Content uploaded by Lorenzo Alvarez-Filip
Author content
All content in this area was uploaded by Lorenzo Alvarez-Filip on Nov 25, 2017
Content may be subject to copyright.
LETTER Evaluating life-history strategies of reef corals from species
traits
Emily S. Darling,
1
*
Lorenzo Alvarez-Filip,
1,2
Thomas A. Oliver,
3
Timothy R. McClanahan
4
and
Isabelle M. Co
ˆte
´
1
Abstract
Classifying the biological traits of organisms can test conceptual frameworks of life-history strategies and
allow for predictions of how different species may respond to environmental disturbances. We apply a
trait-based classification approach to a complex and threatened group of species, scleractinian corals. Using
hierarchical clustering and random forests analyses, we identify up to four life-history strategies that appear
globally consistent across 143 species of reef corals: competitive, weedy, stress-tolerant and generalist taxa,
which are primarily separated by colony morphology, growth rate and reproductive mode. Documented
shifts towards stress-tolerant, generalist and weedy species in coral reef communities are consistent with
the expected responses of these life-history strategies. Our quantitative trait-based approach to classifying
life-history strategies is objective, applicable to any taxa and a powerful tool that can be used to evaluate
theories of community ecology and predict the impact of environmental and anthropogenic stressors on
species assemblages.
Keywords
Community assembly, coral reefs, C–S–R triangle, ecological strategies, functional diversity, selection
pressure.
Ecology Letters (2012)
INTRODUCTION
Ecologists are increasingly realising that trait-based views of species
assemblages may be more meaningful than comparisons of species
richness or composition (Cadotte et al. 2011). Trait-based
approaches can provide general and predictable rules for community
ecology, as well as a more mechanistic understanding of community
assembly and disassembly, habitat filtering and species coexistence,
particularly in the context of global climate change and biodiversity
loss (McGill et al. 2006). Species traits also provide important infor-
mation about life-history strategies, which can broadly define how
organisms interact with one another and their environment.
Life-history strategies describe consistent and context-independent
characteristics of organisms. The classic two-strategy life-history
framework of r–Kmodels (Pianka 1970) is now generally seen as
oversimplified as species can occur along a continuum of ‘fast’ (r)
to ‘slow’ (K) life histories (Stearns 1977). Three-strategy frameworks
resolve some difficulties of r–Kselection by adding a third ‘beyond
K’ group of stress-adapted species that can persist in unfavourable
habitats (i.e., via adversity selection, Greenslade 1983). For example,
Grime’s C–S–R triangle describes three life-history strategies in
plants, in which species are hypothesised to evolve strategies that
promote competitive (C), stress-tolerant (S) or ruderal (R) life histo-
ries (Grime 1977; Grime & Pierce 2012). Similar triangular continu-
ums of life-history strategies have also been proposed for insects
(Greenslade 1983) and fishes (Winemiller & Rose 1992). Life-history
models with three strategies are generally recognised as having more
predictive power than the two-strategy r–Kframework that can fail
to recognise additional axes of life-history variation (Stearns 1977;
Winemiller & Rose 1992). Although more than two or three life-his-
tory strategies likely exist in any given community, such multi-strat-
egy frameworks are rarely quantitatively described. Moreover, many
of these life-history frameworks are based on the responses of
organisms to different environmental or habitat conditions, instead
of on inherent characteristics of species (Westoby 1998). Further-
more, such classifications are often subjective (Gitay et al. 1997),
which can bias tests of evolutionary and ecological theories that are
better addressed with quantitatively and objectively identified life-
history strategies.
Here, we evaluate if life-history strategies can be directly inferred
from species traits in a complex and threatened group of organisms,
scleractinian corals. Climate change, overexploitation, pollution and
disease have resulted in global declines of coral cover and architec-
tural complexity on coral reefs, one of the world’s most complex
and biodiverse ecosystems (Gardner et al. 2003; Pandolfi et al. 2003;
Alvarez-Filip et al. 2009). Human impacts are altering coral commu-
nities in ways that are unprecedented from the historical record
(Pandolfi & Jackson 2006; van Woesik et al. 2012) as historically
dominant corals are replaced by more persistent and opportunistic
species (Knowlton 2001; McClanahan et al. 2007; Green et al. 2008;
Alvarez-Filip et al. 2011). However, shifts in coral species can often
be overlooked for a variety of reasons: they may be subtle –
species-level coral identification is challenging –and changes may
occur on decadal or millennial timescales (Pandolfi & Jackson 2006;
van Woesik et al. 2012). Changes may also go unnoticed as the
baseline for what is considered a ‘normal’ community composition
is often unknown and shifts in species composition can occur
slowly over time (Pauly 1995). More importantly, a general lack of
1
Department of Biological Sciences, Simon Fraser University, Burnaby, British
Columbia, V5A 1S6, Canada
2
Healthy Reefs Initiative, 1755 Coney Drive, Belize City, Belize
3
Hawaiian Institute for Marine Biology, University of Hawai’i Ma
¯noa, 46-007
Lilipuna Rd, Ka
¯ne’ohe, HI, 96744, USA
4
Marine Programs, Wildlife Conservation Society, Bronx, NY, 10460, USA
*Correspondence: E-mail: edarling@sfu.ca
©2012 Blackwell Publishing Ltd/CNRS
Ecology Letters, (2012) doi: 10.1111/j.1461-0248.2012.01861.x
long-term abundance information for individual coral species makes
it difficult to identify species responses to environmental change
and anthropogenic stress, and whether these responses are predict-
able.
There is currently no framework of life-history strategies for scle-
ractinian corals, possibly because data on species traits are too
sparse or scattered for classification or because corals do not fit well
into existing frameworks (e.g. corals are clonal invertebrates with a
complicated life cycle, Jackson & Hughes 1985). A few studies have
considered how some coral traits may relate to life-history strategies.
For example, small corals with brooding reproduction, fast growth
rates and high population turnover are expected to be ‘weedy’
(Knowlton 2001), while large, slow-growing colonies of massive
corals are expected to be more tolerant to chronically stressful or
variable environments (Jackson & Hughes 1985; Soong 1993;
Rachello-Dolmen & Cleary 2007). Similarly, variation in colony
morphology and reproductive mode are thought to suggest three
primary life histories (competitors, stress-tolerators and ruderals;
Edinger & Risk 2000; Murdoch 2007). Observations of increasing
abundances of ‘weedy’ species (Green et al. 2008) and the persis-
tence of massive species on disturbed Caribbean (Alvarez-Filip et al.
2011) and Indo-Pacific reefs (McClanahan et al. 2007; Rachello-Dol-
men & Cleary 2007) suggest that life-history traits can predict which
corals are ‘winners’ or a ‘losers’ in the face of environmental change
(Loya et al. 2001; van Woesik et al. 2012). For example, branching
and plating acroporid corals are dominant species that are very sen-
sitive to stress and disturbance (i.e., ‘losers’), while massive species
and ‘weedy’ species are more likely to be ‘winners’ and persist in
unfavourable and disturbed environments (Loya et al. 2001; McCl-
anahan et al. 2007). However, the underlying species characteristics
that may predict these responses are difficult to evaluate without a
comprehensive understanding of coral traits and associated life-his-
tory strategies.
In this study, we describe a novel, quantitative method that can
be used to evaluate and identify life-history strategies from species
traits using hierarchical clustering and random forests analyses. We
compile a global database of species traits for reef-building corals
and classify taxa into life-history strategies that can be used to eval-
uate ongoing community shifts on coral reefs. Our approach to
objectively classify life-history strategies is applicable to any species
group and can be used to establish trait-based life-history frame-
works, find general rules of community ecology and predict the
impacts of environmental and anthropogenic impacts on ecological
communities.
MATERIAL AND METHODS
Coral species traits
To evaluate life-history strategies in reef corals, we collected infor-
mation on 11 commonly available species traits: colony growth
form, solitary colony formation, reproductive mode and fecundity,
maximum colony size, corallite diameter, depth range, generation
time, growth rate, skeletal density and symbiotic zooxanthellae
(Symbiodinium) associations (Table 1; see also Appendix S1 in Sup-
porting Information). We specifically focused on traits that were
expected to affect coral population dynamics, and for which quanti-
tative data were available at a global scale. All life-history traits are
variable and phenotypic plasticity within species is ubiquitous. Cor-
als, in particular, have notoriously variable life-history traits, and
there can be extensive phenotypic plasticity within populations and
species (Todd 2008). However, the goal of our study was to provide
a broad, global comparison of life-history traits across species and
not focus on intraspecific variation. Thus, we conducted an exten-
sive literature survey to provide the most accurate ‘average’ descrip-
tion of species-specific traits, and we leave in-depth assessments of
intraspecific variability to future investigations.
We collected trait information for 847 scleractinian corals, com-
prising 101 western Atlantic and Caribbean (hereafter referred to as
Atlantic) species and 746 Indo-Pacific species. This information
came from 236 sources, including taxonomic monographs, regional
identification guides, published literature, secondary sources and
online databases (see Appendices S1–S2). Only taxa with informa-
tion for more than 60% of the species traits were included in the
analysis (i.e., species with data for at least 7 of 11 traits, n=143
species: 32 Atlantic and 111 Indo-Pacific species from 51 genera
and 19 families). Our cut-off of 60% trait coverage was a trade-off
between the number of species included in the analysis and compre-
hensive trait information for each species. For example, only a small
subset of species (n=20; 8 Atlantic and 12 Indo-Pacific species)
had complete (100%) trait information. We conducted a sensitivity
analysis and found clusters to be qualitatively unchanged from 100
to 60% trait information. With information on less than 60% of
traits, clustering patterns became more variable, although generally
similar to clustering analyses with more trait coverage.
Classifying life-history strategies
We took a data-driven approach to classify life-history strategies
using a posteriori group classification of Ward’s hierarchical clustering
analyses (e.g. Gitay et al. 1997), followed by random forests analyses
to identify influential traits, and Principal Coordinates ordination to
visually show species life-history strategies and traits in multivariate
space.
We used hierarchical clustering and non-parametric multivariate
analyses of variance (MANOVAs) as an objective and quantitative
method to identify life-history strategies. First, we established a trait
dendrogram of species relationships using Ward’s hierarchical clus-
tering of a Gower dissimilarity matrix. We chose the Gower dissimi-
larity index to compare the multivariate trait distance between
species because it allows for mixed types of data, missing values
and can weight individual traits differently (Laliberte´ & Legendre
2010). The 11 traits were independent characteristics (variance infla-
tion factors <5 for all continuous variables) and all traits were
weighted equally in the analysis, except two traits, colony growth
form and reproductive mode, which were each reclassified into sep-
arate binary variables (branching, plating and domed colony growth
forms; brooding and spawning reproductive modes; see Appendix
S1). The three binary states for growth form were weighted by 0.33
and the two binary states for reproductive mode were weighted by
0.5, to avoid artificially inflating the effect of these traits (Laliberte´
& Legendre 2010). Dissimilarity indices and clustering were per-
formed using the ‘FD’ package (Laliberte´ & Legendre 2010) in R
(2012) R Development Core Team (2010).
We evaluated the ‘best-fit’ number of clusters by testing how
many clusters maximised both within-cluster homogeneity and
between-cluster dissimilarity. We used non-parametric MANOVAs to
evaluate different groupings of species by comparing the coefficient
©2012 Blackwell Publishing Ltd/CNRS
2E. S. Darling et al. Letter
of determination (R
2
) across nine different clustering scenarios
(between two and 10 clusters). We then identified a cut-off or
‘elbow’ where further clustering resulted in a sharp decrease
(>15%) in the amount of explained variance; clusters above this
cut-off were deemed to be the ‘best-fit’ clusters (see Fig. S1). This
is comparable to the use of scree plots to evaluate the number of
principal components to include in a principal components analysis.
After identifying the number of ‘best-fit’ clusters, we calculated the
mean and standard deviation of trait values within each cluster and
tested for mean differences in multivariate traits across clusters
using non-parametric MANOVAs. We also compared the variability of
species strategies within life-history clusters using a multivariate
homogeneity of variance test, which calculates the average distance
of each species to median group trait values, followed by Tukey’s
Honestly Significant Differences tests to identify pairwise differ-
ences between groups (Oksanen et al. 2011).
In addition to classifying species into life-history strategies, we
also wanted to identify traits that were the best predictors of coral
life-history strategies. We used random forests analyses to evaluate
the relative importance of different traits. Random forests are a
machine-based learning method that estimates variable importance
by combining many classification trees through bootstrap sampling
and model averaging, and can also account for multicollinearity
among variables (Cutler et al. 2007). During each iteration of the
classification tree, a random subset of three traits (out of 11) was
used to classify species into life-history groups, and the final tree
was then compared with the original ‘blueprint’ trait dendrogram
produced by Ward’s hierarchical clustering of the full set of 11
traits. After 20 000 iterations, we compared the increase in the
cluster misclassification rate for each trait when it was excluded,
and all other traits were held constant, to assess the relative
importance of each trait; traits that resulted in the greatest increase
in misclassifications when excluded were identified as the most
important (Cutler et al. 2007). Prior to random forests analysis,
missing species trait values were filled in using standard data impu-
tation through proximity (Liaw & Wiener 2002; Cutler et al. 2007)
–for each cluster, missing values of continuous traits were
replaced with the median of the trait within the cluster, and miss-
ing values of categorical traits were replaced with the most fre-
quent level of the trait in the cluster. In total, 288 missing values
out of 2002 data points, or 14.4%, were replaced by imputation
for the random forests analysis. We assessed the model accuracy
of classification trees and random forests using standard metrics
(e.g. out-of-bag error rate; Cutler et al. 2007). We used the
‘randomForest’ package (Liaw & Wiener 2002) in R (2012) R
Development Core Team (2010) for these analyses.
Finally, we used a Principal Coordinates Analysis (PCoA)
ordination of the Gower dissimilarity matrix to visually show the
life-history groups and species traits in multivariate space. However,
this ordination is purely descriptive and does not provide statistical
tests of clustering structure among species (Borcard et al. 2011).
Following species ordination, we projected each trait onto the plot
a posteriori using a double projection method based on correlations
of the species traits with the PCoA axes; species traits were standar-
dised prior to ordination and negative eigenvectors were adjusted
using the Cailliez correction (Borcard et al. 2011). Multivariate analy-
ses were performed using the ‘vegan’ (Oksanen et al. 2011) and
‘FD’ (Laliberte´ & Legendre 2010) packages in R (2012) R Develop-
ment Core Team (2010).
RESULTS AND DISCUSSION
Life-history strategies of reef corals
Models with two, three and four clusters of coral species received
the most support in the hierarchical clustering analyses (Table 2,
Fig. 1, see Figs. S1–S3). The two-cluster scenario identified a ‘fast’
life history of large, branching and plating corals and a ‘slower’ life
history of smaller, slower-growing corals (see Table S1; Fig S2). The
three-cluster scenario distinguished a third group of small, brooding
corals, which would fit with a ruderal or opportunistic life history
of three-strategy frameworks (see Table S2, Fig. S3; Grime 1977;
Grime & Pierce 2012). The four-strategy clustering scenario (Table
2; Fig. 1) added a ‘generalist’ group of species that appear to have
some traits of each life history identified in the three-cluster sce-
nario. We focus on the four-cluster scenario for a discussion of reef
coral life histories. These four clusters describe significantly differ-
ent groupings of species traits (non-parametric MANOVA,P=0.001)
and random forests analyses identified colony growth form, growth
rate and reproductive mode as the most influential traits responsible
for clustering the different groups (Fig. 2; see also Fig. S4).
Cluster 1 describes large, branching and plating species that grow
quickly, occur at shallow depths and reproduce by broadcast spawn-
ing (Table 2). This group includes the staghorn and bottlebrush cor-
als in the genus Acropora in the Atlantic and Indo-Pacific, Dendrogyra
cylindricus in the Atlantic and species of Montipora, Pocillopora and Tur-
binaria in the Indo-Pacific. We suggest that cluster 1 describes a
Table 1 Summary of 11 species traits used to classify life-history strategies of
scleractinian reef corals. Detailed descriptions and source references for each trait
can be found in Appendices S1 and S2
Species trait Description Type of data
Colony
growth form
Branching,
plating, domed*
,
†
Categorical
Solitary colonies Solitary or
colonial colonies
Categorical
Reproductive
mode
Brooding,
broadcast spawning*
Categorical
Colony size Largest recorded
size of colony, cm
Continuous
Corallite
diameter
Average size
of corallites, cm
Continuous
Depth range Median depth
(m) of reported
depth range
Continuous
Fecundity Number of eggs
per polyp or
mesentery
Continuous
Generation
length
Number of years
between generations
Continuous
Growth rate Average annual
growth rate,
mm year
!1
Continuous
Skeletal density Average density
of CaCO
3
skeleton,
g cm
!3
Continuous
Symbiont
richness
Rarefaction curve of
Symbiodinium genotypic
richness corrected for
sampling effort
Continuous
*Colonies can have more than one characteristic.
†
Domed corals include massive, submassive and encrusting morphologies.
©2012 Blackwell Publishing Ltd/CNRS
Letter Life-history strategies of reef corals 3
‘competitive’ life-history strategy that is typically efficient at using
resources and can dominate communities in productive environ-
ments (Grime 1977; Grime & Pierce 2012). Branching and plating
corals often grow quickly into large, arborescent colonies that can
create canopies to shade out competitors for light and plankton
prey, making them effective competitors in shallow, high light and
low water flow environments (Baird & Hughes 2000). However,
these species are extremely sensitive to breakage and dislodgement
during storms (Madin 2005) and often exhibit high mortality follow-
ing temperature anomalies and coral bleaching (McClanahan et al.
2007), suggesting they are only dominant in ideal environments.
Interestingly, previous qualitative studies have classified species of
Acropora and Montipora as a ruderal or weedy life history because of
their fast growth and good colonising characteristics (Edinger &
Risk 2000; Rachello-Dolmen & Cleary 2007). However, the ability
of these species to competitively dominate coral assemblages in the
Indo-Pacific (Baird & Hughes 2000; Hughes et al. 2012) and their
sensitivity to environmental change and marked declines from his-
torical dominance in the Atlantic (Alvarez-Filip et al. 2011) suggest,
instead, that acroporids may be associated with a competitively
dominant life-history strategy.
Species within cluster 2 can reproduce by brooding and have
smaller colony sizes. This cluster includes branching Porites species
in the Atlantic and Indo-Pacific, species of Madracis and Agaricia in
the Atlantic, some pocilloporids (Pocillopora damicornis,Stylophora pistil-
lata,Seriatopora hystrix) and some faviids (e.g., species of Cyphastrea,
Goniastrea and Leptastrea) in the Indo-Pacific. Cluster 2 appears to
describe a weedy or ruderal strategy of species that can opportunis-
tically colonise recently disturbed habitats (Grime 1977; Grime &
Pierce 2012). Ecological theory suggests that successful weeds
reproduce faster and survive better than non-weedy species. While
species in cluster 2 displayed some traits of fast reproduction (e.g.,
shorter generation time, on average), they did not have higher
fecundity (as measured by eggs per polyp) than the other clusters
(Table 2). Moreover, this lack of a fecundity advantage could be
compounded by their smaller colony sizes (Table 2; Soong 1993).
The reproductive advantage of weedy corals may instead lie in a
correlate of their brooding reproductive mode. Although brooders
produce relatively large offspring (Knowlton 2001), which is associ-
ated with high parental investment and would be an unusual charac-
teristic of weeds, some brooding corals can produce larvae via
parthenogenesis (Ayre & Miller 2004). Parthenogenesis might allow
for successful reproduction at low population densities, which may
often occur on recently disturbed reefs. In contrast, many common
broadcast spawning species are vulnerable to Allee effects and can
fail to reproduce in small populations (Knowlton 2001). Further-
more, as population size increases, brooders can also disperse sexu-
ally produced larvae that may favour genetically diverse colonists
(Ayre & Resing 1986). Weedy corals may also be better survivors
than non-weedy corals because they show the most variation in
their species traits compared the other life-history strategies (multi-
variate homogeneity of variances test, P=0.009, Fig. 3), which may
allow these taxa to colonise a variety of disturbed environments,
such as heavily fished reefs or shallow back reef lagoons.
If clusters 1 and 2 describe competitive and weedy life-history
strategies, theory would suggest that one of the remaining clusters
(3 or 4) might represent a stress-tolerant strategy. Species in cluster
4 appear to be the best candidates for a stress-tolerant life history.
This cluster includes slow-growing species that reproduce by broad-
Table 2 Summary of species traits across four life-history strategies of (a) global (n=143 species), (b) Atlantic (n=32) and (c) Indo-Pacific (n=111 species) reef corals. For categorical traits (colony growth form,
solitary colonies and reproductive mode), the per cent of species with each trait characteristic is presented; mean (standard deviation) is reported for continuous traits
Life history
No.
species
%
Branching
%
Domed
%
Plating
%
Solitary
%
Brooding
%
Spawning
Colony
size (cm)
Corallite
diameter
(cm)
Depth
range (m)
Fecundity
(eggs polyp
!1
)
Generation
length
(years)
Growth rate
(mm year
!1
)
Skeletal
density
(gem
!3
)
Symbiont
diversity
(m-value)
(a) Global
1–Competitive 46 95.7 0 39.1 0 0 100 259.32 (314.68) 2.70 (1.31) 12.53 (5.50) 20.04 (32.39) 9.72(1.11) 49.83(41.28) 1.51 (0.34) 1.78 (1.28)
2–Weedy 23 39.1 56.5 21.7 0 100 18.2 105.54 (138.54) 4.45 (5.86) 20.58 (14.91) 19.70 (14.26) 9.41 (1.56) 11.35 (10.04) 1.73(0.41) 1.70 (1.61)
3–Generalist 14 35.7 100 100 0 0 100 248.77 (243.39) 3.56 (3.55) 17.95 (5.05) 18.46 (9.26) 10.00(0.00) 19.18 (11.58) 1.62 (0.34) 1.79(0.63)
4–Stress-tolerant 60 0 100 1.7 5 3.4 100 137.81 (285.26) 8.57 (9.83) 19.50 (10.45) 372.45 (694.15) 10.17(1.52) 7.98 (6.69) 1.54 (0.28) 1.42(0.86)
(b) Atlantic
1–Competitive 3 100 0 0 0 0 100 304.80 (61.00) 2.52 (3.02) 21.10 (8.67) 63.34 (83.71) 10.00 (0.00) 71.09 (50.13) 1.45 (0.58) 1.07(0.56)
2–Weedy 15 26.7 60 33.3 0 100 0 69.80 (97.12) 5.30 (6.85) 26.26 (16.13) 23.00 (16.61) 10.00 (0.00) 6.99 (7.14) 1.72 (0.50) 1.19(0.68)
3–Generalist 2 0 100 100 0 0 100 166.97(46.70) 2.58 (0.02) 23.88 (5.13) 17.31 (13.59) 10.00 (0.00) 7.40(1.07) 1.31 (0.18) 2.50 (0.30)
4–Stress-tolerant 12 0 100 8.3 0 0 100 105.75(48.55) 7.15 (5.18) 32.42 (14.62) 100.68 (123.74) 10.00 (0.00) 5.53(2.00) 1.51 (0.43) 1.26(0.84)
(c) Indo-Pacific
1–Competitive 43 95.3 0 41.9 0 0 100 252.14 (338.70) 2.71 (1.15) 11.93 (4.82) 14.39 (16.31) 9.70 (1.15) 47.18(40.52) 1.52 (0.32) 1.83(1.30)
2–Weedy 8 62.5 50 0 0 100 50 177.02 (186.01) 2.86 (3.24) 10.62 (2.37) 14.74 (9.76) 8.38 (2.33) 16.97 (10.90) 1.73 (0.27) 2.67 (2.37)
3–Generalist 12 41.7 100 100 0 0 100 263.64 (263.22) 3.73 (3.86) 16.96 (4.50) 19.60 (8.20) 10.00(0.00) 23.11 (10.65) 1.77 (0.30) 1.68(0.60)
4–Stress-tolerant 48 0 100 0 6.2 4.3 100 146.97 (322.75) 9.01 (10.89) 16.28 (5.82) 472.58 (790.45) 10.21 (1.70) 8.82 (7.53) 1.55 (0.19) 1.46(0.87)
©2012 Blackwell Publishing Ltd/CNRS
4E. S. Darling et al. Letter
Figure 1 Hierarchical cluster analysis of global scleractinian corals based on species traits. The analysis includes 143 species with "60% trait coverage (i.e., information
for at least 7 of 11 traits). Clustering scenarios with two, three and four groups received the strongest model support (see Figs. S1–S3); a four-cluster scenario is shown
here. Boxes indicate Atlantic (Atl) or Indo-Pacific (IP) distributions.
©2012 Blackwell Publishing Ltd/CNRS
Letter Life-history strategies of reef corals 5
cast spawning and have primarily domed morphologies, large coral-
lites and high fecundity. Examples include the lobe and brain corals
Montastraea annularis and M. cavernosa,Colpophyllia natans and species
of Diploria in the Atlantic, and massive Porites species and many
faviids (e.g. Favia,Favites,Platygyra,Goniastrea) in the Indo-Pacific.
Slow growth, longer generation times, large corallites, which pro-
mote energy storage (van Woesik et al. 2012), and high fecundity
during episodic spawning events may all be advantageous traits in
chronically harsh environments with, for example, low light or high
sedimentation. These traits have previously been assigned to a sub-
dominant, stress-tolerant life history (Edinger & Risk 2000; Rac-
hello-Dolmen & Cleary 2007). Long-lived corals can also persist in
the absence of recruitment and can withstand sustained recruitment
failure for decades, which may increase their long-term survival in
stressful environments (Hughes & Tanner 2000). Interestingly, clus-
ter 4 is the most species-rich group, possibly because stress can be
caused by many environmental factors, including depth and low
light, wave exposure, sedimentation and temperature variability,
which may create many ecological axes that can promote speciation
within this group.
The remaining group, cluster 3, includes an assortment of species
that show some overlap with the competitive, weedy and stress-toler-
ant life histories in the Principal Coordinates ordination (Table 2;
Fig. 3). These taxa occur as domed and plating colonies (but can also
have branching growth forms), with moderate growth rates and can
reach large colony sizes. This group includes species of Echinopora,
Hydnophora, Montipora, Turbinaria and Pachyseris in the Indo-Pacific, and
two Atlantic species, Montastraea faveolata and M. franksi. These species
may represent a ‘generalist’ life-history strategy that can do well in
habitats where competition is limited by low levels of stress (Grime
1977; Winemiller & Rose 1992). Alternatively, this cluster may repre-
sent a subgroup of stress-tolerant taxa that are somewhat more com-
petitive than average for that group, with horizontally spreading
plating colonies and faster growth rates. For now, we label this cluster
as a ‘generalist’ life-history strategy and more trait information is
needed to discern whether species within this cluster are a separate
life-history strategy, a type of stress-tolerant life history, or even a
slower-growing subset of ‘subdominant’ competitive taxa.
Three of the coral clusters emerging from our analysis resemble
three-strategy life-history frameworks, such as Grime’s triangle of
714
Disturbance
Stress
Ideal
8
12
–0.4
–0.3
–0.2
–0.1
0.0
0.1
0.2
–0.3 –0.2 –0.1 0.0 0.1 0.2 0.3 0.4
PCoA axis 2
PCoA axis 1
C
S
R
PCoA axis 2
(4.1%)
Competitive
Stress-toleran
t
PCoA axis 1 (6.0%)
1
2
6
3
4
9
11
5
Weedy
714
Disturbance
Stress
Ideal
8
12 Fun fung
Hel act
Pav dec
Fav fr
Alv fen
Aga lam
Fun gr
(a) (b)
10
13
Figure 3 Grime’s triangle of life-history strategies applied to reef corals. (a) The conceptual framework of Grime’s three primary life-history strategies and the
hypothetical environmental conditions associated with each strategy. (b) Principal Coordinates ordination of 143 coral species with four life-history strategies (red:
competitive, green: weedy, blue: stress-tolerant, grey: generalist). Open circles are Atlantic taxa and filled circles are Indo-Pacific taxa. Arrows indicate trait loadings; traits
are numbered from most important to least important in differentiating the life-history strategies: (1) domed morphology; (2) growth rate; (3) brooding reproduction; (4)
fecundity; (5) broadcast spawning reproduction; (6) branching morphology; (7) colony size; (8) skeletal density; (9) plating morphology; (10) corallite diameter; (11) depth;
(12) symbiont diversity; (13) generation length and (14) solitary colonies.
Solitary colonies
Generation length
Symbiont diversity
Depth range
Corallite diameter
Growth form, plating
Skeletal density
Colony size
Growth form, branching
Reproduction, spawning
Fecundity
Reproduction, brooding
Growth rate
Growth form, domed
Global
Atlantic
Indo-Pacific
Figure 2 Importance of individual traits for differentiating the four clusters of
global, Atlantic and Indo-Pacific reef corals as determined by random forests
analyses (see also Fig. S4). Circle size indicates rank of trait importance; larger
circles represent more influential traits than smaller circles. A summary of
species traits is given in Table 1.
©2012 Blackwell Publishing Ltd/CNRS
6E. S. Darling et al. Letter
primary plant life-history strategies (Grime 1977; Grime & Pierce
2012; Fig. 3). It is perhaps not surprising that reef-building corals
and plants can share some similarities in their life-history strategies.
Corals and plants are both sessile and clonal organisms (Jackson &
Hughes 1985), structured by competition for space (Lang et al. 1990;
Karlson & Hurd 1993), gradients of environmental productivity via
light and food availability (Anthony & Connolly 2004), and distur-
bance (Connell 1978; Karlson & Hurd 1993). Previous studies inves-
tigating coral morphology (Edinger & Risk 2000; Murdoch 2007),
the recruitment of juvenile corals (Bak & Engel 1979) and the
response of coral assemblages to depth gradients (Jackson & Hughes
1985), bleaching stress (Obura 2001) and pollution (Rachello-Dol-
men & Cleary 2007) have also suggested three major life-history
strategies of reef corals, which are similar to those we describe quan-
titatively here, although with different species classifications (e.g.,
acroporids were classified as ruderal, and domed and plating species
with moderate growth rates were classified as competitive).
Complexities of coral life histories
Although Grime’s triangle can reflect important life-history patterns
across species, this framework has been criticised for being too gen-
eral and for failing to consider the many complexities associated
with life-history strategies (e.g., the Grime–Tilman debate summar-
ised by Aerts 1999; see also Westoby 1998). These drawbacks apply
here, even though the C–S–R framework appears to have some
application to reef corals. For example, ‘faster’ or ‘slower’ life histo-
ries, and competitive, stress-tolerant and weedy life histories, are
end point life-history strategies (e.g., located at the ‘tips’ of Grime’s
triangle), but species can have suites of traits that place them along
a continuum between primary strategies (Grime 1977; Winemiller &
Rose 1992; Grime & Pierce 2012). This appears to be the case for
species we have labelled as generalists (cluster 3), which display
traits that occur in between competitive, weedy and stress-tolerant
life histories (i.e., ‘C–S–R’ strategy, Grime 1977; Grime & Pierce
2012; Fig. 3). It is also the case for a few species that were placed
in one of the three conventional strategies. For example, four
stress-tolerant species (Alveopora fenestrata,Favia fragum and two fun-
giids, Fungia fungites and Heliofungia actiniformis) overlap with the
weedy group, suggesting these species may be part of a secondary
strategy of species adapted to lightly disturbed and unproductive
habitats (i.e., ‘S–R’ strategy; Grime 1977; Grime & Pierce 2012).
‘Mixing and matching’ of traits in the generalist life history, and
overlap between primary strategies, suggests there may be a variety
of environmental selection pressures and trade-offs shaping life-
history strategies in reef corals. This may explain why we observed
some trait combinations in coral life histories that are not usually
associated with these strategies. For example, fast growth and high
fecundity are expected to be weedy traits (Grime 1977; Grime &
Pierce 2012), but competitive corals have the fastest growth rates
and stress-tolerant corals, the highest fecundities (Table 2). Brood-
ing reproduction is another example of an unusual trait in coral life
histories. Brooding is associated with parental investment, which is
usually considered a competitive (K) trait (Winemiller & Rose 1992).
However, in reef corals, brooding predominates among weedy,
opportunistic species (Table 2). Brooding in weedy corals may
reflect the need for asexual reproduction by isolated colonies
following disturbance, but may also entail a trade-off between
reproductive output and genetic heterozygosity in environments
where competition is low. Conversely, broadcast spawning in com-
petitive, stress-tolerant and generalist species may increase genetic
heterozygosity where competition is high, but might also result in a
limited ability to recover from small population sizes following dis-
turbance (e.g. Allee effects, Knowlton 2001). Consequently, the
reproductive mode of corals may be a response to the trade-off
between colonisation ability and competition, thus giving rise to
intermediate strategies and seemingly unusual trait combinations
that attempt to optimise these trade-offs.
Corals do not occupy all of the trait space of the hypothetical tri-
angle framework (Fig. 3b). This may be because most reef corals are
good competitors, compared with other benthic components of coral
reefs (Lang et al. 1990), moving species towards the competitive side
of the triangle. We hypothesise that other components of the coral
reef benthic community, such as sponges, soft corals and algae, might
primarily occupy trait space within the weedy and stress-tolerant
zones to further complete the life-history triangle. Future efforts to
quantify life-history strategies for the entire benthic community
would also be more comparable to the larger groups of taxa (i.e.,
plants, Grime 1977; fishes, Winemiller & Rose 1992; Grime & Pierce
2012) for which life-history theories were originally developed.
Understanding the life-history strategies of the entire coral reef ben-
thic community might also reveal whether ongoing shifts towards
non-coral assemblages (Norstro
¨met al. 2009) are associated with dif-
ferent life-history strategies. Overall, this highlights the need for
more trait information within coral species and other benthic groups.
Biogeographic and phylogenetic patterns
The life-history strategies we have identified appear largely
independent of biogeography as Atlantic and Indo-Pacific species
occur within each cluster (Fig. 1). For example, two lower-level
clusters of weedy taxa are apparent in the PCoA ordination and
each sub-cluster includes Atlantic and Indo-Pacific species (Fig. 1;
Fig. 3b). One sub-cluster includes shallow, fast-growing, branching
taxa, such as Pocillopora damicornis,Seriatopora hystrix,Stylophora pistillata
and species of digitate Porites from the Indo-Pacific (P. cylindrica,
P. rus) and two Atlantic taxa (P. furcata and P. porites), while a
slower-growing domed and plating sub-cluster includes species of
Agaricia,Porites astreoides,Siderastrea radians in the Atlantic, and several
faviid species from the Indo-Pacific (Cyphastrea ocellina,Goniastrea
aspera,Leptastrea purpurea). However, different traits nevertheless
separated regional taxa within the four clusters. Random forests
identified growth rate and colony growth form as the most impor-
tant traits differentiating clusters of Indo-Pacific species, while
reproductive mode and colony size were the most important traits
differentiating Atlantic species (Fig. 2; see also Fig. S4).
One of the more remarkable differences observed between the
regions was the larger proportion of weedy taxa in the Atlantic (15
weedy species of 32 total species, ~47%) than the Indo-Pacific (8 of
111, ~7%). This pattern likely reflects the overabundance of brood-
ers (a trait of weedy corals) in the Atlantic, which has been hypoth-
esised to be associated with colonisation history: more brooding
species may have been able to colonise Atlantic reefs from the east-
ern Pacific because brooded autotrophic larvae can be effective
long-distance dispersers (Baird et al. 2009). The overabundance of
brooders could also reflect the higher survival of brooding taxa fol-
lowing regional patterns of extinction over longer time scales
(Edinger & Risk 1995; van Woesik et al. 2012).
©2012 Blackwell Publishing Ltd/CNRS
Letter Life-history strategies of reef corals 7
Many life-history traits are phylogenetically conserved and our
classifications may in part reflect evolutionary relationships. For
example, most life-history classifications are consistent within gen-
era. Of the 22 genera with more than two species included in our
analysis, only seven genera (Cyphastrea,Goniastrea,Montipora,
Pocillopora,Porites, Siderastrea and Turbinaria) have species in different
life-history groups. This suggests that for most taxa (15 of 22, or
68%), genus-level identification may be sufficient to classify coral
life histories. There was more variability within families. Of the 15
families (based on Fukami et al. 2008) with more than one species,
10 had multiple life-history classifications. This suggests that sclerac-
tinian families often include species with diverse life histories,
although continued reconstruction of reef coral phylogenies may
also resolve some of this variability.
Forecasting the future of coral communities and the usefulness of
a trait-based approach
A key question arising from our results is whether we can predict
shifts in coral communities on the basis of life-history strategies. We
hypothesise that while competitive species can dominate less
impacted reefs, increasing stress and disturbance from human
impacts (fishing, pollution and sedimentation) or environmental con-
ditions (thermal stress and ocean acidification) can lead to the loss of
these sensitive competitive corals and their replacement with stress-
tolerant, weedy and generalist species, which may be better able to
persist in unproductive conditions and recolonise disturbed reefs.
There is some evidence that shifts from competitive to stress-tol-
erant, weedy and generalist life histories have occurred on contem-
porary coral reefs. Following the precipitous loss of competitive
acroporid corals from Caribbean reefs in the 1970s, stress-tolerant
and generalist Montastraea corals dominated communities until dis-
ease, bleaching and other disturbances led to high mortality of
Montastraea and other stress-tolerant species (Alvarez-Filip et al.
2011). Currently, reefs are often dominated by weedy Agaricia and
Porites corals (Aronson et al. 2004; Green et al. 2008), although even
these species have experienced population declines, likely associated
with recruitment failure (Hughes & Tanner 2000). Furthermore,
weedy and generalist species went regionally extinct from the Carib-
bean during large-scale climate disturbances of the Plio–Pleistocene
(van Woesik et al. 2012), suggesting even more ‘hardy’ life histories
may be vulnerable to population losses. Such shifts away from
architecturally complex competitive species towards the simpler
morphologies of stress-tolerant, weedy and generalist life histories
may underlie the ‘flattening’ of Caribbean coral reefs observed in
the past decades (Alvarez-Filip et al. 2009, 2011) and may also be
associated with region-wide declines in Caribbean reef fish popula-
tions linked to habitat degradation (Paddack et al. 2009). Compara-
ble community shifts have been reported on Indo-Pacific reefs
(McClanahan et al. 2007; Rachello-Dolmen & Cleary 2007; Hughes
et al. 2012), although further studies are needed to fully evaluate
hypotheses of life-history replacement within coral assemblages.
Species traits are implicit in life-history theory, yet the lack of
quantitative, trait-based methods to classify life-history strategies has
impeded the evaluation and adoption of life-history frameworks
(Westoby 1998). There are advantages to identifying such frame-
works. Trait-based life-history strategies may contribute to the gen-
erality of community ecology through a more mechanistic
understanding of community assembly and species coexistence (e.g.,
McGill et al. 2006). Furthermore, objectively quantifying life-history
strategies directly from species traits may be a pragmatic approach
to predict the increasing impacts of environmental and anthropo-
genic stressors on diverse species assemblages. Here, we have pro-
vided an objective, statistical approach to identify life-history
strategies from species traits that can be applied to any ecological
community. Our findings in relation to corals are consistent with
findings for communities of plants, insects and fishes, which sug-
gests that there may be a limited number of life-history strategies
available to organisms (Stearns 1977; Grime & Pierce 2012). The
next step may be to simplify the assignment of species to concep-
tual strategies described by multiple traits by defining the axes of
life-history variation using single traits (e.g. the leaf–height–seed
strategy; Westoby 1998), which can allow for even more direct glo-
bal comparisons across species. Establishing simple and universal
frameworks of species life histories can be used to understand fun-
damental axes of life-history variation and identify general patterns
in community ecology.
ACKNOWLEDGEMENTS
N. Dulvy, C. Harley, D. Braun, S. Green, S. Anderson and the
Earth to Ocean research group at Simon Fraser University contrib-
uted to the conceptual framework and statistical analyses of this
project. We thank Andrew Baird, Sean Connolly, and three anon-
ymous reviewers for comments on the manuscript. ESD and IMC
were supported by the Natural Sciences and Engineering Research
Council of Canada; ESD was also supported by graduate funding
from Simon Fraser University. The Mexican Council of Science and
Technology (CONACYT; 160230) supported LAF. TAO acknowl-
edges the Palumbi research group at Stanford University. TRM was
supported by the Wildlife Conservation Society through grants from
the John D. and Catherine T. MacArthur Foundation.
AUTHORSHIP
ESD, LAF, TRM and IMC designed the study, ESD and TAO
collected data, ESD analysed the data; all authors contributed to
writing and revising the manuscript.
REFERENCES
Aerts, R. (1999). Interspecific competition in natural plant communities:
mechanisms, trade-offs and plant–soil feedbacks. J. Exp. Bot., 50, 29–37.
Alvarez-Filip, L., Dulvy, N.K., Gill, J.A., Co
ˆte´, I.M. & Watkinson, A.R. (2009).
Flattening of Caribbean coral reefs: region-wide declines in architectural
complexity. Proc. R. Soc. Lond., Ser. B: Biol. Sci., 276, 3019–3025.
Alvarez-Filip, L., Dulvy, N.K., Co
ˆte´, I.M. & Watkinson, A.R. (2011). Coral
identity underpins reef complexity on Caribbean reefs. Ecol. App., 21, 2223–
2231.
Anthony, K. & Connolly, S. (2004). Environmental limits to growth:
physiological niche boundaries of corals along turbidity and light gradients.
Oecologia, 141, 373–384.
Aronson, R.B., Macintyre, I.G., Wapnick, C.M. & O’Neill, M.W. (2004). Phase
shifts, alternative states, and the unprecedented convergence of two reef
systems. Ecology, 85, 1876–1891.
Ayre, D.J. & Miller, K.J. (2004). Where do clonal coral larvae go? Adult
genotypic diversity conflicts with reproductive effort in the brooding coral,
Pocillopora damicornis.Mar. Ecol. Prog. Ser., 277, 95–105.
Ayre, D. & Resing, J. (1986). Sexual and asexual production of planulae larvae in
reef corals. Mar. Biol., 90, 187–190.
©2012 Blackwell Publishing Ltd/CNRS
8E. S. Darling et al. Letter
Baird, A.H. & Hughes, T.P. (2000). Competitive dominance by tabular corals: an
experimental analysis of recruitment and survival of understorey assemblages.
J. Exp. Mar. Biol. Ecol., 251, 117–132.
Baird, A.H., Guest, J.R. & Willis, B.L. (2009). Systematic and biogeographical
patterns in the reproductive biology of scleractinian corals. Annu. Rev. Ecol.
Evol. Syst., 40, 551–571.
Bak, R.P.M. & Engel, M.S. (1979). Distribution, abundance and survival of
juvernile hermatypical corals (Scleractinia) and the important of life history
strategies in the parent coral community. Mar. Biol., 54, 341–352.
Borcard, D., Gillet, F. & Legendre, P. (2011). Numerical Ecology with R. Springer,
New York, NY.
Cadotte, M.W., Carscadden, K. & Mirotchnick, N. (2011). Beyond species:
functional diversity and the maintenance of ecological processes and services.
J. Appl. Ecol., 48, 1079–1087.
Connell, J.H. (1978). Diversity in tropical rain forests and coral reefs. Science, 199,
1302–1310.
Cutler, D.R., Edwards, T.C.J., Beard, K.H., Cutler, A. & Hess, K.T. (2007).
Random forests for classification in ecology. Ecology, 88, 2783–2792.
Edinger, E.N. & Risk, M.J. (1995). Preferential survivorship of brooding corals
in a regional extinction. Paleobiology, 21, 200–219.
Edinger, E.N. & Risk, M.J. (2000). Reef classification by coral morphology
predicts coral reef conservation value. Biol. Conserv., 92, 1–13.
Fukami, H., Chen, C.A., Budd, A.F., Collins, A., Wallace, C., Chuang, Y.-Y. et al.
(2008). Mitochondrial and nuclear genes suggest that stony corals are
monophyletic but most families of stony corals are not (Order Scleractinia,
Class Anthozoa, Phylum Cnidaria). PLoS ONE, 3, e3222.
Gardner, T., Co
ˆte´, I.M., Gill, J., Grant, A. & Watkinson, A. (2003). Long-term
region-wide declines in Caribbean corals. Science, 301, 958–960.
Gitay, H. & Noble, I. (1997). What are functional types and how should we seek
them? In: Plant Functional Types: Their Relevance to Ecosystem Properties and Global
Change (edsSmith, T., Shugart, H. & Woodward, F.I.). Cambridge University
Press, Cambridge, UK, pp. 3–19.
Green, D.H., Edmunds, P.J. & Carpenter, R.C. (2008). Increasing relative
abundance of Porites astreoides on Caribbean reefs mediated by an overall
decline in coral cover. Mar. Ecol. Prog. Ser., 359, 1–10.
Greenslade, P.J.M. (1983). Adversity selection and the habitat templet. Am. Nat.,
122, 352–365.
Grime, J.P. (1977). Evidence for the existence of three primary strategies in
plants and its relevance to ecological and evolutionary theory. Am. Nat., 111,
1169–1194.
Grime, J.P. & Pierce, S. (2012). The Evolutionary Strategies that Shape Ecosystems, 7th
edn. Wiley-Blackwell, Oxford, UK.
Hughes, T.P. & Tanner, J.E. (2000). Recruitment failure, life histories, and long-
term decline of Caribbean corals. Ecology, 81, 2250–2263.
Hughes, T.P., Baird, A.H., Dinsdale, E.A., Moltschaniwskyj, N.A., Pratchett, M.
S., Tanner, J.E. et al. (2012). Assembly rules of reef corals are flexible along a
steep climatic gradient. Curr. Biol., 22, 736–741.
Jackson, J. & Hughes, T. (1985). Adaptive strategies of coral-reef invertebrates.
Am. Sci., 73, 265–274.
Karlson, R.H. & Hurd, L.E. (1993). Disturbance, coral reef communities, and
changing ecological paradigms. Coral Reefs, 12, 117–125.
Knowlton, N. (2001). The future of coral reefs. Proc. Natl. Acad. Sci. USA, 98,
5419–5425.
Laliberte´, E. & Legendre, P. (2010). A distance-based framework for measuring
functional diversity from multiple traits. Ecology, 91, 299–305.
Lang, J. & Chornesky, E. (1990). Competition between scleractinian reef
corals —a review of the mechanisms and effects. In: Ecosystems of the World,
Vol. 25. Coral Reefs (ed. Dubinsky, Z.). Elsevier, Amsterdam, Netherlands, pp.
133–207.
Liaw, A. & Wiener, M. (2002). Classification and regression by randomForest. R
News, 2/3, 18–22.
Loya, Y., Sakai, K., Yamazato, K. & Nakano, Y. (2001). Coral bleaching: the
winners and the losers. Ecol. Lett., 4, 122–131.
Madin, J.S. (2005). Mechanical limitations of reef corals during hydrodynamic
disturbances. Coral Reefs, 24, 630–635.
McClanahan, T.R., Ateweberhan, M., Graham, N.A.J., Wilson, S.K., Ruiz
Sebastia´n, C., Guillaume, M. et al. (2007). Western Indian Ocean coral
communities: bleaching responses and susceptibility to extinction. Mar. Ecol.
Prog. Ser., 337, 1–13.
McGill, B.J., Enquist, B., Weiher, E. & Westoby, M. (2006). Rebuilding
community ecology from functional traits. Trends Ecol. Evol., 21, 178–185.
Murdoch, T.J.T. (2007). A Functional Group Approach for Predicting the
Composition of Hard Coral Assemblages in Florida and Bermuda. PhD
Dissertation, Department of Marine Science. University of South Alabama,
Alabama.
Norstro
¨m, A., Nystro
¨m, M., Lokrantz, J. & Folke, C. (2009). Alternative states
on coral reefs: beyond coral–macroalgal phase shifts. Mar. Ecol. Prog. Ser., 376,
295–306.
Obura, D.O. (2001). Can differential bleaching and mortality among coral
species offer useful indicators for assessment and management of reefs under
stress? Bull. Mar. Sci., 69, 421–442.
Oksanen, J., Blanchet, F., Kindt, R., Legendre, P., O’Hara, R., Simpson, G. et al.
(2011). vegan: Community Ecology Package. R package version 1.17-6. Available
at: http://CRAN.R-project.org/package=vegan. Last accessed 10 January 2011.
Paddack, M.J., Reynolds, J.D., Aguilar, C., Appeldoorn, R.S., Beets, J., Burkett,
E.W. et al. (2009). Recent region-wide declines in Caribbean reef fish
abundance. Curr. Biol., 19, 590–595.
Pandolfi, J.M. & Jackson, J.B.C. (2006). Ecological persistence interrupted in
Caribbean coral reefs. Ecol. Lett., 9, 818–826.
Pandolfi, J.M., Bradbury, R.H., Sala, E., Hughes, T.P., Bjorndal, K.A., Cooke, R.
G. et al. (2003). Global trajectories of the long-term decline of coral reef
ecosystems. Science, 301, 955–958.
Pauly, D. (1995). Anecdotes and the shifting baseline syndrome of fisheries.
Trends Ecol. Evol., 10, 430.
Pianka, E. (1970). On r- and K- selection. Am. Nat., 104, 592–597.
R (2012) R Development Core Team (2010). R: A language and environment for
statistical computing. R Foundation for Statistical Computing, Vienna,
AustriaAvailable at: http://www.R-project.org.
Rachello-Dolmen, P. & Cleary, D. (2007). Relating coral species traits to
environmental conditions in the Jakarta Bay/Pulau Seribu reef system,
Indonesia. Estuar. Coast. Shelf Sci., 73, 816–826.
Soong, K. (1993). Colony size as a species character in massive reef corals. Coral
Reefs, 12, 77–83.
Stearns, S. (1977). The evolution of life history traits: a critique of the theory
and a review of the data. Annu. Rev. Ecol. Evol. Syst., 8, 145–171.
Todd, P.A. (2008). Morphological plasticity in scleractinian corals. Biol. Rev., 83,
315–337.
Westoby, M. (1998). A leaf-height-seed (LHS) plant ecology strategy scheme.
Plant Soil, 199, 213–227.
Winemiller, K.O. & Rose, K.A. (1992). Patterns of life-history diversification in
North American fishes: implications for population regulation. Can. J. Fish.
Aquat. Sci., 49, 2196–2218.
van Woesik, R., Franklin, E., O’Leary, J., McClanahan, T., Klaus, J. & Budd, A.
(2012). Hosts of the Plio-Pleistocene past reflect modern-day coral
vulnerability. Proc. R. Soc. Lond., Ser. B: Biol. Sci., 279, 2448–2456.
SUPPORTING INFORMATION
Additional Supporting Information may be downloaded via the online
version of this article at Wiley Online Library (www.ecologyletters.com).
As a service to our authors and readers, this journal provides sup-
porting information supplied by the authors. Such materials are
peer-reviewed and may be re-organised for online delivery, but are
not copy-edited or typeset. Technical support issues arising from
supporting information (other than missing files) should be
addressed to the authors.
Editor, David Bellwood
Manuscript received 12 March 2012
First decision made 18 July 2012
Manuscript accepted 31 July 2012
©2012 Blackwell Publishing Ltd/CNRS
Letter Life-history strategies of reef corals 9
