resolved trees. All analyses were performed in R v.3.2.1 (R Development Core Team, 2015).
Results
Associations between bark thickness traits and plant size
Species ranged from 0.1 cm to > 2 m in SD, and from 6 cm to
47 m in height. TBT varied accordingly, ranging from 0.2 mm in
the smallest shrubs to over 50 mm in a eucalypt from a fire-prone
woodland and a Symplocos species from a savanna. Most thickness
traits were correlated very closely with stem size (Fig. 3; Table 2).
SD explained 72% of variation in TBT (Fig. 3a), 68% of variation in IBT (Fig. 3b), and 27% of variation in OBT (Fig. 3c). To
take stem size into account in further analyses, I used the residuals of these thickness–SD regressions.
Associations between bark thickness traits and
environment
Reflecting the very wide variety of environments, the mean
annual precipitation varied from 90 to 4312 mm, the mean
annual temperature from 4.5 to 27.3°C, and the annual
New Phytologist (2016)
www.newphytologist.com
New
Phytologist
6 Research
Total bark thickness (mm)
(a) 50.0
20.0
10.0
5.0
2.0
1.0
0.5
0.2
0.1
r 2 = 0.72 ***
Inner bark thickness (mm)
(b) 50.0
20.0
10.0
5.0
2.0
1.0
0.5
0.2
0.1
r 2 = 0.68 ***
Outer bark thickness (mm)
(c) 50.0
20.0
10.0
5.0
r 2 = 0.27 ***
2.0
1.0
0.5
0.2
0.1
0.2 0.5
2.0 5.0 20.0 50.0 200.0
Stem diameter (cm)
Fig. 3 Bark thickness is strongly predicted by stem diameter. Regressions
of (a) total bark thickness, (b) inner bark thickness, and (c) outer bark
thickness against stem diameter are presented, showing that outer bark
thickness is not correlated with stem size as closely as inner bark thickness.
Solid lines, linear fits; dashed lines, 95% confidence intervals. All variables
were log10 transformed. ***, P < 0.001.
temperature range from 12.8 to 35.1°C. Based on the groups
formed by WorldClim variables, I built environmental indices
reflecting precipitation of the wet and dry seasons, mean temperature, and temperature seasonality. These indices summarized
variation very well, explaining from 84% to 93% of variation
within each variable group (Table S2).
In general, thicker bark tended to be associated with drier, hotter, and more seasonal environmental conditions, although associations were modest. Residual TBT and IBT were both associated
negatively with dry season precipitation (Spearman r = 0.30 and
0.29, respectively; P < 0.001; Table S2; Fig. 4a,c) and positively
with mean temperature (Spearman r = 0.15 and 0.23, respectively;
New Phytologist (2016)
www.newphytologist.com
P < 0.001; Fig. 4b,d). In addition, TBT and OBT were also positively associated with temperature seasonality (Spearman r = 0.18
and r = 0.14, respectively; P < 0.001), indicating that total and
outer bark had slight tendencies to be thicker in sites with wider
temperature ranges. OBT tended to be loosely or nonsignificantly
correlated with precipitation variables (Table S2).
In contrast with precipitation and temperature indices, residual OBT was more closely associated with fire return interval
(Spearman r = 0.33; P < 0.001; n = 506; Fig. 4f), but raw OBT
was not (Spearman r = 0.14; P < 0.005; n = 506; Fig. 4e). In turn,
the correlation between fire and TBT was weaker (Spearman
r = 0.23; P < 0.001; n = 537), and that between fire and IBT was
very weak (Spearman r = 0.10; P = 0.03; n = 506). When only
large plants were included (> 2 m), associations with fire were
slightly stronger, increasing to 0.38 (P < 0.001; n = 357) for
OBT, to 0.36 (P < 0.001; n = 364) for TBT, and to 0.19
(P < 0.001; n = 357) for IBT. For small plants (≤ 2 m) these correlations were considerably weaker between fire and OBT
(r = 0.17; P = 0.036; n = 149), TBT (r = 0.09; P > 0.05;
n = 173), and IBT (r = 0.10; P > 0.05; n = 149).
Roles of plant size, precipitation, temperature, and fire in
explaining bark thickness traits
I fitted multiple regression models predicting log10 TBT, IBT or
OBT based on log10 SD, dry season precipitation, wet season
precipitation, mean temperature, temperature seasonality, and
fire return interval. There was convergence in the set of predictors
in all model selection procedures. The model predicting TBT fitted the data very well, explaining 79% of the variation in TBT.
The model included as predictors SD, dry season precipitation,
fire return interval, and temperature seasonality, in addition to a
dry season precipitation 9 temperature seasonality interaction
and an SD 9 fire interaction, suggesting that TBT–SD scaling
was affected by fire regime (Table 3). SD was by far the most
important variable (squared standardized coefficient = 0.810),
followed far behind by fire return interval (0.032), dry season
precipitation (0.018), and temperature seasonality (0.003;
Table 3). The model for IBT had the same predictors with the
exception of temperature seasonality, and explained 72% of variation in IBT. The most important variable was again SD (0.712),
followed by dry season precipitation (0.020). Fire return interval
made only a very small contribution (0.009). Finally, the model
for OBT explained 42% of the variation and included SD, dry
season precipitation, fire return interval, temperature seasonality,
a dry season precipitation 9 fire interaction, and an SD 9 fire
interaction, suggesting again different OBT–SD scaling with fire
regime (Table 3). For explaining OBT, SD was important, but
less so in comparison with the other models (0.372). By contrast,
fire return was more important in explaining OBT (0.110) than
IBT, followed by temperature seasonality (0.013), and dry season
precipitation was the least important predictor (0.009; Table 3).
As expected, fire gained importance in all multiple models for
species > 2 m in height (Table 3). This was particularly noticeable
in the model for OBT, in which SD and fire return interval were
almost equally important (0.202 and 0.163, respectively),
Ó 2016 The Author
New Phytologist Ó 2016 New Phytologist Trust
New
Phytologist
Research 7
Table 2 Regression models predicting total, inner, and outer bark thicknesses (log10 transformed) based on log10 stem diameter
Number of species
r2
Model ANOVA
Intercept
Stem diameter
Total bark thickness
Inner bark thickness
Outer bark thickness
640
0.72
F(1638) = 1633***
0.032 (0.017)ns
0.700 (0.017)***
592
0.68
F(1590) = 1251***
0.201 (0.019)***
0.702 (0.020)***
592
0.27
F(1590) = 223.1***
0.633 (0.036)***
0.545 (0.037)***
***, P < 0.005; ns, P ≥ 0.05. Standard errors are shown in parentheses.
0.5
0.0
−0.5
r = −0.30 ***
−1.0
−2
Residual inner bark thickness (mm)
(c)
0
2
4
6
8
Dry season precipitation index
0.0
−0.5
−1.0
r = −0.29 ***
−1.5
r = 0.15 ***
−1.0
−6
−4
−2
0
Mean temperature index
2
1.0
0.5
0.0
−0.5
−1.0
r = 0.23 ***
−1.5
−6
−4
−2
0
Mean temperature index
2
Residual outer bark thickness (mm)
(f)
40
Outer bark thickness (mm)
−0.5
10
(e)
r = 0.14 ***
30
20
10
0
>100
70–100
20
5–20
Fire return interval (yr)
followed by temperature seasonality (0.011) and dry season precipitation (0.020; Table 3).
Variation in thickness traits within and across plant
communities
Variation within sites far exceeded variation across sites in all
thickness traits. For TBT and OBT, variation within sites was
76% (95% CI 66–80%) and 77% (95% CI 67–81%),
Ó 2016 The Author
New Phytologist Ó 2016 New Phytologist Trust
0.0
(d)
0.5
0
2
4
6
8
Dry season precipitation index
0.5
10
1.0
−2
Fig. 4 Bark thickness is significantly
correlated with climate indices and fire return
interval. (a, b) Correlations between residual
total bark thickness and (a) the dry season
precipitation index and (b) the mean
temperature index. (c, d) Correlations
between residual inner bark thickness and (c)
the dry season precipitation index and (d) the
mean temperature index. (e, f) Correlations
between fire return interval and (e) outer
bark thickness and (f) residual outer bark
thickness. Solid lines, linear fits; dashed lines,
95% confidence intervals. Spearman
correlation coefficients are reported. ***,
P < 0.001.
Residual total bark thickness (mm)
(b)
Residual inner bark thickness (mm)
Residual total bark thickness (mm)
(a)
2–5
1.5
1.0
0.5
0.0
−0.5
−1.0
−1.5
r = 0.33 ***
>100
70–100
20
5–20
Fire return interval (yr)
2–5
respectively, whereas variation across sites was 24% (95% CI 20–
34%) and 23% (95% CI 19–33%), respectively. Within-site
variation for IBT was the highest at 85% (95% CI 73–88%),
leaving 15% for across-site variation (95% CI 12–27%).
There was very strong overlap in thickness traits across the 18
plant communities. In agreement with the traditional view of fire
as the main selective force acting on bark thickness, TBT was
greatest in the species of the frequently burned savannas (Fig. 5a).
However, these sites were followed by rainforests, seasonally dry
New Phytologist (2016)
www.newphytologist.com
New
Phytologist
8 Research
Table 3 Multiple regression models predicting total, inner, and outer bark thicknesses (log10 transformed) based on log10 stem diameter, climate indices,
and fire return interval
Total bark thickness
Model terms and
goodness of fit
Intercept
Stem diameter
Dry season precipitation
Fire return interval
Temperature seasonality
Stem diameter 9
fire return interval
Dry season precipitation 9
temperature seasonality
Dry season precipitation 9
fire return interval
R2adj
Model ANOVA
Species n (all species)
Coefficient (SE)
Inner bark thickness
Squared
standardized
coefficient, all
species
(or species > 2 m)
0.110 (0.032)***
0.609 (0.029)***
0.048 (0.008)***
0.020 (0.011)ns
0.009 (0.009)ns
0.065 (0.011)***
–
0.810 (0.646)
0.018 (0.031)
0.032 (0.092)
0.003 (0.003)
–
0.018 (0.007)*
–
–
–
0.79
F(6530) = 343.5***
537
Coefficient (SE)
0.250 (0.038) ***
0.673 (0.036)***
0.044 (0.008)***
0.009 (0.014)ns
Outer bark thickness
Squared
standardized
coefficient, all
species
(or species > 2 m)
Coefficient (SE)
Squared
standardized
coefficient,
all species
(or species > 2 m)
–
0.712 (0.613)
0.020 (0.026)
0.009 (0.024)
–
–
0.776 (0.067)***
0.405 (0.064)***
0.020 (0.023)ns
0.060 (0.025)*
0.054 (0.018)**
0.097 (0.026)***
–
0.372 (0.202)
0.009 (0.020)
0.110 (0.163)
0.013 (0.011)
–
–
–
–
–
–
–
0.031 (0.012)**
–
–
0.029 (0.014)*
0.72
F(4501) = 331.8***
506
0.42
F(6499) = 62.9***
506
***, P < 0.005; **, P < 0.01; *, P < 0.05; ns, P ≥ 0.05.
forests, and a xerophytic shrubland, all fire-free habitats
(Table 1). Thinner barks were observed in sites subject to fire
such as the Mediterranean shrubland and sclerophyll forests and
woodlands, as well as sites that do not have fire as a selective
agent, such as a desert and a sarcocaulescent shrubland (Fig. 5a).
When size was taken into account, the order of sites changed.
The thickest residual total bark was still observed in the savannas,
but the rainforests became the sites with the thinnest residual
bark (Fig. 5b). Species of sites with long fire return intervals or
no fire present still had thick bark despite their smaller size
(Table 1), such as species of the xerophytic shrubland and the seasonally dry forests. Species of the desert and the sarcocaulescent
shrubland, which had medium to thin raw total bark, appeared
among the sites with the thickest residual bark (Fig. 5b). In general, species of fire-prone areas, as well as those of warm, dry, firefree areas tended to have thicker bark.
Regarding the proportion of TBT represented by inner bark,
sites with very frequent fires, such as the savannas, or with frequent fires, such as sclerophyll systems, tended to have lower
inner bark proportions. By contrast, species in sites with practically no fire such as seasonally dry forests, rainforests, and sarcocaulescent shrublands had very high proportions of inner bark
(Figs 5c, S2).
Inner and outer bark thickness variation and covariation
Although outer bark is usually thought of as the main driver of
TBT, my data showed high variation in the inner living portion
as well. Standard deviations of IBT and OBT were very similar
(4.69 and 4.16, respectively), and the coefficients of variation
were relatively close (115.9 and 216.5, respectively). Species with
thicker inner bark also had thicker outer bark (Spearman
New Phytologist (2016)
www.newphytologist.com
r = 0.51; P < 0.001; inset in Fig. 6), but this association weakened
significantly once stem size was taken into account (Spearman
r = 0.14; P < 0.001). This weak association suggests that,
although broadly associated with plant size, IBT and TBT can
vary widely across species. This variation can be seen in Fig. 6, in
which data for bark thicker than 10 mm are plotted, showing the
widely varying proportions of inner and outer bark.
Evolutionary lability of bark thickness traits
Thickness traits were found to be evolutionarily labile. For raw
and residual TBT, the phylogenetic signal was low (K < 0.17)
and nonsignificant (P > 0.09) in all the 1000 calculations based
on fully resolved trees. The same applied to raw and residual IBT
(K < 0.19; P > 0.06) and OBT (K < 0.10; P > 0.35).
Discussion
Total bark thickness was mainly driven by plant size
My global sampling showed that the very wide variation in TBT,
from < 1 mm to several centimeters (Paine et al., 2010; Rosell &
Olson, 2014b; Pausas, 2015), was associated first and foremost
with SD. Although the TBT–SD association has often been documented (Adams & Jackson, 1995; Pinard & Huffman, 1997;
Lawes et al., 2013; Poorter et al., 2014), it remained unclear how
much variation in TBT is available for explanation by other factors once SD is taken into account. In my data set, SD explained
a very substantial 72% of TBT variation, and in the multiple
regression model, SD was > 25 times more important than fire
return interval and dry season precipitation, the second and third
most important predictors in the model, respectively (Table 3).
Ó 2016 The Author
New Phytologist Ó 2016 New Phytologist Trust
New
Phytologist
Research 9
(a)
(b)
Total bark thickness (mm)
0.2 0.5
2.0 5.0 10.0
Residual total bark thickness (mm)
–1.0
50.0
Temperate
sclerophyll forest
Cool
sclerophyll forest
Desert
–0.5
0.0
0.5
1.0
Temperate
rainforest
Tropical
rainforest 3
Tropical
rainforest 1
Temp. broad-leaved
deciduous forest
Mediterranean
shrubland
Tropical
rainforest 2
Temperate
sclerophyll forest
Cool
sclerophyll forest
Coastal Mediter.
woodland
Seasonally dry
forest 1
Cool
temp. woodland
Seasonally dry
forest 2
Desert
Mediterranean
shrubland
Temp. broad-leaved
deciduous forest
Tropical
rainforest 1
Cool
temp. woodland
Coastal Mediter.
woodland
Sarcocaulescent
shrubland
Temperate
woodland
Seasonally dry
forest 1
Temperate
rainforest
Tropical
rainforest 2
Seasonally dry
forest 2
Xerophytic
shrubland
Tropical
rainforest 3
Savanna 1
Sarcocaulescent
shrubland
Xerophytic
shrubland
Temperate
woodland
Savanna 2
Savanna 1
Savanna 2
(c)
Inner bark proportion
0.2
0.4
0.6
0.8
1.0
Savanna 1
Savanna 2
Coastal Mediter.
woodland
Desert
Temperate
sclerophyll forest
Mediterranean
shrubland
Xerophytic
shrubland
Temperate
rainforest
Cool
sclerophyll forest
Temperate
woodland
Temp. broad-leaved
deciduous forest
Cool
temp. woodland
Seasonally dry
forest 2
Sarcocaulescent
shrubland
Tropical
rainforest 2
Seasonally dry
forest 1
Tropical
rainforest 3
Tropical
rainforest 1
Fig. 5 Variation in total bark thickness and inner bark proportion across 18 plant communities. (a) Total bark thickness, (b) residual total bark thickness, and
(c) proportion of total bark thickness represented by inner bark. Homogeneous groups are in gray to the right of boxplots (P < 0.05). The significance of
differences across sites was determined using Tukey post hoc tests for total bark thickness and nonparametric tests for residual bark thickness and inner
bark proportion. Boxplots are centered around the median.
Ó 2016 The Author
New Phytologist Ó 2016 New Phytologist Trust
New Phytologist (2016)
www.newphytologist.com
New
Phytologist
10 Research
40
30
50.0
Outer bark thickness (mm)
Total bark thickness (mm)
50
10.0
r = 0.51 ***
2.0
Allocasuarina Eucalyptus
torulosa
punctata
Total bark thickness: 29.9 mm
30.2 mm
Outer bark thickness: 26.4 mm
0.3 mm
Inner bark thickness: 3.5 mm
29.9 mm
0.5
0.1
0.05 0.20 1.00 5.00 20.00
Inner bark thickness (mm)
20
Inner bark
Outer bark
10
0
Species with total bark thickness > 10 mm
Evolutionarily, the very strong TBT–SD association implies that
selective pressures acting on plant size (Moles et al., 2009) will
lead to changes in bark thickness, and vice versa. Methodologically, this association means that TBT across-species comparisons
must take plant size into account (Hempson et al., 2014; cf.
Rosell & Olson, 2014a). This strong association with plant size
leaves little room for explanation of TBT by environmental conditions, although environmental conditions and their associations
with thickness were informative when separating inner from
outer bark.
Inner bark thickness reflected photosynthate translocation
and storage needs
The evolution of IBT might have been significantly influenced
by one of its main functions, photosynthate translocation (Ryan
& Asao, 2014). Inner bark includes the secondary phloem, the
tissue translocating photosynthates from the leaves to the rest of
the plant. Although only a small fraction of secondary phloem
near the cambium is active in translocation (Fig. 1), the thick
layer of nonconductive phloem includes a large fraction of living
cells (Esau et al., 1957). Given that it is largely living, the amount
of conductive and nonconductive phloem would be expected to
scale with plant size. Such scaling between supply and sink areas
is pervasive in plants (Olson et al., 2009; Sperry et al., 2012), and
has been well documented between the supply area of xylem and
plant size (Mencuccini et al., 2011; Olson et al., 2014). In a similar way, the cross-sectional area of phloem in the stem should
scale with photosynthetic production and volume of sink tissue.
Thicker secondary phloem, and thus thicker inner bark, should
thus reflect the metabolic current and past needs of an increasingly larger plant in terms of photosynthate translocation and
also storage.
Given its very large fraction of living parenchyma, inner bark
is thought to be key in the storage of water, sugars, starch, and
other compounds (Srivastava, 1964). The associations observed
here between IBT and environmental variables support this
important role. Although not very strong after taking SD into
account, correlations were significant and in the expected direction. Inner bark tended to be thicker in areas with less rainfall,
with dry season rainfall being more strongly associated with IBT
New Phytologist (2016)
www.newphytologist.com
Fig. 6 Wide variation in inner and outer bark
thicknesses across species with bark
> 10 mm, and covariation between inner and
outer bark thicknesses of all species (inset),
showing that thick total bark can be achieved
by different combinations of inner and outer
bark. ***, P < 0.001.
than wet season rainfall (Table S2; Fig. 4c). Also, inner bark was
thicker in sites with higher mean temperature (Table S2; Fig. 4d),
which would also suggest a storage role in hotter environments
with higher evapotranspiration. Water in inner bark could reach
the xylem via phloem rays (Pfautsch et al., 2015) to restore the
transpiration stream interrupted by daily and seasonal variations
in water availability (Chapotin et al., 2006b; Nardini et al.,
2011). Such interruptions are more frequent in hotter, drier, and
more seasonal environments. Water in inner bark could also contribute to support leaf and flower flushes, which have usually
been regarded as fueled by water in wood (Borchert, 1994;
Chapotin et al., 2006a). Thicker inner bark would not translate
into higher water storage if the water-storing capacity of bark varied widely. However, it has been shown that bark water storage is
driven mainly by bark quantity (thickness) and not quality
(Rosell et al., 2014). In contrast with inner bark, association patterns with size and environment suggest a different functional
profile for outer bark.
Outer bark thickness reflected protection
The patterns of association recovered here support a protection
role for outer bark (Romero & Bolker, 2008). OBT did not correlate with SD as closely as IBT did. It could be argued that this
loose association could result from outer bark shedding or abrasion during the life of a plant. However, the markedly different
proportions of outer bark in plants of similar size strongly suggest
that OBT does not scale with plant size as IBT does. For example, for trees of SD c. 100 cm, OBT ranged from < 0.1 mm to
almost 50 mm (see vertical dispersion on right side of Fig. 3c),
and while abrasion or other damage might account for a small
amount of this variation, species-specific differences are marked
and biologically real. Given this lack of strong proportionality
with the rest of the stem, rather than metabolic reasons, other
selective factors probably underlie OBT variation. One of these
factors could be thermal insulation (Pasztory & Ronyecz, 2013),
given that residual outer bark had a slight tendency to be thicker
in more seasonal environments (Table S2). However, fire return
interval was a much more important factor explaining OBT variation (Table 3). Congruently, investment in OBT in comparison
with IBT was highest in the frequently burned savannas, and in
Ó 2016 The Author
New Phytologist Ó 2016 New Phytologist Trust
New
Phytologist
the fire-prone coastal Mediterranean shrubland (Fig. 5c). These
observations support previous claims that it is the thickness of its
outer region and not TBT that is the main trait responding evolutionarily to fire (Graves et al., 2014). This observation seems to
be particularly true in large species, for which thick bark has been
regarded as an adaptation permitting stem persistence through
fire (Keeley et al., 2011; Charles-Dominique et al., 2015).
Evolutionary responses to fire are strongly linked with plant
size (Clarke et al., 2013; Dantas & Pausas, 2013). Small plants
tend to lose their stems and to reseed or resprout from underground organs after fires. By contrast, larger plants tend to have
fire-resistant stems (Clarke et al., 2013), with some exceptions;
for example, obligate reseeders > 2 m in height can be found in
some Mediterranean systems (Keeley et al., 2012). Congruent
with expectations, the effect of fire on bark thickness was stronger
in species > 2 m. The correlation of OBT with fire regime was
0.38 for large species, whereas for smaller plants it was 0.17.
Likewise, for larger plants fire was as important as SD for predicting OBT (Table 3). Most studies examining the role of thickness
in fire protection have focused on TBT (but see Graves et al.,
2014), a trait for which thresholds have been identified as indicators of stem survival after fire (Lawes et al., 2013). If OBT is the
trait more closely involved in fire protection, it will be crucial to
examine the role that outer bark plays in these thresholds. In any
case, small-statured species do not usually meet these critical firesurviving thresholds (Hoffmann & Solbrig, 2003). As a result,
selection seems to favor the loss of small stems to fire, so bark in
these plants cannot be regarded as reflecting selection favoring
fire resistance. TBT in small species should reflect the effects of
selection on all functions other than protection from fire and
could thus be a very useful system for understanding thickness in
the context of functions other than fire resistance.
Inner and outer bark drove variation in total thickness
Congruent with the notion that OBT drives variation in TBT
(Paine et al., 2010), the coefficient of variation of OBT was
larger than that of IBT. However, IBT also varied markedly
and had a similar standard deviation to OBT, suggesting that
both regions are important drivers of TBT variation. Interestingly, IBT and OBT covaried only to a limited degree with
one another (inset in Fig. 6), especially considering residual
thicknesses. This relatively modest covariation could be
explained by the very different functional roles highlighted
here for inner and outer bark, and also by developmental factors. Inner and outer bark production are regulated by different meristems, the vascular cambium and the phellogen,
respectively (Roth, 1981; Fig. 1), so inner and outer bark can
potentially be produced at different rates. Differences in function and origin would predict that contrasting combinations
of inner and outer bark amounts would be observable across
species. This was exactly the case in my data set and is illustrated in the different amounts of inner (gray) and outer
(black) bark in species of similar TBT in Fig. 6. For example,
a thick bark of 30 mm could be made up of mostly inner living tissue, as in Eucalyptus punctata, or of mostly outer dead
Ó 2016 The Author
New Phytologist Ó 2016 New Phytologist Trust
Research 11
cells, as in Allocasuarina torulosa (arrows in Fig. 6). These
species with extreme bark proportions were growing side by
side in the same temperate sclerophyll woodland, highlighting
that not only total thickness but also bark construction varies
markedly within sites.
Bark thickness varied markedly within plant communities
and was evolutionarily labile
If TBT is correlated with SD, high within-site variation in raw
thickness would be expected, given that plant size varies widely
within communities (Westoby et al., 2002; Falster & Westoby,
2003). However, within-site variation was also marked in residual TBT (Fig. 5b) and was mirrored by high variation across even
closely related species (low and nonsignificant phylogenetic signal). Marked within-site variation has been documented for other
plant traits (Gleason et al., 2012; Olson & Rosell, 2013), and
seems to reflect the coexistance of divergent ecological strategies
(Marks & Lechowicz, 2006; Reich, 2014).
Conclusions
TBT seems affected evolutionarily by plant size and the different functions of bark, such as photosynthate translocation and
storage for inner bark, and fire protection for outer bark. Other
functions not examined here, such as photosynthesis and protection from herbivory, probably also play a role in generating differences in IBT and OBT (Ferrenberg & Mitton, 2014; Rosell
et al., 2015). Given that TBT variation is associated with different functions carried out by inner and outer bark, it is hard to
justify invoking a single function such as fire resistance to
explain TBT variation at a global scale. Moreover, a single environmental factor as a global cause is impossible to entertain
given the marked variation observed within any given community. Further examination of the patterns and causes of bark
diversity, especially in fire-free systems, will be crucial to understand the range of bark ecological strategies. To this end, multifactorial approaches seem to be the most profitable way of
understanding the ecological and evolutionary significance of
bark, which represents an evolutionary response to pressures far
beyond fire.
Acknowledgements
This project was supported by CONACYT (nos. 237061 &
132404), UNAM-DGAPA-PAPIIT (no. IA201415), a Young
Scientist Award from the MAB program (UNESCO), and the
Daintree Rainforest Observatory (James Cook University). I
thank M. Olson, C. Marcati, M. Westoby, A. Crivellaro, R.
Lima, R. Mendez, N. Martınez, C. Sorce, S. Carlquist, L.
Alvarado, P. Byrnes, M. Castorena, Y. Chang, S. Gleason, C.
Blackman, A. Cook, J. Cooke, A. Ford, M. Garcıa, L. Hutley, S.
Isnard, R. Kooyman, C. Laws, C. Leon, I. Letocart, D. Letocart,
J. Olson, E. Ramırez, M. Scalon, S. Stuart, A. Thompson, W.
Tozer, S. Trueba, J. Vega and K. Zieminska for their kind help
with field work and helpful discussions.
New Phytologist (2016)
www.newphytologist.com
12 Research
Author contributions
J.A.R. designed the study, carried out fieldwork, designed the
analyses, and wrote the manuscript.
References
Adams DC, Jackson JF. 1995. Estimating the allometry of tree bark. American
Midland Naturalist 134: 99–106.
Barton K. 2015. MuMIn: Multi-model inference: R package v.1.15.1. [WWW
document] URL http://CRAN.R-project.org/package=MuMIn [accessed 3
July 2015].
Blomberg SP, Garland T, Ives AR. 2003. Testing for phylogenetic signal in
comparative data: behavioral traits are more labile. Evolution 57: 717–745.
Borchert R. 1994. Soil and stem water storage determine phenology and
distribution of tropical dry forest trees. Ecology 75: 1437–1449.
Brando PM, Nepstad DC, Balch JK, Bolker B, Christman MC, Coe M, Putz
FE. 2012. Fire-induced tree mortality in a neotropical forest: the roles of
bark traits, tree size, wood density and fire behavior. Global Change Biology 18:
630–641.
Cantarello E, Newton AC, Hill RA, Tejedor-Garavito N, Williams-Linerab G,
Lopez-Barrera F, Manon RH, Golicher DJ. 2011. Simulating the potential for
ecological restoration of dryland forests in Mexico under different disturbance
regimes. Ecological Modelling 222: 1112–1128.
Cernusak LA, Hutley LB. 2011. Stable isotopes reveal the contribution of
corticular photosynthesis to growth in branches of Eucalyptus miniata. Plant
Physiology 155: 515–523.
Chapotin SM, Razanameharizaka JH, Holbrook NM. 2006a. Baobab trees
(Adansonia) in Madagascar use stored water to flush new leaves but not to
support stomatal opening before the rainy season. New Phytologist 169: 549–559.
Chapotin SM, Razanameharizaka JH, Holbrook NM. 2006b. Water relations of
baobab trees (Adansonia spp. L.) during the rainy season: does stem water
buffer daily water deficits? Plant, Cell & Environment 29: 1021–1032.
Charles-Dominique T, Beckett H, Midgley GF, Bond WJ. 2015. Bud
protection: a key trait for species sorting in a forest–savanna mosaic.
New Phytologist 207: 1052–1060.
Clarke PJ, Lawes MJ, Midgley JJ, Lamont BB, Ojeda F, Burrows GE, Enright
NJ, Knox KJ. 2013. Resprouting as a key functional trait: how buds,
protection and resources drive persistence after fire. New Phytologist 197:
19–35.
Dantas VdL, Pausas JG. 2013. The lanky and the corky: fire-escape strategies in
savanna woody species. Journal of Ecology 101: 1265–1272.
Esau K, Currier HB, Cheadle VI. 1957. Physiology of phloem. Annual Review of
Plant Physiology 8: 349–374.
ESRI. 2006. ArcGIS v. 9.2. Redlands, CA, USA: Environmental Systems
Research Institute.
Evert RF, Eichhorn SE. 2006. Esau’s plant anatomy: meristems, cells, and tissues of
the plant body: their structure, function, and development. Hoboken, NJ, USA:
John Wiley & Sons.
Falster DS, Westoby M. 2003. Plant height and evolutionary games. Trends in
Ecology & Evolution 18: 337–343.
Ferrenberg S, Mitton JB. 2014. Smooth bark surfaces can defend trees against
insect attack: resurrecting a ‘slippery’ hypothesis. Functional Ecology 28:
837–845.
Giraudoux P. 2015. pgirmess: Data analysis in ecology: R package v.1.6.2. [WWW
document] URL http://CRAN.R-project.org/package=pgirmess [accessed 7
June 2015].
Gleason SM, Butler DW, Zieminska K, Waryszak P, Westoby M. 2012. Stem
xylem conductivity is key to plant water balance across Australian angiosperm
species. Functional Ecology 26: 343–352.
Graves SJ, Rifai SW, Putz FE. 2014. Outer bark thickness decreases more with
height on stems of fire-resistant than fire-sensitive Floridian oaks (Quercus spp.;
Fagaceae). American Journal of Botany 101: 2183–2188.
Gr€
omping U. 2006. Relative importance for linear regression in R: the package
relaimpo. Journal of Statistical Software 17: 1–27.
New Phytologist (2016)
www.newphytologist.com
New
Phytologist
Hempson GP, Midgley JJ, Lawes MJ, Vickers KJ, Kruger LM. 2014.
Comparing bark thickness: testing methods with bark–stem data from two
South African fire-prone biomes. Journal of Vegetation Science 25: 1247–1256.
Hijmans RJ, Cameron SE, Parra JL, Jones PG, Jarvis A. 2005. Very high
resolution interpolated climate surfaces for global land areas. International
Journal of Climatology 25: 1965–1978.
Hoffmann WA, Geiger EL, Gotsch SG, Rossatto DR, Silva LCR, Lau OL,
Haridasan M, Franco AC. 2012. Ecological thresholds at the savanna-forest
boundary: how plant traits, resources and fire govern the distribution of
tropical biomes. Ecology Letters 15: 759–768.
Hoffmann WA, Solbrig OT. 2003. The role of topkill in the differential
response of savanna woody species to fire. Forest Ecology and Management
180: 273–286.
Janzen DH. 2002. Tropical dry forest: area de Conservacion Guanacaste,
northwestern Costa Rica. In: Perrow MR, Davy AJ, eds. Handbook of ecological
restoration, volume 2, restoration in practice. Cambridge, UK: Cambridge
University Press, 559–583.
Jensen KH, Liesche J, Bohr T, Schulz A. 2012. Universality of phloem transport
in seed plants. Plant, Cell & Environment 35: 1065–1076.
Junikka L, Koek-Noorman J. 2007. Anatomical structure of barks in neotropical
genera of Annonaceae. Annales Botanici Fennici 44: 79–132.
Kattge J, Diaz S, Lavorel S, Prentice C, Leadley P, Bonisch G, Garnier E,
Westoby M, Reich PB, Wright IJ et al. 2011. TRY – a global database of plant
traits. Global Change Biology 17: 2905–2935.
Keeley JE, Bond WJ, Bradstock RA, Pausas JG, Rundel PW. 2012. Fire in
Mediterranean ecosystems. New York, NY, USA: Cambridge University Press.
Keeley JE, Pausas JG, Rundel PW, Bond WJ, Bradstock RA. 2011. Fire as
an evolutionary pressure shaping plant traits. Trends in Plant Science 16:
406–411.
Lawes MJ, Midgley JJ, Clarke PJ. 2013. Costs and benefits of relative bark
thickness in relation to fire damage: a savanna/forest contrast. Journal of Ecology
101: 517–524.
Manly BFJ. 1997. Randomization, bootstrap and Monte Carlo methods in biology,
2nd edn. London, UK: Chapman and Hall.
Marks CO, Lechowicz MJ. 2006. Alternative designs and the evolution of
functional diversity. The American Naturalist 167: 55–66.
Mencuccini M, H€olltt€a T, Martınez-Vilalta J. 2011. Comparative criteria for
models of the vascular transport systems of tall trees. In: Meinzer FC,
Lachenbruch B, Dawson TE, eds. Size- and age-related changes in tree structure
and function. Dordrecht, the Netherlands: Springer, 309–339.
Messier J, McGill BJ, Lechowicz MJ. 2010. How do traits vary across ecological
scales? A case for trait-based ecology. Ecology Letters 13: 838–848.
Midgley JJ, Lawes MJ, Chamaille-Jammes S. 2010. Savanna woody plant
dynamics: the role of fire and herbivory, separately and synergistically.
Australian Journal of Botany 58: 1–11.
Moles AT, Warton DI, Warman L, Swenson NG, Laffan SW, Zanne AE,
Pitman A, Hemmings FA, Leishman MR. 2009. Global patterns in plant
height. Journal of Ecology 97: 923–932.
Murphy BP, Bradstock RA, Boer MM, Carter J, Cary GJ, Cochrane MA,
Fensham RJ, Russell-Smith J, Williamson GJ, Bowman DMJS. 2013. Fire
regimes of Australia: a pyrogeographic model system. Journal of Biogeography
40: 1048–1058.
Nardini A, Lo Gullo MA, Salleo S. 2011. Refilling embolized xylem conduits: is
it a matter of phloem unloading? Plant Science 180: 604–611.
Niklas KJ. 1999. The mechanical role of bark. American Journal of Botany 86:
465–469.
Olson ME, Aguirre-Herna ndez R, Rosell JA. 2009. Universal foliage-stem
scaling across environments and species in dicot trees: plasticity, biomechanics
and Corner’s Rules. Ecology Letters 12: 210–219.
Olson ME, Anfodillo T, Rosell JA, Petit G, Crivellaro A, Isnard S, Leon-Gomez
C, Alvarado-Ca rdenas LO, Castorena M. 2014. Universal hydraulics of the
flowering plants: vessel diameter scales with stem length across angiosperm
lineages, habits and climates. Ecology Letters 17: 988–997.
Olson ME, Rosell JA. 2013. Vessel diameter–stem diameter scaling across woody
angiosperms and the ecological causes of xylem vessel diameter variation.
New Phytologist 197: 1204–1213.
Ó 2016 The Author
New Phytologist Ó 2016 New Phytologist Trust
New
Phytologist
Paine CET, Stahl C, Courtois EA, Pati~
no S, Sarmiento C, Baraloto C. 2010.
Functional explanations for variation in bark thickness in tropical rain forest
trees. Functional Ecology 24: 1202–1210.
Paradis E, Claude J, Strimmer K. 2004. APE: analyses of phylogenetics and
evolution in R language. Bioinformatics 20: 289–290.
Pa sztory Z, Ronyecz I. 2013. The thermal insulation capacity of tree bark. Acta
Silvatica et Lignaria Hungarica 9: 111–117.
Pausas JG. 2015. Bark thickness and fire regime. Functional Ecology 29: 315–327.
Pennington RT, Lavin M, Oliveira-Filho A. 2009. Woody plant diversity,
evolution, and ecology in the tropics: perspectives from seasonally dry tropical
forests. Annual Review of Ecology Evolution and Systematics 40: 437–457.
Pfanz H, Aschan G, Langenfeld-Heyser R, Wittmann C, Loose M. 2002.
Ecology and ecophysiology of tree stems: corticular and wood photosynthesis.
Naturwissenschaften 89: 147–162.
Pfautsch S, Renard J, Tjoelker MG, Salih A. 2015. Phloem as capacitor: radial
transfer of water into xylem of tree stems occurs via symplastic transport in ray
parenchyma. Plant Physiology 167: 963–971.
Pinard MA, Huffman J. 1997. Fire resistance and bark properties of trees in a
seasonally dry forest in eastern Bolivia. Journal of Tropical Ecology 13: 727–740.
Pinheiro J, Bates D, DebRoy S, Sarkar D, R Development Core Team. 2015.
nlme: Linear and nonlinear mixed effects models. R package v.3.1-120. [WWW
document] URL http://CRAN.R-project.org/package=nlme.
Poorter L, McNeil A, Hurtado VH, Prins H, Putz J. 2014. Bark traits and life
history strategies of tropical dry- and moist forest trees. Functional Ecology 28:
232–242.
R Development Core Team. 2015. R: a language and environment for statistical
computing. Vienna, Austria: R Foundation for Statistical Computing. v.3.2.1.
[WWW document] URL http://www.rproject.org.
Reich PB. 2014. The world-wide ‘fast–slow’ plant economics spectrum: a traits
manifesto. Journal of Ecology 102: 275–301.
Revell LJ. 2012. phytools: an R package for phylogenetic comparative biology
(and other things). Methods in Ecology and Evolution 3: 217–223.
Rodrıguez Trejo DA. 2008. Fire regimes, fire ecology, and fire management in
Mexico. Ambio 37: 548–556.
Romero C. 2014. Bark structure and functional ecology. In: Cunningham AB,
Campbell BM, Luckert MK, eds. Bark: use, management, and commerce in
Africa. New York, NY, USA: The New York Botanical Garden Press, 5–25.
Romero C, Bolker BM. 2008. Effects of stem anatomical and structural traits on
responses to stem damage: an experimental study in the Bolivian Amazon.
Canadian Journal of Forest Research 38: 611–618.
Romero C, Bolker BM, Edwards CE. 2009. Stem responses to damage: the
evolutionary ecology of Quercus species in contrasting fire regimes. New
Phytologist 182: 261–271.
Rosell JA, Castorena M, Laws C, Westoby M. 2015. Bark ecology of twigs vs
main stems: functional traits across 85 species of angiosperms. Oecologia 178:
1033–1043.
Rosell JA, Gleason SM, Mendez-Alonzo R, Chang Y, Westoby M. 2014. Bark
functional ecology: evidence for tradeoffs, functional coordination, and
environment producing bark diversity. New Phytologist 201: 486–497.
Rosell JA, Olson ME. 2014a. Do lianas really have wide vessels? Vessel diameterstem length scaling in non-self-supporting plants. Perspectives in Plant Ecology,
Evolution and Systematics 16: 288–295.
Ó 2016 The Author
New Phytologist Ó 2016 New Phytologist Trust
Research 13
Rosell JA, Olson ME. 2014b. The evolution of bark mechanics and storage across
habitats in a clade of tropical trees. American Journal of Botany 101: 764–777.
Roth I. 1981. Structural patterns of tropical barks. Berlin, Germany: Gebruder
Borntraeger.
Ryan MG, Asao S. 2014. Phloem transport in trees. Tree Physiology 34: 1–4.
Scholz FG, Bucci SJ, Goldstein G, Meinzer FC, Franco AC, Miralles-Wilhelm
F. 2007. Biophysical properties and functional significance of stem water
storage tissues in Neotropical savanna trees. Plant, Cell & Environment 30:
236–248.
Schwilk DW, Gaetani MS, Poulos HM. 2013. Oak bark allometry and fire
survival strategies in the Chihuahuan Desert Sky Islands, Texas, USA. PLoS
ONE 8: e79285.
Sperry JS, Smith DD, Savage VM, Enquist BJ, McCulloh KA, Reich PB,
Bentley LP, von Allmen EI. 2012. A species-level model for metabolic scaling
in trees I. Exploring boundaries to scaling space within and across species.
Functional Ecology 26: 1054–1065.
Srivastava LM. 1964. Anatomy, chemistry and physiology of bark. International
Review of Forestry Research 1: 203–277.
Vines RG. 1968. Heat transfer through bark and the resistance of trees to fire.
Australian Journal of Botany 16: 499–514.
Westoby M, Falster DS, Moles AT, Vesk PA, Wright IJ. 2002. Plant ecological
strategies: some leading dimensions of variation between species. Annual
Review of Ecology and Systematics 33: 125–159.
Supporting Information
Additional supporting information may be found in the online
version of this article.
Fig. S1 Phylogenetic relationships of the 640 sampled species.
Fig. S2 Inner and outer bark thickness variation across the 18
sampled plant communities.
Table S1 Fire return interval for each of the 18 sampled plant
communities
Table S2 Environmental indices and correlations with thickness
traits
Please note: Wiley Blackwell are not responsible for the content
or functionality of any supporting information supplied by the
authors. Any queries (other than missing material) should be
directed to the New Phytologist Central Office.
New Phytologist (2016)
www.newphytologist.com
New Phytologist Supporting Information
Article title: Bark thickness across the angiosperms: more than just fire
Authors: Julieta A. Rosell
Article acceptance date: 06 January 2016
The following Supporting Information is available for this article:
Fig. S1 Phylogenetic relationships of the 640 sampled species.
Fig. S2 Inner and outer bark thickness variation across the 18 sampled plant communities.
Table S1 Fire return interval for each of the 18 sampled plant communities.
Table S2 Environmental indices and correlations with thickness traits.
Fig. S1. Phylogenetic relationships of the 640 sampled species. Numbers indicate
where branches of the phylogeny connect.
Figure S1 cont.
Figure S1 cont.
Figure S1 cont.
Fig. S2. Inner and outer bark thickness variation across 18 plant communities.
Residual (a) inner bark thickness and (b) outer bark thickness. Homogeneous groups
in gray to the right of boxplots estimated with non-parametric comparisons (P <0.05).
Boxplots centered around the median.
Table S1. Fire return interval and coding for each of the 18 plant communities.
Vegetation
Locality
Sarcocaulescent
shrubland
Desert
San José del Cabo,
Mexico
Morongo Valley,
California
Stunt Ranch,
California
Cariri, Brazil
Mediterranean
shrubland
Seasonally dry tropical
forest 1
Cool temperate
woodland
Temperate woodland
Seasonally dry tropical
forest 2
Xerophytic shrubland
Coastal Mediterranean
woodland
Temperate sclerophyll
forest
Temperate broadleaved deciduous
forest
Savanna 1
Tropical rainforest 1
Cool sclerophyll forest
Savanna 2
Temperate rainforest
Tropical rainforest 2
Tropical rainforest 3
Fire return References
interval
(yr)1 and
coding
>100 (1)
Rodríguez Trejo (2008), León
de la Luz et al. (2000)
>100 (1)
Brooks and Pyke (2001)
70–100
(2)
>100 (1)
National Park Service (2015)
Bothwell, Tasmania
5–20 (4)
Murphy et al. (2013)
Yengo National Park,
Australia
Chamela-Cuixmala
Reserve, Mexico
Pedregal de San
Ángel Reserve,
Mexico
Migliarino San
Rossore Park, Italy
Sydney area,
Australia
Pordenone, Italy
5–20 (4)
Murphy et al. (2013)
>100 (1)
>100 (1)
Maas et al. (2002);
Pennington et al. (2009)
Rodríguez Trejo (2008)
20 (3)
Mouillot et al. (2005)
5–20 (4)
Murphy et al. (2013)
>100 (1)
Conedera et al. (2002);
Kaltenrieder et al. (2010)
Botucatu, Brazil
Atherton Tablelands,
Australia
Mount Field National
Park, Tasmania
Howard Springs
Nature Park,
Australia
New South Wales
rainforests, Australia
Daintree National
Park, Australia
Los Tuxtlas Reserve,
Mexico
3–10 (5)
>100 (1)
Coutinho (1990)
Murphy et al. (2013)
>100 (1)
2–5 (5)
Murphy et al. (2013),
Kirkpatrick & Bridle (2013)
Murphy et al. (2013)
>100 (1)
Murphy et al. (2013)
>100 (1)
Murphy et al. (2013)
>100 (1)
Rodríguez Trejo (2008)
Pennington et al. (2009)
REFERENCES
Brooks ML, Pyke DA 2001. Invasive plants and fire in the deserts of North America.In
Galley KEM, Wilson TP. Proceedings of the invasive species workshop: the role of
fire in the control and spread of invasive species. Fire Conference 2000: the First
National Congress on Fire Ecology, Prevention, and Management. Miscellaneous
Publication No. 11. Tall Timbers Station, Tallahassee, Florida. 1–14.
Conedera M, Moretti M, Tinner W 2002. Storia ed ecologia degli incendi boschivi al sud
delle Alpi della Svizzera. In: Anfodillo T, Carraro V eds. Fuoco in Foresta:
Ecologia e Controllo. Atti del XXXIX Corso di Cultura in Ecologia: Università degli
Studi di Padova, Padua, Italy, 15–30.
Coutinho LM 1990. Fire in the Ecology of the Brazilian Cerrado. In: Goldammer J ed.
Fire in the Tropical Biota: Springer Berlin Heidelberg, 82–105.
Kaltenrieder P, Procacci G, Vannière B, Tinner W. 2010. Vegetation and fire history of
the Euganean Hills (Colli Euganei) as recorded by Lateglacial and Holocene
sedimentary series from Lago della Costa (northeastern Italy). The Holocene 20:
679–695.
Kirkpatrick JB, Bridle KL. 2013. Natural and cultural histories of fire differ between
Tasmanian and mainland Australian alpine vegetation. Australian Journal of Botany
61: 465–474.
León de la Luz JL, Pérez Navarro JJ, Breceda A. 2000. A transitional xerophytic
tropical plant community of the Cape Region, Baja California. Journal of
Vegetation Science 11: 555–564.
Maas JM, Martínez-Yrízar A, Patiño C, Sarukhán J. 2002. Distribution and annual net
accumulation of above-ground dead phytomass and its influence on throughfall
quality in a Mexican tropical deciduous forest ecosystem Journal of Tropical
Ecology 18: 821–834.
Mouillot F, Ratte J-P, Joffre R, Mouillot D, Serge Rambal a. 2005. Long-term forest
dynamic after land abandonment in a fire prone Mediterranean landscape (central
Corsica, France). Landscape Ecology 20: 101–112.
Murphy BP, Bradstock RA, Boer MM, Carter J, Cary GJ, Cochrane MA, Fensham
RJ, Russell-Smith J, Williamson GJ, Bowman DMJS. 2013. Fire regimes of
Australia: a pyrogeographic model system. Journal of Biogeography 40: 1048–
1058.
National Park Service 2015. Santa Monica Mountains. Retrieved August 2015.
http://www.nsp.gov/samo/learn/management/fireecology.htm.
Pennington RT, Lavin M, Oliveira-Filho A. 2009. Woody plant diversity, evolution, and
ecology in the tropics: perspectives from seasonally dry tropical forests. Annual
Review of Ecology Evolution and Systematics 40: 437–457.
Rodríguez Trejo DA. 2008. Fire regimes, fire ecology, and fire management in Mexico.
Ambio 37: 548–556.
Table S2. Environmental indices and Spearman correlations with residual total (TBT),
inner (IBT) and outer bark thickness (OBT). ***P <0.005, nsP  0.05
Variable
PC1
Wet season precipitation index
Variable
PC1
Dry season precipitation index
Annual precipitation
0.50
Precipitation driest month
0.53
Precipitation wettest month
0.51
Precipitation driest quarter
0.53
Precipitation wettest quarter
0.51
Precipitation coldest quarter
0.48
Precipitation warmest quarter
0.48
Precipitation seasonality
–0.45
Variance explained
93%
Variance explained
85%
rSp with residual TBT (N = 624)
–0.12*** rSp with residual TBT (N = 624)
–0.30***
rSp with residual IBT (N = 582)
–0.03ns
rSp with residual IBT (N = 582)
–0.29***
rSp with residual OBT (N = 582)
–0.07ns
rSp with residual OBT (N = 582)
–0.12***
Mean temperature index
Temperature seasonality index
Annual mean temperature
0.44
Mean diurnal temperature range
0.71
Maximum temperature warmest
0.39
Annual temperature range
0.71
month
Mean temperature wettest quarter
0.39
Mean temperature warmest quarter
0.43
Mean temperature driest quarter
0.38
Mean temperature coldest quarter
0.42
Variance explained
85%
Variance explained
84%
rSp with residual TBT (N = 624)
0.15***
rSp with residual TBT (N = 624)
0.18***
rSp with residual IBT (N = 582)
0.23***
rSp with residual IBT (N = 582)
0.07ns
rSp with residual OBT (N = 582)
–0.02ns
rSp with residual OBT (N = 582)
0.14***