demography beyond the population functional traits as ... · demography beyond the population...
TRANSCRIPT
DEMOGRAPHY BEYOND THE POPULATION
Functional traits as predictors of vital rates across thelife cycle of tropical treesMarco D. Visser*,1,2,3, Marjolein Bruijning1,2, S. Joseph Wright3, Helene C. Muller-Landau3,Eelke Jongejans1,2, Liza S. Comita4 and Hans de Kroon1,2
1Department of Experimental Plant Ecology & Animal Ecology, Institute for Water and Wetland Research, RadboudUniversity, Heyendaalseweg 135, 6525 AJ Nijmegen, The Netherlands; 2Department of Physiology, Institute for Waterand Wetland Research, Radboud University, Heyendaalseweg 135, 6525 AJ, Nijmegen, The Netherlands;3Smithsonian Tropical Research Institute, Box 0843-03092, Balboa, Anc�on, Panama; and 4School of Forestry andEnvironmental Studies, Yale University, New Haven, CT 06511, USA
Summary
1. The ‘functional traits’ of species have been heralded as promising predictors for species’
demographic rates and life history. Multiple studies have linked plant species’ demographic
rates to commonly measured traits. However, predictive power is usually low – raising ques-
tions about the practical usefulness of traits – and analyses have been limited to size-indepen-
dent univariate approaches restricted to a particular life stage.
2. Here we directly evaluated the predictive power of multiple traits simultaneously across the
entire life cycle of 136 tropical tree species from central Panama. Using a model-averaging
approach, we related wood density, seed mass, leaf mass per area and adult stature (maximum
diameter) to onset of reproduction, seed production, seedling establishment, and growth and
survival at seedling, sapling and adult stages.
3. Three of the four traits analysed here (wood density, seed mass and adult stature) typically
explained 20–60% of interspecific variation at a given vital rate and life stage. There were
strong shifts in the importance of different traits throughout the life cycle of trees, with seed
mass and adult stature being most important early in life, and wood density becoming most
important after establishment. Every trait had opposing effects on different vital rates or at dif-
ferent life stages; for example, seed mass was associated with higher seedling establishment and
lower initial survival, wood density with higher survival and lower growth, and adult stature
with decreased juvenile but increased adult growth and survival.
4. Forest dynamics are driven by the combined effects of all demographic processes across the
full life cycle. Application of a multitrait and full-life cycle approach revealed the full role of
key traits, and illuminated how trait effects on demography change through the life cycle. The
effects of traits on one life stage or vital rate were sometimes offset by opposing effects at
another stage, revealing the danger of drawing broad conclusions about functional trait–demography relationships from analysis of a single life stage or vital rate. Robust ecological
and evolutionary conclusions about the roles of functional traits rely on an understanding of
the relationships of traits to vital rates across all life stages.
Key-words: adult stature, leaf mass per area, model averaging, seed size, tree growth, tree
mortality, wood density
Introduction
Functional biology has raised the possibility that morpho-
logical and physiological traits, henceforth functional
traits, might be strongly related to interspecific variation in
*Correspondence author. E-mail: [email protected]
This article forms part of the British Ecological Society
journals’ Demography Beyond the Population special feature
http://wileyonlinelibrary.com/BES_demography
© 2016 The Authors. Functional Ecology © 2016 British Ecological Society
Functional Ecology 2016, 30, 168–180 doi: 10.1111/1365-2435.12621
vital rates and serve as proxies for life history variation
(McIntyre et al. 1999; Westoby et al. 2002; Westoby &
Wright 2006). However, the predictive power of functional
traits is often very low, raising questions about how ‘func-
tional’ the selected traits really are (Paine et al. 2015). For
example, coefficients of determination (R2) average just
0�08 for predictions of growth and mortality rates of
tropical trees (Poorter et al. 2008; Wright et al. 2010; Iida
et al. 2014a,b). Such low predictive power strongly limits
the potential of traits to serve as proxies for life history
variation or inform global vegetation models (Cox et al.
2013; Friedlingstein et al. 2014).
Previous studies in tropical forest trees have provided an
incomplete picture of the role of traits in tree demography.
Limitations include restriction of analyses to particular life
stages, ignoring size dependence (Poorter et al. 2008; Wright
et al. 2010), and/or consideration of only one trait at a time
(Iida et al. 2014a,b). First, most previous studies consider a
single vital rate and/or life stage, even though traits will gen-
erally have different roles and different predictive power for
different demographic rates. For instance, seed mass is
strongly negatively correlated with seed production (Muller-
Landau et al. 2008), strongly positively correlated with
seedling establishment rates (Moles & Westoby 2006), and
has weaker relationships with growth and survival later in
life (Wright et al. 2010). Secondly, most previous studies
ignore the size dependency of demographic rates, focusing
instead on mean survival and growth over broad size classes
(Poorter et al. 2008; Wright et al. 2010; but see Iida et al.
2014b). This overlooks important interactions between
traits and size, as taller trees experience strikingly different
resource and competitive conditions (Poorter et al. 2005;
H�erault et al. 2011; Falster, FitzJohn & Westoby 2016).
Thirdly, many studies consider only a single trait, thus
inherently limiting the total explanatory power of traits
(Muller-Landau et al. 2008; Iida et al. 2014a,b). Many traits
will influence plant function simultaneously and each trait
can be involved in multiple trade-offs with contradictory
effects on vital rates (Marks & Lechowicz 2006). To improve
understanding of how traits influence plant demography, we
consider size dependence, evaluate trait–demography rela-
tionships across the entire life cycle, and consider multiple
traits in a model-averaging framework.
We evaluate to what degree four key traits can explain
variation in demography among tropical tree species. The
four key traits are adult stature, wood density, seed mass
and specific leaf area, which provide largely independent
information about plant strategies (Table 1; Westoby et al.
2002; Wright et al. 2007). In contrast to previous work,
our analyses are comprehensive including not only growth
and mortality of trees (Poorter et al. 2008; Wright et al.
2010; R€uger et al. 2012; Iida et al. 2014a), but also repro-
ductive schedules; seed production; and seedling establish-
ment, growth and mortality. We use a model-averaging
approach (Burnham & Anderson 2002; Grueber et al.
2011) to simultaneously weigh the effects of all four traits
on each vital rate for up to 136 tree and shrub species from
Barro Colorado Island, Panama. We aim to quantify (i)
which traits explain variation at different life stages; (ii)
the predictive power of all traits combined to explain vari-
ation in all components of the life cycle; and (iii) the rela-
tive effect sizes of each trait at each life stage while
mapping out contradictory effects across life stages.
Materials and methods
STUDY S ITE
Our data are from the moist tropical forest of the 50-ha Forest
Dynamics Plot (FDP) on Barro Colorado Island (BCI; 9°90 N,
79°510 W), Panama. Annual rainfall averages 2650 mm (since
1929), with a dry season between January and April, and tempera-
ture averages 27 °C (see Leigh 1999, for details).
V ITAL RATES
We used five data sets to quantify vital rates for the entire life
cycle.
1. Trees. In the FDP, all free-standing woody stems ≥1 cm diame-
ter breast height (dbh, measured at 1�3 m) were censused in
1981–1982, 1985 and every 5 years thereafter. In each census,
diameters of every stem are measured, and all new individuals
are tagged, mapped, and identified to species. These censuses
Table 1. Details on the functional trait data, including units, mean and standard deviation, range, common ecological associations and
references
Trait Mean (SD) Range Associations Sources
Wood density
(WD: g cm�3)
0�59 (0�14) 0�29, 0�91 Mechanical strength, vulnerability to
hydraulic failure, defence against decay,
growth–survival trade-off
Chave et al. (2009), Anten &
Schieving (2010) and Larjavaara &
Muller-Landau (2010)
Seed mass (SM: g) 0�06 (8�14)* 4�9e-05, 22�87 Seed production, seed dispersal,
seedling tolerance to stress,
seedling competitive ability
Westoby et al. (2002), Moles &
Westoby (2006) and
Muller-Landau (2010)
Leaf mass per area
(LMA: g m�2)
231�84 (314�89) 9�4, 1891�35 Leaf construction cost, photosynthetic
capacity, respiration rates,
leaf life span, leaf herbivory
Westoby et al. (2002),
Wright et al. (2004) and
Osnas et al. (2013)
Adult stature
(Dmax: mm)
398�97 (334�35) 12, 1775�67 Life history variation Kohyama (1993) and
Westoby et al. (2002)
*Geometric mean.
© 2016 The Authors. Functional Ecology © 2016 British Ecological Society, Functional Ecology, 30, 168–180
Traits affect whole life-cycle vital rates 169
provide information on growth and survival for individuals
≥1 cm dbh (hereafter ‘trees’). We analysed data from the 1990
to 2010 censuses, excluding earlier censuses because of small
but important differences in measurement methods (Condit
et al. 1999; R€uger et al. 2009).
2. Seed rain. Seed rain has been recorded in 200 0�5-m2 seed traps
since January 1987 (Wright et al. 2005). Traps are located in a
stratified random manner along trails within the FDP. All
reproductive parts (seeds, flowers, fruits and capsules) are iden-
tified to species and counted weekly (presence is recorded for
flowers). We used seed data from 1993 to 2012, as these years
correspond to records of newly recruiting seedlings (data set 3,
below).
3. Small seedlings. All seedlings and small saplings <1 cm dbh
(with no limits on height) were censused annually in 600 1-m2
seedling plots from 1994 through 2012. These plots are located
2 m from three sides of each of the 200 seed traps (Wright
et al. 2005).
4. Large seedlings. Free-standing woody plants ≥20 cm in height
and<1 cm dbh were censused in 20 000 1-m2 seedling plots each
year from 2001 through 2013, with the exceptions of 2005, 2007
and 2010 (Comita et al. 2007; Comita & Hubbell 2009). These
plots are located in a 5-m grid across the FDP. In each census,
the status (i.e. alive/dead) of previously tagged seedlings is
checked, all individuals are measured for height (except in 2002,
when only new recruits were measured), and new individuals are
tagged and identified to species. In analyses of growth and sur-
vival all census intervals which include missing years were
dropped (i.e. no intervals of>1 year are included for data set 4).
5. Reproductive status. We assessed the reproductive status of
13 358 individual trees to quantify size-dependent probabilities
of reproduction. For each species, a size-stratified sample of
trees was randomly selected and visited during species-specific
reproductive seasons. Reproductive status (fertile or sterile)
was evaluated from the ground using binoculars. For eight
dioecious species, we evaluated sex expression of all individuals
within the FDP. Data were collected between January 1995
and January 1996 for 31 species (Wright et al. 2005), between
2005 and 2007 for 51 wind dispersed species, and between April
2011 and September 2014 for 81 species.
TRA IT DATA
Trait data include seed mass (SM), leaf mass per area (LMA),
adult size (Dmax) and wood density (WD; Table 1; Wright et al.
2010). SM refers to endosperm and embryo dry mass determined
after dissecting diaspores to isolate the endosperm and embryo.
LMA was determined for shade leaves collected from the upper
canopy of the six smallest individuals of each species in the FDP.
We could not use sun-exposed leaves as a basis of comparison
because most FDP species are treelets that complete their entire
life cycle in the shaded understorey (King, Wright & Connell
2006). Dmax is the mean dbh of the six largest individuals in the
FDP (2005 census) and an additional 150 ha of mapped tree
plots located within 30 km and mostly within 10 km of BCI.
Dmax is well correlated with maximum tree height (r = 0�95 on a
log–log scale). Species-specific WD was estimated from tree cores
collected within 15 km of BCI, and was calculated as oven-dried
(60 °C) mass divided by fresh volume (technically wood specific
gravity). Further details can be found in Wright et al. (2010).
The four traits are largely independent of one another, with coef-
ficients of determination (R2 values) of 0�00068, 0�0056, 0�017,0�052, 0�12 and 0�13 for LMA-Dmax, SM-WD, WD-Dmax, SM-
Dmax, LMA-SM and LMA-WD relationships, respectively
(Wright et al. 2010).
We normalized trait values to enable model averaging, and
facilitate comparison of effect sizes among traits with very differ-
ent levels of interspecific variation (Grueber et al. 2011), using all
136 species evaluated here. Species-level trait values were normal-
ized, with SM and Dmax first log-transformed, by subtracting
mean trait values and then dividing by the standard deviation of
the trait values (Table 1).
STUDY SPEC IES
For each life stage and vital rate, we analysed all species with trait
data and sufficient demographic data to ensure reasonable preci-
sion of species-specific vital rate estimates. Table 2 gives exact
selection criteria and the number of species in each analysis.
Table S1 (Supporting information) gives the identities of the spe-
cies in each analysis. Figure S1 shows the distribution of trait val-
ues across all species within each analysis.
F ITT ING TRAIT -BASED MODELS FOR VITAL RATES
We evaluated relationships between size-dependent vital rates and
traits, including trait–size interactions, using generalized linear
mixed models (GLMMs), with species and individual included as
random effects. The most complex full model had the following
form:
Table 2. Species selection criteria and sample sizes for each analysis
Analysis Selection criteria Number of species Years
Reproduction (mm dbh) Reproductive status assessed for >20 trees. Species with too wide
confidence intervals were excluded after visual inspection of fit
60 (8891) 1995–2014
Seed production Species had at least 30 seeds captured in traps, and must have been
included in the reproductive analysis to estimate reproductive basal area
38 (NA) 1993–2010
Seedling establishment 30 or more seedling recruits observed between 1995 and 2011, with
>30 seeds observed for the fruiting years corresponding to
1995–2011 seedling recruitment (taking account of species-specific
germination delays)
68 (NA) 1994–2011
Seedling survival (mm height) >100 individuals in data set 80 (93 082) 2001–2013Seedling growth (mm height) >100 individuals in data set 80 (75 990) 2001–2013Tree survival (mm dbh) >100 individuals in data set 117 (267 469) 1990–2010Tree growth (mm2 basal area) >100 individuals in data set 117 (214 373) 1990–2010
Each vital rate under the column ‘Analysis’ is defined in the text, with corresponding units for size given in parenthesis when the analysis
was carried out on individual data. Under ‘Number of species’, the total number of species in each analysis is given with the total number
of individuals between parentheses. This value is ‘NA’ when the corresponding analysis concerns species-level data. The column ‘Years’
gives the time span of data used in each analysis. A total of 136 unique species were included across all analyses.
© 2016 The Authors. Functional Ecology © 2016 British Ecological Society, Functional Ecology, 30, 168–180
170 M. D. Visser et al.
y� bo þ b1sþX4
i¼1
biþ1Ti þ biþ5Tis� �þ esp þ eind þ eresidual
eqn1
where y is growth, (logit) survival or (logit) reproductive fraction; and s is
size in mm height, dbh or mm2 basal area for analyses of seedlings, repro-
ductive size or tree survival, and growth, respectively. Trait effects including
their interactions with size are given by the expression in parentheses, where
Ti represents trait i (corresponding to SM, WD, LMA or Dmax). The ran-
dom effects of species and individuals are denoted by esp and eind, respec-tively, and eresidual is the residual error. For each size-dependent vital rate,
we fit 82 possible models including eqn. 1 and all subsets involving different
combinations of the trait and trait by size effects (Table S2).
Two vital rates, seed production and seedling establishment, were mea-
sured and analysed at species level, and we related these to traits directly
using generalized linear models (GLMs) without size effects. Here too, we
evaluated a suite of models including all subsets of SM, LMA, Dmax and
WD (16 models per vital rate; Tables S3 and S4). Details of model fitting
for each vital rate follow.
Reproduction
The size-dependent probability of reproduction was evaluated
with a logistic GLMM (eqn. 1 with binomial error) using data set
5.
Seed production
Species-specific seed production (fseeds, seeds per year per m2 of
reproductive basal area) was quantified as the mean flux of
seeds arriving (seeds per year per m2 of trap area) divided by
mean reproductive basal area density (m2 of reproductive basal
area per m2 of plot area). We used seed trap and tree census
data from 1993 through 2012. Reproductive basal area was cal-
culated from the tree census data in combination with the fitted
logistic models for size-dependent probability of reproduction.
The logistic models predicted each individual’s reproductive
probability as a function of its size. We then weighted each
individual’s basal area by its reproductive probability to calcu-
late total reproductive basal area. Total reproductive basal area
was interpolated between FDP censuses to calculate annual val-
ues of fseeds, which were then averaged over years to obtain a
single mean value for each species. These simple estimates of
seed production were qualitatively similar to more sophisticated
estimates obtained using inverse modelling (Text S1, Fig. S2).
We chose to use the simple estimates because they were avail-
able for more species. Estimates of fseeds were then related to
traits using linear regression.
Seedling establishment
Species-specific mean seed to seedling establishment probabilities
were calculated as the mean flux of newly recruiting seedlings
per year per m2 in seedling plots in years 1995 to 2012 (data set
3) divided by the mean flux of seeds arriving per year per m2 in
seed traps for the corresponding fruiting years after accounting
for germination delays (Wright et al. 2005; data set 2). Seedling
establishment rates were related to traits using GLMs (i.e. logit
transform).
Growth
Growth was modelled as height growth for seedlings (mm per
year) and basal area growth for trees (mm2 per year) using LMMs
(eqn 1). We used basal area growth because general additive mod-
els (GAMs) showed that basal area growth was generally linearly
related to size (mm2 basal area). Growth rates were calculated as
the difference in sizes divided by the time in years between cen-
suses (data sets 1, 3 and 4). For data set 1, we excluded individuals
marked as ‘resprout’, ‘buttressed’, ‘leaning’ and ‘broken above
1�3 m’ in each census, as well as those with growth rates more
than four 4 standard deviations from the mean. These are likely
measurement errors (R€uger et al. 2011).
Survival
The size-dependent (mm height and mm dbh for seedlings and
trees, respectively) probability of survival was evaluated with a
logistic GLMM (eqn 1 with binomial error) using data sets 1, 3
and 4.
We used model averaging to calculate average parameters
(Burnham & Anderson 2002). All models were assigned a weight
based on their AIC score and fitted parameters were averaged
over the full set of models using these weights to obtain a final
average model. The final average model provides a basis to com-
pare effect sizes. Model averaging is superior to selecting the best
model because models with similar fits are not ignored (Burnham
& Anderson 2002; Whittingham et al. 2006; Bolker et al. 2009).
For this reason, model averaging provides a more robust basis for
inference and prediction, reducing bias in estimation of effect sizes,
especially in cases where multiple variables influence the response
variable (Grueber et al. 2011). This contrasts with stepwise multi-
ple regression, which is seen as poorly suited to disentangle contri-
butions of multiple traits to vital rates (Wittingham et al. 2006).
We averaged parameters over all models having AIC weights >0using the ‘zero method’ in which parameters are assigned the value
zero where absent from models. This is a conservative approach
to model averaging (i.e. leading to lower effect sizes) and is recom-
mended when comparing effect sizes among variables (Burnham &
Anderson 2002; Grueber et al. 2011). Confidence intervals for
each weighted parameter were estimated following Buckland,
Burnham & Augustin (1997).
We evaluated model fits to the full data sets, including varia-
tion among individuals, using marginal and conditional R2 val-
ues developed for mixed-effects models (Nakagawa & Schielzeth
2013). These R2 values provide information on how well the
trait-based hierarchical models (the GLMMs) explain individual-
level variation in vital rates over all species in the community.
To evaluate linearity, we plotted model residuals against size for
each model (Fig. S3). Residuals deviated from linearity for seed-
lings taller than 2�5 m and for trees with dbh >50 cm in the
growth and survival analyses. To ensure linearity, we therefore
excluded trees with dbh >50 cm and seedlings taller than 2�5 m
(corresponding to 0�41% and 0�47% of the data, respectively).
Nonlinearity was not detected in our reproduction analysis. All
analyses were performed in R 3.1.1 (R-core 2014), making use
of the LME4 package for mixed-effects models (Bates et al.
2015). An example R-code is provided for model averaging
(Appendix S2).
THE POWER OF TRA ITS TO EXPLA IN INTERSPEC IF IC
VAR IAT ION
We performed a second set of analyses to estimate the contribu-
tions of traits to explaining interspecific variation in demographic
rates at particular sizes, and thereby to enable more direct com-
parisons of our results with earlier studies based on species-level
estimates of vital rates (Poorter et al. 2008; Wright et al. 2010;
Iida et al. 2012). We first calculated trait-based predictions for
each species and each vital rate based on the fitted average mod-
els. We then compared these predictions with observed mean
© 2016 The Authors. Functional Ecology © 2016 British Ecological Society, Functional Ecology, 30, 168–180
Traits affect whole life-cycle vital rates 171
vital rates and calculated associated R2 values. For seed produc-
tion and seedling establishment, these comparisons used observed
species-specific mean rates. For size-dependent vital rates
(growth, survival, reproduction), we estimated species-specific
moving averages using generalized additive models (GAMs;
example code in Appendix S2), which make no prior assumptions
on the functional shape of the relationship (Figs S4–S8). We then
estimated size-specific R2 values in three steps: (i) species-specific
mean rates were given from GAM predictions for each vital rate
at size points ranging from 0 to 1�5 m height for seedlings and
from 1 to 25 cm dbh for trees (at these size ranges, analyses
always included 15 or more species); (ii) demographic rates for
these sizes were predicted using only fixed effects from each aver-
aged trait models; and (iii) we calculated R2 values for correla-
tions between trait-based and GAM estimated mean rates at each
size. To evaluate the predictive power of individual single traits,
we repeated the calculation of R2 values for single-trait average
models for every trait, where single-trait models were based on
averaging over models including only size, the single focal trait
and/or the trait–size interaction as predictors. Single-trait models
represent the best-case scenario, in terms of R2, when using a sin-
gle trait.
SEPARATE ANALYSES AT EVERY SIZE
Out of an abundance of caution, we performed a final analysis to
guard against the possibility that underlying assumptions of lin-
ear relationships with size or the random effects structure in the
mixed-effects model may unduly influence results. The (G)LMMs
allow only linear relationships between vital rates and size, and
might misstate the influence of predictors if there are underlying
nonlinearities with size. Additionally, to improve computational
feasibility, we included only random species intercepts, but
ignored random slopes with size, which potentially may impact
effect sizes (Schielzeth & Forstmeier 2009). To address these con-
cerns, we evaluated relationships between traits and species-speci-
fic vital rates separately for each size. We fit GLMs (sets of 16
models, as shown in Tables S3 and S4) to species mean rates (es-
timated from the GAMs) for seedlings between 0 and 1�5 m tall
and trees between 1 and 25 cm dbh. We then compared effect
sizes and R2 values with the mixed-effect models (eqn 1). This
analysis allows for varying intercepts, slopes and functional
shapes between size and vital rates for each species. However, the
analysis is also far less parsimonious and weights all species
equally regardless of sample size and hence is likely to overesti-
mate the strength of trait–vital rate relationships. When results
are qualitatively similar this indicates no major problems with
our assumptions. In this case, our initial analyses (eqn 1) are less
biased for inference and yield the greatest predictive accuracy
(Gelman 2006).
Results
Our analyses included 136 different species, with 38 to
117 species for each vital rate (Table 2). The full range
of trait values observed among BCI trees was well repre-
sented for each vital rate (Fig. S1). The species included
a wide range of growth forms (shrubs to understorey
and canopy trees), seed dispersal mechanisms (ballistic,
wind, mammals and/or birds), and relative abundances
[from 13�5% (Faramea occidentalis) to 0�29% (Hampea
appendiculata) of all stems in the FDP]. WD, SM, LMA
and Dmax varied by 0�5, 5�7, 2�3 and 2�2 orders of mag-
nitude, respectively (Table 1).
SEED PRODUCT ION
The average model for seed production was based on nine
models with nonzero weights (weights above 0�001;Table S3), included all four traits and explained 65% of
interspecific variation (Table 3). Seed production was neg-
atively related to SM (slope �1�13; Table 3, Fig. 1a) with
all other traits having approximately 7–50 times smaller
effect sizes (Table 3, Fig. 1a).
SEEDL ING ESTABL ISHMENT
The average model for seedling establishment was based
on four models with nonzero weights (weights above
0�001; Table S4), included all four traits and explained
66% of interspecific variation (Table 3). Seedling establish-
ment increased with seed mass (slope 1�32; Table 3,
Fig. 1b) and decreased with Dmax (slope �1�06; Table 3,
Fig. 1b). LMA and WD hardly influenced seedling estab-
lishment rates (Table 3, Fig. 1b).
SEEDL ING GROWTH RATES
The average model for seedling growth rates was based on
nine models with nonzero weights (weights above 0�001;Table S5). It included all four size–trait interactions and
explained 10% of the individual-level variation, with the
fixed effects of size, traits and their interactions explaining
only 1% of the individual growth variation (Table 3). Seed-
ling growth rates decreased with SM and WD, with smaller
effects for larger seedlings (Table 3, Fig. 2a,b). LMA and
Dmax hardly influenced seedling growth rates (Fig. 2c,d).
SEEDL ING SURV IVAL RATES
The full model, including all size–trait interactions, con-
tained 100% of the weight for seedling survival (Table S6).
This model explained 25% of the individual-level variation
in seedling survival, with the fixed effects of size, traits and
their interactions explaining 16% of the individual varia-
tion (Table 3). Seedling survival increased with WD and
decreased with Dmax, and these effects diminished with
seedling size (Fig. 2f,h). The direction of relationships
between seedling survival and SM and LMA changed with
seedling size (Fig. 2e,g). For small seedlings, survival was
greater for species with smaller seeds and larger LMA. For
larger seedlings, survival was greater for species with larger
seeds and smaller LMA.
TREE GROWTH RATES
The full model, including all size–trait interactions, also
contained 100% of the weight for tree growth rates
(Table S7). This model explained 66% of the individual-
level variation in tree growth, with the fixed effects of size,
traits and their interactions explaining 26% of the individ-
ual variation (Table 3). Tree growth rates decreased with
© 2016 The Authors. Functional Ecology © 2016 British Ecological Society, Functional Ecology, 30, 168–180
172 M. D. Visser et al.
Table
3.Coeffi
cients
(andstandard
errors)from
thefullaveraged
model
foreach
oftheevaluatedvitalrates(columns)
Seedproduction
Seedlingestablishment
Seedlinggrowth
Treegrowth
Seedlingsurvival
Treesurvival
Reproduction
Intercept
�2�95
(0�07
81)
�3�21
(0�08
16)
42(2�07
)108(8�7)
1�31
(0�05
82)
2�19
(0�06
63)
�3�71
(0�09
39)
SM
�1�13
(0�06
18)
1�32
(0�08
66)
�11�1
(2�31
)�2
2�5
(9�5)
�0�22
(0�06
5)
0�38
7(0�07
22)
�0�22
6(0�07
24)
WD
0�01
52(0�03
12)
0�01
37(0�03
49)
�13(2�2)
�70�4
(9�79
)0�45
1(0�06
07)
0�53
6(0�07
48)
0�12
3(0�06
93)
LMA
�0�16
6(0�08
44)
�0�00
413(0�03
5)
�0�06
05(2�32
)�2
�19(9�41
)0�20
4(0�06
65)
0�24
9(0�07
24)
�0�28
1(0�08
75)
Dmax
�0�02
47(0�04
48)
�1�06
(0�08
26)
0�43
4(1�96
)36�6
(9�97
)�0
�348(0�05
78)
�0�29
7(0�07
58)
�0�38
6(0�11
)
Size
�0�01
24(0�00
0477)
0�00
934(7�63
e-05)
0�00
116(2�34
e-05)
�0�00
282(0�00
0116)
0�01
74(0�00
0224)
Size:SM
0�00
479(0�00
0332)
�0�00
031(3�14
e-05)
0�00
0458(2�1e
-05)
�0�00
0976(9�73
e-05)
0�00
0675(5�99
e-05)
Size:WD
0�00
496(0�00
0596)
0�00
0858(4�83
e-05)
8�65
e-06(1�52
e-05)
�0�00
121(0�00
0107)
�0�00
0109(7�53
e-05)
Size:LMA
0�00
112(0�00
0772)
�0�00
0946(5�03
e-05)
�0�00
0529(3�04
e-05)
�0�00
0642(0�00
0104)
�0�00
0143(9�62
e-05)
Size:Dmax
�1�98
e-05(2�29
e-05)
0�00
255(9�66
e-05)
0�00
0272(2�25
e-05)
0�00
581(0�00
0161)
�0�00
877(0�00
0197)
R2 size
0�00
15
0�16
76
0�03
21
1e-04
0�17
75
R2 fixed
0�00
84
0�25
91
0�15
81
0�06
54
0�24
69
R2 species
0�07
76
0�40
15
0�25
07
0�18
70�50
63
R2 individual
0�03
43
0�51
72
0�15
81
0�06
54
0�34
51
R2 full
0�65
0�66
0�10
35
0�65
95
0�25
07
0�18
70�60
45
Bold
face
indicatesconfidence
intervalsnotincludingzero
(a<0�05
).Themeasure
ofsize
washeight(m
m)forseedlings,dbh(m
m)forsurvivalandreproduction,andbasalareaforgrowth
(mm
2).
Complete
listsofevaluatedmodelswithAIC
values
andAIC
weights
are
given
inTablesS4–S
9.Withtheexceptionofseed
productionandseedlingestablishment,theR2values
inthetable
are
specificto
mixed-effects
models(N
akagawa&
Schielzeth2013),andreflectthefitofthemodelincludingfixed
effects
ofsize
only
(‘R
2 size’),fixed
effects
ofsize
andtraits(‘R
2 fixed’),fixed
andspeciesran-
dom
effects
(‘R
2 species’),fixed
andindividualrandom
effects
(‘R
2 individual’),andthefullmixed
model
(‘R
2 full’).Traitvariableswerenorm
alizedpriorto
modelfits.
© 2016 The Authors. Functional Ecology © 2016 British Ecological Society, Functional Ecology, 30, 168–180
Traits affect whole life-cycle vital rates 173
WD and increased with Dmax (Table 3, Fig. 2j,l). The effect of
WD decreased with tree size (Fig. 2j), and the effect of Dmax
increased with tree size (Table 3, Fig. 2l). SM and LMA had
much smaller effects on tree growth rates (Fig. 2I,k).
TREE SURV IVAL RATES
The full model, including all size–trait interactions, con-
tained 99�7% of the weight for tree survival rates
(Table S8). The average model explained 19% of the indi-
vidual-level variation in seedling survival, with the fixed
effects of size, traits and their interactions explaining 7%
of the individual variation (Table 3). Tree survival was
strongly influenced by Dmax, and was generally larger for
species with larger Dmax (Fig. 2p). Tree survival was also
greater for species with larger SM, WD and LMA over
nearly the full range of tree sizes, with the exception of the
very largest individuals (Fig. 2m–o).
REPRODUCT ION
Nine models with nonzero weights contributed to the aver-
age model for reproduction (Table S9), which included all
size–trait interactions and explained 60% of the individ-
ual-level variation, with the fixed effects of size, traits and
their interactions explaining 25% of the variation (Table 3
and Table S9). Dmax had the largest effect on reproductive
status, with larger-statured species becoming reproductive
at larger sizes than smaller-statured species (Fig. 2t). The
threshold size at which 50% of individuals are reproduc-
tive is well predicted by the following simple equation:
R50 = ½Dmax (R2 = 0�81; Fig. S9).
THE POWER OF TRA ITS TO EXPLA IN INTERSPEC IF IC
VAR IAT ION
Figure 3 summarizes the proportion of interspecific varia-
tion in vital rates explained by Dmax, LMA, SM and WD
throughout tree life cycles on BCI. These proportions are
consistently higher than the proportions explained by traits
and size in the GLMMs (Table 3) because the latter includes
additional variation among individuals. The R2 values for
each trait separately are presented in Tables S10–12.
SEPARATE ANALYSES AT EVERY SIZE
Traits had qualitatively similar influences on vital rates
(Fig. S10) and explained similar proportions of interspecific
−2 −1 0 1 20·005
0·010
0·020
0·050
0·100
0·200
0·500
1·000Seed production
See
ds p
er u
nit b
asal
are
a (N
/mm
−2/y
r)
SMWDLMADmax
−2 −1 0 1 20·001
0·002
0·005
0·010
0·020
0·050
0·100
0·200
0·500 Seedling establishment
See
dlin
g re
crui
ts p
er s
eed
Standardized trait value
(a)
(b)
Fig. 1. Fitted effects of deviations in seed
mass (SM), wood density (WD), leaf mass
per area (LMA) and maximum stature
(Dmax) from their mean values on seed pro-
duction (a), and the rate of seedling estab-
lishment (b) when other traits are held at
their mean values. Predictions are plotted
against standardized trait values (standard
deviations from the mean). Observed trait
value means and standard deviations are
given in Table 1. Panels a and b show that
traits can have opposing effects not only
between life stages (SM in a and b) but also
between different traits within a single life
stage (Dmax and SM in Panel b).
© 2016 The Authors. Functional Ecology © 2016 British Ecological Society, Functional Ecology, 30, 168–180
174 M. D. Visser et al.
variation (Fig. S11) in analyses that fit separate models for
every size and in our main analyses.
Discussion
We systematically quantified trait–demography relation-
ships across the entire life cycle of multiple co-occurring
species of tropical trees for the first time while incorporat-
ing individual-level, size-dependent variation in growth,
survival and reproduction (Table 3, Fig. 2). Full models,
including random effects for individuals and species,
explained 10–25% of the overall variation for seedling
growth, seedling survival and tree survival (Table 3). Fac-
tors missing from our models clearly affect these three vital
rates. Likely candidates include abiotic and biotic environ-
mental variation associated with soils, local competitive
effects and plant pests. The full models performed much
better for tree growth rates and reproductive status,
explaining 66% and 60% of overall variation, respectively.
The fixed effects of size, wood density (WD), seed mass
(SM), leaf mass per area (LMA), adult size (Dmax) and
interactions between size and traits explained 1–26% of
the variation observed over all individuals of all species for
the five size-dependent vital rates (Table 3). Our trait-
based average models explained between 4% and 65% of
interspecific variation in mean size-specific demographic
rates, depending on the size and demographic rate, with
more variation explained for small than for large size
classes (Fig. 3). In comparison with previous studies of
species-level trait–demography relationships among tropi-
cal trees (Poorter et al. 2008; Wright et al. 2010; Iida et al.
2012, 2014a), our analyses provide clear improvements in
predictive power and new insights into how effects vary
with size.
(a)
20
30
40
50(b) (c) (d)
(e)
0 300 600 900 1200 15000·700·750·800·850·900·95
(f)
0 300 600 900 1200 1500
(g)
0 300 600 900 1200 1500
(h)
0 300 600 900 1200 1500
(i)
0500
1000150020002500 (j) (k) (l)
(m)
0·85
0·90
0·95
1·00(n) (o) (p)
(q)
0 100 200 300 400 5000·00·20·40·60·81·0
(r)
0 100 200 300 400 500
(s)
0 100 200 300 400 500
(t)
0 100 200 300 400 500
Rep
rodu
ctio
nTr
ee s
urvi
val
Tree
gro
wth
Sdl
. sur
viva
lS
dl. g
row
thfra
ctio
n re
prod
uctiv
ean
nual
rate
mm
^2/y
ear
annu
al ra
tem
m/y
ear
Seedmass Wood density Leaf mass per area Max. d.b.h.
Size (mm)
Fig. 2. Fitted effects of each trait (columns) on size-dependent vital rates (rows) in the averaged models. The black lines present the vital
rate–size relationships with all traits set to their mean values. The red plus signs and blue minus lines present the same relationships with
one trait set to its mean plus or minus one standard deviation, respectively, and the three remaining traits set to their mean values. The
trait whose value varies among the blue, black and red lines is named at the top of each column. Observed trait value means and standard
deviations are given in Table 1. Grey text at the right outermost column gives the corresponding measure of size (in mm) for each row of
panels. The figure shows that the effect sizes corresponding to different traits differ greatly throughout the life cycle of trees among species,
as shown by the + and � lines from the top to the bottom of the graph.
© 2016 The Authors. Functional Ecology © 2016 British Ecological Society, Functional Ecology, 30, 168–180
Traits affect whole life-cycle vital rates 175
WOOD DENS ITY
Higher WD is associated with higher resistance to hydrau-
lic failure and to decay and with higher structural strength
for a given diameter, but at the cost of slower diameter
growth rates (reviewed by Chave et al. 2009). Previous
studies concur that wood density (WD) is the single trait
best able to predict growth and survival among tropical
tree species, with coefficients of determination (R2) averag-
ing 0�093 (�0�077 SD) and 0�076 (�0�079 SD) for relation-
ships with growth and survival, respectively (using the
maximum R2 values reported in Poorter et al. 2008;
Wright et al. 2010; Iida et al. 2014a,b). Our size-specific
coefficients of determination, which were always greater
for wood density than for other traits for analyses of
growth and survival (Table 3), are consistent with this con-
clusion. Wood density had progressively less predictive
power for the survival and growth of larger trees, which is
consistent with previous comparisons of broad sizes classes
(Poorter et al. 2008; Wright et al. 2010). The benefits and
costs associated with variation in WD affect basal area
growth and survival directly, but have negligible effects on
reproduction, seed production and seedling establishment
(Fig. 3).
In interpreting the relationship of WD to growth and sur-
vival at constant diameter or constant height in this and
other studies, it is important to keep in mind that these rela-
tionships are dependent upon the size (and growth) cur-
rency used as the basis for comparison. The same ‘size’ in
diameter or height is associated with larger biomass in
higher WD species, and the same biomass growth translates
to less diameter, basal area, and height growth in higher
WD species (Larjavaara & Muller-Landau 2010). Relation-
ships of growth with WD may weaken or even disappear
when growth is expressed on the basis of biomass instead of
diameter (R€uger et al. 2012). Thus, the strong relationships
of growth with WD might in part be seen as an artefact of
our choice of currency, much as mass-normalized leaf traits
show stronger interrelationships than area-normalized leaf
traits (Osnas et al. 2013). Similarly, the use of height and
diameter as measures of size might introduce bias towards
positive relationships between survival and WD, because
the same ‘size’ in diameter or height is associated with lar-
ger biomass in higher WD species, and survival increases
with biomass on BCI (Muller-Landau et al. 2006). Future
analyses should evaluate how much variation in growth
and survival is explained by WD when these currency
effects are eliminated.
(a)
(b)
(c)
Fig. 3. The proportion of interspecific variation in various demographic rates explained by the four functional traits throughout the life
cycle, as measured by R2 values. The rows show results for (top to bottom) size-dependent growth, size-dependent survival, and vital rates
associated with reproduction. The R2 value at the upper edge of the stacked colours represents the proportion of the total variation among
species explained by the fixed effect terms in the full averaged model (i.e. including traits Dmax, LMA, SM and WD), where the observed
values for each species are based on species-specific GAMs. The relative importance of different traits is indicated by the relative height of
each colour band as a proportion of the total, with height scaled to the R2 values for averaged models including only one trait (i.e. includ-
ing only Dmax, LMA, SM or WD; Tables S10–12). The number of species included in the analyses varies with size; for reference, the num-
bers of species included at various sizes are shown in solid circles above each graph. Variation in vital rates explained among species
shows that seed mass is initially influential but diminishes in importance with wood density becoming more important.
© 2016 The Authors. Functional Ecology © 2016 British Ecological Society, Functional Ecology, 30, 168–180
176 M. D. Visser et al.
SEED MASS
Previous studies and our analyses (Fig. 3) concur that seed
mass (SM) explains minimal variation in growth and sur-
vival among tropical trees (R2 averages 0�002 � 0�0022 SD
and 0�101 � 0�041 SD, respectively; Fig. 3; Poorter et al.
2008; Wright et al. 2010). This is unsurprising because seed
reserves and direct effects of seed size are exhausted well
before the large minimum sizes (≥1 cm dbh) used to deli-
mit trees. Those relationships with SM that remain at these
large sizes reflect indirect effects, whose unknown causa-
tion must involve unrecognized correlations among traits
and life histories.
SM has much stronger effects at the earliest stages of
regeneration. SM was strongly negatively related to seed
production and strongly positively related to seedling
establishment on BCI (Fig. 1). This is consistent with the
well documented trade-off between seed quantity and per-
seed investment (Henery & Westoby 2001; Muller-Landau
2010). SM also influenced seedling performance. Small-
seeded species consistently grew faster than large-seeded
species, although this effect diminished with seedling size
(Fig. 2a). Small-seeded species also had lower survival
rates than large-seeded species, but only at larger seedling
sizes. Among the smallest seedlings, small-seeded species
actually had higher survival rates than large-seeded species
(Fig. 2e). Environmental variation associated with germi-
nation sites likely confounds all of these relationships
(Lichstein et al. 2010; Muller-Landau 2010). Small-seeded
species tend to establish in the least stressful locations, and
those locations improve subsequent performance. Large-
seeded species are able to establish in more competitive
environments, which limit subsequent performance. On a
population level, this causes smaller-seeded species to have
lower establishment rates, because appropriate resource-
rich sites are infrequent, and larger growth and survival
rates than larger-seeded species that establish widely in less
favourable environments. Clearly, observed vital rates are
influenced not only by species traits, but by the habitats in
which individuals are found (i.e. environmental filtering;
Lasky et al. 2013). We discuss this issue in more detail
below.
LEAF MASS PER AREA
Previous studies and our results concur that LMA explains
minimal variation in growth and survival among tropical
trees (R2 averages 0�08 � 0�025 SD and 0�11 � 0�071 SD,
respectively; Fig. 3, this study; Poorter et al. 2008; Wright
et al. 2010; Iida et al. 2014b). LMA is thought to be a
minor factor affecting carbon gain in larger trees where
crown architecture determines light interception (Sterck &
Bongers 2001). Surprisingly, LMA was also vanishingly
unimportant for seedling establishment, growth and sur-
vival (Figs 1b, 2c,g and 3). LMA had non-negligible corre-
lations only with reproductive status, with species with
lower LMA tending to reproduce at smaller sizes (Figs 2
and 3). This relationship between LMA and reproductive
status provides another example, as with seed mass, of an
indirect effect, whose unknown causation must involve
unrecognized correlations among traits and life histories.
There are at least two possible reasons for the negligible
relationships of LMA with seedling growth and survival
(Figs 2c,g and 3). First, our LMA values are for saplings
(≥1 cm dbh), and seedling LMA values may differ. Previ-
ous studies show that leaf traits at juvenile and adult
stages are generally strongly correlated (Iida et al. 2014a);
however, as LMA measurements for recently established
seedlings remain rare, we cannot discount ontogenetic
changes in LMA between seedlings and saplings (Spasoje-
vic et al. 2014). LMA values determined for seedlings
might yet yield stronger relationships between LMA and
seedling performance. A second possible cause of weak
relationships between LMA and growth and survival
applies to both seedlings and trees. Costs and benefits
associated with LMA variation might balance, yielding
similar growth and survival rates on a population level.
Low LMA species tend to have low construction costs and
large leaf turnover rates, whereas high LMA species tend
to have larger construction costs and lower leaf turnover
rates (Wright et al. 2004). This may result in similar net
carbon gain over time, minimizing potential relationships
between LMA and demographic rates.
ADULT STATURE
Previous studies and our results concur that larger-statured
species have larger growth and survival rates among tropi-
cal trees (Fig. 2l,p) R2 values average 0�11 � 0�039 SD
and 0�2 � 0�066 SD, respectively (using the maximum of
reported R2 values; Fig. 3; Poorter et al. 2008; Wright
et al. 2010; Iida et al. 2014a,b). This led to the conclusion
that smaller-statured species have lower survival and
growth rates as they may have less access to light (Poorter
et al. 2008; Wright et al. 2010; Iida et al. 2014a). We see
opposing effects on seedlings, however, with taller species
at an inherent disadvantage in early life (King, Wright &
Connell 2006). Seedling establishment and survival
decreased with increasing adult stature (Figs 1b and 2h).
These opposing relationships between maximum size and
performance between life stages are a condition for coexis-
tence in the ‘forest architecture hypothesis’. Large-statured
species will out-compete smaller species, if not handi-
capped during establishment (Kohyama 1993). The handi-
cap observed among large-statured species during
establishment might be related to a trade-off between allo-
cation to traits and architectures that enable survival in
the forest understorey vs. rapid vertical growth towards
the canopy (for a more detailed discussion see Kohyama
et al. 2003; Poorter et al. 2005).
Dmax was the single most influential trait explaining inter-
specific variation in reproductive size thresholds (cf.
Fig. 2q–s vs. 2t). Species that grow larger only begin to
reproduce at larger sizes. The threshold size for reproduction
© 2016 The Authors. Functional Ecology © 2016 British Ecological Society, Functional Ecology, 30, 168–180
Traits affect whole life-cycle vital rates 177
also increased for taller species in a Malaysian forest
(Thomas 1996; Davies & Ashton 1999). For BCI trees, the
simple equation R50 = ½Dmax explains 81% of interspecific
variation in reproductive size thresholds (Fig. S9). That is,
at the time individuals attain half of their species’ maximum
observed diameter, they have a 50% probability of being
reproductive. This is consistent with expectations from game
theory models (reviewed by Falster & Westoby 2003), which
predict that to maximize reproductive output individuals
should first invest heavily in growth, and then only after
reaching an optimal size start to invest in reproduction.
Extending this, we would expect large-statured species to
become reproductive only when they attain a position in the
forest canopy (Thomas 1996; Zuidema & Boot 2002), while
smaller understorey species will likely reproduce when they
obtain optimal crown depth (Kohyama et al. 2003) or foli-
age cover. However, whether this optimal size on average
corresponds to the ½Dmax size threshold reported here
remains to be tested.
TRA ITS AS QUANT ITAT IVE PRED ICTORS OF TREE
DEMOGRAPHY
Westoby (1998) suggested that specific leaf area (the
inverse of LMA), seed mass and adult stature are three
readily measurable traits that represent important dimen-
sions of variation in plant ecology. Our results suggest
that, at least for tropical forests, a more promising combi-
nation would be adult stature, seed mass and wood den-
sity. In our study system, adult stature, seed mass and
wood density (but not LMA) each explained substantial
interspecific variation in particular vital rates or particular
life stages (Fig. 3). Nevertheless, a large proportion of
interspecific and individual variation remained unexplained
(Table 3 and Fig. 3). Why is the explained variation not
higher, and what are the implications for the functional
trait research agenda?
Variance partitioning suggests that a considerable fraction
of the unexplained interindividual variation is due to species
effects not captured by the functional traits included in this
study. Variance partitioning quantifies the unexplained vari-
ation and, thus, the potential for additional factors to
explain variation at each grouping level. Mixed-effects mod-
els allow variance partitioning (Bolker et al. 2009), and we
calculated conditional R2 values to quantify variation
explained at the individual and species levels (Nakagawa &
Schielzeth 2013). The addition of species random effects
increased conditional R2 values on average by a factor of 3�2(compare R2
fixed with R2species in Table 3). This demonstrates
that substantial unexplained interspecific variation remains,
variation that might potentially be explained by additional
traits. However, species–environment associations might
also contribute to unexplained interspecific variation (Mes-
sier, McGill & Lechowicz 2010; Lasky et al. 2013).
Variation in plant performance among individuals
depends strongly on local environment as well as on spe-
cies traits and their interaction (Uriarte et al. 2016). Local
environmental variation includes both abiotic factors such
as soil nutrients (Condit et al. 2013) and water availability
(Comita & Engelbrecht 2009), and biotic factors such as
local competitive neighbourhoods (Uriarte et al. 2004).
Such environmental variation is not explicitly included in
our models, and thus contributes to variation among indi-
viduals captured here by individual-level random effects.
Environmental variation can also confound interspecific
comparisons of vital rates (Lichstein et al. 2010; McMa-
hon, Metcalf & Woodall 2011; Baraloto et al. 2012). For
example, in closed canopy tropical forests, small-seeded
species only establish successfully in relatively high light
microsites and higher light levels then contribute to higher
initial growth and survival rates (Fig. 2a,e). Inclusion of
more information on key environmental covariates for
each individual would make it possible to control for any
systematic differences in environments among species, and
thereby better estimate the true effects of traits on perfor-
mance (Paine et al. 2011; Lasky et al. 2014). Other sources
of variation will remain stochastic and unpredictable.
These include negative height growth due to stem breakage
for seedlings (e.g. from falling branches), which happens
regularly on BCI (Paciorek et al. 2000).
Our full-life cycle approach shows that individual traits
can have opposing effects on different vital rates. This
raises the possibility that effects at one life stage or vital
rate may be offset by opposite effects at another life stage
or vital rate. Robust ecological and evolutionary conclu-
sions, based on findings at single life stages or vital rates,
will therefore depend on how effect sizes translate to net
effects over the full life cycle. Trait-based models that map
full-life cycle demographic patterns across trait axes
(Figs 1 and 2) may help resolve full-life cycle effects. For
instance, when an empirical study finds an effect of
increased seed production – which comes at the cost of
reduced seed size – trait-based models can be used to cal-
culate whether a net positive effect on seedling recruitment
can be expected. A trait-based framework may even be
used in a population modelling context (sensu Visser et al.
2011; Merow et al. 2014), to calculate expected net effects
when information on the whole life cycle is lacking. In this
context, a full-life cycle trait-based approach may add
value to ecological research by enabling robust assessments
of relationships between traits and population fitness.
The holy grail of the functional traits research agenda is
the identification of easily measured traits that are good
predictors of life history and demographic performance,
and the parameterization of associated models for inferring
life history and demography from these traits. With increas-
ing amounts of vital rate variation across species explained,
researchers start daring to ask the exciting question of
whether trait-based vital rate models can be used to inter-
polate vital rates for species for which they have trait infor-
mation but lack demographic data. Such interpolations
would increase the number of species that can be included
in Earth system models and community-wide studies. How
certain do we have to be about these interpolations for the
© 2016 The Authors. Functional Ecology © 2016 British Ecological Society, Functional Ecology, 30, 168–180
178 M. D. Visser et al.
resulting multispecies analyses to be trustworthy? Our
analyses still show a large proportion of unexplained inter-
specific variation, implying that trait-based models should
be used with caution. Nevertheless, we have shown that a
single individual measurement (size) and three species-level
traits (Dmax, SM and WD) explained on average 41% of
interspecific variation in vital rates (mean R2 ranged
between 0�11 and 0�66; Fig. 3), despite unquantified envi-
ronmental effects on vital rates. This is quite remarkable
and represents a substantial improvement over earlier stud-
ies. Functional biology may yet improve understanding of
tropical forest dynamics.
Acknowledgements
We thank Timothy Paine and an anonymous reviewer for helpful com-
ments. This study was supported by the Netherlands Organization for Sci-
entific Research (NWO-ALW 801-01-009; MDV), the Smithsonian Tropical
Research Institute (MDV, MB) and the HSBC Climate Partnership
(HCM). The data sets were collected with funding from the National
Science Foundation (DEB 0425651 & 0948585 to SPH and 1242622 &
1464389 to LSC), the Smithsonian Tropical Research Institute, the Centre
for Tropical Forest Science, the John D. and Catherine T. MacArthur
Foundation, the Mellon Foundation and the Small World Institute Fund.
Data accessibility
Previously archived BCI data set is available from Dryad (datadryad.org)
or the Smithsonian DSpace repository (repository.si.edu). Data sets 1, 2, 3
and 5 are achieved at DSpace with DOIs 10.5479/data.bci.20130603,
10.5479/si.data.201511251137, 10.5479/si.data.201511251134 and 10.5479/
si.data.201511251100, respectively, while data set 4 is archived at Dryad
http://dx.doi.org/10.5061/dryad.fm654 (Visser et al. 2016).
References
Anten, N.P.R. & Schieving, F. (2010) The role of wood mass density and
mechanical constraints in the economy of tree architecture. The Ameri-
can Naturalist, 175, 250–260.Baraloto, C., Hardy, O.J., Paine, C.E.T., Dexter, K.G., Cruaud, C., Dun-
ning, L.T. et al. (2012) Using functional traits and phylogenetic trees to
examine the assembly of tropical tree communities. Journal of Ecology,
100, 690–701.Bates, D., Maechler, M., Bolker, B. & Walker, S. (2015) Fitting linear
mixed-effects models using lme4. Journal of Statistical Software, 67, 1–48.
Bolker, B.M., Brooks, M.E., Clark, C.J., Geange, S.W., Poulsen, J.R., Stevens,
M.H.H. et al. (2009) Generalized linear mixed models: a practical guide for
ecology and evolution. Trends in Ecology and Evolution, 24, 127–135.Buckland, S.T., Burnham, K.P. & Augustin, N.H. (1997) Model selection:
an integral part of inference. Biometrics, 53, 603–618.Burnham, K.P. & Anderson, D.R. (2002) Model selection and multimode
inference: A practical information-theoretic approach. Springer-Verlag,
New York, NY, USA.
Chave, J., Coomes, D., Jansen, S., Lewis, S.L., Swenson, N.G. & Zanne,
A.E. (2009) Towards a worldwide wood economics spectrum. Ecology
Letters, 12, 351–366.Comita, L.S. & Engelbrecht, B.M.J. (2009) Seasonal and spatial variation
in water availability drive habitat associations in a tropical forest. Ecol-
ogy, 90, 2755–2765.Comita, L.S. & Hubbell, S.P. (2009) Local neighborhood and species’ shade
tolerance influence survival in a diverse seedling bank. Ecology, 90, 328–334.
Comita, L.S., Aguilar, S., P�erez, R., Lao, S. & Hubbell, S.P. (2007) Pat-
terns of woody plant species abundance and diversity in the seedling
layer of a tropical forest. Journal of Vegetation Science, 18, 163–174.Condit, R., Ashton, P.S., Manokaran, M., LaFrankie, J.V., Hubbell, S.P.
& Foster, R.B. (1999) Dynamics of the forest communities at Pasoh and
Barro Colorado: comparing two 50-ha plots. Philosophical Transactions
of the Royal Society of London, Series B: Biological Sciences, 354, 1739–1748.
Condit, R., Engelbrecht, B.M.J., Pino, D., P�erez, R. & Turner, B.L. (2013)
Species distributions in response to individual soil nutrients and seasonal
drought across a community of tropical trees. Proceedings of the
National Academy of Sciences of the United States of America, 110,
5064–5068.Cox, P.M., Pearson, D., Booth, B.B., Friedlingstein, P., Huntingford, C.,
Jones, C.D. et al. (2013) Sensitivity of tropical carbon to climate change
constrained by carbon dioxide variability. Nature, 494, 341–344.Davies, S.J. & Ashton, P.S. (1999) Phenology and fecundity in 11 sympatric
pioneer species of Macaranga (Euphorbiaceae) in Borneo. American
Journal of Botany, 86, 1786–1795.Falster, D.S., FitzJohn, R.G. & Westoby, M. (2016) Plant: A package for
modelling forest trait ecology & evolution. Methods in Ecology and Evo-
lution, 7, 136–146.Falster, D.S. & Westoby, M. (2003) Plant height and evolutionary games.
Trends in Ecology and Evolution, 18, 337–343.Friedlingstein, P., Meinshausen, M., Arora, V.K., Jones, C.D., Anav, A.,
Liddicoat, S.K. et al. (2014) Uncertainties in CMIP5 climate projections
due to carbon cycle feedbacks. Journal of Climate, 27, 511–526.Gelman, A. (2006) Multilevel (Hierarchical) Modeling: what it can and can-
not do. Technometrics, 48, 432–435.Grueber, C.E., Nakagawa, S., Laws, R.J. & Jamieson, I.G. (2011) Multi-
model inference in ecology and evolution: challenges and solutions. Jour-
nal of Evolutionary Biology, 24, 699–711.Henery, M.L. & Westoby, M. (2001) Seed mass and seed nutrient content
as predictors of seed output variation between species. Oikos, 92, 479–490.
H�erault, B., Bachelot, B., Poorter, L., Rossi, V., Bongers, F., Chave, J.
et al. (2011) Functional traits shape ontogenetic growth trajectories of
rain forest tree species. Journal of Ecology, 99, 1431–1440.Iida, Y., Poorter, L., Sterck, F.J., Kassim, A.R., Kubo, T., Potts, M.D.
et al. (2012) Wood density explains architectural differentiation
across 145 co-occurring tropical tree species. Functional Ecology, 26,
274–282.Iida, Y., Kohyama, T.S., Swenson, N.G., Su, S.H., Chen, C.T., Chiang,
J.M. et al. (2014a) Linking functional traits and demographic rates in a
subtropical tree community: the importance of size dependency. Journal
of Ecology, 102, 641–650.Iida, Y., Poorter, L., Sterck, F., Kassim, A.R., Potts, M.D., Kubo, T. et al.
(2014b) Linking size-dependent growth and mortality with architectural
traits across 145 co-occurring tropical tree species. Ecology, 95, 353–363.King, D.A., Wright, S.J. & Connell, J.J.H. (2006) The contribution of inter-
specific variation in maximum tree height to tropical and temperate
diversity. Journal of Tropical Ecology, 22, 11–24.Kohyama, T. (1993) Size-structured tree populations in gap-dynamic forest
– the forest architecture hypothesis for the stable coexistence of species.
Journal of Ecology, 81, 131–143.Kohyama, T., Suzuki, E., Partomihardjo, T., Yamada, T. & Kubo, T.
(2003) Tree species differentiation in growth, recruitment and allometry
in relation to maximum height in a Bornean mixed dipterocarp forest.
Journal of Ecology, 91, 797–806.Larjavaara, M. & Muller-Landau, H.C. (2010) Rethinking the value of high
wood density. Functional Ecology, 24, 701–705.Lasky, J.R., Sun, I.F., Su, S.H., Chen, Z.S. & Keitt, T.H. (2013) Trait-
mediated effects of environmental filtering on tree community dynamics.
Journal of Ecology, 101, 722–733.Lasky, J.R., Yang, J., Zhang, G., Cao, M., Tang, Y. & Keitt, T.H. (2014)
The role of functional traits and individual variation in the co-occur-
rence of Ficus species. Ecology, 95, 978–990.Leigh, E.G. (1999) Tropical Forest Ecology a View from Barro Colorado
Island. Oxford University Press, NY, USA.
Lichstein, J.W., Dushoff, J., Ogle, K., Chen, A., Purves, D.W., Caspersen,
J.P. et al. (2010) Unlocking the forest inventory data: relating individual
tree performance to unmeasured environmental factors. Ecological Appli-
cations, 20, 684–699.Marks, C.O. & Lechowicz, M.J. (2006) Alternative designs and the evolu-
tion of functional diversity. The American Naturalist, 167, 55–66.McIntyre, S., Lavorel, S., Landsberg, J. & Forbes, T.D.A. (1999) Distur-
bance response in vegetation–towards a global perspective on functional
traits. Journal of Vegetation Science, 10, 621–630.McMahon, S.M., Metcalf, C.J.E. & Woodall, C.W. (2011) High-
dimensional coexistence of temperate tree species: functional traits,
© 2016 The Authors. Functional Ecology © 2016 British Ecological Society, Functional Ecology, 30, 168–180
Traits affect whole life-cycle vital rates 179
demographic rates, life-history stages, and their physical context. PLoS
One, 6, e16253.
Merow, C., Dahlgren, J.P., Metcalf, C.J.E., Childs, D.Z., Evans, M.E.K.,
Jongejans, E. et al. (2014) Advancing population ecology with integral
projection models: a practical guide. Methods in Ecology and Evolution,
5, 99–110.Messier, J., McGill, B.J. & Lechowicz, M.J. (2010) How do traits vary
across ecological scales? A case for trait-based ecology. Ecology Letters,
13, 838–848.Moles, A.A.T. & Westoby, M. (2006) Seed size and plant strategy across
the whole life cycle. Oikos, 113, 91–105.Muller-Landau, H.C. (2010) The tolerance–fecundity trade-off and the
maintenance of diversity in seed size. Proceedings of the National Acad-
emy of Sciences of the United States of America, 107, 4242–4247.Muller-Landau, H.C., Condit, R.S., Chave, J., Thomas, S.C., Bohlman,
S.A., Bunyavejchewin, S. et al. (2006) Testing metabolic ecology theory
for allometric scaling of tree size, growth and mortality in tropical for-
ests. Ecology Letters, 9, 575–588.Muller-Landau, H.C., Wright, S.J., Calder�on, O., Condit, R. & Hubbell,
S.P. (2008) Interspecific variation in primary seed dispersal in a tropical
forest. Journal of Ecology, 96, 653–667.Nakagawa, S. & Schielzeth, H. (2013) A general and simple method for
obtaining R2 from generalized linear mixed-effects models. Methods in
Ecology and Evolution, 4, 133–142.Osnas, J.L.D., Lichstein, J.W., Reich, P.B. & Pacala, S.W. (2013) Global
leaf trait relationships: mass, area, and the leaf economics spectrum.
Science, 340, 741–744.Paciorek, C.J., Condit, R., Hubbell, S.P. & Foster, R.B. (2000) The demo-
graphics of resprouting in tree and shrub species of a moist tropical for-
est. Journal of Ecology, 88, 765–777.Paine, C.E.T., Baraloto, C., Chave, J. & H�erault, B. (2011) Functional
traits of individual trees reveal ecological constraints on community
assembly in tropical rain forests. Oikos, 120, 720–727.Paine, C.E.T., Amissah, L., Auge, H., Baraloto, C., Baruffol, M., Bourland,
N. et al. (2015) Globally, functional traits are weak predictors of juvenile
tree growth, and we do not know why. Journal of Ecology, 103, 978–989.Poorter, L., Bongers, F., Sterck, F.J. & W€oll, H. (2005) Beyond the regen-
eration phase: differentiation of height light trajectories among tropical
tree species. Journal of Ecology, 93, 256–267.Poorter, L., Wright, S.J., Paz, H., Ackerly, D.D., Condit, R., Ibarra-Man-
r�ıquez, G. et al. (2008) Are functional traits good predictors of demo-
graphic rates? Evidence from five Neotropical forests. Ecology, 89, 1908–1920.
R Core Team. (2014) R: A language and environment for statistical comput-
ing. R Foundation for Statistical Computing, Vienna, Austria. https://
www.R-project.org/.
R€uger, N., Huth, A., Hubbell, S.P. & Condit, R. (2009) Response of
recruitment to light availability across a tropical lowland rain forest
community. Journal of Ecology, 97, 1360–1368.R€uger, N., Berger, U., Hubbell, S.P., Vieilledent, G. & Condit, R. (2011)
Growth strategies of tropical tree species: disentangling light and size
effects. PLoS One, 6, e25330.
R€uger, N., Wirth, C., Wright, S.J. & Condit, R. (2012) Functional traits
explain light and size response of growth rates in tropical tree species.
Ecology, 93, 2626–2636.Schielzeth, H. & Forstmeier, W. (2009) Conclusions beyond support: over-
confident estimates in mixed models. Behavioral Ecology, 20, 416–420.Spasojevic, M., Yablon, E., Oberle, B. & Myers, J. (2014) Ontogenetic trait
variation influences tree community assembly across environmental gra-
dients. Ecosphere, 5, art129.
Sterck, F.J. & Bongers, F. (2001) Crown development in tropical rain forest
trees: patterns with tree height and light availability. Journal of Ecology,
89, 1–13.
Thomas, S.C. (1996) Relative size at onset of maturity in rain forest trees: a
comparative analysis of 37 Malaysian species. Oikos, 76, 145–154.Uriarte, M., Canham, C.D., Thompson, J. & Zimmerman, J.K. (2004) A
neighborhood analysis of tree growth and survival in a hurricane-driven
tropical forest. Ecological Monographs, 74, 591–614.Uriarte, M., Lasky, J.R., Boukili, V.K. & Chazdon, R.L. (2016) A trait-
mediated, neighbourhood approach to quantify climate impacts on
successional dynamics of tropical rainforests. Functional Ecology, 30,
157–167.Visser, M.D., Jongejans, E., van Breugel, M., Zuidema, P.A., Chen, Y.-Y.,
Kassim, A.R. et al. (2011) Strict mast fruiting for a tropical dipterocarp
tree: a demographic cost-benefit analysis of delayed reproduction and
seed predation. Journal of Ecology, 99, 1033–1044.Visser, M.D., Bruijning, M., Wright, S.J., Muller-Landau, H.C., Jongejans,
E., Comita, L.S. & de Kroon, H. (2016) Functional traits as predictors of
vital rates across the life-cycle of tropical trees. Functional Ecology, 30,
168–180.Westoby, M. (1998) A leaf-height-seed (LHS) plant ecology strategy
scheme. Plant and Soil, 199, 213–227.Westoby, M. & Wright, I.J.I. (2006) Land-plant ecology on the basis of
functional traits. Trends in Ecology and Evolution, 21, 261–268.Westoby, M., Falster, D.D.S., Moles, A.A.T., Vesk, P.A. & Wright, I.J.
(2002) Plant ecological strategies: some leading dimensions of variation
between species. Annual Review of Ecology, Evolution, and Systematics,
33, 125–159.Whittingham, M.J., Stephens, P.A., Bradbury, R.B. & Freckleton, R.P.
(2006) Why do we still use stepwise modelling in ecology and behaviour?
Journal of Animal Ecology, 75, 1182–1189.Wright, I.J., Reich, P.B., Westoby, M., Ackerly, D.D., Baruch, Z., Bongers,
F. et al. (2004) The worldwide leaf economics spectrum. Nature, 428,
821–827.Wright, S.J., Muller-Landau, H.C., Calder�on, O. & Hernandez, A. (2005)
Annual and spatial variation in seedfall and seeding recruitment in a
neotropical forest. Ecology, 86, 848–860.Wright, I.J., Ackerly, D.D., Bongers, F., Harms, K.E., Ibarra-Manriquez,
G., Martinez-Ramos, M. et al. (2007) Relationships among ecologically
important dimensions of plant trait variation in seven Neotropical for-
ests. Annals of Botany, 99, 1003–1015.Wright, S.J., Kitajima, K., Kraft, N.J.B., Reich, P.P.B., Wright, I.J., Bun-
ker, D.E. et al. (2010) Functional traits and the growth-mortality trade-
off in tropical trees. Ecology, 91, 3664–3674.Zuidema, P.A. & Boot, R.G.A. (2002) Demography of the Brazil nut tree
(Bertholletia excelsa) in the Bolivian Amazon: impact of seed extraction
on recruitment and population dynamics. Journal of Tropical Ecology,
18, 1–31.
Received 2 June 2015; accepted 4 December 2015
Handling Editor: Cory Merow
Supporting Information
Additional Supporting information may be found in the online
version of this article:
Appendix S1. Including Tables S1–S12, Figures S1–S11 and
Text S1.
Appendix S2. R-code example for model averaging.
© 2016 The Authors. Functional Ecology © 2016 British Ecological Society, Functional Ecology, 30, 168–180
180 M. D. Visser et al.
SI: Functional traits as predictors of vital rates
across the life-cycle of tropical trees.
Marco D. Visser∗1,2, Marjolein Bruijning†1, Stuart J. Wright‡2,Helene C. Muller-Landau§2, Eelke Jongejans¶1, Liza S. Comita‖3,
and Hans de Kroon∗∗1
1Radboud University, Departments of Experimental Plant Ecology& Animal Ecology and Physiology, The Netherlands
2Smithsonian Tropical Research Institute, Republic of Panama3Yale University, School of Forestry and Environmental Studies
December 1, 2015
Contents
S1 TEXT: Comparison of results using inverse modeling 2
S2 Tables 4
S3 Figures 21
S1 TEXT: Comparison of results using
inverse modeling
When multiple adult trees contribute seeds to a single location it becomes com-plicated to assign individual seeds to their respective sources. A now commonlyused method pioneered by Ribbens et al (1994) to overcome this problem isinverse modeling (IM) of seed shadows (Clark et al, 1999; Muller-Landau etal, 2008). A seed shadow describes the distribution of seeds with distance from
∗[email protected]†[email protected]‡[email protected]§[email protected]¶[email protected]‖[email protected]∗∗[email protected]
1
S1 TEXT: COMPARISON OF RESULTS USING INVERSE MODELING
their parent, and consists of two elements; (1) a component describing the fecun-dity of a tree, usually as a function of size and (2) a dispersal kernel describingthe distribution of seeds as a function of distance from the source. The seedshadow is the product of these two elements (Clark et al., 1999).
Seed shadows from individual trees will overlap within a stand and the seedrain of these multiple sources can be viewed as a smoothed version of the individ-ual shadows (Clark et al., 1999). This smoothed shadow can be expressed as thesummed contributions of each individual seed shadow (hereafter the summedseed shadow SSS). The number of seeds in a certain quadrat is predicted by sum-ming the densities the quadrat receives from each of the surrounding sources, asprojected by each individual seed shadow. IM has been shown to be a promisingestimate of dispersal distances, as it is not inherently biased to long or shortdistance dispersal (Jones and Muller-Landau, 2008; Muller-Landau et al, 2008).With IM the predicted seed density in a sample quadrat at location j is givenby the summed seed shadow of all N trees at location j (SSSj);
SSSj = α ·T∑
i=1
F (pibi|β) ·Q(rij |µ, σ) (1)
where SSSj is the predicted seed number in trap j, which is the sum of seedcontributions from T trees multiplied by trap size α (0.5 m2). For each treei, the seed shadow is a function of the fecundity F , and a two-dimensionaldispersal kernel Q. Q depends on the distance rij between tree i and trapj, while F is a function of basal area bi (calculated as [dbh/2]2π) and seedproduction probabilities pi (obtained from the reproductive census as describedin the main text). Only females were included for dioecious species.
The parameters β, µ and σ were fitted per species. We tested four differentdispersal kernels: the Exponential, 2dt (Clark et al., 1999), 2dt1k (which is the2dt distribution with the degrees of freedom parameter set to 3) (Muller-Landauet al., 2008) and the Cauchy density distribution.
Off plot integrationTraps within 20 meter from the edge of the plot were excluded since they couldbe highly influenced by trees outside the plot (Muller-Landau et al., 2008). Datafrom the remaining 188 traps were used. We then estimated the contribution tothe seed rain from unknown sources outside the mapped plot. Here it is assumedthat the per unit area seed production off plot is uniform to that within theplot (see Muller-Landau et al., 2008, for details). The total expected seed count
ˆSSSj in trap j then becomes the sum of the expected contributions from treesinside the plot SSSj and the expected contributions from trees outside the plot
(immigrant seeds) Ij to seed trap j.
ˆSSSj = SSSj + Ij (2)
2
S2 TABLES
Parameter estimationModels were fit using maximum likelihood, assuming seed counts in traps weredistributed according to a negative binomial distribution, with an additinalclumping parameter k (Muller-Landau et al., 2008), as previous studies haveshown that seed arrival tends to be clumped (Muller-Landau et al., 2008).
The results from the above described inverse modeling approach are comparedto the alternative estimates (see maintext) in figure S2.
S2 Tables
3
S2
TABLES
Table S1. Overview of species used in each analysis, ordered alphabetically. True indicates a species was included.
Species Reproduction Seed production Seedlingestablishment
Seedlingsurvival
Seedlinggrowth
Tree survival Treegrowth
Adelia triloba FALSE FALSE FALSE FALSE FALSE TRUE TRUEAlchornea costaricensi TRUE TRUE TRUE TRUE TRUE TRUE TRUEAlibertia edulis FALSE FALSE TRUE TRUE TRUE TRUE TRUEAllophylus psilospermus FALSE FALSE FALSE FALSE FALSE TRUE TRUEAlseis blackiana TRUE TRUE TRUE TRUE TRUE TRUE TRUEAnacardium excelsum FALSE TRUE FALSE TRUE TRUE FALSE FALSEAndira inermis FALSE FALSE FALSE FALSE FALSE TRUE TRUEAnnona acuminata FALSE FALSE FALSE TRUE TRUE TRUE TRUEAnnona spraguei FALSE FALSE FALSE TRUE TRUE TRUE TRUEApeiba membranacea TRUE FALSE TRUE TRUE TRUE TRUE TRUEApeiba tibourbou TRUE FALSE FALSE FALSE FALSE FALSE FALSEAspidosperma spruceanum TRUE FALSE TRUE TRUE TRUE TRUE TRUEAstronium graveolens TRUE TRUE FALSE FALSE FALSE TRUE TRUEBeilschmiedia pendula TRUE TRUE TRUE TRUE TRUE TRUE TRUEBrosimum alicastrum TRUE TRUE TRUE TRUE TRUE TRUE TRUECalophyllum longifolium FALSE FALSE TRUE TRUE TRUE TRUE TRUECapparis frondosa FALSE FALSE TRUE TRUE TRUE TRUE TRUECasearia aculeata TRUE FALSE FALSE FALSE FALSE TRUE TRUECasearia arborea FALSE FALSE FALSE FALSE FALSE TRUE TRUECassipourea elliptica TRUE FALSE TRUE TRUE TRUE TRUE TRUECecropia insignis FALSE FALSE FALSE FALSE FALSE FALSE FALSEChrysochlamy eclipes FALSE FALSE FALSE FALSE FALSE TRUE TRUEChrysophyllum argenteum TRUE TRUE TRUE TRUE TRUE TRUE TRUEChrysophyllum cainito TRUE TRUE TRUE TRUE TRUE TRUE TRUECoccoloba coronata FALSE FALSE FALSE FALSE FALSE TRUE TRUECoccoloba manzinellens FALSE FALSE FALSE FALSE FALSE TRUE TRUECordia alliodora TRUE TRUE TRUE TRUE TRUE TRUE TRUECordia bicolor TRUE TRUE TRUE TRUE TRUE TRUE TRUECoussarea curvigemmia TRUE TRUE TRUE TRUE TRUE TRUE TRUECupania rufescens FALSE FALSE FALSE FALSE FALSE TRUE TRUECupania seemannii FALSE FALSE FALSE FALSE FALSE TRUE TRUE
Continued on next page4
S2
TABLES
Table S1. Overview of species used in each analysis, ordered alphabetically. True indicates a species was included.
Species Reproduction Seed production Seedlingestablishment
Seedlingsurvival
Seedlinggrowth
Tree survival Treegrowth
Dendropanax arboreus FALSE FALSE TRUE FALSE FALSE FALSE FALSEDesmopsis panamensis TRUE TRUE TRUE TRUE TRUE TRUE TRUEDipteryx oleifera TRUE TRUE FALSE TRUE TRUE FALSE FALSEDrypetes standleyi FALSE TRUE TRUE TRUE TRUE TRUE TRUEErythroxylum panamense FALSE FALSE FALSE FALSE FALSE TRUE TRUEEugenia coloradoensi FALSE FALSE TRUE TRUE TRUE TRUE TRUEEugenia galalonensis FALSE FALSE FALSE TRUE TRUE TRUE TRUEEugenia nesiotica FALSE FALSE TRUE TRUE TRUE TRUE TRUEEugenia oerstediana TRUE TRUE TRUE TRUE TRUE TRUE TRUEFaramea occidentalis TRUE TRUE TRUE TRUE TRUE TRUE TRUEFicus insipida FALSE FALSE TRUE FALSE FALSE FALSE FALSEGarcinia intermedia TRUE TRUE TRUE TRUE TRUE TRUE TRUEGarcinia madruno FALSE FALSE FALSE FALSE FALSE TRUE TRUEGenipa americana FALSE FALSE TRUE FALSE FALSE FALSE FALSEGuapira standleyana FALSE FALSE TRUE TRUE TRUE TRUE TRUEGuarea guidonia TRUE TRUE TRUE TRUE TRUE TRUE TRUEGuatteria dumetorum TRUE FALSE TRUE TRUE TRUE TRUE TRUEGustavia superba TRUE TRUE TRUE TRUE TRUE TRUE TRUEHamelia axillaris FALSE FALSE FALSE TRUE TRUE TRUE TRUEHampea appendiculat FALSE FALSE TRUE TRUE TRUE TRUE TRUEHasseltia floribunda TRUE FALSE TRUE FALSE FALSE TRUE TRUEHeisteria acuminata FALSE FALSE TRUE FALSE FALSE TRUE TRUEHeisteria concinna TRUE TRUE TRUE TRUE TRUE TRUE TRUEHerrania purpurea FALSE FALSE FALSE FALSE FALSE TRUE TRUEHirtella triandra TRUE TRUE TRUE TRUE TRUE TRUE TRUEHybanthus prunifolius FALSE FALSE TRUE TRUE TRUE TRUE TRUEHieronyma alchorneoide TRUE TRUE FALSE FALSE FALSE TRUE TRUEInga goldmanii FALSE FALSE FALSE FALSE FALSE TRUE TRUEInga marginata TRUE TRUE TRUE TRUE TRUE TRUE TRUEInga nobilis FALSE FALSE FALSE FALSE FALSE TRUE TRUEInga acuminata FALSE FALSE TRUE TRUE TRUE TRUE TRUE
Continued on next page5
S2
TABLES
Table S1. Overview of species used in each analysis, ordered alphabetically. True indicates a species was included.
Species Reproduction Seed production Seedlingestablishment
Seedlingsurvival
Seedlinggrowth
Tree survival Treegrowth
Inga sapindoides FALSE FALSE FALSE FALSE FALSE TRUE TRUEInga thibaudiana FALSE FALSE FALSE TRUE TRUE TRUE TRUEInga umbellifera FALSE FALSE FALSE TRUE TRUE TRUE TRUEJacaranda copaia TRUE FALSE TRUE FALSE FALSE TRUE TRUELacistema aggregatum FALSE FALSE FALSE TRUE TRUE TRUE TRUELacmellea panamensis FALSE FALSE TRUE TRUE TRUE TRUE TRUELaetia thamnia TRUE FALSE FALSE FALSE FALSE TRUE TRUELicania hypoleuca FALSE FALSE FALSE FALSE FALSE TRUE TRUELicania platypus FALSE FALSE FALSE FALSE FALSE TRUE TRUELonchocarpus heptaphyllus TRUE FALSE TRUE TRUE TRUE TRUE TRUELuehea seemannii TRUE TRUE TRUE TRUE TRUE TRUE TRUEMosannona garwoodii TRUE FALSE FALSE FALSE FALSE TRUE TRUEMaquira guianensis FALSE FALSE FALSE FALSE FALSE TRUE TRUEMouriri myrtilloides TRUE TRUE TRUE TRUE TRUE TRUE TRUENectandra cissiflora FALSE FALSE FALSE FALSE FALSE TRUE TRUEOcotea cernua FALSE FALSE FALSE TRUE TRUE TRUE TRUEOcotea puberula FALSE FALSE TRUE TRUE TRUE TRUE TRUETrophis caucana FALSE FALSE FALSE TRUE TRUE TRUE TRUEOrmosia coccinea FALSE FALSE FALSE FALSE FALSE TRUE TRUEOrmosia macrocalyx FALSE FALSE FALSE FALSE FALSE TRUE TRUEOuratea lucens FALSE FALSE TRUE TRUE TRUE TRUE TRUEPalicourea guianensis FALSE FALSE TRUE TRUE TRUE TRUE TRUEPentagonia macrophylla FALSE FALSE FALSE FALSE FALSE TRUE TRUEPerebea xanthochyma FALSE FALSE FALSE FALSE FALSE TRUE TRUECinnamomum triplinerve FALSE FALSE TRUE TRUE TRUE FALSE FALSEPicramnia latifolia FALSE FALSE TRUE TRUE TRUE TRUE TRUEPlatymiscium pinnatum FALSE FALSE FALSE TRUE TRUE TRUE TRUEPlatypodium elegans TRUE FALSE FALSE FALSE FALSE TRUE TRUEPoulsenia armata TRUE FALSE TRUE FALSE FALSE TRUE TRUEPourouma bicolor FALSE FALSE FALSE FALSE FALSE TRUE TRUEPouteria reticulata TRUE TRUE TRUE TRUE TRUE TRUE TRUE
Continued on next page6
S2
TABLES
Table S1. Overview of species used in each analysis, ordered alphabetically. True indicates a species was included.
Species Reproduction Seed production Seedlingestablishment
Seedlingsurvival
Seedlinggrowth
Tree survival Treegrowth
Prioria copaifera TRUE TRUE FALSE TRUE TRUE TRUE TRUEProtium panamense TRUE FALSE FALSE TRUE TRUE TRUE TRUEProtium tenuifolium TRUE TRUE TRUE TRUE TRUE TRUE TRUEPseudobombax septenatum TRUE FALSE FALSE FALSE FALSE FALSE FALSEPsychotria acuminata FALSE FALSE TRUE TRUE TRUE FALSE FALSEPsychotria deflexa FALSE FALSE FALSE TRUE TRUE FALSE FALSEPsychotria horizontalis FALSE FALSE TRUE TRUE TRUE TRUE TRUEPsychotria marginata FALSE FALSE TRUE TRUE TRUE TRUE TRUEPsychotria racemosa FALSE FALSE TRUE TRUE TRUE FALSE FALSEPterocarpus rohrii TRUE FALSE FALSE FALSE FALSE TRUE TRUEQuararibea asterolepis TRUE TRUE TRUE TRUE TRUE TRUE TRUEQuassia amara FALSE FALSE FALSE FALSE FALSE TRUE TRUERandia armata TRUE TRUE TRUE TRUE TRUE TRUE TRUESimarouba amara TRUE TRUE TRUE TRUE TRUE TRUE TRUESloanea terniflora TRUE FALSE FALSE FALSE FALSE TRUE TRUESorocea affinis FALSE FALSE TRUE TRUE TRUE TRUE TRUESpondias mombin FALSE FALSE FALSE FALSE FALSE TRUE TRUESpondias radlkoferi FALSE FALSE FALSE TRUE TRUE TRUE TRUESterculia apetala FALSE FALSE FALSE TRUE TRUE FALSE FALSEStylogyne turbacensis FALSE FALSE FALSE TRUE TRUE TRUE TRUESwartzia simplex FALSE FALSE FALSE TRUE TRUE FALSE FALSESymphonia globulifera FALSE FALSE FALSE FALSE FALSE TRUE TRUETabebuia guayacan FALSE FALSE TRUE FALSE FALSE FALSE FALSETabebuia rosea TRUE TRUE TRUE TRUE TRUE TRUE TRUETabernaemont arborea TRUE FALSE TRUE TRUE TRUE TRUE TRUETachigali versicolor FALSE FALSE FALSE TRUE TRUE TRUE TRUETalisia nervosa TRUE FALSE FALSE FALSE FALSE TRUE TRUETalisia princeps FALSE FALSE FALSE TRUE TRUE TRUE TRUETerminalia amazonia TRUE TRUE FALSE FALSE FALSE FALSE FALSETerminalia oblonga TRUE FALSE FALSE FALSE FALSE FALSE FALSETetragastris panamensis TRUE TRUE TRUE TRUE TRUE TRUE TRUE
Continued on next page7
S2
TABLES
Table S1. Overview of species used in each analysis, ordered alphabetically. True indicates a species was included.
Species Reproduction Seed production Seedlingestablishment
Seedlingsurvival
Seedlinggrowth
Tree survival Treegrowth
Thevetia ahouai FALSE FALSE FALSE FALSE FALSE TRUE TRUETrattinnicki aspera TRUE TRUE FALSE FALSE FALSE FALSE FALSETrichilia pallida TRUE TRUE TRUE FALSE FALSE TRUE TRUETrichilia tuberculata TRUE TRUE TRUE TRUE TRUE TRUE TRUETriplaris cumingiana TRUE TRUE TRUE FALSE FALSE TRUE TRUETurpinia occidentalis FALSE FALSE FALSE FALSE FALSE TRUE TRUEUnonopsis pittieri TRUE FALSE TRUE TRUE TRUE TRUE TRUEVirola sebifera TRUE FALSE TRUE TRUE TRUE TRUE TRUEVirola surinamensis FALSE FALSE FALSE FALSE FALSE TRUE TRUEVochysia ferruginea TRUE FALSE TRUE FALSE FALSE FALSE FALSEXylopia macrantha TRUE FALSE FALSE TRUE TRUE TRUE TRUEZanthoxylum panamense FALSE FALSE TRUE TRUE TRUE TRUE TRUE
8
S2 TABLES
Table S2. All evaluated forms of the models for demographic rates (Y) that arepotentially size-dependent. The variable Y refers to seedling growth, seedlingsurvival, tree growth, tree survival, or tree reproduction. The variable size refersto seedling height, tree dbh (when Y is tree reproduction or survival) or treebasal area (when Y is tree growth).The variables wd, sm, lma and dmax referto wood density, seed mass, leaf mass per area and maximum dbh, respectively.Random effects are for species (1 | species) and individual (1 | tag).
formula
1 Y ∼ 1 + (1|species) + (1|tag)2 Y ∼ size + (1|species) + (1|tag)3 Y ∼ size + lma + (1|species) + (1|tag)4 Y ∼ size ∗ lma + (1|species) + (1|tag)5 Y ∼ size + sm + (1|species) + (1|tag)6 Y ∼ size + sm + lma + (1|species) + (1|tag)7 Y ∼ size + sm + size ∗ lma + (1|species) + (1|tag)8 Y ∼ size ∗ sm + (1|species) + (1|tag)9 Y ∼ size ∗ sm + lma + (1|species) + (1|tag)
10 Y ∼ size ∗ sm + size ∗ lma + (1|species) + (1|tag)11 Y ∼ size + wd + (1|species) + (1|tag)12 Y ∼ size + wd + lma + (1|species) + (1|tag)13 Y ∼ size + wd + size ∗ lma + (1|species) + (1|tag)14 Y ∼ size + wd + sm + (1|species) + (1|tag)15 Y ∼ size + wd + sm + lma + (1|species) + (1|tag)16 Y ∼ size + wd + sm + size ∗ lma + (1|species) + (1|tag)17 Y ∼ size + wd + size ∗ sm + (1|species) + (1|tag)18 Y ∼ size + wd + size ∗ sm + lma + (1|species) + (1|tag)19 Y ∼ size + wd + size ∗ sm + size ∗ lma + (1|species) + (1|tag)20 Y ∼ size ∗ wd + (1|species) + (1|tag)21 Y ∼ size ∗ wd + lma + (1|species) + (1|tag)22 Y ∼ size ∗ wd + size ∗ lma + (1|species) + (1|tag)23 Y ∼ size ∗ wd + sm + (1|species) + (1|tag)24 Y ∼ size ∗ wd + sm + lma + (1|species) + (1|tag)25 Y ∼ size ∗ wd + sm + size ∗ lma + (1|species) + (1|tag)26 Y ∼ size ∗ wd + size ∗ sm + (1|species) + (1|tag)27 Y ∼ size ∗ wd + size ∗ sm + lma + (1|species) + (1|tag)28 Y ∼ size ∗ wd + size ∗ sm + size ∗ lma + (1|species) + (1|tag)29 Y ∼ size + dmax + (1|species) + (1|tag)30 Y ∼ size + lma + dmax + (1|species) + (1|tag)31 Y ∼ size ∗ lma + dmax + (1|species) + (1|tag)32 Y ∼ size + sm + dmax + (1|species) + (1|tag)33 Y ∼ size + sm + lma + dmax + (1|species) + (1|tag)34 Y ∼ size + sm + size ∗ lma + dmax + (1|species) + (1|tag)35 Y ∼ size ∗ sm + dmax + (1|species) + (1|tag)36 Y ∼ size ∗ sm + lma + dmax + (1|species) + (1|tag)37 Y ∼ size ∗ sm + size ∗ lma + dmax + (1|species) + (1|tag)38 Y ∼ size + wd + dmax + (1|species) + (1|tag)39 Y ∼ size + wd + lma + dmax + (1|species) + (1|tag)40 Y ∼ size + wd + size ∗ lma + dmax + (1|species) + (1|tag)41 Y ∼ size + wd + sm + dmax + (1|species) + (1|tag)42 Y ∼ size + wd + sm + lma + dmax + (1|species) + (1|tag)43 Y ∼ size + wd + sm + size ∗ lma + dmax + (1|species) + (1|tag)44 Y ∼ size + wd + size ∗ sm + dmax + (1|species) + (1|tag)45 Y ∼ size + wd + size ∗ sm + lma + dmax + (1|species) + (1|tag)46 Y ∼ size + wd + size ∗ sm + size ∗ lma + dmax + (1|species) + (1|tag)47 Y ∼ size ∗ wd + dmax + (1|species) + (1|tag)48 Y ∼ size ∗ wd + lma + dmax + (1|species) + (1|tag)49 Y ∼ size ∗ wd + size ∗ lma + dmax + (1|species) + (1|tag)50 Y ∼ size ∗ wd + sm + dmax + (1|species) + (1|tag)51 Y ∼ size ∗ wd + sm + lma + dmax + (1|species) + (1|tag)52 Y ∼ size ∗ wd + sm + size ∗ lma + dmax + (1|species) + (1|tag)53 Y ∼ size ∗ wd + size ∗ sm + dmax + (1|species) + (1|tag)54 Y ∼ size ∗ wd + size ∗ sm + lma + dmax + (1|species) + (1|tag)55 Y ∼ size ∗ wd + size ∗ sm + size ∗ lma + dmax + (1|species) + (1|tag)56 Y ∼ size ∗ dmax + (1|species) + (1|tag)57 Y ∼ size + lma + size ∗ dmax + (1|species) + (1|tag)58 Y ∼ size ∗ lma + size ∗ dmax + (1|species) + (1|tag)59 Y ∼ size + sm + size ∗ dmax + (1|species) + (1|tag)60 Y ∼ size + sm + lma + size ∗ dmax + (1|species) + (1|tag)61 Y ∼ size + sm + size ∗ lma + size ∗ dmax + (1|species) + (1|tag)62 Y ∼ size ∗ sm + size ∗ dmax + (1|species) + (1|tag)63 Y ∼ size ∗ sm + lma + size ∗ dmax + (1|species) + (1|tag)64 Y ∼ size ∗ sm + size ∗ lma + size ∗ dmax + (1|species) + (1|tag)65 Y ∼ size + wd + size ∗ dmax + (1|species) + (1|tag)66 Y ∼ size + wd + lma + size ∗ dmax + (1|species) + (1|tag)67 Y ∼ size + wd + size ∗ lma + size ∗ dmax + (1|species) + (1|tag)68 Y ∼ size + wd + sm + size ∗ dmax + (1|species) + (1|tag)69 Y ∼ size + wd + sm + lma + size ∗ dmax + (1|species) + (1|tag)70 Y ∼ size + wd + sm + size ∗ lma + size ∗ dmax + (1|species) + (1|tag)71 Y ∼ size + wd + size ∗ sm + size ∗ dmax + (1|species) + (1|tag)72 Y ∼ size + wd + size ∗ sm + lma + size ∗ dmax + (1|species) + (1|tag)73 Y ∼ size + wd + size ∗ sm + size ∗ lma + size ∗ dmax + (1|species) + (1|tag)74 Y ∼ size ∗ wd + size ∗ dmax + (1|species) + (1|tag)75 Y ∼ size ∗ wd + lma + size ∗ dmax + (1|species) + (1|tag)76 Y ∼ size ∗ wd + size ∗ lma + size ∗ dmax + (1|species) + (1|tag)77 Y ∼ size ∗ wd + sm + size ∗ dmax + (1|species) + (1|tag)78 Y ∼ size ∗ wd + sm + lma + size ∗ dmax + (1|species) + (1|tag)79 Y ∼ size ∗ wd + sm + size ∗ lma + size ∗ dmax + (1|species) + (1|tag)80 Y ∼ size ∗ wd + size ∗ sm + size ∗ dmax + (1|species) + (1|tag)81 Y ∼ size ∗ wd + size ∗ sm + lma + size ∗ dmax + (1|species) + (1|tag)82 Y ∼ size ∗ wd + size ∗ sm + size ∗ lma + size ∗ dmax + (1|species) + (1|tag)
9
S2
TABLES
Table S3. All evaluated model forms, model weights, and ∆AIC values for the seed production analysis (fseeds).
formulas weight ∆AIC
fec ∼ sm + dmax + lma 0.212 0.000fec ∼ sm + dmax + wd 0.203 0.091fec ∼ sm + lma 0.178 0.354fec ∼ sm + dmax 0.120 1.144fec ∼ sm + wd 0.087 1.778fec ∼ sm + dmax + wd + lma 0.084 1.859fec ∼ sm + lma + wd 0.072 2.166fec ∼ sm 0.043 3.211fec ∼ lma 0.001 11.975fec ∼ lma + wd 0.000 12.628fec ∼ wd 0.000 13.034fec ∼ dmax + wd 0.000 13.535fec ∼ dmax + lma 0.000 13.779fec ∼ dmax + lma + wd 0.000 14.241fec ∼ 1 0.000 19.228fec ∼ dmax 0.000 20.645
10
S2
TABLES
Table S4. All evaluated model forms, model weights, and ∆AIC values for the seedling establishment analysis.
formulas weight ∆AIC
qlogis(p) ∼ sm + dmax 0.484 0.000qlogis(p) ∼ sm + dmax + wd 0.246 1.356qlogis(p) ∼ sm + dmax + lma 0.178 1.999qlogis(p) ∼ sm + dmax + wd + lma 0.093 3.303qlogis(p) ∼ dmax + wd 0.000 31.452qlogis(p) ∼ dmax + lma + wd 0.000 33.443qlogis(p) ∼ dmax 0.000 34.448qlogis(p) ∼ dmax + lma 0.000 35.980qlogis(p) ∼ sm + wd 0.000 39.262qlogis(p) ∼ sm + lma 0.000 39.317qlogis(p) ∼ sm + lma + wd 0.000 39.588qlogis(p) ∼ sm 0.000 40.278log(p) ∼ 1 0.000 47.684qlogis(p) ∼ wd 0.000 48.810qlogis(p) ∼ lma + wd 0.000 49.492qlogis(p) ∼ lma 0.000 51.928
11
S2
TABLES
Table S5. Evaluated model forms, model weights, and ∆AIC values for the seedling growth analysis, for the top 30 models(out of 82). A full list of model forms can be found in table S2.
formulas weight ∆AIC
growth ∼ height ∗ wd + height ∗ sm + lma + dmax + (1|species) + (1|tag) 0.599 0.000growth ∼ height ∗ wd + height ∗ sm + lma + (1|species) + (1|tag) 0.166 2.567growth ∼ height ∗ wd + height ∗ sm + dmax + (1|species) + (1|tag) 0.143 2.870growth ∼ height ∗ wd + height ∗ sm + (1|species) + (1|tag) 0.041 5.368growth ∼ height ∗ wd + height ∗ sm + height ∗ lma + dmax + (1|species) + (1|tag) 0.021 6.727growth ∼ height ∗ wd + height ∗ sm + height ∗ lma + height ∗ dmax + (1|species) + (1|tag) 0.020 6.752growth ∼ height ∗ wd + height ∗ sm + height ∗ lma + (1|species) + (1|tag) 0.006 9.293growth ∼ height ∗ wd + height ∗ sm + lma + height ∗ dmax + (1|species) + (1|tag) 0.004 10.255growth ∼ height ∗ wd + height ∗ sm + height ∗ dmax + (1|species) + (1|tag) 0.001 13.125growth ∼ height + wd + height ∗ sm + lma + dmax + (1|species) + (1|tag) 0.000 16.569growth ∼ height + wd + height ∗ sm + lma + (1|species) + (1|tag) 0.000 19.128growth ∼ height + wd + height ∗ sm + dmax + (1|species) + (1|tag) 0.000 19.429growth ∼ height + wd + height ∗ sm + (1|species) + (1|tag) 0.000 21.918growth ∼ height + wd + height ∗ sm + lma + height ∗ dmax + (1|species) + (1|tag) 0.000 22.950growth ∼ height ∗ sm + lma + dmax + (1|species) + (1|tag) 0.000 24.422growth ∼ height + wd + height ∗ sm + height ∗ dmax + (1|species) + (1|tag) 0.000 25.810growth ∼ height ∗ sm + lma + (1|species) + (1|tag) 0.000 27.096growth ∼ height ∗ sm + dmax + (1|species) + (1|tag) 0.000 27.599growth ∼ height ∗ sm + (1|species) + (1|tag) 0.000 30.311growth ∼ height ∗ sm + lma + height ∗ dmax + (1|species) + (1|tag) 0.000 30.802growth ∼ height + wd + height ∗ sm + height ∗ lma + dmax + (1|species) + (1|tag) 0.000 31.209growth ∼ height + wd + height ∗ sm + height ∗ lma + (1|species) + (1|tag) 0.000 33.768growth ∼ height ∗ sm + height ∗ dmax + (1|species) + (1|tag) 0.000 33.972growth ∼ height + wd + height ∗ sm + height ∗ lma + height ∗ dmax + (1|species) + (1|tag) 0.000 34.538growth ∼ height ∗ sm + height ∗ lma + dmax + (1|species) + (1|tag) 0.000 39.062growth ∼ height ∗ sm + height ∗ lma + (1|species) + (1|tag) 0.000 41.735growth ∼ height ∗ sm + height ∗ lma + height ∗ dmax + (1|species) + (1|tag) 0.000 42.380growth ∼ height ∗ wd + sm + height ∗ lma + dmax + (1|species) + (1|tag) 0.000 61.905growth ∼ height ∗ wd + sm + height ∗ lma + (1|species) + (1|tag) 0.000 64.478growth ∼ height ∗ wd + sm + lma + dmax + (1|species) + (1|tag) 0.000 65.205
12
S2
TABLES
Table S6. Evaluated model forms, model weights, and ∆AIC values for the seedling survival analysis, for the top 30 models(out of 82). A full list of model forms can be found in table S2.
formulas weight ∆AIC
survive ∼ height ∗ wd + height ∗ sm + height ∗ dmax + height ∗ lma + (1|species) + (1|tag) 1.000 0.000survive ∼ height ∗ wd + height ∗ sm + height ∗ dmax + (1|species) + (1|tag) 0.000 81.243survive ∼ height ∗ wd + height ∗ sm + height ∗ dmax + lma + (1|species) + (1|tag) 0.000 83.232survive ∼ height + wd + height ∗ sm + height ∗ lma + dmax + (1|species) + (1|tag) 0.000 114.721survive ∼ height + wd + height ∗ sm + height ∗ lma + (1|species) + (1|tag) 0.000 120.806survive ∼ height + wd + height ∗ lma + height ∗ dmax + (1|species) + (1|tag) 0.000 205.165survive ∼ height ∗ wd + height ∗ lma + height ∗ dmax + (1|species) + (1|tag) 0.000 206.284survive ∼ height + wd + sm + height ∗ lma + height ∗ dmax + (1|species) + (1|tag) 0.000 207.109survive ∼ height ∗ wd + sm + height ∗ lma + height ∗ dmax + (1|species) + (1|tag) 0.000 208.230survive ∼ height + sm + height ∗ lma + height ∗ dmax + (1|species) + (1|tag) 0.000 231.883survive ∼ height ∗ wd + (1|species) + height ∗ dmax + (1|tag) 0.000 279.614survive ∼ height ∗ wd + sm + height ∗ dmax + (1|species) + (1|tag) 0.000 281.542survive ∼ height ∗ wd + lma + height ∗ dmax + (1|species) + (1|tag) 0.000 281.589survive ∼ height ∗ wd + sm + lma + height ∗ dmax + (1|species) + (1|tag) 0.000 283.529survive ∼ height + lma + height ∗ dmax + (1|species) + (1|tag) 0.000 314.015survive ∼ height + wd + height ∗ lma + dmax + (1|species) + (1|tag) 0.000 517.184survive ∼ height + wd + sm + height ∗ lma + dmax + (1|species) + (1|tag) 0.000 519.107survive ∼ height + wd + height ∗ lma + (1|species) + (1|tag) 0.000 525.973survive ∼ height + wd + sm + height ∗ lma + (1|species) + (1|tag) 0.000 526.077survive ∼ height ∗ wd + lma + (1|species) + (1|tag) 0.000 526.516survive ∼ height ∗ lma + dmax + (1|species) + (1|tag) 0.000 542.056survive ∼ height + sm + height ∗ lma + dmax + (1|species) + (1|tag) 0.000 543.147survive ∼ height + lma + (1|species) + (1|tag) 0.000 549.065survive ∼ height ∗ lma + (1|species) + (1|tag) 0.000 549.670survive ∼ height + sm + height ∗ lma + (1|species) + (1|tag) 0.000 551.647survive ∼ 1 + (1|species) + (1|tag) 0.000 2721.140survive ∼ height + wd + height ∗ sm + height ∗ lma + height ∗ dmax + (1|species) + (1|tag) 0.000 4640.336survive ∼ height + wd + height ∗ sm + lma + height ∗ dmax + (1|species) + (1|tag) 0.000 4752.207survive ∼ height + wd + height ∗ sm + height ∗ dmax + (1|species) + (1|tag) 0.000 4763.285survive ∼ height + wd + height ∗ sm + lma + dmax + (1|species) + (1|tag) 0.000 4820.751
13
S2
TABLES
Table S7. Evaluated model forms, model weights, and ∆AIC values for the tree growth analysis, for the top 30 models (outof 82). A full list of model forms can be found in table S2. The variable BA refers to basal area (mm2).
formulas weight ∆AICgrowth ∼ (BA) ∗ wd + (BA) ∗ sm + (BA) ∗ lma + BA ∗ dmax + (1|species) + (1|tag) 1.000 0.000growth ∼ (BA) ∗ wd + (BA) ∗ sm + lma + BA ∗ dmax + (1|species) + (1|tag) 0.000 722.011growth ∼ (BA) ∗ wd + (BA) ∗ sm + BA ∗ dmax + (1|species) + (1|tag) 0.000 727.452growth ∼ (BA) ∗ wd + (BA) ∗ sm + (BA) ∗ lma + dmax + (1|species) + (1|tag) 0.000 1434.112growth ∼ (BA) ∗ wd + (BA) ∗ sm + (BA) ∗ lma + (1|species) + (1|tag) 0.000 1455.667growth ∼ (BA) ∗ wd + sm + (BA) ∗ lma + BA ∗ dmax + (1|species) + (1|tag) 0.000 1732.065growth ∼ (BA) ∗ wd + (BA) ∗ lma + BA ∗ dmax + (1|species) + (1|tag) 0.000 1745.412growth ∼ (BA) + wd + (BA) ∗ sm + (BA) ∗ lma + BA ∗ dmax + (1|species) + (1|tag) 0.000 1809.328growth ∼ (BA) ∗ sm + (BA) ∗ lma + BA ∗ dmax + (1|species) + (1|tag) 0.000 1857.663growth ∼ (BA) ∗ wd + sm + lma + BA ∗ dmax + (1|species) + (1|tag) 0.000 2075.792growth ∼ (BA) ∗ wd + sm + BA ∗ dmax + (1|species) + (1|tag) 0.000 2080.841growth ∼ (BA) ∗ wd + lma + BA ∗ dmax + (1|species) + (1|tag) 0.000 2088.759growth ∼ (BA) ∗ wd + BA ∗ dmax + (1|species) + (1|tag) 0.000 2094.665growth ∼ (BA) + wd + (BA) ∗ sm + lma + BA ∗ dmax + (1|species) + (1|tag) 0.000 2680.176growth ∼ (BA) + wd + (BA) ∗ sm + BA ∗ dmax + (1|species) + (1|tag) 0.000 2685.318growth ∼ (BA) ∗ sm + lma + BA ∗ dmax + (1|species) + (1|tag) 0.000 2730.689growth ∼ (BA) ∗ sm + BA ∗ dmax + (1|species) + (1|tag) 0.000 2736.455growth ∼ (BA) ∗ wd + (BA) ∗ sm + lma + dmax + (1|species) + (1|tag) 0.000 2785.345growth ∼ (BA) ∗ wd + (BA) ∗ sm + dmax + (1|species) + (1|tag) 0.000 2791.423growth ∼ (BA) ∗ wd + (BA) ∗ sm + lma + (1|species) + (1|tag) 0.000 2807.186growth ∼ (BA) ∗ wd + (BA) ∗ sm + (1|species) + (1|tag) 0.000 2812.767growth ∼ (BA) + wd + sm + (BA) ∗ lma + BA ∗ dmax + (1|species) + (1|tag) 0.000 3643.162growth ∼ (BA) + wd + (BA) ∗ lma + BA ∗ dmax + (1|species) + (1|tag) 0.000 3656.590growth ∼ (BA) + sm + (BA) ∗ lma + BA ∗ dmax + (1|species) + (1|tag) 0.000 3689.670growth ∼ (BA) ∗ lma + BA ∗ dmax + (1|species) + (1|tag) 0.000 3707.314growth ∼ (BA) + wd + sm + lma + BA ∗ dmax + (1|species) + (1|tag) 0.000 4086.731growth ∼ (BA) + wd + sm + BA ∗ dmax + (1|species) + (1|tag) 0.000 4091.493growth ∼ (BA) + wd + lma + BA ∗ dmax + (1|species) + (1|tag) 0.000 4099.658growth ∼ (BA) + wd + BA ∗ dmax + (1|species) + (1|tag) 0.000 4105.179growth ∼ (BA) + sm + lma + BA ∗ dmax + (1|species) + (1|tag) 0.000 4135.374
14
S2
TABLES
Table S8. Evaluated model forms, model weights, and ∆AIC values for the tree survival analysis, for the top 30 models (outof 82). A full list of model forms can be found in table S2.
formulas weight ∆AIC
surv ∼ dbh ∗ wd + dbh ∗ sm + dbh ∗ lma + dbh ∗ dmax + (1|species) + (1|tag) 0.997 0.000surv ∼ dbh ∗ wd + dbh ∗ sm + lma + dbh ∗ dmax + (1|species) + (1|tag) 0.003 11.965surv ∼ dbh + wd + dbh ∗ sm + dbh ∗ lma + dbh ∗ dmax + (1|species) + (1|tag) 0.000 75.464surv ∼ dbh + wd + dbh ∗ sm + lma + dbh ∗ dmax + (1|species) + (1|tag) 0.000 83.261surv ∼ dbh ∗ wd + sm + dbh ∗ lma + dbh ∗ dmax + (1|species) + (1|tag) 0.000 155.016surv ∼ dbh ∗ wd + sm + lma + dbh ∗ dmax + (1|species) + (1|tag) 0.000 167.228surv ∼ dbh + wd + sm + dbh ∗ lma + dbh ∗ dmax + (1|species) + (1|tag) 0.000 274.457surv ∼ dbh + wd + sm + lma + dbh ∗ dmax + (1|species) + (1|tag) 0.000 279.971surv ∼ dbh ∗ wd + dbh ∗ sm + dbh ∗ dmax + (1|species) + (1|tag) 0.000 606.679surv ∼ dbh + wd + dbh ∗ sm + dbh ∗ dmax + (1|species) + (1|tag) 0.000 685.003surv ∼ dbh ∗ wd + sm + dbh ∗ dmax + (1|species) + (1|tag) 0.000 771.537surv ∼ dbh + wd + sm + dbh ∗ dmax + (1|species) + (1|tag) 0.000 898.364surv ∼ dbh ∗ wd + dbh ∗ sm + lma + (1|species) + (1|tag) 0.000 1003.192surv ∼ dbh ∗ wd + dbh ∗ sm + dbh ∗ lma + (1|species) + (1|tag) 0.000 1003.911surv ∼ dbh ∗ wd + dbh ∗ sm + lma + dmax + (1|species) + (1|tag) 0.000 1004.600surv ∼ dbh ∗ wd + dbh ∗ sm + dbh ∗ lma + dmax + (1|species) + (1|tag) 0.000 1005.322surv ∼ dbh ∗ wd + sm + lma + (1|species) + (1|tag) 0.000 1140.664surv ∼ dbh ∗ wd + sm + dbh ∗ lma + (1|species) + (1|tag) 0.000 1140.977surv ∼ dbh ∗ wd + sm + lma + dmax + (1|species) + (1|tag) 0.000 1142.302surv ∼ dbh ∗ wd + sm + dbh ∗ lma + dmax + (1|species) + (1|tag) 0.000 1142.616surv ∼ dbh + wd + dbh ∗ sm + dbh ∗ lma + (1|species) + (1|tag) 0.000 1198.664surv ∼ dbh + wd + dbh ∗ sm + dbh ∗ lma + dmax + (1|species) + (1|tag) 0.000 1200.094surv ∼ dbh + wd + dbh ∗ sm + lma + (1|species) + (1|tag) 0.000 1216.958surv ∼ dbh + wd + dbh ∗ sm + lma + dmax + (1|species) + (1|tag) 0.000 1218.383surv ∼ dbh + wd + sm + dbh ∗ lma + (1|species) + (1|tag) 0.000 1366.128surv ∼ dbh + wd + sm + dbh ∗ lma + dmax + (1|species) + (1|tag) 0.000 1367.800surv ∼ dbh + wd + sm + lma + (1|species) + (1|tag) 0.000 1385.911surv ∼ dbh + wd + sm + lma + dmax + (1|species) + (1|tag) 0.000 1387.589surv ∼ dbh ∗ wd + dbh ∗ sm + (1|species) + (1|tag) 0.000 1589.261surv ∼ dbh ∗ wd + dbh ∗ sm + dmax + (1|species) + (1|tag) 0.000 1590.885
15
S2
TABLES
Table S9. Evaluated model forms, model weights, and ∆AIC values for the reproduction analysis, for the top 30 models (outof 82). A full list of model forms can be found in table S2.
formulas weight ∆AIC
repr ∼ dbh ∗ sm + dbh ∗ dmax + lma + (1|species) + (1|tag) 0.232 0.000repr ∼ dbh ∗ sm + dbh ∗ dmax + dbh ∗ lma + (1|species) + (1|tag) 0.158 0.775repr ∼ dbh ∗ wd + dbh ∗ sm + dbh ∗ lma + dbh ∗ dmax + (1|species) + (1|tag) 0.149 0.895repr ∼ dbh + wd + dbh ∗ sm + dbh ∗ dmax + lma + (1|species) + (1|tag) 0.104 1.601repr ∼ dbh ∗ wd + dbh ∗ sm + lma + dbh ∗ dmax + (1|species) + (1|tag) 0.097 1.741repr ∼ dbh ∗ sm + dbh ∗ dmax + (1|species) + (1|tag) 0.077 2.222repr ∼ dbh + wd + dbh ∗ sm + dbh ∗ dmax + dbh ∗ lma + (1|species) + (1|tag) 0.071 2.361repr ∼ dbh + wd + dbh ∗ sm + dbh ∗ dmax + (1|species) + (1|tag) 0.055 2.878repr ∼ dbh ∗ wd + dbh ∗ sm + dbh ∗ dmax + (1|species) + (1|tag) 0.055 2.885repr ∼ dbh ∗ lma + dbh ∗ dmax + (1|species) + (1|tag) 0.000 13.651repr ∼ dbh ∗ dmax + lma + (1|species) + (1|tag) 0.000 13.751repr ∼ dbh + wd + dbh ∗ lma + dbh ∗ dmax + (1|species) + (1|tag) 0.000 15.238repr ∼ dbh + wd + lma + dbh ∗ dmax + (1|species) + (1|tag) 0.000 15.352repr ∼ dbh + sm + dbh ∗ lma + dbh ∗ dmax + (1|species) + (1|tag) 0.000 15.601repr ∼ dbh + sm + lma + dbh ∗ dmax + (1|species) + (1|tag) 0.000 15.680repr ∼ dbh ∗ wd + dbh ∗ lma + dbh ∗ dmax + (1|species) + (1|tag) 0.000 16.003repr ∼ dbh ∗ dmax + (1|species) + (1|tag) 0.000 16.543repr ∼ dbh ∗ wd + lma + dbh ∗ dmax + (1|species) + (1|tag) 0.000 17.059repr ∼ dbh + wd + dbh ∗ dmax + (1|species) + (1|tag) 0.000 17.228repr ∼ dbh + wd + sm + dbh ∗ lma + dbh ∗ dmax + (1|species) + (1|tag) 0.000 17.231repr ∼ dbh + wd + sm + lma + dbh ∗ dmax + (1|species) + (1|tag) 0.000 17.335repr ∼ dbh ∗ wd + sm + dbh ∗ lma + dbh ∗ dmax + (1|species) + (1|tag) 0.000 17.999repr ∼ dbh + sm + dbh ∗ dmax + (1|species) + (1|tag) 0.000 18.538repr ∼ dbh ∗ wd + dbh ∗ dmax + (1|species) + (1|tag) 0.000 18.896repr ∼ dbh ∗ wd + sm + lma + dbh ∗ dmax + (1|species) + (1|tag) 0.000 19.044repr ∼ dbh + wd + sm + dbh ∗ dmax + (1|species) + (1|tag) 0.000 19.204repr ∼ dbh ∗ wd + sm + dbh ∗ dmax + (1|species) + (1|tag) 0.000 20.871repr ∼ dbh ∗ wd + dbh ∗ sm + dbh ∗ lma + dmax + (1|species) + (1|tag) 0.000 260.580repr ∼ dbh + wd + dbh ∗ sm + dmax + dbh ∗ lma + (1|species) + (1|tag) 0.000 263.641repr ∼ dbh ∗ wd + dbh ∗ sm + lma + dmax + (1|species) + (1|tag) 0.000 264.562
16
S2 TABLES
Table S10. R2 values for average models including only single traits for seedproduction and seedling establishment.
dmax lma sm wd
Fecundity 0.01 0.04 0.63 0.00Establishment 0.20 0.05 0.33 0.10
17
S2
TABLES
Table S11. R2 values at different sizes for average models including only single traits for seedlings. Dmax, LMA, SM and WDrepresent adult stature, leaf mass per area, seed mass and wood density.
Dmax LMA SM WD
Size (mm hght) Growth Survival Growth Survival Growth Survival Growth Survival200 0.283 0.283 0.002 0.022 0.103 0.053 0.151 0.219260 0.261 0.261 0.002 0.032 0.114 0.036 0.14 0.263320 0.234 0.234 0.003 0.041 0.104 0.021 0.142 0.305380 0.126 0.126 0 0.018 0.086 0.002 0.173 0.357440 0.1 0.1 0.001 0.017 0.081 0.01 0.174 0.385500 0.08 0.08 0.002 0.016 0.074 0.021 0.168 0.405560 0.063 0.063 0.003 0.014 0.066 0.036 0.161 0.418620 0.064 0.064 0.005 0.015 0.058 0.042 0.152 0.42680 0.054 0.054 0.008 0.014 0.037 0.06 0.141 0.42740 0.041 0.041 0.01 0.012 0.031 0.074 0.124 0.431800 0.04 0.04 0.009 0.01 0.026 0.082 0.116 0.429860 0.032 0.032 0.008 0.008 0.019 0.092 0.101 0.423920 0.027 0.027 0.01 0.008 0.012 0.1 0.086 0.42980 0.028 0.028 0.011 0.02 0.008 0.094 0.068 0.4081040 0.025 0.025 0.015 0.019 0.004 0.105 0.06 0.41100 0.024 0.024 0.007 0.015 0.008 0.11 0.067 0.4071160 0.022 0.022 0.011 0.016 0.006 0.109 0.042 0.41220 0.019 0.019 0.004 0.015 0.005 0.104 0.027 0.3931280 0.021 0.021 0.008 0.019 0.004 0.097 0.039 0.3921340 0.03 0.03 0.003 0.011 0.003 0.087 0.009 0.4511400 0.028 0.028 0.001 0.019 0.002 0.081 0.001 0.5081460 0.032 0.032 0.011 0.021 0.011 0.068 0.004 0.5021490 0.036 0.036 0.01 0.022 0.013 0.061 0.01 0.495
18
S2
TABLES
Table S12. R2 values at different sizes from average models for growth, survival and reproduction including only single traitsfor trees (>1cm dbh). Dmax, LMA, SM and WD represent adult stature, leaf mass per area, seed mass and wood density.
Dmax LMA SM WD
Size (mm dbh) Growth Survival Repr. Growth Survival Repr. Growth Survival Repr. Growth Survival Repr.10 0.1108 0.0172 0.4429 0.0258 3e-04 0.4525 0.0564 0.1666 0.0269 0.3446 0.2765 0.384240 0.1187 0.0062 0.4429 0.0091 4e-04 0.4525 0.0863 0.0938 0.0269 0.3596 0.2609 0.384270 0.0701 0.0344 0.5423 0.0206 0.0068 0.3898 0.0799 0.0544 0.0035 0.4181 0.2547 0.1874100 0.0012 6e-04 0.5808 0.0257 0.0191 0.433 0.0532 0.0411 0.003 0.4112 0.2253 0.426130 0.025 0.0024 0.3845 0.0196 0.0214 0.1971 0.0208 0.0143 1e-04 0.371 0.193 0.1509160 0.0139 0.0164 0.3976 0.0046 0.018 0.0618 0.0184 0.0048 0.0095 0.2693 0.1712 0.0534190 0.0035 0.1314 0.2418 0.011 0.008 0.0423 0.0343 7e-04 7e-04 0.2859 0.1176 0.0175220 0.0096 0.1105 0.195 8e-04 0.0028 0.0611 0.0812 0.0048 0.0082 0.3293 0.0922 0.0028250 0.0418 0.2496 0.2374 0.0216 0.0025 0.0693 0.0716 1e-04 0.009 0.2707 0.0922 0.0046
19
S3 FIGURES
S3 Figures
20
S3 FIGURES
●
−2
−1
0
1
2Wood Density
●
●
●●
●
●
●
●
● ●
●
●
●
●
●
●
●
●●
●
●
●
●●
●
−3−2−1
0123 Seed Mass
●
●
●●
●
●
●
●●●
●
●●
●
●●
●
●
●
●
●
●
−2
−1
0
1
2
Leaf Mass per Area
●
●
●
●
●●●
●●●
●
Fec
undi
ty
Rep
rodu
ctio
n
See
d es
tabl
ishm
ent
Sdl
. gro
wth
Sdl
. sur
viva
l
Tree
gro
wth
Tree
Sur
viva
l
−3
−2
−1
0
1
2Adult Stature
Sta
ndar
dize
d tr
ait u
nits
(Vital rates)
Figure S1. Boxplots showing the distribution of trait values (in standardizedunits) for each vital rate analysis. Each box plot within a panel corresponds tothe distribution of trait values for the species used within that specific analysis.
21
S3 FIGURES
Ful
l
SM
WD
LMA
Dm
ax
A
R2 fo
r se
ed p
rodu
ctio
n
0.0
0.2
0.4
0.6
0.8
1.0
0.64
0.54
0.04
0.09
0.06
Ful
l
SM
WD
LMA
Dm
ax
B
0.0
0.2
0.4
0.6
0.8
1.0
0.55
0.54 0
0.05
0.04
Figure S2. R2 values for interspecific variation in seed production from averagedmodels for a single trait and for the full averaged model for all four traits,compared between seed production estimates obtained through inverse modeling(A, 18 species) or by estimating seed production per unit basal area as describedin the main text (B, 38 species).
22
S3 FIGURES
Figure S3. Plots of residuals of the full averaged model against size for treegrowth (A,B), tree survival (C,D), seedling growth (E,F), seedling survival (G,H,and reproduction (I). Black dots indicate residual values from each model-averaged mixed-effect model (predicted - observed), while the red line is themoving average. The first column (A,C,E,G,I) shows residuals for models fit tothe full datasets, while the second column (B,D,F,H) shows residuals for modelsfit to datasets truncated above at 500 mm diameter (B,D) or 2500 mm height(F, H) to avoid non-linearities with size.
23
S3
FIG
URES
0 100 200 300 400 500
0.0
0.4
0.8
Alchornea costaricensi
Dbh (mm)
Rep
rodu
ctio
n pr
obab
ility
Dbh (mm)
Rep
rodu
ctio
n pr
obab
ility
0 100 200 300 400 500
0.0
0.4
0.8
Alseis blackiana
Dbh (mm)
Rep
rodu
ctio
n pr
obab
ility
Dbh (mm)
Rep
rodu
ctio
n pr
obab
ility
0 100 200 300 400 500
0.0
0.4
0.8
Apeiba membranacea
Dbh (mm)
Rep
rodu
ctio
n pr
obab
ility
Dbh (mm)
Rep
rodu
ctio
n pr
obab
ility
0 100 200 300 400 500
0.0
0.4
0.8
Apeiba tibourbou
Dbh (mm)
Rep
rodu
ctio
n pr
obab
ility
Dbh (mm)
Rep
rodu
ctio
n pr
obab
ility
0 100 200 300 400 500
0.0
0.4
0.8
Aspidosperma spruceanum
Dbh (mm)
Rep
rodu
ctio
n pr
obab
ility
Dbh (mm)
Rep
rodu
ctio
n pr
obab
ility
0 100 200 300 400 500
0.0
0.4
0.8
Astronium graveolens
Dbh (mm)
Rep
rodu
ctio
n pr
obab
ility
Dbh (mm)
Rep
rodu
ctio
n pr
obab
ility
0 100 200 300 400 500
0.0
0.4
0.8
Beilschmiedi pendula
Dbh (mm)
Rep
rodu
ctio
n pr
obab
ility
Dbh (mm)
Rep
rodu
ctio
n pr
obab
ility
0 100 200 300 400 500
0.0
0.4
0.8
Brosimum alicastrum
Dbh (mm)
Rep
rodu
ctio
n pr
obab
ility
Dbh (mm)
Rep
rodu
ctio
n pr
obab
ility
0 100 200 300 400 500
0.0
0.4
0.8
Casearia aculeata
Dbh (mm)
Rep
rodu
ctio
n pr
obab
ility
Dbh (mm)
Rep
rodu
ctio
n pr
obab
ility
0 100 200 300 400 500
0.0
0.4
0.8
Cassipourea elliptica
Dbh (mm)
Rep
rodu
ctio
n pr
obab
ility
Dbh (mm)
Rep
rodu
ctio
n pr
obab
ility
0 100 200 300 400 500
0.0
0.4
0.8
Chrysophyllu argenteum
Dbh (mm)R
epro
duct
ion
prob
abili
ty
Dbh (mm)
Rep
rodu
ctio
n pr
obab
ility
0 100 200 300 400 500
0.0
0.4
0.8
Chrysophyllu cainito
Dbh (mm)
Rep
rodu
ctio
n pr
obab
ility
Dbh (mm)
Rep
rodu
ctio
n pr
obab
ility
0 100 200 300 400 500
0.0
0.4
0.8
Cordia alliodora
Dbh (mm)
Rep
rodu
ctio
n pr
obab
ility
Dbh (mm)
Rep
rodu
ctio
n pr
obab
ility
0 100 200 300 400 500
0.0
0.4
0.8
Cordia bicolor
Dbh (mm)
Rep
rodu
ctio
n pr
obab
ility
Dbh (mm)
Rep
rodu
ctio
n pr
obab
ility
0 100 200 300 400 500
0.0
0.4
0.8
Coussarea curvigemmia
Dbh (mm)
Rep
rodu
ctio
n pr
obab
ility
Dbh (mm)
Rep
rodu
ctio
n pr
obab
ility
0 100 200 300 400 500
0.0
0.4
0.8
Desmopsis panamensis
Dbh (mm)
Rep
rodu
ctio
n pr
obab
ility
Dbh (mm)
Rep
rodu
ctio
n pr
obab
ility
Figure S4. Observed and fitted proportion of individuals that are reproductive as a function of tree diameter (mm) for eachspecies. Black dots show observations, red dashed lines show species-specific moving averages (from a Generalized AdditiveModel using a loess smoother), blue dashed lines show the fitted trait based average model, and grey lines show a fitted modelbased only on size (not on traits, thus identical for all species).
24
S3
FIG
URES
0 100 200 300 400 500
0.0
0.4
0.8
Dipteryx oleifera
Dbh (mm)
Rep
rodu
ctio
n pr
obab
ility
Dbh (mm)
Rep
rodu
ctio
n pr
obab
ility
0 100 200 300 400 500
0.0
0.4
0.8
Eugenia oerstediana
Dbh (mm)
Rep
rodu
ctio
n pr
obab
ility
Dbh (mm)
Rep
rodu
ctio
n pr
obab
ility
0 100 200 300 400 500
0.0
0.4
0.8
Faramea occidentalis
Dbh (mm)
Rep
rodu
ctio
n pr
obab
ility
Dbh (mm)
Rep
rodu
ctio
n pr
obab
ility
0 100 200 300 400 500
0.0
0.4
0.8
Garcinia intermedia
Dbh (mm)
Rep
rodu
ctio
n pr
obab
ility
Dbh (mm)
Rep
rodu
ctio
n pr
obab
ility
0 100 200 300 400 500
0.0
0.4
0.8
Guarea guidonia
Dbh (mm)
Rep
rodu
ctio
n pr
obab
ility
Dbh (mm)
Rep
rodu
ctio
n pr
obab
ility
0 100 200 300 400 500
0.0
0.4
0.8
Guatteria dumetorum
Dbh (mm)
Rep
rodu
ctio
n pr
obab
ility
Dbh (mm)
Rep
rodu
ctio
n pr
obab
ility
0 100 200 300 400 500
0.0
0.4
0.8
Gustavia superba
Dbh (mm)
Rep
rodu
ctio
n pr
obab
ility
Dbh (mm)
Rep
rodu
ctio
n pr
obab
ility
0 100 200 300 400 500
0.0
0.4
0.8
Hasseltia floribunda
Dbh (mm)
Rep
rodu
ctio
n pr
obab
ility
Dbh (mm)
Rep
rodu
ctio
n pr
obab
ility
0 100 200 300 400 500
0.0
0.4
0.8
Heisteria concinna
Dbh (mm)
Rep
rodu
ctio
n pr
obab
ility
Dbh (mm)
Rep
rodu
ctio
n pr
obab
ility
0 100 200 300 400 500
0.0
0.4
0.8
Hirtella triandra
Dbh (mm)
Rep
rodu
ctio
n pr
obab
ility
Dbh (mm)
Rep
rodu
ctio
n pr
obab
ility
0 100 200 300 400 500
0.0
0.4
0.8
Hieronyma alchorneoide
Dbh (mm)R
epro
duct
ion
prob
abili
ty
Dbh (mm)
Rep
rodu
ctio
n pr
obab
ility
0 100 200 300 400 500
0.0
0.4
0.8
Inga marginata
Dbh (mm)
Rep
rodu
ctio
n pr
obab
ility
Dbh (mm)
Rep
rodu
ctio
n pr
obab
ility
0 100 200 300 400 500
0.0
0.4
0.8
Jacaranda copaia
Dbh (mm)
Rep
rodu
ctio
n pr
obab
ility
Dbh (mm)
Rep
rodu
ctio
n pr
obab
ility
0 100 200 300 400 500
0.0
0.4
0.8
Laetia thamnia
Dbh (mm)
Rep
rodu
ctio
n pr
obab
ility
Dbh (mm)
Rep
rodu
ctio
n pr
obab
ility
0 100 200 300 400 500
0.0
0.4
0.8
Lonchocarpus heptaphyllus
Dbh (mm)
Rep
rodu
ctio
n pr
obab
ility
Dbh (mm)
Rep
rodu
ctio
n pr
obab
ility
0 100 200 300 400 500
0.0
0.4
0.8
Luehea seemannii
Dbh (mm)
Rep
rodu
ctio
n pr
obab
ility
Dbh (mm)
Rep
rodu
ctio
n pr
obab
ility
Figure S4. Continued.
25
S3
FIG
URES
0 100 200 300 400 500
0.0
0.4
0.8
Mosannona garwoodii
Dbh (mm)
Rep
rodu
ctio
n pr
obab
ility
Dbh (mm)
Rep
rodu
ctio
n pr
obab
ility
0 100 200 300 400 500
0.0
0.4
0.8
Mouriri myrtilloides
Dbh (mm)
Rep
rodu
ctio
n pr
obab
ility
Dbh (mm)
Rep
rodu
ctio
n pr
obab
ility
0 100 200 300 400 500
0.0
0.4
0.8
Platypodium elegans
Dbh (mm)
Rep
rodu
ctio
n pr
obab
ility
Dbh (mm)
Rep
rodu
ctio
n pr
obab
ility
0 100 200 300 400 500
0.0
0.4
0.8
Poulsenia armata
Dbh (mm)
Rep
rodu
ctio
n pr
obab
ility
Dbh (mm)
Rep
rodu
ctio
n pr
obab
ility
0 100 200 300 400 500
0.0
0.4
0.8
Pouteria reticulata
Dbh (mm)
Rep
rodu
ctio
n pr
obab
ility
Dbh (mm)
Rep
rodu
ctio
n pr
obab
ility
0 100 200 300 400 500
0.0
0.4
0.8
Prioria copaifera
Dbh (mm)
Rep
rodu
ctio
n pr
obab
ility
Dbh (mm)
Rep
rodu
ctio
n pr
obab
ility
0 100 200 300 400 500
0.0
0.4
0.8
Protium panamense
Dbh (mm)
Rep
rodu
ctio
n pr
obab
ility
Dbh (mm)
Rep
rodu
ctio
n pr
obab
ility
0 100 200 300 400 500
0.0
0.4
0.8
Protium tenuifolium
Dbh (mm)
Rep
rodu
ctio
n pr
obab
ility
Dbh (mm)
Rep
rodu
ctio
n pr
obab
ility
0 100 200 300 400 500
0.0
0.4
0.8
Pseudobombax septenatum
Dbh (mm)
Rep
rodu
ctio
n pr
obab
ility
Dbh (mm)
Rep
rodu
ctio
n pr
obab
ility
0 100 200 300 400 500
0.0
0.4
0.8
Pterocarpus rohrii
Dbh (mm)
Rep
rodu
ctio
n pr
obab
ility
Dbh (mm)
Rep
rodu
ctio
n pr
obab
ility
0 100 200 300 400 500
0.0
0.4
0.8
Quararibea asterolepis
Dbh (mm)R
epro
duct
ion
prob
abili
ty
Dbh (mm)
Rep
rodu
ctio
n pr
obab
ility
0 100 200 300 400 500
0.0
0.4
0.8
Randia armata
Dbh (mm)
Rep
rodu
ctio
n pr
obab
ility
Dbh (mm)
Rep
rodu
ctio
n pr
obab
ility
0 100 200 300 400 500
0.0
0.4
0.8
Simarouba amara
Dbh (mm)
Rep
rodu
ctio
n pr
obab
ility
Dbh (mm)
Rep
rodu
ctio
n pr
obab
ility
0 100 200 300 400 500
0.0
0.4
0.8
Sloanea terniflora
Dbh (mm)
Rep
rodu
ctio
n pr
obab
ility
Dbh (mm)
Rep
rodu
ctio
n pr
obab
ility
0 100 200 300 400 500
0.0
0.4
0.8
Tabebuia rosea
Dbh (mm)
Rep
rodu
ctio
n pr
obab
ility
Dbh (mm)
Rep
rodu
ctio
n pr
obab
ility
0 100 200 300 400 500
0.0
0.4
0.8
Tabernaemont arborea
Dbh (mm)
Rep
rodu
ctio
n pr
obab
ility
Dbh (mm)
Rep
rodu
ctio
n pr
obab
ility
Figure S4. Continued.
26
S3
FIG
URES
0 100 200 300 400 500
0.0
0.4
0.8
Talisia nervosa
Dbh (mm)
Rep
rodu
ctio
n pr
obab
ility
Dbh (mm)
Rep
rodu
ctio
n pr
obab
ility
0 100 200 300 400 500
0.0
0.4
0.8
Terminalia amazonia
Dbh (mm)
Rep
rodu
ctio
n pr
obab
ility
Dbh (mm)
Rep
rodu
ctio
n pr
obab
ility
0 100 200 300 400 500
0.0
0.4
0.8
Terminalia oblonga
Dbh (mm)
Rep
rodu
ctio
n pr
obab
ility
Dbh (mm)
Rep
rodu
ctio
n pr
obab
ility
0 100 200 300 400 500
0.0
0.4
0.8
Tetragastris panamensis
Dbh (mm)
Rep
rodu
ctio
n pr
obab
ility
Dbh (mm)
Rep
rodu
ctio
n pr
obab
ility
0 100 200 300 400 500
0.0
0.4
0.8
Trattinnicki aspera
Dbh (mm)
Rep
rodu
ctio
n pr
obab
ility
Dbh (mm)
Rep
rodu
ctio
n pr
obab
ility
0 100 200 300 400 500
0.0
0.4
0.8
Trichilia pallida
Dbh (mm)
Rep
rodu
ctio
n pr
obab
ility
Dbh (mm)
Rep
rodu
ctio
n pr
obab
ility
0 100 200 300 400 500
0.0
0.4
0.8
Trichilia tuberculata
Dbh (mm)
Rep
rodu
ctio
n pr
obab
ility
Dbh (mm)
Rep
rodu
ctio
n pr
obab
ility
0 100 200 300 400 500
0.0
0.4
0.8
Triplaris cumingiana
Dbh (mm)
Rep
rodu
ctio
n pr
obab
ility
Dbh (mm)
Rep
rodu
ctio
n pr
obab
ility
0 100 200 300 400 500
0.0
0.4
0.8
Unonopsis pittieri
Dbh (mm)
Rep
rodu
ctio
n pr
obab
ility
Dbh (mm)
Rep
rodu
ctio
n pr
obab
ility
0 100 200 300 400 500
0.0
0.4
0.8
Virola sebifera
Dbh (mm)
Rep
rodu
ctio
n pr
obab
ility
Dbh (mm)
Rep
rodu
ctio
n pr
obab
ility
0 100 200 300 400 500
0.0
0.4
0.8
Vochysia ferruginea
Dbh (mm)R
epro
duct
ion
prob
abili
ty
Dbh (mm)
Rep
rodu
ctio
n pr
obab
ility
0 100 200 300 400 500
0.0
0.4
0.8
Xylopia macrantha
Dbh (mm)
Rep
rodu
ctio
n pr
obab
ility
Dbh (mm)
Rep
rodu
ctio
n pr
obab
ility
Figure S4. Continued.
27
S3
FIG
URES
500 1000 1500 2000 2500
−20
00
100
Alchornea costaricensi
Height (mm)
Year
ly g
row
th (
mm
hei
ght)
Height (mm)
Year
ly g
row
th (
mm
hei
ght)
500 1000 1500 2000 2500
−20
00
100
Alibertia edulis
Height (mm)
Year
ly g
row
th (
mm
hei
ght)
Height (mm)
Year
ly g
row
th (
mm
hei
ght)
500 1000 1500 2000 2500
−20
00
100
Alseis blackiana
Height (mm)
Year
ly g
row
th (
mm
hei
ght)
Height (mm)
Year
ly g
row
th (
mm
hei
ght)
500 1000 1500 2000 2500
−20
00
100
Anacardium excelsum
Height (mm)
Year
ly g
row
th (
mm
hei
ght)
Height (mm)
Year
ly g
row
th (
mm
hei
ght)
500 1000 1500 2000 2500
−20
00
100
Annona acuminata
Height (mm)
Year
ly g
row
th (
mm
hei
ght)
Height (mm)
Year
ly g
row
th (
mm
hei
ght)
500 1000 1500 2000 2500
−20
00
100
Annona spraguei
Height (mm)
Year
ly g
row
th (
mm
hei
ght)
Height (mm)
Year
ly g
row
th (
mm
hei
ght)
500 1000 1500 2000 2500
−20
00
100
Apeiba membranacea
Height (mm)
Year
ly g
row
th (
mm
hei
ght)
Height (mm)
Year
ly g
row
th (
mm
hei
ght)
500 1000 1500 2000 2500
−20
00
100
Aspidosperma spruceanum
Height (mm)
Year
ly g
row
th (
mm
hei
ght)
Height (mm)
Year
ly g
row
th (
mm
hei
ght)
500 1000 1500 2000 2500
−20
00
100
Beilschmiedi pendula
Height (mm)
Year
ly g
row
th (
mm
hei
ght)
Height (mm)
Year
ly g
row
th (
mm
hei
ght)
500 1000 1500 2000 2500
−20
00
100
Brosimum alicastrum
Height (mm)
Year
ly g
row
th (
mm
hei
ght)
Height (mm)
Year
ly g
row
th (
mm
hei
ght)
500 1000 1500 2000 2500
−20
00
100
Calophyllum longifolium
Height (mm)Ye
arly
gro
wth
(m
m h
eigh
t)
Height (mm)
Year
ly g
row
th (
mm
hei
ght)
500 1000 1500 2000 2500
−20
00
100
Capparis frondosa
Height (mm)
Year
ly g
row
th (
mm
hei
ght)
Height (mm)
Year
ly g
row
th (
mm
hei
ght)
500 1000 1500 2000 2500
−20
00
100
Cassipourea elliptica
Height (mm)
Year
ly g
row
th (
mm
hei
ght)
Height (mm)
Year
ly g
row
th (
mm
hei
ght)
500 1000 1500 2000 2500
−20
00
100
Chrysophyllu argenteum
Height (mm)
Year
ly g
row
th (
mm
hei
ght)
Height (mm)
Year
ly g
row
th (
mm
hei
ght)
500 1000 1500 2000 2500
−20
00
100
Chrysophyllu cainito
Height (mm)
Year
ly g
row
th (
mm
hei
ght)
Height (mm)
Year
ly g
row
th (
mm
hei
ght)
500 1000 1500 2000 2500
−20
00
100
Cordia alliodora
Height (mm)
Year
ly g
row
th (
mm
hei
ght)
Height (mm)
Year
ly g
row
th (
mm
hei
ght)
Figure S5. Observed and predicted annual seedling height growth (mm/yr) as a function of initial height (mm). Black dotsshow observations, red dashed lines show species-specific moving averages (from a Generalized Additive Model using a loesssmoother), blue dashed lines show the fitted trait based average model, and grey lines show a fitted model based only on size(not on traits, thus identical for all species).
28
S3
FIG
URES
500 1000 1500 2000 2500
−20
00
100
Cordia bicolor
Height (mm)
Year
ly g
row
th (
mm
hei
ght)
Height (mm)
Year
ly g
row
th (
mm
hei
ght)
500 1000 1500 2000 2500
−20
00
100
Coussarea curvigemmia
Height (mm)
Year
ly g
row
th (
mm
hei
ght)
Height (mm)
Year
ly g
row
th (
mm
hei
ght)
500 1000 1500 2000 2500
−20
00
100
Desmopsis panamensis
Height (mm)
Year
ly g
row
th (
mm
hei
ght)
Height (mm)
Year
ly g
row
th (
mm
hei
ght)
500 1000 1500 2000 2500
−20
00
100
Dipteryx oleifera
Height (mm)
Year
ly g
row
th (
mm
hei
ght)
Height (mm)
Year
ly g
row
th (
mm
hei
ght)
500 1000 1500 2000 2500
−20
00
100
Drypetes standleyi
Height (mm)
Year
ly g
row
th (
mm
hei
ght)
Height (mm)
Year
ly g
row
th (
mm
hei
ght)
500 1000 1500 2000 2500
−20
00
100
Eugenia coloradoensi
Height (mm)
Year
ly g
row
th (
mm
hei
ght)
Height (mm)
Year
ly g
row
th (
mm
hei
ght)
500 1000 1500 2000 2500
−20
00
100
Eugenia galalonensis
Height (mm)
Year
ly g
row
th (
mm
hei
ght)
Height (mm)
Year
ly g
row
th (
mm
hei
ght)
500 1000 1500 2000 2500
−20
00
100
Eugenia nesiotica
Height (mm)
Year
ly g
row
th (
mm
hei
ght)
Height (mm)
Year
ly g
row
th (
mm
hei
ght)
500 1000 1500 2000 2500
−20
00
100
Eugenia oerstediana
Height (mm)
Year
ly g
row
th (
mm
hei
ght)
Height (mm)
Year
ly g
row
th (
mm
hei
ght)
500 1000 1500 2000 2500
−20
00
100
Faramea occidentalis
Height (mm)
Year
ly g
row
th (
mm
hei
ght)
Height (mm)
Year
ly g
row
th (
mm
hei
ght)
500 1000 1500 2000 2500
−20
00
100
Garcinia intermedia
Height (mm)Ye
arly
gro
wth
(m
m h
eigh
t)
Height (mm)
Year
ly g
row
th (
mm
hei
ght)
500 1000 1500 2000 2500
−20
00
100
Guapira standleyana
Height (mm)
Year
ly g
row
th (
mm
hei
ght)
Height (mm)
Year
ly g
row
th (
mm
hei
ght)
500 1000 1500 2000 2500
−20
00
100
Guarea guidonia
Height (mm)
Year
ly g
row
th (
mm
hei
ght)
Height (mm)
Year
ly g
row
th (
mm
hei
ght)
500 1000 1500 2000 2500
−20
00
100
Guatteria dumetorum
Height (mm)
Year
ly g
row
th (
mm
hei
ght)
Height (mm)
Year
ly g
row
th (
mm
hei
ght)
500 1000 1500 2000 2500
−20
00
100
Gustavia superba
Height (mm)
Year
ly g
row
th (
mm
hei
ght)
Height (mm)
Year
ly g
row
th (
mm
hei
ght)
500 1000 1500 2000 2500
−20
00
100
Hamelia axillaris
Height (mm)
Year
ly g
row
th (
mm
hei
ght)
Height (mm)
Year
ly g
row
th (
mm
hei
ght)
Figure S5. Continued.
29
S3
FIG
URES
500 1000 1500 2000 2500
−20
00
100
Hampea appendiculat
Height (mm)
Year
ly g
row
th (
mm
hei
ght)
Height (mm)
Year
ly g
row
th (
mm
hei
ght)
500 1000 1500 2000 2500
−20
00
100
Heisteria concinna
Height (mm)
Year
ly g
row
th (
mm
hei
ght)
Height (mm)
Year
ly g
row
th (
mm
hei
ght)
500 1000 1500 2000 2500
−20
00
100
Hirtella triandra
Height (mm)
Year
ly g
row
th (
mm
hei
ght)
Height (mm)
Year
ly g
row
th (
mm
hei
ght)
500 1000 1500 2000 2500
−20
00
100
Hybanthus prunifolius
Height (mm)
Year
ly g
row
th (
mm
hei
ght)
Height (mm)
Year
ly g
row
th (
mm
hei
ght)
500 1000 1500 2000 2500
−20
00
100
Inga marginata
Height (mm)
Year
ly g
row
th (
mm
hei
ght)
Height (mm)
Year
ly g
row
th (
mm
hei
ght)
500 1000 1500 2000 2500
−20
00
100
Inga acuminata
Height (mm)
Year
ly g
row
th (
mm
hei
ght)
Height (mm)
Year
ly g
row
th (
mm
hei
ght)
500 1000 1500 2000 2500
−20
00
100
Inga thibaudiana
Height (mm)
Year
ly g
row
th (
mm
hei
ght)
Height (mm)
Year
ly g
row
th (
mm
hei
ght)
500 1000 1500 2000 2500
−20
00
100
Inga umbellifera
Height (mm)
Year
ly g
row
th (
mm
hei
ght)
Height (mm)
Year
ly g
row
th (
mm
hei
ght)
500 1000 1500 2000 2500
−20
00
100
Lacistema aggregatum
Height (mm)
Year
ly g
row
th (
mm
hei
ght)
Height (mm)
Year
ly g
row
th (
mm
hei
ght)
500 1000 1500 2000 2500
−20
00
100
Lacmellea panamensis
Height (mm)
Year
ly g
row
th (
mm
hei
ght)
Height (mm)
Year
ly g
row
th (
mm
hei
ght)
500 1000 1500 2000 2500
−20
00
100
Lonchocarpus heptaphyllus
Height (mm)Ye
arly
gro
wth
(m
m h
eigh
t)
Height (mm)
Year
ly g
row
th (
mm
hei
ght)
500 1000 1500 2000 2500
−20
00
100
Luehea seemannii
Height (mm)
Year
ly g
row
th (
mm
hei
ght)
Height (mm)
Year
ly g
row
th (
mm
hei
ght)
500 1000 1500 2000 2500
−20
00
100
Mouriri myrtilloides
Height (mm)
Year
ly g
row
th (
mm
hei
ght)
Height (mm)
Year
ly g
row
th (
mm
hei
ght)
500 1000 1500 2000 2500
−20
00
100
Ocotea cernua
Height (mm)
Year
ly g
row
th (
mm
hei
ght)
Height (mm)
Year
ly g
row
th (
mm
hei
ght)
500 1000 1500 2000 2500
−20
00
100
Ocotea puberula
Height (mm)
Year
ly g
row
th (
mm
hei
ght)
Height (mm)
Year
ly g
row
th (
mm
hei
ght)
500 1000 1500 2000 2500
−20
00
100
Trophis caucana
Height (mm)
Year
ly g
row
th (
mm
hei
ght)
Height (mm)
Year
ly g
row
th (
mm
hei
ght)
Figure S5. Continued.
30
S3
FIG
URES
500 1000 1500 2000 2500
−20
00
100
Ouratea lucens
Height (mm)
Year
ly g
row
th (
mm
hei
ght)
Height (mm)
Year
ly g
row
th (
mm
hei
ght)
500 1000 1500 2000 2500
−20
00
100
Palicourea guianensis
Height (mm)
Year
ly g
row
th (
mm
hei
ght)
Height (mm)
Year
ly g
row
th (
mm
hei
ght)
500 1000 1500 2000 2500
−20
00
100
Cinnamomum triplinerve
Height (mm)
Year
ly g
row
th (
mm
hei
ght)
Height (mm)
Year
ly g
row
th (
mm
hei
ght)
500 1000 1500 2000 2500
−20
00
100
Picramnia latifolia
Height (mm)
Year
ly g
row
th (
mm
hei
ght)
Height (mm)
Year
ly g
row
th (
mm
hei
ght)
500 1000 1500 2000 2500
−20
00
100
Platymiscium pinnatum
Height (mm)
Year
ly g
row
th (
mm
hei
ght)
Height (mm)
Year
ly g
row
th (
mm
hei
ght)
500 1000 1500 2000 2500
−20
00
100
Pouteria reticulata
Height (mm)
Year
ly g
row
th (
mm
hei
ght)
Height (mm)
Year
ly g
row
th (
mm
hei
ght)
500 1000 1500 2000 2500
−20
00
100
Prioria copaifera
Height (mm)
Year
ly g
row
th (
mm
hei
ght)
Height (mm)
Year
ly g
row
th (
mm
hei
ght)
500 1000 1500 2000 2500
−20
00
100
Protium panamense
Height (mm)
Year
ly g
row
th (
mm
hei
ght)
Height (mm)
Year
ly g
row
th (
mm
hei
ght)
500 1000 1500 2000 2500
−20
00
100
Protium tenuifolium
Height (mm)
Year
ly g
row
th (
mm
hei
ght)
Height (mm)
Year
ly g
row
th (
mm
hei
ght)
500 1000 1500 2000 2500
−20
00
100
Psychotria acuminata
Height (mm)
Year
ly g
row
th (
mm
hei
ght)
Height (mm)
Year
ly g
row
th (
mm
hei
ght)
500 1000 1500 2000 2500
−20
00
100
Psychotria deflexa
Height (mm)Ye
arly
gro
wth
(m
m h
eigh
t)
Height (mm)
Year
ly g
row
th (
mm
hei
ght)
500 1000 1500 2000 2500
−20
00
100
Psychotria horizontalis
Height (mm)
Year
ly g
row
th (
mm
hei
ght)
Height (mm)
Year
ly g
row
th (
mm
hei
ght)
500 1000 1500 2000 2500
−20
00
100
Psychotria marginata
Height (mm)
Year
ly g
row
th (
mm
hei
ght)
Height (mm)
Year
ly g
row
th (
mm
hei
ght)
500 1000 1500 2000 2500
−20
00
100
Psychotria racemosa
Height (mm)
Year
ly g
row
th (
mm
hei
ght)
Height (mm)
Year
ly g
row
th (
mm
hei
ght)
500 1000 1500 2000 2500
−20
00
100
Quararibea asterolepis
Height (mm)
Year
ly g
row
th (
mm
hei
ght)
Height (mm)
Year
ly g
row
th (
mm
hei
ght)
500 1000 1500 2000 2500
−20
00
100
Randia armata
Height (mm)
Year
ly g
row
th (
mm
hei
ght)
Height (mm)
Year
ly g
row
th (
mm
hei
ght)
Figure S5. Continued.
31
S3
FIG
URES
500 1000 1500 2000 2500
−20
00
100
Simarouba amara
Height (mm)
Year
ly g
row
th (
mm
hei
ght)
Height (mm)
Year
ly g
row
th (
mm
hei
ght)
500 1000 1500 2000 2500
−20
00
100
Sorocea affinis
Height (mm)
Year
ly g
row
th (
mm
hei
ght)
Height (mm)
Year
ly g
row
th (
mm
hei
ght)
500 1000 1500 2000 2500
−20
00
100
Spondias radlkoferi
Height (mm)
Year
ly g
row
th (
mm
hei
ght)
Height (mm)
Year
ly g
row
th (
mm
hei
ght)
500 1000 1500 2000 2500
−20
00
100
Sterculia apetala
Height (mm)
Year
ly g
row
th (
mm
hei
ght)
Height (mm)
Year
ly g
row
th (
mm
hei
ght)
500 1000 1500 2000 2500
−20
00
100
Stylogyne turbacensis
Height (mm)
Year
ly g
row
th (
mm
hei
ght)
Height (mm)
Year
ly g
row
th (
mm
hei
ght)
500 1000 1500 2000 2500
−20
00
100
Swartzia simplex_var.
Height (mm)
Year
ly g
row
th (
mm
hei
ght)
Height (mm)
Year
ly g
row
th (
mm
hei
ght)
500 1000 1500 2000 2500
−20
00
100
Tabebuia rosea
Height (mm)
Year
ly g
row
th (
mm
hei
ght)
Height (mm)
Year
ly g
row
th (
mm
hei
ght)
500 1000 1500 2000 2500
−20
00
100
Tabernaemont arborea
Height (mm)
Year
ly g
row
th (
mm
hei
ght)
Height (mm)
Year
ly g
row
th (
mm
hei
ght)
500 1000 1500 2000 2500
−20
00
100
Tachigali versicolor
Height (mm)
Year
ly g
row
th (
mm
hei
ght)
Height (mm)
Year
ly g
row
th (
mm
hei
ght)
500 1000 1500 2000 2500
−20
00
100
Talisia croatii
Height (mm)
Year
ly g
row
th (
mm
hei
ght)
Height (mm)
Year
ly g
row
th (
mm
hei
ght)
500 1000 1500 2000 2500
−20
00
100
Tetragastris panamensis
Height (mm)Ye
arly
gro
wth
(m
m h
eigh
t)
Height (mm)
Year
ly g
row
th (
mm
hei
ght)
500 1000 1500 2000 2500
−20
00
100
Trichilia tuberculata
Height (mm)
Year
ly g
row
th (
mm
hei
ght)
Height (mm)
Year
ly g
row
th (
mm
hei
ght)
500 1000 1500 2000 2500
−20
00
100
Unonopsis pittieri
Height (mm)
Year
ly g
row
th (
mm
hei
ght)
Height (mm)
Year
ly g
row
th (
mm
hei
ght)
500 1000 1500 2000 2500
−20
00
100
Virola sebifera
Height (mm)
Year
ly g
row
th (
mm
hei
ght)
Height (mm)
Year
ly g
row
th (
mm
hei
ght)
500 1000 1500 2000 2500
−20
00
100
Xylopia macrantha
Height (mm)
Year
ly g
row
th (
mm
hei
ght)
Height (mm)
Year
ly g
row
th (
mm
hei
ght)
500 1000 1500 2000 2500
−20
00
100
Zanthoxylum panamense
Height (mm)
Year
ly g
row
th (
mm
hei
ght)
Height (mm)
Year
ly g
row
th (
mm
hei
ght)
Figure S5. Continued.
32
S3
FIG
URES
500 1000 1500 2000 2500
0.0
0.4
0.8
Alchornea costaricensi
Height (mm)
Year
ly s
urvi
val p
roba
bilit
y
Height (mm)
Year
ly s
urvi
val p
roba
bilit
y
500 1000 1500 2000 2500
0.0
0.4
0.8
Alibertia edulis
Height (mm)
Year
ly s
urvi
val p
roba
bilit
y
Height (mm)
Year
ly s
urvi
val p
roba
bilit
y
500 1000 1500 2000 2500
0.0
0.4
0.8
Alseis blackiana
Height (mm)
Year
ly s
urvi
val p
roba
bilit
y
Height (mm)
Year
ly s
urvi
val p
roba
bilit
y
500 1000 1500 2000 2500
0.0
0.4
0.8
Anacardium excelsum
Height (mm)
Year
ly s
urvi
val p
roba
bilit
y
Height (mm)
Year
ly s
urvi
val p
roba
bilit
y
500 1000 1500 2000 2500
0.0
0.4
0.8
Annona acuminata
Height (mm)
Year
ly s
urvi
val p
roba
bilit
y
Height (mm)
Year
ly s
urvi
val p
roba
bilit
y
500 1000 1500 2000 2500
0.0
0.4
0.8
Annona spraguei
Height (mm)
Year
ly s
urvi
val p
roba
bilit
y
Height (mm)
Year
ly s
urvi
val p
roba
bilit
y
500 1000 1500 2000 2500
0.0
0.4
0.8
Apeiba membranacea
Height (mm)
Year
ly s
urvi
val p
roba
bilit
y
Height (mm)
Year
ly s
urvi
val p
roba
bilit
y
500 1000 1500 2000 2500
0.0
0.4
0.8
Aspidosperma spruceanum
Height (mm)
Year
ly s
urvi
val p
roba
bilit
y
Height (mm)
Year
ly s
urvi
val p
roba
bilit
y
500 1000 1500 2000 2500
0.0
0.4
0.8
Beilschmiedi pendula
Height (mm)
Year
ly s
urvi
val p
roba
bilit
y
Height (mm)
Year
ly s
urvi
val p
roba
bilit
y
500 1000 1500 2000 2500
0.0
0.4
0.8
Brosimum alicastrum
Height (mm)
Year
ly s
urvi
val p
roba
bilit
y
Height (mm)
Year
ly s
urvi
val p
roba
bilit
y
500 1000 1500 2000 2500
0.0
0.4
0.8
Calophyllum longifolium
Height (mm)Ye
arly
sur
viva
l pro
babi
lity
Height (mm)
Year
ly s
urvi
val p
roba
bilit
y
500 1000 1500 2000 2500
0.0
0.4
0.8
Capparis frondosa
Height (mm)
Year
ly s
urvi
val p
roba
bilit
y
Height (mm)
Year
ly s
urvi
val p
roba
bilit
y
500 1000 1500 2000 2500
0.0
0.4
0.8
Cassipourea elliptica
Height (mm)
Year
ly s
urvi
val p
roba
bilit
y
Height (mm)
Year
ly s
urvi
val p
roba
bilit
y
500 1000 1500 2000 2500
0.0
0.4
0.8
Chrysophyllu argenteum
Height (mm)
Year
ly s
urvi
val p
roba
bilit
y
Height (mm)
Year
ly s
urvi
val p
roba
bilit
y
500 1000 1500 2000 2500
0.0
0.4
0.8
Chrysophyllu cainito
Height (mm)
Year
ly s
urvi
val p
roba
bilit
y
Height (mm)
Year
ly s
urvi
val p
roba
bilit
y
500 1000 1500 2000 2500
0.0
0.4
0.8
Cordia alliodora
Height (mm)
Year
ly s
urvi
val p
roba
bilit
y
Height (mm)
Year
ly s
urvi
val p
roba
bilit
y
Figure S6. Observed and predicted annual seedling survival as a function of initial height (mm). Black dots show observations,red dashed lines show species-specific moving averages (from a Generalized Additive Model using a loess smoother), bluedashed lines show the fitted trait based average model, and grey lines show a fitted model based only on size (not on traits,thus identical for all species).
33
S3
FIG
URES
500 1000 1500 2000 2500
0.0
0.4
0.8
Cordia bicolor
Height (mm)
Year
ly s
urvi
val p
roba
bilit
y
Height (mm)
Year
ly s
urvi
val p
roba
bilit
y
500 1000 1500 2000 2500
0.0
0.4
0.8
Coussarea curvigemmia
Height (mm)
Year
ly s
urvi
val p
roba
bilit
y
Height (mm)
Year
ly s
urvi
val p
roba
bilit
y
500 1000 1500 2000 2500
0.0
0.4
0.8
Desmopsis panamensis
Height (mm)
Year
ly s
urvi
val p
roba
bilit
y
Height (mm)
Year
ly s
urvi
val p
roba
bilit
y
500 1000 1500 2000 2500
0.0
0.4
0.8
Dipteryx oleifera
Height (mm)
Year
ly s
urvi
val p
roba
bilit
y
Height (mm)
Year
ly s
urvi
val p
roba
bilit
y
500 1000 1500 2000 2500
0.0
0.4
0.8
Drypetes standleyi
Height (mm)
Year
ly s
urvi
val p
roba
bilit
y
Height (mm)
Year
ly s
urvi
val p
roba
bilit
y
500 1000 1500 2000 2500
0.0
0.4
0.8
Eugenia coloradoensi
Height (mm)
Year
ly s
urvi
val p
roba
bilit
y
Height (mm)
Year
ly s
urvi
val p
roba
bilit
y
500 1000 1500 2000 2500
0.0
0.4
0.8
Eugenia galalonensis
Height (mm)
Year
ly s
urvi
val p
roba
bilit
y
Height (mm)
Year
ly s
urvi
val p
roba
bilit
y
500 1000 1500 2000 2500
0.0
0.4
0.8
Eugenia nesiotica
Height (mm)
Year
ly s
urvi
val p
roba
bilit
y
Height (mm)
Year
ly s
urvi
val p
roba
bilit
y
500 1000 1500 2000 2500
0.0
0.4
0.8
Eugenia oerstediana
Height (mm)
Year
ly s
urvi
val p
roba
bilit
y
Height (mm)
Year
ly s
urvi
val p
roba
bilit
y
500 1000 1500 2000 2500
0.0
0.4
0.8
Faramea occidentalis
Height (mm)
Year
ly s
urvi
val p
roba
bilit
y
Height (mm)
Year
ly s
urvi
val p
roba
bilit
y
500 1000 1500 2000 2500
0.0
0.4
0.8
Garcinia intermedia
Height (mm)Ye
arly
sur
viva
l pro
babi
lity
Height (mm)
Year
ly s
urvi
val p
roba
bilit
y
500 1000 1500 2000 2500
0.0
0.4
0.8
Guapira standleyana
Height (mm)
Year
ly s
urvi
val p
roba
bilit
y
Height (mm)
Year
ly s
urvi
val p
roba
bilit
y
500 1000 1500 2000 2500
0.0
0.4
0.8
Guarea guidonia
Height (mm)
Year
ly s
urvi
val p
roba
bilit
y
Height (mm)
Year
ly s
urvi
val p
roba
bilit
y
500 1000 1500 2000 2500
0.0
0.4
0.8
Guatteria dumetorum
Height (mm)
Year
ly s
urvi
val p
roba
bilit
y
Height (mm)
Year
ly s
urvi
val p
roba
bilit
y
500 1000 1500 2000 2500
0.0
0.4
0.8
Gustavia superba
Height (mm)
Year
ly s
urvi
val p
roba
bilit
y
Height (mm)
Year
ly s
urvi
val p
roba
bilit
y
500 1000 1500 2000 2500
0.0
0.4
0.8
Hamelia axillaris
Height (mm)
Year
ly s
urvi
val p
roba
bilit
y
Height (mm)
Year
ly s
urvi
val p
roba
bilit
y
Figure S6. Continued.
34
S3
FIG
URES
500 1000 1500 2000 2500
0.0
0.4
0.8
Hampea appendiculat
Height (mm)
Year
ly s
urvi
val p
roba
bilit
y
Height (mm)
Year
ly s
urvi
val p
roba
bilit
y
500 1000 1500 2000 2500
0.0
0.4
0.8
Heisteria concinna
Height (mm)
Year
ly s
urvi
val p
roba
bilit
y
Height (mm)
Year
ly s
urvi
val p
roba
bilit
y
500 1000 1500 2000 2500
0.0
0.4
0.8
Hirtella triandra
Height (mm)
Year
ly s
urvi
val p
roba
bilit
y
Height (mm)
Year
ly s
urvi
val p
roba
bilit
y
500 1000 1500 2000 2500
0.0
0.4
0.8
Hybanthus prunifolius
Height (mm)
Year
ly s
urvi
val p
roba
bilit
y
Height (mm)
Year
ly s
urvi
val p
roba
bilit
y
500 1000 1500 2000 2500
0.0
0.4
0.8
Inga marginata
Height (mm)
Year
ly s
urvi
val p
roba
bilit
y
Height (mm)
Year
ly s
urvi
val p
roba
bilit
y
500 1000 1500 2000 2500
0.0
0.4
0.8
Inga acuminata
Height (mm)
Year
ly s
urvi
val p
roba
bilit
y
Height (mm)
Year
ly s
urvi
val p
roba
bilit
y
500 1000 1500 2000 2500
0.0
0.4
0.8
Inga thibaudiana
Height (mm)
Year
ly s
urvi
val p
roba
bilit
y
Height (mm)
Year
ly s
urvi
val p
roba
bilit
y
500 1000 1500 2000 2500
0.0
0.4
0.8
Inga umbellifera
Height (mm)
Year
ly s
urvi
val p
roba
bilit
y
Height (mm)
Year
ly s
urvi
val p
roba
bilit
y
500 1000 1500 2000 2500
0.0
0.4
0.8
Lacistema aggregatum
Height (mm)
Year
ly s
urvi
val p
roba
bilit
y
Height (mm)
Year
ly s
urvi
val p
roba
bilit
y
500 1000 1500 2000 2500
0.0
0.4
0.8
Lacmellea panamensis
Height (mm)
Year
ly s
urvi
val p
roba
bilit
y
Height (mm)
Year
ly s
urvi
val p
roba
bilit
y
500 1000 1500 2000 2500
0.0
0.4
0.8
Lonchocarpus heptaphyllus
Height (mm)Ye
arly
sur
viva
l pro
babi
lity
Height (mm)
Year
ly s
urvi
val p
roba
bilit
y
500 1000 1500 2000 2500
0.0
0.4
0.8
Luehea seemannii
Height (mm)
Year
ly s
urvi
val p
roba
bilit
y
Height (mm)
Year
ly s
urvi
val p
roba
bilit
y
500 1000 1500 2000 2500
0.0
0.4
0.8
Mouriri myrtilloides
Height (mm)
Year
ly s
urvi
val p
roba
bilit
y
Height (mm)
Year
ly s
urvi
val p
roba
bilit
y
500 1000 1500 2000 2500
0.0
0.4
0.8
Ocotea cernua
Height (mm)
Year
ly s
urvi
val p
roba
bilit
y
Height (mm)
Year
ly s
urvi
val p
roba
bilit
y
500 1000 1500 2000 2500
0.0
0.4
0.8
Ocotea puberula
Height (mm)
Year
ly s
urvi
val p
roba
bilit
y
Height (mm)
Year
ly s
urvi
val p
roba
bilit
y
500 1000 1500 2000 2500
0.0
0.4
0.8
Trophis caucana
Height (mm)
Year
ly s
urvi
val p
roba
bilit
y
Height (mm)
Year
ly s
urvi
val p
roba
bilit
y
Figure S6. Continued.
35
S3
FIG
URES
500 1000 1500 2000 2500
0.0
0.4
0.8
Ouratea lucens
Height (mm)
Year
ly s
urvi
val p
roba
bilit
y
Height (mm)
Year
ly s
urvi
val p
roba
bilit
y
500 1000 1500 2000 2500
0.0
0.4
0.8
Palicourea guianensis
Height (mm)
Year
ly s
urvi
val p
roba
bilit
y
Height (mm)
Year
ly s
urvi
val p
roba
bilit
y
500 1000 1500 2000 2500
0.0
0.4
0.8
Cinnamomum triplinerve
Height (mm)
Year
ly s
urvi
val p
roba
bilit
y
Height (mm)
Year
ly s
urvi
val p
roba
bilit
y
500 1000 1500 2000 2500
0.0
0.4
0.8
Picramnia latifolia
Height (mm)
Year
ly s
urvi
val p
roba
bilit
y
Height (mm)
Year
ly s
urvi
val p
roba
bilit
y
500 1000 1500 2000 2500
0.0
0.4
0.8
Platymiscium pinnatum
Height (mm)
Year
ly s
urvi
val p
roba
bilit
y
Height (mm)
Year
ly s
urvi
val p
roba
bilit
y
500 1000 1500 2000 2500
0.0
0.4
0.8
Pouteria reticulata
Height (mm)
Year
ly s
urvi
val p
roba
bilit
y
Height (mm)
Year
ly s
urvi
val p
roba
bilit
y
500 1000 1500 2000 2500
0.0
0.4
0.8
Prioria copaifera
Height (mm)
Year
ly s
urvi
val p
roba
bilit
y
Height (mm)
Year
ly s
urvi
val p
roba
bilit
y
500 1000 1500 2000 2500
0.0
0.4
0.8
Protium panamense
Height (mm)
Year
ly s
urvi
val p
roba
bilit
y
Height (mm)
Year
ly s
urvi
val p
roba
bilit
y
500 1000 1500 2000 2500
0.0
0.4
0.8
Protium tenuifolium
Height (mm)
Year
ly s
urvi
val p
roba
bilit
y
Height (mm)
Year
ly s
urvi
val p
roba
bilit
y
500 1000 1500 2000 2500
0.0
0.4
0.8
Psychotria acuminata
Height (mm)
Year
ly s
urvi
val p
roba
bilit
y
Height (mm)
Year
ly s
urvi
val p
roba
bilit
y
500 1000 1500 2000 2500
0.0
0.4
0.8
Psychotria deflexa
Height (mm)Ye
arly
sur
viva
l pro
babi
lity
Height (mm)
Year
ly s
urvi
val p
roba
bilit
y
500 1000 1500 2000 2500
0.0
0.4
0.8
Psychotria horizontalis
Height (mm)
Year
ly s
urvi
val p
roba
bilit
y
Height (mm)
Year
ly s
urvi
val p
roba
bilit
y
500 1000 1500 2000 2500
0.0
0.4
0.8
Psychotria marginata
Height (mm)
Year
ly s
urvi
val p
roba
bilit
y
Height (mm)
Year
ly s
urvi
val p
roba
bilit
y
500 1000 1500 2000 2500
0.0
0.4
0.8
Psychotria racemosa
Height (mm)
Year
ly s
urvi
val p
roba
bilit
y
Height (mm)
Year
ly s
urvi
val p
roba
bilit
y
500 1000 1500 2000 2500
0.0
0.4
0.8
Quararibea asterolepis
Height (mm)
Year
ly s
urvi
val p
roba
bilit
y
Height (mm)
Year
ly s
urvi
val p
roba
bilit
y
500 1000 1500 2000 2500
0.0
0.4
0.8
Randia armata
Height (mm)
Year
ly s
urvi
val p
roba
bilit
y
Height (mm)
Year
ly s
urvi
val p
roba
bilit
y
Figure S6. Continued.
36
S3
FIG
URES
500 1000 1500 2000 2500
0.0
0.4
0.8
Simarouba amara
Height (mm)
Year
ly s
urvi
val p
roba
bilit
y
Height (mm)
Year
ly s
urvi
val p
roba
bilit
y
500 1000 1500 2000 2500
0.0
0.4
0.8
Sorocea affinis
Height (mm)
Year
ly s
urvi
val p
roba
bilit
y
Height (mm)
Year
ly s
urvi
val p
roba
bilit
y
500 1000 1500 2000 2500
0.0
0.4
0.8
Spondias radlkoferi
Height (mm)
Year
ly s
urvi
val p
roba
bilit
y
Height (mm)
Year
ly s
urvi
val p
roba
bilit
y
500 1000 1500 2000 2500
0.0
0.4
0.8
Sterculia apetala
Height (mm)
Year
ly s
urvi
val p
roba
bilit
y
Height (mm)
Year
ly s
urvi
val p
roba
bilit
y
500 1000 1500 2000 2500
0.0
0.4
0.8
Stylogyne turbacensis
Height (mm)
Year
ly s
urvi
val p
roba
bilit
y
Height (mm)
Year
ly s
urvi
val p
roba
bilit
y
500 1000 1500 2000 2500
0.0
0.4
0.8
Swartzia simplex_var.
Height (mm)
Year
ly s
urvi
val p
roba
bilit
y
Height (mm)
Year
ly s
urvi
val p
roba
bilit
y
500 1000 1500 2000 2500
0.0
0.4
0.8
Tabebuia rosea
Height (mm)
Year
ly s
urvi
val p
roba
bilit
y
Height (mm)
Year
ly s
urvi
val p
roba
bilit
y
500 1000 1500 2000 2500
0.0
0.4
0.8
Tabernaemont arborea
Height (mm)
Year
ly s
urvi
val p
roba
bilit
y
Height (mm)
Year
ly s
urvi
val p
roba
bilit
y
500 1000 1500 2000 2500
0.0
0.4
0.8
Tachigali versicolor
Height (mm)
Year
ly s
urvi
val p
roba
bilit
y
Height (mm)
Year
ly s
urvi
val p
roba
bilit
y
500 1000 1500 2000 2500
0.0
0.4
0.8
Talisia croatii
Height (mm)
Year
ly s
urvi
val p
roba
bilit
y
Height (mm)
Year
ly s
urvi
val p
roba
bilit
y
500 1000 1500 2000 2500
0.0
0.4
0.8
Tetragastris panamensis
Height (mm)Ye
arly
sur
viva
l pro
babi
lity
Height (mm)
Year
ly s
urvi
val p
roba
bilit
y
500 1000 1500 2000 2500
0.0
0.4
0.8
Trichilia tuberculata
Height (mm)
Year
ly s
urvi
val p
roba
bilit
y
Height (mm)
Year
ly s
urvi
val p
roba
bilit
y
500 1000 1500 2000 2500
0.0
0.4
0.8
Unonopsis pittieri
Height (mm)
Year
ly s
urvi
val p
roba
bilit
y
Height (mm)
Year
ly s
urvi
val p
roba
bilit
y
500 1000 1500 2000 2500
0.0
0.4
0.8
Virola sebifera
Height (mm)
Year
ly s
urvi
val p
roba
bilit
y
Height (mm)
Year
ly s
urvi
val p
roba
bilit
y
500 1000 1500 2000 2500
0.0
0.4
0.8
Xylopia macrantha
Height (mm)
Year
ly s
urvi
val p
roba
bilit
y
Height (mm)
Year
ly s
urvi
val p
roba
bilit
y
500 1000 1500 2000 2500
0.0
0.4
0.8
Zanthoxylum panamense
Height (mm)
Year
ly s
urvi
val p
roba
bilit
y
Height (mm)
Year
ly s
urvi
val p
roba
bilit
y
Figure S6. Continued.
37
S3
FIG
URES
0 10000 20000 30000
−10
000
500
Adelia triloba
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
0 20000 40000 60000
−10
000
500
Alchornea costaricensi
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
200 400 600 800 1200
−10
000
500
Alibertia edulis
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
0 5000 10000 15000 20000
−10
000
500
Allophylus
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
0 2000 6000 10000
−10
000
500
Alseis blackiana
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
0 2000 4000 6000 8000
−10
000
500
Andira inermis
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
200 400 600 800 1200
−10
000
500
Annona acuminata
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
0 10000 20000 30000
−10
000
500
Annona spraguei
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
0e+00 4e+04 8e+04
−10
000
500
Apeiba membranacea
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
0 2000 4000 6000 8000
−10
000
500
Aspidosperma spruceanum
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
0 20000 40000 60000
−10
000
500
Astronium graveolens
Basal area (mm2)Ye
arly
gro
wth
(m
m2 )
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
0 2000 6000 10000
−10
000
500
Beilschmiedi pendula
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
0 2000 6000 10000
−10
000
500
Brosimum alicastrum
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
0 1000 2000 3000
−10
000
500
Calophyllum longifolium
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
200 400 600 800 1200
−10
000
500
Capparis frondosa
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
0 1000 2000 3000
−10
000
500
Casearia aculeata
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
Figure S7. Observed and predicted annual tree growth (mm2 basal area/yr) as a function of initial size (mm2 basal area).Black dots show observations, red dashed lines show species-specific moving averages (from a Generalized Additive Model usinga loess smoother), blue dashed lines show the fitted trait based average model, and grey lines show a fitted model based onlyon size (not on traits, thus identical for all species).
38
S3
FIG
URES
0 20000 60000
−10
000
500
Casearia arborea
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
0 2000 6000 10000
−10
000
500
Cassipourea elliptica
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
0 1000 2000 3000
−10
000
500
Chrysochlamy eclipes
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
0 5000 10000 15000 20000
−10
000
500
Chrysophyllu argenteum
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
0 2000 4000 6000 8000
−10
000
500
Chrysophyllu cainito
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
0 5000 10000 15000 20000
−10
000
500
Coccoloba coronata
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
0 1000 2000 3000
−10
000
500
Coccoloba manzinellens
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
0e+00 4e+04 8e+04
−10
000
500
Cordia alliodora
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
0 20000 40000 60000
−10
000
500
Cordia bicolor
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
0 1000 2000 3000
−10
000
500
Coussarea curvigemmia
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
200 400 600 800 1200
−10
000
500
Cupania rufescens
Basal area (mm2)Ye
arly
gro
wth
(m
m2 )
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
0 1000 2000 3000
−10
000
500
Cupania seemannii
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
200 400 600 800 1200
−10
000
500
Desmopsis panamensis
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
0 5000 10000 15000 20000
−10
000
500
Drypetes standleyi
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
200 400 600 800 1200
−10
000
500
Erythroxylum panamense
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
0 2000 6000 10000
−10
000
500
Eugenia coloradoensi
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
Figure S7. Continued.
39
S3
FIG
URES
200 400 600 800 1200
−10
000
500
Eugenia galalonensis
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
0 2000 6000 10000
−10
000
500
Eugenia nesiotica
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
0 2000 4000 6000 8000
−10
000
500
Eugenia oerstediana
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
0 2000 4000 6000 8000
−10
000
500
Faramea occidentalis
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
0 1000 2000 3000
−10
000
500
Garcinia intermedia
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
0 1000 2000 3000
−10
000
500
Garcinia madruno
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
0 20000 40000 60000
−10
000
500
Guapira standleyana
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
0 10000 20000 30000
−10
000
500
Guarea guidonia
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
0 5000 10000 15000 20000
−10
000
500
Guatteria dumetorum
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
0 10000 30000 50000
−10
000
500
Gustavia superba
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
50 60 70 80 90 100
−10
000
500
Hamelia axillaris
Basal area (mm2)Ye
arly
gro
wth
(m
m2 )
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
0 10000 30000 50000
−10
000
500
Hampea appendiculat
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
0 5000 15000 25000
−10
000
500
Hasseltia floribunda
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
0 2000 4000 6000 8000
−10
000
500
Heisteria acuminata
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
0 10000 30000 50000
−10
000
500
Heisteria concinna
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
200 400 600 800 1200
−10
000
500
Herrania purpurea
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
Figure S7. Continued.
40
S3
FIG
URES
0 5000 10000 15000 20000
−10
000
500
Hirtella triandra
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
200 400 600 800 1200
−10
000
500
Hybanthus prunifolius
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
5000 15000 25000
−10
000
500
Hieronyma alchorneoide
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
0 2000 4000 6000 8000
−10
000
500
Inga goldmanii
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
0 2000 6000 10000
−10
000
500
Inga marginata
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
0 2000 4000 6000 8000
−10
000
500
Inga nobilis
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
0 2000 6000 10000
−10
000
500
Inga acuminata
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
0 5000 10000 15000 20000
−10
000
500
Inga sapindoides
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
0 5000 15000 25000
−10
000
500
Inga thibaudiana
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
0 1000 2000 3000
−10
000
500
Inga umbellifera
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
0 50000 100000 150000
−10
000
500
Jacaranda copaia
Basal area (mm2)Ye
arly
gro
wth
(m
m2 )
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
0 1000 2000 3000
−10
000
500
Lacistema aggregatum
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
0 10000 30000 50000
−10
000
500
Lacmellea panamensis
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
0 2000 4000 6000 8000
−10
000
500
Laetia thamnia
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
0 2000 6000 10000
−10
000
500
Licania hypoleuca
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
0 1000 2000 3000
−10
000
500
Licania platypus
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
Figure S7. Continued.
41
S3
FIG
URES
0 5000 10000 15000 20000
−10
000
500
Lonchocarpus heptaphyllus
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
0 5000 15000 25000
−10
000
500
Luehea seemannii
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
0 1000 2000 3000
−10
000
500
Mosannona garwoodii
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
0 2000 4000 6000 8000
−10
000
500
Maquira guianensis
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
50 60 70 80 90 100
−10
000
500
Mouriri myrtilloides
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
0 5000 10000 15000 20000
−10
000
500
Nectandra cissiflora
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
0 2000 6000 10000
−10
000
500
Ocotea cernua
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
0 5000 10000 15000 20000
−10
000
500
Ocotea puberula
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
0 2000 4000 6000 8000
−10
000
500
Trophis caucana
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
0 1000 2000 3000
−10
000
500
Ormosia coccinea
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
200 400 600 800 1200
−10
000
500
Ormosia macrocalyx
Basal area (mm2)Ye
arly
gro
wth
(m
m2 )
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
200 400 600 800 1200
−10
000
500
Ouratea lucens
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
200 400 600 800 1200
−10
000
500
Palicourea guianensis
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
200 400 600 800 1200
−10
000
500
Pentagonia macrophylla
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
0 2000 4000 6000 8000
−10
000
500
Perebea xanthochyma
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
0 1000 2000 3000
−10
000
500
Picramnia latifolia
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
Figure S7. Continued.
42
S3
FIG
URES
0 10000 30000 50000
−10
000
500
Platymiscium pinnatum
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
0 2000 6000 10000
−10
000
500
Platypodium elegans
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
0 5000 10000 15000 20000
−10
000
500
Poulsenia armata
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
0 5000 15000 25000
−10
000
500
Pourouma bicolor
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
0 2000 6000 10000
−10
000
500
Pouteria reticulata
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
0 5000 10000 15000 20000
−10
000
500
Prioria copaifera
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
0 1000 2000 3000
−10
000
500
Protium panamense
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
0 5000 10000 15000 20000
−10
000
500
Protium tenuifolium
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
50 60 70 80 90 100
−10
000
500
Psychotria horizontalis
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
50 60 70 80 90 100
−10
000
500
Psychotria marginata
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
0 1000 2000 3000
−10
000
500
Pterocarpus rohrii
Basal area (mm2)Ye
arly
gro
wth
(m
m2 )
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
0 5000 10000 15000 20000
−10
000
500
Quararibea asterolepis
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
1500 2500 3500
−10
000
500
Quassia amara
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
0 2000 6000 10000
−10
000
500
Randia armata
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
0 5000 10000 15000 20000
−10
000
500
Simarouba amara
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
0 2000 6000 10000
−10
000
500
Sloanea terniflora
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
Figure S7. Continued.
43
S3
FIG
URES
0 1000 2000 3000
−10
000
500
Sorocea affinis
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
0 5000 10000 15000 20000
−10
000
500
Spondias mombin
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
0 5000 15000 25000
−10
000
500
Spondias radlkoferi
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
200 400 600 800 1200
−10
000
500
Stylogyne turbacensis
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
0 2000 4000 6000 8000
−10
000
500
Symphonia globulifera
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
0 5000 10000 15000 20000
−10
000
500
Tabebuia rosea
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
0 2000 6000 10000
−10
000
500
Tabernaemont arborea
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
0 1000 2000 3000
−10
000
500
Tachigali versicolor
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
200 400 600 800 1200
−10
000
500
Talisia nervosa
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
200 400 600 800 1200
−10
000
500
Talisia croatii
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
0 2000 6000 10000
−10
000
500
Tetragastris panamensis
Basal area (mm2)Ye
arly
gro
wth
(m
m2 )
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
0 1000 2000 3000
−10
000
500
Thevetia ahouai
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
0 5000 10000 15000 20000
−10
000
500
Trichilia pallida
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
0 5000 15000 25000
−10
000
500
Trichilia tuberculata
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
0 10000 20000 30000
−10
000
500
Triplaris cumingiana
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
0 10000 20000 30000
−10
000
500
Turpinia occidentalis
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
Figure S7. Continued.
44
S3
FIG
URES
0 5000 15000 25000
−10
000
500
Unonopsis pittieri
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
0 10000 30000 50000
−10
000
500
Virola sebifera
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
0 10000 30000 50000
−10
000
500
Virola surinamensis
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
0 2000 6000 10000
−10
000
500
Xylopia macrantha
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
0 10000 30000 50000
−10
000
500
Zanthoxylum panamense
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
Basal area (mm2)
Year
ly g
row
th (
mm
2 )
Figure S7. Continued.
45
S3
FIG
URES
0 100 200 300 400 500
0.0
0.4
0.8
Adelia triloba
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
0 100 200 300 400 500
0.0
0.4
0.8
Alchornea costaricensi
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
0 100 200 300 400 500
0.0
0.4
0.8
Alibertia edulis
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
0 100 200 300 400 500
0.0
0.4
0.8
Allophylus
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
0 100 200 300 400 500
0.0
0.4
0.8
Alseis blackiana
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
0 100 200 300 400 500
0.0
0.4
0.8
Andira inermis
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
0 100 200 300 400 500
0.0
0.4
0.8
Annona acuminata
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
0 100 200 300 400 500
0.0
0.4
0.8
Annona spraguei
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
0 100 200 300 400 500
0.0
0.4
0.8
Apeiba membranacea
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
0 100 200 300 400 500
0.0
0.4
0.8
Aspidosperma spruceanum
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
0 100 200 300 400 500
0.0
0.4
0.8
Astronium graveolens
Dbh (mm)Ye
arly
sur
viva
l pro
babi
lity
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
0 100 200 300 400 500
0.0
0.4
0.8
Beilschmiedi pendula
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
0 100 200 300 400 500
0.0
0.4
0.8
Brosimum alicastrum
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
0 100 200 300 400 500
0.0
0.4
0.8
Calophyllum longifolium
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
0 100 200 300 400 500
0.0
0.4
0.8
Capparis frondosa
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
0 100 200 300 400 500
0.0
0.4
0.8
Casearia aculeata
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
Figure S8. Observed and predicted annual tree survival as a function of initial dbh (mm). Black dots show observations, reddashed lines show species-specific moving averages (from a Generalized Additive Model using a loess smoother), blue dashedlines show the fitted trait based average model, and grey lines show a fitted model based only on size (not on traits, thusidentical for all species).
46
S3
FIG
URES
0 100 200 300 400 500
0.0
0.4
0.8
Casearia arborea
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
0 100 200 300 400 500
0.0
0.4
0.8
Cassipourea elliptica
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
0 100 200 300 400 500
0.0
0.4
0.8
Chrysochlamy eclipes
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
0 100 200 300 400 500
0.0
0.4
0.8
Chrysophyllu argenteum
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
0 100 200 300 400 500
0.0
0.4
0.8
Chrysophyllu cainito
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
0 100 200 300 400 500
0.0
0.4
0.8
Coccoloba coronata
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
0 100 200 300 400 500
0.0
0.4
0.8
Coccoloba manzinellens
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
0 100 200 300 400 500
0.0
0.4
0.8
Cordia alliodora
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
0 100 200 300 400 500
0.0
0.4
0.8
Cordia bicolor
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
0 100 200 300 400 500
0.0
0.4
0.8
Coussarea curvigemmia
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
0 100 200 300 400 500
0.0
0.4
0.8
Cupania rufescens
Dbh (mm)Ye
arly
sur
viva
l pro
babi
lity
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
0 100 200 300 400 500
0.0
0.4
0.8
Cupania seemannii
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
0 100 200 300 400 500
0.0
0.4
0.8
Desmopsis panamensis
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
0 100 200 300 400 500
0.0
0.4
0.8
Drypetes standleyi
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
0 100 200 300 400 500
0.0
0.4
0.8
Erythroxylum panamense
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
0 100 200 300 400 500
0.0
0.4
0.8
Eugenia coloradoensi
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
Figure S8. Continued.
47
S3
FIG
URES
0 100 200 300 400 500
0.0
0.4
0.8
Eugenia galalonensis
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
0 100 200 300 400 500
0.0
0.4
0.8
Eugenia nesiotica
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
0 100 200 300 400 500
0.0
0.4
0.8
Eugenia oerstediana
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
0 100 200 300 400 500
0.0
0.4
0.8
Faramea occidentalis
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
0 100 200 300 400 500
0.0
0.4
0.8
Garcinia intermedia
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
0 100 200 300 400 500
0.0
0.4
0.8
Garcinia madruno
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
0 100 200 300 400 500
0.0
0.4
0.8
Guapira standleyana
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
0 100 200 300 400 500
0.0
0.4
0.8
Guarea guidonia
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
0 100 200 300 400 500
0.0
0.4
0.8
Guatteria dumetorum
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
0 100 200 300 400 500
0.0
0.4
0.8
Gustavia superba
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
0 100 200 300 400 500
0.0
0.4
0.8
Hamelia axillaris
Dbh (mm)Ye
arly
sur
viva
l pro
babi
lity
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
0 100 200 300 400 500
0.0
0.4
0.8
Hampea appendiculat
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
0 100 200 300 400 500
0.0
0.4
0.8
Hasseltia floribunda
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
0 100 200 300 400 500
0.0
0.4
0.8
Heisteria acuminata
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
0 100 200 300 400 500
0.0
0.4
0.8
Heisteria concinna
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
0 100 200 300 400 500
0.0
0.4
0.8
Herrania purpurea
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
Figure S8. Continued.
48
S3
FIG
URES
0 100 200 300 400 500
0.0
0.4
0.8
Hirtella triandra
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
0 100 200 300 400 500
0.0
0.4
0.8
Hybanthus prunifolius
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
0 100 200 300 400 500
0.0
0.4
0.8
Hieronyma alchorneoide
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
0 100 200 300 400 500
0.0
0.4
0.8
Inga goldmanii
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
0 100 200 300 400 500
0.0
0.4
0.8
Inga marginata
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
0 100 200 300 400 500
0.0
0.4
0.8
Inga nobilis
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
0 100 200 300 400 500
0.0
0.4
0.8
Inga acuminata
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
0 100 200 300 400 500
0.0
0.4
0.8
Inga sapindoides
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
0 100 200 300 400 500
0.0
0.4
0.8
Inga thibaudiana
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
0 100 200 300 400 500
0.0
0.4
0.8
Inga umbellifera
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
0 100 200 300 400 500
0.0
0.4
0.8
Jacaranda copaia
Dbh (mm)Ye
arly
sur
viva
l pro
babi
lity
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
0 100 200 300 400 500
0.0
0.4
0.8
Lacistema aggregatum
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
0 100 200 300 400 500
0.0
0.4
0.8
Lacmellea panamensis
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
0 100 200 300 400 500
0.0
0.4
0.8
Laetia thamnia
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
0 100 200 300 400 500
0.0
0.4
0.8
Licania hypoleuca
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
0 100 200 300 400 500
0.0
0.4
0.8
Licania platypus
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
Figure S8. Continued.
49
S3
FIG
URES
0 100 200 300 400 500
0.0
0.4
0.8
Lonchocarpus heptaphyllus
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
0 100 200 300 400 500
0.0
0.4
0.8
Luehea seemannii
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
0 100 200 300 400 500
0.0
0.4
0.8
Mosannona garwoodii
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
0 100 200 300 400 500
0.0
0.4
0.8
Maquira guianensis
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
0 100 200 300 400 500
0.0
0.4
0.8
Mouriri myrtilloides
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
0 100 200 300 400 500
0.0
0.4
0.8
Nectandra cissiflora
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
0 100 200 300 400 500
0.0
0.4
0.8
Ocotea cernua
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
0 100 200 300 400 500
0.0
0.4
0.8
Ocotea puberula
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
0 100 200 300 400 500
0.0
0.4
0.8
Trophis caucana
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
0 100 200 300 400 500
0.0
0.4
0.8
Ormosia coccinea
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
0 100 200 300 400 500
0.0
0.4
0.8
Ormosia macrocalyx
Dbh (mm)Ye
arly
sur
viva
l pro
babi
lity
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
0 100 200 300 400 500
0.0
0.4
0.8
Ouratea lucens
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
0 100 200 300 400 500
0.0
0.4
0.8
Palicourea guianensis
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
0 100 200 300 400 500
0.0
0.4
0.8
Pentagonia macrophylla
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
0 100 200 300 400 500
0.0
0.4
0.8
Perebea xanthochyma
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
0 100 200 300 400 500
0.0
0.4
0.8
Picramnia latifolia
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
Figure S8. Continued.
50
S3
FIG
URES
0 100 200 300 400 500
0.0
0.4
0.8
Platymiscium pinnatum
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
0 100 200 300 400 500
0.0
0.4
0.8
Platypodium elegans
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
0 100 200 300 400 500
0.0
0.4
0.8
Poulsenia armata
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
0 100 200 300 400 500
0.0
0.4
0.8
Pourouma bicolor
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
0 100 200 300 400 500
0.0
0.4
0.8
Pouteria reticulata
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
0 100 200 300 400 500
0.0
0.4
0.8
Prioria copaifera
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
0 100 200 300 400 500
0.0
0.4
0.8
Protium panamense
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
0 100 200 300 400 500
0.0
0.4
0.8
Protium tenuifolium
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
0 100 200 300 400 500
0.0
0.4
0.8
Psychotria horizontalis
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
0 100 200 300 400 500
0.0
0.4
0.8
Psychotria marginata
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
0 100 200 300 400 500
0.0
0.4
0.8
Pterocarpus rohrii
Dbh (mm)Ye
arly
sur
viva
l pro
babi
lity
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
0 100 200 300 400 500
0.0
0.4
0.8
Quararibea asterolepis
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
0 100 200 300 400 500
0.0
0.4
0.8
Quassia amara
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
0 100 200 300 400 500
0.0
0.4
0.8
Randia armata
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
0 100 200 300 400 500
0.0
0.4
0.8
Simarouba amara
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
0 100 200 300 400 500
0.0
0.4
0.8
Sloanea terniflora
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
Figure S8. Continued.
51
S3
FIG
URES
0 100 200 300 400 500
0.0
0.4
0.8
Sorocea affinis
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
0 100 200 300 400 500
0.0
0.4
0.8
Spondias mombin
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
0 100 200 300 400 500
0.0
0.4
0.8
Spondias radlkoferi
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
0 100 200 300 400 500
0.0
0.4
0.8
Stylogyne turbacensis
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
0 100 200 300 400 500
0.0
0.4
0.8
Symphonia globulifera
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
0 100 200 300 400 500
0.0
0.4
0.8
Tabebuia rosea
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
0 100 200 300 400 500
0.0
0.4
0.8
Tabernaemont arborea
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
0 100 200 300 400 500
0.0
0.4
0.8
Tachigali versicolor
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
0 100 200 300 400 500
0.0
0.4
0.8
Talisia nervosa
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
0 100 200 300 400 500
0.0
0.4
0.8
Talisia croatii
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
0 100 200 300 400 500
0.0
0.4
0.8
Tetragastris panamensis
Dbh (mm)Ye
arly
sur
viva
l pro
babi
lity
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
0 100 200 300 400 500
0.0
0.4
0.8
Thevetia ahouai
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
0 100 200 300 400 500
0.0
0.4
0.8
Trichilia pallida
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
0 100 200 300 400 500
0.0
0.4
0.8
Trichilia tuberculata
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
0 100 200 300 400 500
0.0
0.4
0.8
Triplaris cumingiana
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
0 100 200 300 400 500
0.0
0.4
0.8
Turpinia occidentalis
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
Figure S8. Continued.
52
S3
FIG
URES
0 100 200 300 400 500
0.0
0.4
0.8
Unonopsis pittieri
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
0 100 200 300 400 500
0.0
0.4
0.8
Virola sebifera
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
0 100 200 300 400 500
0.0
0.4
0.8
Virola surinamensis
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
0 100 200 300 400 500
0.0
0.4
0.8
Xylopia macrantha
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
0 100 200 300 400 500
0.0
0.4
0.8
Zanthoxylum panamense
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
Dbh (mm)
Year
ly s
urvi
val p
roba
bilit
y
Figure S8. Continued.
53
S3 FIGURES
●
●
●
●
●
●●
●●●
●
●●
●
●
● ●
●
●●
●
●●
●
●●
● ● ●
●●
● ●●
●●●
●
●●
●
●
●
●
●●
●
●
●
●
●● ●
●
●
●
●●
●●
●
●
●
●
●
●
●
●
●●
●●
●
●●
●
●
●
●
●●
●
●
●
●
●●
50 100
200
500
1000
2000
5
10
20
50
100
200
500
1000
Maximum diameter (mm)
Rep
rodu
ctiv
e th
resh
old
size
(m
m d
iam
eter
)
R2=0.807
Figure S9. Relationship between species’ maximum observed tree diameter(Dmax) and the onset of reproduction (50 percent quantile of reproduction;R50). The dashed line indicates the relationship R50 = 0.5Dmax).
54
S3
FIG
URES
200 400 600 800 1200
2040
60SM
size
sdl g
rowt
h −−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−++++++++++++++++
++++++++++++++++++++++++++++
200 400 600 800 1000 1400
2040
60
WD
size
−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−++++++++++++++++++++++++++
++++++++++++++++++
200 400 600 800 1000 1400
2040
60
LMA
size
−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−
++++++++++++++++++++++++++++++++++++++++++++
200 400 600 800 1000 1400
2040
60
Dmax
size
−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−
++++++++++++++++++++++++++++++++++++++++++++
200 400 600 800 1200
0.75
0.85
0.95
SM
size
sdl s
urviv
al
−−−−−−
−−−−−−−−−−−−
−−−−−−−−−−−−−−−−−−−−−−−−−−
+++++
+++++++
++++++++++++++++++
++++++++++++++
200 400 600 800 1000 1400
0.75
0.85
0.95
WD
size
−−−−−
−−−−−−−−
−−−−−−−−−−−−−−−−−−−−−
−−−−−−−−−−+++++
++++++++
+++++++++++++++++++++++++++++++
200 400 600 800 1000 1400
0.75
0.85
0.95
LMA
size
−−−−−
−−−−−−−
−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−
+++++
+++++++++
++++++++++++++++++++++++++++++
200 400 600 800 1000 1400
0.75
0.85
0.95
Dmax
size
−−−−−−−−−
−−−−−−−−−−−−−−−−−−−−−−−
−−−−−−−−−−−−
++++
++++
++++++++++
++++++++++++++++++++++++++
50 100 150 200 250
200
600
1000 SM
size
tree
grow
th
− − − −− −
− −−
+ + ++
+ ++ +
+
50 100 150 200 250
200
600
1000 WD
size
− − −−
− −− − −
+ + + ++ +
+ +
+
50 100 150 200 250
200
600
1000 LMA
size
− − −−
− −− −
−
+ + ++
+ ++ +
+
50 100 150 200 250
200
600
1000 Dmax
size
− − −−
− −− −
−
+ + + + ++
++
+
50 100 150 200 250
0.88
0.92
0.96
SM
size
tree
surv
ival
−−
− − − − − −−+ + + + + + + ++
50 100 150 200 250
0.88
0.92
0.96
WD
size
− −− − − − − −
−+ + + + + + + + +
50 100 150 200 250
0.88
0.92
0.96
LMA
size
− − − − − − − −−
+ + + + + + + ++
50 100 150 200 250
0.88
0.92
0.96
Dmax
size
− − − − −− − −
−+ + + + + + + + +
50 100 150 200
0.0
0.4
0.8
SM
size
Repr
. fra
ctio
n
− − − − − − −−
+ + + + + + ++
50 100 150 200
0.0
0.4
0.8
WD
size
−− −
− − − −−
+ ++ + + + +
+
50 100 150 200
0.0
0.4
0.8
LMA
size
−− − − − − −
−
+ + + ++ + + +
50 100 150 200
0.0
0.4
0.8
Dmax
size
−− − −
− − −−
+ + + + + + + +
Figure S10. Separate Analyses at Every Size. Fitted effects of each trait (columns) on size-dependent vital rates (rows), asin Figure 2, but based on separate models for each size (the robustness analysis). The black lines present the vital rate-sizerelationships with all traits set to their mean values (with the corresponding predictions from the hierarchical model in grey).The blue and red lines present the same relationships with one trait set to its mean plus or minus one standard deviation,respectively, and the two remaining traits set to their mean values (with exception of Dmax, see below). The trait whose valuevaries among the blue, black and red lines is named at the top of each column. Actual mean values and standard deviations foreach trait are given in table 1. An important difference from figure 3 is that for this analysis here, small statured species dropout at larger sizes and thus a different subset of species are used at each size (mean Dmax increases); in contrast, the result infigure 3 based on the hierarchical model simply shows the main effects as fitted across sizes. The size ranges of the axis differfrom Fig. 2 also because species drop out with increasing size, and we limited the analysis include at least 15 species.
55
S3
FIG
URES
Figure S11. Separate Analyses at Every Size. Among-species variance explained by traits (as measured by R2 values) throughout thelife-cycle of tree species from Barro Colorado Island, as in Figure 3, but based on separate models for each size (the robustness analysis).The rows show results for (top to bottom) size-dependent growth, size-dependent survival, and vital rates associated with reproduction.The R2 value at the upper edge of the stacked colors represents the proportion of the total variation among species explained by traits,as reflect in the averaged model when species-specific values for that vital rate are regressed on all four traits (i.e. including traits Dmax,LMA, SM and WD) and model averaging is applied, where the observed values for each species are based on species-specific GAMs. Therelative importance of different traits is indicated by the relative height of each color band as a proportion of the total, with height scaledto the R2 values for averaged models including only one trait (i.e. including only Dmax, LMA, SM or WD). The number of speciesincluded in the analyses depends on size; for reference, species numbers are shown in solid circles above each graph at fixed intervals,with their diameters proportional to log species number.
56
REFERENCES
References
James. S Clark, Miles R Silman, Ruth. Kern, Eric. Macklin, and Janneke. Hil-leRisLambers. Seed dispersal near and far: patterns across temperate andtropical forests. Ecology, 80(5):1475–1494, 1999.
H C Muller-Landau, S J Wright, O Calderon, R Condit, and S P Hubbell.Interspecific variation in primary seed dispersal in a tropical forest. Journalof Ecology, 96:653–667, 2008.
57