Functional Ecology. 2021;00:1–13.
Received: 18 February 2021
Accepted: 23 June 2021
DOI : 10.1111/136 5-243 5.13 874
Predator abundance drives the association between
exploratory personality and foraging habitat risk in a wild
marine meso- predator
Félicie Dhellemmes1,2,3 | Matthew J. Smukall3,4 | Tristan L. Guttridge3,5 |
Jens Krause1,2 | Nigel E. Hussey6
1Leibniz- Institute of Freshwater Ecolog y and Inland Fisheries (IGB), Berlin, Germany; 2Faculty of Life Sciences, Albrecht Danie l Thaer- Institute of A gricultural
and Hor ticultural Sciences, Humboldt- Universität zu B erlin, Berlin, Germany; 3Bimini Biological F ield Station Foundation, Sout h Bimini, Bahamas; 4College of
Fisheries and Ocean Sciences, University of Alaska Fairbanks , Fairbanks, AK , USA; 5Saving the blue, Cooper City, FL, USA and 6Integrative Biolog y, University
of Windsor, Windsor, ON, Canada
This is an op en access article under the ter ms of the Creat ive Commo ns Attri bution License, which pe rmits use, distribution and reproduction in any medium,
provide d the original wor k is properly cited.
© 2021 The Authors. Functional Ecology published by John Wiley & Sons Ltd o n behalf of British Ecological Society
Save Our Seas Foundation, Grant/
Award Number: SOS 367; Elsa Neumann
Stipendium des L andes; Natural Sciences
and Engineering Research Council of Canada
Handling Editor: Zacchaeus Compson
1. In recent years, the incorporation of lower levels of organization to the under-
standing of population ecology, has led to an increase in interest for animal per-
sonality and individual foraging specialization. Despite these topics investigating
comparable phenomena, that is, individual consistency in behaviour and in food
resource use respectively, they have rarely been investigated together.
2. Food resource use is thought to be at the interface between personality and life
history. More explorative individuals in a population, for example, are thought
to increase encounter rates with food resources and consequently have faster
growth than less explorative conspecifics.
3. Such hypotheses have so far only received partial support, and the link between
personality and life history is increasingly speculated to be plastic and dependent
on spatio- temporal variation in ecological conditions. Intraspecific competition
and/or predation risk, for example are known to influence foraging specialization.
4. Here, we investigated the relationship between exploration personality of juve-
nile lemon sharks Negaprion brevirostris (measured in captivity) and foraging habi-
tat use (high risk vs. low risk; measured via stable isotope analysis in the field) in
relation to conspecific and predator abundance.
5. We identified predator abundance as the main driver for the association between
foraging habitat and exploration personality. When predators were less abundant,
increased exploratio n was associate d with foraging in riskier ha bitat s. When pred-
ator abundance increased, an inverse relationship occurred, with less explorative
individuals using more dangerous habitat.
DHELLEMM ES Et aL .
1 | INTRODUCTION
Population ecology is often studied while identifying and incor-
porating sources of variation within populations. Age (Polis, 1984)
and sex (Shine, 1989) have, for instance, long been recognized as
significant sources of ecological variation. But variation can also
be observed at the individual level (i.e. within sex and age groups),
a phenomenon that is essential for understanding population and
ecosystem dynamics (Toscano et al., 2016). In behavioural sciences,
animal personalit y research has emerged as a means to investigate
individual level differences in behaviour that are consistent across
time and situation (i.e. set of conditions at one point in time; Réale
et al., 2007). Such consistency in individual behaviour has been
hypothesized to have multiple ecological and evolutionary conse-
quences ( Wolf & Weissing, 2012), which include implications for life
history (Dammhahn et al., 2018). Individual bluegill sunfish Lepomis
macrochirus, for example, that are consistently bolder (i.e. faster at
emerging from a refuge) have been shown to have greater maximum
metabolic rates than shyer individuals (Binder et al., 2016), while
more explorative female wild cavies Cavia aperea grow faster than
their less explorative conspecifics (Guenther, 2018). The evolution
and maintenance of animal personalities are hypothesized to be fa-
voured by trade- offs that promote a range of behavioural pheno-
types with similar fitness (Mangel & Stamps, 2011).
The field of food web ecology has concurrently intensified its
focus on the concept of ‘individual resource specialization’, defined
as consistent inter- individual differences in food resource use within
populations (Bolnick et al., 2002). Resource acquisition is governed
by the need for animals to maximize their rate of energy intake
while minimizing other costs such as energy expenditure, preda-
tion risk and intraspecific competition (see optimal foraging theory,
Pyke, 1984). Individual resource specialization arises when multiple
specialists with different resource acquisition strategies coexist
within a population (Powell & Taylor, 2017). These resource acquisi-
tion strategies can be obse rved on a geogr aphical level, wit h individ-
uals consistently foraging at different locations (Harris et al., 2020;
Wakefield et al., 2015), at a prey level, with individuals specializing in
particular prey types (Ratclif fe et al., 2018) or a combination of both.
Such specializations can be mediated through morphological traits
(e.g. bill length in hummingbirds, Trochilidae, Tinoco et al., 2017),
physiological state (e.g. hunger, reproductive state, developmen-
tal stage, as reviewed in Bedoya- Perez et al., 2013) and behaviour
(e.g. boldness in black- legged kittiwakes, Rissa tridactyla, Harris
et al., 2020). The emergence of individual foraging specialization
is hypothesized to be dictated by factors that promote consistent
variation in resource use among individuals (Carneiro et al., 2017).
Intraspecific competition, for example, may limit the abundance of
available prey leading to the emergence of foraging specialization
(Araújo et al., 2008, 2011; Svanbäck & Bolnick, 2007). This is illus-
trated in gentoo penguins Pygoscelis papua, where an increase in
intraspecific competition led to the specialization of individuals for
one resource (krill), or another (fish) (Ratclif fe et al., 2018). Predation
risk has also been hypothesized to promote the emergence of in-
dividual foraging specialization within populations if perceived risk
varies among individuals (Araújo et al., 2011). For instance, increased
predation from ants and birds led to a reduction in individual diet
breadth in caterpillars, Lepidoptera (Singer et al., 2019).
Despite parallel growth of the ‘animal personality’ and the ‘in-
dividual foraging specialization’ research areas, and overlap in key
features (e.g. individual consistency), these topics have developed in
almost complete isolation (Kalinkat, 2014, Toscano et al., 2016, but
see Harris et al. (2020), for personality driven foraging specialization
in black- legged kittiwakes). This is surprising because studies sug-
gest that resource acquisition is at the interface between person-
ality and life history (Spiegel et al., 2017). For instance, risk- inclined
behaviours are predicted to facilitate greater resource acquisition,
leading to a subsequent increase in growth, but also to higher prob-
abilities of predation mortality (Réale et al., 2010; Stamps, 2007).
Understanding how personalit y, foraging specialization and life his-
tory are causally linked is likely crucial, as the association between
personality and life history is still not well understood and empirical
studies often have inconclusive results (Moirón et al., 2020; Royauté
et al., 2018). Such ambiguity may be due to environmental condi-
tions (e.g. predator or resource abundance) inconsistently favouring
the covariance between observed personality and life- history trait s
across time (Royauté et al., 2018). In this situation, we argue that it
is necessary to investigate how personality covaries with foraging
behaviour under var ying ecological conditions as an important step
to bridge the gap between personalit y, resource specialization and
The paucity of studies investigating animal personality in parallel
with individual foraging specialization may in part be due to a meth-
odological divide. Animal personality studies involve standardized
designs that allow for repeated measures of behaviour (Dingemanse
& Wright, 2020). For example, the novel open- field test, a popular
6. We conclude that the relationship between personality and resource use is plastic
and context dependent, which could explain the inconclusive results of previous
studies investigating links between personality and life history.
ecological drivers, pace- of- life syndrome, shark, stable isotope analysis, trait covariance
DHELLEMM ES Et aL .
personality test, assesses the willingness of individuals to explore
a novel arena (Perals et al., 2017). The need for standardized, re-
peatable tests has led to an abundance of studies on captive bred
animals and/or studies conducted in captivity, as these tests would
be impractical or impossible to achieve in the wild (Archard &
Braithwaite, 2010). In contrast, individual foraging specialization
typically relies on monitoring animals' foraging in the field (Toscano
et al ., 2016). Stable isotope analysis has be come an incre asingly pop-
ular tool to investigate animals' foraging habitat and trophic level
(Carneiro et al., 2017). Carbon (13C/12C, measured as δ13C) and ni-
trogen (15N/14N, measured as δ15N) stable isotopes are integrated
into consumers' tissues (e.g. skin, blood) from assimilated resources,
creating a record of their diet and foraging ecology (Boucher
et al., 2020). Nitrogen stable isotopes are typically enriched in con-
sumers relative to their food, serving as an indicator of trophic po-
sition (Cherel & Hobson, 2007). In contrast, carbon stable isotopes
are relatively conserved across trophic levels, allowing the determi-
nation of consumer foraging habitat related to the primary carbon
sources within a given trophic network (Cherel & Hobson, 20 07).
In Bimini, The Bahamas (Figure 1), juvenile lemon sharks
Negaprion brevirostris inhabit coastal shallow water lagoon hab-
itats characterized by a mangrove- fringed shore (red mangrove:
Rhizophora mangle and black mangrove: Avicennia germinans) and
shallow seagrass beds (turtle grass: Thalassia testudinum and shoal
grass: Halodule wrightii) interspersed by exposed sediment patches.
Previous work demonstrated that carbon isotope ratios of individual
lemon sharks were consistent over time (i.e. individual foraging spe-
cialization) representing sharks' foraging in habitats on a continuum
between protected low- risk mangrove (low δ13C) and presumably
riskier exposed seagrass (high δ13C) (Hussey et al., 2017). In addi-
tion, exploration personality (measured in a novel open- field test)
was found to predict distance from the mangrove shore (measured
via acoustic telemetry), growth and mortality in one of two subpop-
ulations (Dhellemmes, Finger, Smukall, et al., 2020). Previous work
on this system found that predator and conspecific abundance were
crucial drivers for trait associations involving personality and life
history (Dhellemmes, Finger, Laskowski, et al., 2020; Dhellemmes,
Finger, Smukall, et al., 2020). Here we combined captive person-
ality tests and stable isotope analysis to test how the correlation
between foraging habitat and exploration personalit y varies with
predation and intraspecific competition in two subpopulations of
juvenile lemon sharks (North Sound and Sharkland, Figure 1) across
4 years. Because foraging specialization has been documented to be
stronger in the context of high intraspecific competition (Ratcliffe
et al., 2018; Sheppard et al., 2018), we predicted a stronger trait cor-
relation when competition was high. Similarly, with specializations
being stronger in high predation contexts (Singer et al., 2019) one
could also expect stronger trait correlations when predator abun-
dance was high. Juvenile lemon shark exploration score was, how-
ever, previously found to predict distance from the shore in the
subpopulation with the lowest predator abundance (North Sound),
but not in the neighbouring subpopulation with high predation risk
(Sharkland; Dhellemmes, Finger, Smukall, et al., 2020). Given these
results, we predicted a stronger association between personalit y
and foraging habitat (as indicated by stable isotope values) when
predator abundance was low, with more exploratory individuals for-
aging in riskier seagrass habitat.
1.1 | Study site and sampling
The current study was conducted in Bimini, The Bahamas (Figure 1),
a mangrove- fringed chain of islands located approximately 80 km
off the coast of Florida (U.S.A). Juvenile lemon sharks are known
to use nursery habitats consisting of nearshore protected man-
grove habitats and offshore exposed seagrass beds, until they dis-
perse around 3– 5 years of age (Chapman et al., 2009; DiBattista
et al., 2007). Inhabiting small home ranges (<600 m² for sharks
under 56 cm pre- caudal length, PCL, Morrissey & Gruber, 1993),
recapture probabilities of individuals from 1 year to the next are
high (0.67– 0.85; DiBattista et al., 20 07). Here, we focus on juvenile
lemon shark subpopulations inhabiting two adjacent nursery areas
(North Sound and Sharkland, see Figure 1) which are known to differ
in their predator abundance (Sharkland has higher predator abun-
dance; Dhellemmes, Finger, Laskowski, et al., 2020), and that are
FIGURE 1 Map of the Bimini Islands, The Bahamas
(25.736232°N, −79.267353°W) showing the two principal juvenile
lemon shark subpopulations, the six capture locations and the
locations of fixed acoustic telemetry receivers
DHELLEMM ES Et aL .
almost completely isolated in regard to emigration and immigration
of juvenile sharks <2 years of age due to their small home ranges
(Chapman et al., 2009; 1.5% of the sharks in the current study emi-
grated between 0 and 2 years of age, unpublished data).
Between 2014 and 2017, we captured juvenile lemon sharks
using gillnets (180 m leng th, 10 cm stretch- mesh size) set perpen-
dicular to the shore at six standard locations in the North Sound and
Sharkland (see Figure 1). Sampling was undertaken for six nights
(12 hr each) in each nurser y, with the three nets sampled simultane-
ously in each area, resulting in a total of 12 nights of fishing effort.
Upon capture, sharks were scanned for the presence of a uniquely
coded passive integrated transponder (PIT, Destron Fearing) tag.
Tag ID (if no PIT tag was found, one was implanted subcutaneous
at the base of the dorsal fin), measurements (PCL, nearest mm), sex
(the presence or absence of claspers) and the state of umbilical scar
healing (for age determination; see below) were recorded. When
possible (i.e. when sample collection was safe for the animal and the
operator) we took a sample of the trailing edge of the first dorsal
fin (<5 mm−2) and immediately stored it on ice. Fin samples, used
for subsequent stable isotope analysis, were stored at −18℃ within
12 hr of their collection. Each shark was then housed in semi- captive
arenas temporarily built within the nurser y areas (see below section
1.2 for details on arena construction) until the end of the sampling
campaign, that is, the 12 days of fishing, to avoid repeated captures.
Lemon sharks, a placental viviparous species, are born with
an umbilical wound which heals during the first few weeks of life
(Feldheim et al., 2002). This allowed assignment of age for each
shark according to their umbilical state (opened to any extent:
young- of- the- year (YOY); closed: unknown age from 1 to 5 years).
Given that shark sampling has been systematically undertaken
each year since the 90s as part of a capture– mark– recapture study
(Gruber et al., 2001), most sharks could be precisely aged, as they
had been captured as YOY in previous years. When the umbilical
state of an individual could not be recorded (e.g. the shark was dif-
ficult to handle), or a shark had never been captured as a YOY, we
determined age using a linear regression of age on PCL (accuracy:
91%, see Dhellemmes, Finger, Smukall, et al., 2020 for details).
At the two study sites, YOY and 1- year- old sharks are the most
commonly captured age classes (Dhellemmes, Finger, Smukall,
et al., 2020; DiBat tista et al., 2007). Because stable isotope value of
YOY sharks is initially confounded by the maternal isotopic signature
(Olin et al., 2011), sampling targeted sharks of age 1 year, resulting in
data for 131 individuals (see Table 1).
1.2 | Assessment of exploration personality
At the conclusion of the gillnet survey in each nurser y area, a ran-
domly selected subset of captured lemon sharks was transferred to
a nearby behavioural testing arena (Figure 2A), where they were ac-
climated for 4 days before commencing experiments. We inser ted
T- bar anchor tags (Floy Tag Inc.) in the first and/or second dorsal
fin of each shark in a unique colour combination to allow individual
visual recognition during tests. While in captivity, sharks were fed
every 2 days with approx. 2% of their body weight of locally caught
barracuda Sphyraena barracuda and Sardinella spp. (with the feeder
ensuring every shark received food) to match their estimated daily
ration in the wild (Cortés & Gruber, 1990).
The behavioural testing arena consisted of three parts: (a) an
oblong enclosure (10 × 5 m) divided into three compartments used
to house sharks, (b) a circular arena (diameter, 10 m) that was built
close (4 m) to the housing enclosure to host a social behaviour test
and (c) a rectangular arena (6 × 12 m) that was built 2 m from the
sociability arena to host the novel open- field test, the focus of the
current study. Each arena was connected to the adjoining one via
a channel, allowing sharks to be ushered from one arena to the
next without the need for handling (Figure 2A). The channel that
separated the sociability arena and the novel open- field served as
a start- box, where sharks spent 5 minutes after being ushered to
recuperate from potential stress. All par ts of the behavioural testing
arena were constructed with orange construction mesh (6 cm mesh
size, Tenax Sentr y HD; Tenax Fence), steel rebars, cable ties and cin-
der blocks. We erected wooden towers (~4 m height) on the North
side of the sociability arena and the novel open- field arena to allow
for behavioural observations while limiting shadows from observers
One day prior to tests, we transferred six size- matched sharks
into the sociability arena via the channel. Sharks were then fed to
satiation and left to acclimate overnight. On the following day we
observed sharks in the sociability arena for 20 min followed by im-
mediate individual testing in the novel open- field. The results of
the sociabilit y test are not included in the current analyses, but see
Finger et al. (2018) for repeatability of sociability.
The novel open- field test was conducted as follows. An individ-
ual shark was ushered from the sociability arena to the start- box
(Figure 2B) connecting the sociability arena to the novel open- field
arena. Sharks were ushered opportunistically, accepting that previ-
ous re sea rch fou nd th at th e ord er of tes tin g did not influe nce test re-
sults (Dhellemmes, Finger, Laskowski, et al., 2020). After 5 minutes in
the start- box, a sliding door was remotely opened allowing entrance
to the novel open- field (Figure 2B). Once a shark entered the novel
open- field, the door was closed, and behavioural observation was
conducted for 10 min. The novel open- field arena was divided into
16 sectors (2 × 2 m) by green concrete markers placed on the sub-
strate (Figure 2B). An exploration personality score was derived as
the mean number of sectors visited per minute of the test (including
multiple visits to the same sectors). This exploration score was pre-
viously shown to measure a shark's reaction to a novel environment
TABLE 1 Summary of age 1- year- old individuals sampled for
stable isotopes and tested for personality for each year and each
2014 2015 2 016 2017 Tot a l
North Sound 18 14 15 14 61
Sharkland 15 11 32 16 74
DHELLEMM ES Et aL .
rather than general activity (Finger et al., 2016). Furthermore, ex-
ploration of the novel open- field was found to be repeatable (R ad-
justed for PCL = 0.33, 95% confidence interval (CI) [0.13, 0.52]) in a
sample that included all sharks between 1 and 3 years of age that
were repeatedly tested in North Sound and Sharkland over the years
of the current study (n = 95; 85% 1- year- old; Dhellemmes, Finger,
Laskowski, et al., 2020) and to predict sharks' distance from the
mangrove shore in North Sound but not in Sharkland (Dhellemmes,
Finger, Smukall, et al., 2020). The density of sharks held in captivity
prior to the test was found not to impact the exploration score in a
separate population and a dif ferent year (see Appendix 1.1).
1.3 | Foraging habitat: Seagrass versus mangrove
Carbon stable isotope values (δ13C) have been previously shown
to be consistent within individual juvenile lemon sharks over time
(i.e. between two measurements a year apart) and to represent
differences in individual foraging habitat on a continuum between
low- risk protected mangrove (lower δ13C) and high- risk exposed sea-
grass (higher δ13C) (Hussey et al., 2017). Consequently, we inferred
foraging habitat of each shark through measuring the δ13C values of
fin tissue. In the current study, it was not possible to test for con-
sistency in δ13C values through repeat sampling, consequently our
single δ13C value for each individual (referred to as ‘foraging habitat’
throughout the manuscript) provides a proxy for foraging specializa-
tion (Hussey et al., 2017).
To avoid lipid and urea biases on stable isotope values, we
lipid extracted and water washed fin samples following Kinney
et al. (2011) and Li et al. (2016). Samples were then freeze dried,
weighed (400– 600 mg) and placed into small tin capsules. Carbon
isotope values and the total carbon per cent were determined
usi ng a con ti nuous fl ow isotope r atio mass sp ec trometer (Fi nn igan
MAT Deltaplus; Thermo Fisher Scientific) equipped with an ele-
mental analyser (Costech Analytical Technologies) at the Great
Lakes Institute for Environmental Research (Windsor, Ontario,
The stable isotope ratio is expressed in δ value and represents
the parts per thousand (‰) deviation from a standard according to
the following formula:
where R is the ratio of 13C on 12C. An assessment of the standard de-
viation of replicate analyses of four standards (Standard bovine liver
(NIST1577c), internal laboratory standard (tilapia muscle), USGS 40
and Urea (N = 45 for all)), revealed a precision ≤0.18‰ for all stan-
dards. Accuracy showed a difference of −0.04‰ from the certified
values of USGS 40 (N = 45) analysed throughout runs and not used to
1.4 | Measuring intraspecific competition
We used the annual population size of juvenile lemon sharks in
each nursery area (i.e. North Sound and Sharkland) as a proxy for
intraspecific competition. This acknowledges that intraspecific com-
petition does not solely depend on population density, but also on
resource abundance and distribution. Estimates of resource abun-
dance were not available during the current study, but previous
work has reported that mangrove fish communities are stable across
seasons (Newman et al., 2007), therefore the changes in population
size of juvenile lemon sharks were deemed to provide a reasonable
proxy for changes in intraspecific competition. Shark subpopulation
size was measured during the annual gillnet survey described above
and accounts for ever y shark captured regardless of age. Since all
sharks captured were held in an arena during the sampling campaign,
a sharp decline in capture rate over the duration of the sur vey was
observed, as would be expected. By the sixth night of sampling in
each nursery area, we estimated that the full juvenile lemon shark
population in that area had been captured (96% of the subpopula-
tion captured by the fourth night of fishing as estimated by Gruber
et al., 2001).
FIGURE 2 Behavioural testing arena: (A) aerial view of the
complete set- up, (B) schematic representation of the novel open-
field. The section markers are represented by green dot s, each
section is identified by a unique coordinate as represented by the
numbers and letters on the side of the arena (e.g. the sliding door is
in section b1)
DHELLEMM ES Et aL .
1.5 | Measuring predator abundance
We estimated predator abundance in each area using passive acous-
tic telemetry undertaken for a concurrent project quantifying shark
movement and habitat use around Bimini. Large sharks were cap-
tured monthly using fisheries- independent longlines surveys (for
more information see Hansell et al., 2018) or other shark fishing
methods and acoustic transmitters (V16, 90– 150 s delay, 10 year
life, VEMCO, Bedford, Canada) were surgically implanted following
standard procedures. Sharks' movements were monitored using an
array of ~65 acoustic receivers (VR2W, VEMCO), deployed around
the islands (including two in North Sound and two in Sharkland,
Figure 1). The receivers recorded the date, time and unique trans-
mitter identity of each shark that swam within their range (50%
detection probability at 255 m, see Guttridge et al., 2017). Data on
predator presence collected in 2015, 2016 and 2017 were consid-
ered here, as the receivers in Nor th Sound and Sharkland were only
deployed at the end of 2014. We calculated predator abundance in
each nurser y as the number of large sharks detected in a given nurs-
ery area during a calendar year divided by the total number of sharks
detected around Bimini that same year. Predator abundance was
calculated for all detected sub- adult/adult lemon sharks (N = 34,
PCL = 152 mean ± 62 SD), bull sharks (Carcharhinus leucas, N = 19,
PCL = 183 mean ± 13 SD) and black tip sharks (Carcharhinus limbatus,
N = 13, PCL = 110 mean ± 9 SD) that are known to feed on juve-
nile lemon sharks and other chondrichthyans (Guttridge et al., 2012;
Hoffmayer & Parsons, 2006; Morrissey & Gruber, 1993; Tuma, 1976;
Vorenberg, 1962; Wetherbee et al., 1990).
1.6 | Ethical note
The experimental procedures for this study were approved by
the Department of Marine Resources, Bahamas (Permit no: MAF/
LIA/22). Handling was kept under 5 min (e.g. for measuring and fin
sampling) to limit stress. No sharks died in captivity and upon test
completion, sharks were fed to satiation, all external tags removed,
and sharks released at their site of capture.
1.7 | Statistical methods
Comparisons of mean between groups (i.e. the two subpopulations)
were conducted using two- sample t tests when sample size was
equal between groups. In cases of unequal sample sizes, we con-
ducted a test of variance and used a two- sample t test with equal
To test whether exploration score predicted foraging habitat
in each subpopulation, we constructed a linear mixed model in the
MCMCglmm package (Hadfield, 2010). We used a lowly informa-
tive inverse gamma prior, with 240,000 iterations, a thinning inter-
val of 200 and we discarded the first 40,0 00 iterations resulting in
a Markov chain Monte Carlo with a sample size of 1,000 and low
autocorrelation. The model included foraging habitat (δ13C) as the
response, and exploration score in an interaction with subpopulation
as fixed effects. We accounted for yearly differences by including
‘year’ as a random effect.
To test whether predator abundance and/or population den-
sity drove the correlation between exploration and foraging in
high- versus low- risk habitat, we used meta- analytic methods in the
metafor package (Viechtbauer, 2010). Traditionally, meta- analytic
approaches are used to compare results from different studies (e.g.
correlation coefficient s) by converting values into a common cur-
rency called ‘effect size’ which considers differences in precision
(e.g. sample size) among studies (Lajeunesse & Forbes, 2003). These
approaches also allow for the use of a ‘test of heterogeneity’ to as-
sess whether the effect sizes are different across studies. If this is
the case, we can assess whether the observed heterogeneity can be
attributed to the different variables of interest (i.e. the moderators).
Here we implemented these methods to investigate if the cor-
relation between foraging habitat and exploration varied among
years and subpopulations (i.e. significant test of heterogeneity) and
if predator abundance (i.e. proportion of predators present in each
nurser y) and intraspecific competition (i.e. total size of the juvenile
lemon shark subpopulation in each nursery) could explain the het-
erogeneity. Accordingly, we first created a null model, that included
no moderators to run a test of heterogeneity. For each year and
each subpopulation, we calculated the Spearman's correlation coef-
ficient and computed ef fect sizes using the ‘escalc’ function within
the metafor package, using the ‘ZCOR’ argument to apply a Fisher
z- transf or mation to the coef ficient s and to meet ass umptions of nor-
mality. The obtained effect sizes and their corresponding sampling
variances were used in the model.
If the test of heterogeneity was significant, we tested for the ef-
fect of pred at or ab und anc e and int r asp eci fic com p etition, me an cen-
tred at both the between subpopulation level (i.e. overall mean = 0)
and the within subpopulation level (i.e. North Sound mean = 0 and
Sharkland mean = 0). We did this to tease apart effects that were
due to the subpopulations being different (e.g. x is always higher in
Sharkland than North Sound and so is y, causing a positive x~y re-
lationship) from the effects due to fluctuations in ecological condi-
tions within subpopulation (e.g. when x goes up within Sharkland,
y goes up as well, regardless of what happens in North Sound; see
van de Pol & Wright, 2009 for similar methods). In ef fect, the vari-
ables centred between subpopulations represented a combination
of between and within subpopulation effects, while the variables
centred within the two subpopulations represented only the within-
Because predator abundance was not available for 2014, we al-
ways ran the test of moderators in separate models (containing all
years for intraspecific competition, and missing 2014 for predator
abundance) before testing them together in the same model (exclud-
ing intraspecific competition data from 2014).
We first tested the variables centred between subpopulations.
We then tested the moderators centred at the within- population
level while adding subpopulation as an interaction. We used
DHELLEMM ES Et aL .
log- likelihood ratio tests to compare models with the moderators in
an interaction with subpopulation (i.e. different slope and intercept
for each subpopulation) and with subpopulation as an independent
effect (i.e. different intercept but same slope for each subpopula-
tion) to determine whether each subpopulation required modelling
with a different slope.
All analysis were performed in R, version 3.6.2 (R Core
Team, 2017). The data are available from the Dr yad Digital Repository
https://doi.org/10.5061/dryad.rr4xg xd8f (Dhellemmes et al., 2021).
A R markdown file documenting the step- by- step analytical process
is uploaded as Supporting Information (Appendix 2).
2 | RESULTS
Exploration personality and foraging habitat was determined for
a total of 131 individual lemon sharks age 1 year (female = 85,
male = 91) captured in North Sound and Sharkland subpopulations.
Sharkland had higher predator abundance than North Sound (paired
t test: t3 = −4.26, p = 0.02; Sharkland mean = 28.9 ± 11.1 SD, North
Sound mean = 6.6 ± 4.2 SD; Figure 3A), but juvenile lemon shark
population size, used here as a prox y for intraspecific competition,
was similar (paired t test: t6 = −0.28151, p = 0.7; Figure 3B). Sharks
from North Sound and Sharkland did not differ in their captive ex-
ploration score (t test: t129 = 1.72, p = 0.09; Figure 3C), but sharks
from Sharkland had higher δ13C values than their North Sound
conspecifics suggesting Sharkland sharks used proportionally more
higher risk seagrass habitat than North Sound individuals (t test:
t133 = −5.82, p < 0.0001; Sharkland mean = −11.1, North Sound
mean = −12.3; Figure 3D).
2.1 | Do subpopulations differ in their exploration–
foraging habitat relationship?
When testing for a relationship between exploration personality and
foraging habitat (high- risk seagrass vs. low- risk mangrove, as desig-
nated by δ13C), we found an effect of exploration personality score
on foraging habitat in North Sound (Posterior mean = 13.36 [7.01,
19.83], Figure 4A), but not in Sharkland (Posterior mean = −2 . 57
[−17.85, 12.26], Figure 4B).
2.2 | What drives the association between foraging
habitat and exploration?
When no moderators were included in the meta- analytic model,
significant heterogeneity in effect sizes was observed (Q7 = 27.56 ,
p = 0.0003). Consequently, the correlation coefficients between
exploration score and foraging habitat were different between year
and subpopulation, allowing for a subsequent test of moderators
FIGURE 3 Graphical visualization
of each variable of interest across years
and nurseries. Histograms of (A) the
abundance of predators (predators
detected in the nursery/predators
detected in Bimini × 100) and (B) the
subpopulation size (number of juvenile
lemon sharks in the area) as a prox y for
intraspecific competition in each year and
each subpopulation. Boxplots of (C) the
exploration score of sharks and (D) the
foraging habitat occupied (i.e. protec ted
low- risk mangrove [low δ13C value] versus
exposed high- risk seagrass [high δ13 C
value]) in each year and each nursery.
Data are not available for predator
abundance in 2014 given acoustic
receivers were not deployed in the study
area for most of that year
DHELLEMM ES Et aL .
We found the tests of moderators to be significant when we in-
dividually tested predator abundance and intraspecific competition
centred at the between subpopulation level (i.e. overall mean = 0;
predator abundance: QM1 = 19.35, p < 0.0001; intraspecific compe-
tition: QM1 = 3.92, p = 0.047). This identified that both moderators
were important predictors of the correlation between exploration
ha bit at and fo r agi ng hab itat . In the fi nal mo del in clu din g both mo der-
ators, predator abundance had a negative effect on the relationship
between foraging habitat and exploration score (estimate = −0.05
[−0.07, −0.02], Figure 6A), while intraspecific competition did not in-
fluence the relationship between traits (estimate = 0.0097 [−0.03,
0.01], Figure 6B).
At the within subpopulation level, models were not improved by
the addition of interactions between subpopulation and the mod-
erator of interest (predator abundance: Log- likelihood ratio = 0.63,
p = 0.42; intra- specific competition log- likelihood ratio = 2.02,
p = 0.15). In the absence of interactions, predator abundance was
the only significant moderator (QM1 = 5.28, p = 0.02; intraspecific
competition QM1 = 1.16, p = 0.28), showing a negative effect on the
relationship between foraging habitat and exploration score (esti-
mate = −0.05 [−0.09, −0.007], Figure 7).
Given we compared two distinct sampling sites, one within
each subpopulation, rather than a continuum across the nursery
regions, we repeated the above analyses on our data organized
into three sampling groups. Sharks captured in the two northern
most gillnets in North Sound, were assigned to a ‘North North
Sound’ group, sharks captured in the south of North Sound and
North of Sharkland in a ‘Middle’ group and sharks captured in
th e t wo sou the r nmo st gil lne ts of Sha rkl and in a ‘So uth Sh arkl and ’
group (see Figure 1 for capture locations). The results of this al-
ternative gradient analysis, in terms of the significance and di-
rection of the relationships, supported those obtained using the
original two subpopulations providing further confidence in our
conclusions (Appendix 1.2). The results from this later analy-
sis, however, were limited by low sample sizes within year and
3 | DISCUSSION
In the current study, we aimed to bridge the gap between animal per-
sonality and individual foraging specialization by investigating under
which ecological circumstances personality correlates with low- ver-
sus high- risk foraging habitat in two subpopulations of wild juvenile
lem on shar ks known to differ in predator abundance and wit h vary ing
intraspecific competition. In North Sound, we found an overall posi-
tive relationship bet ween exploration and δ13C values, indicating that
sharks which explored more in captivity also foraged predominantly
in exposed seagrass habitats. In Sharkland, no link between explora-
tion score and δ13C values was observed. When we sub- divided the
data by year and subpopulation, we found that the correlation coef-
ficients bet ween foraging habitat and exploration personality were
significantly heterogeneous, indicating that the relationship between
these traits fluctuated across years and subpopulations. Importantly,
predator abundance was a significant predictor of both the strength
and direction of correlations, with reduced predator abundance asso-
ciated with more positive coefficients (i.e. more explorative sharks in
captivity predominantly foraged in risky seagrass habitats). This result
was retained when predator abundance was centred within subpop-
ulations, indicating that it was not exclusively driven by the known
difference in predator abundance between the areas. Intraspecific
competition was also a significant moderator of the relationship, but
only at the between subpopulation level, and its effect on the coef-
ficients was not different from zero when it was included in a model
along with predator abundance.
With all years pooled, the differences observed between North
Sound and Sharkland were in accordance with previous findings.
Dhellemmes, Finger, Smukall, et al. (2020) found exploration of the
novel open- field to predict the distance sharks swam from shore
and their growth rate in North Sound, with fast growth and off-
shore swimming selected against suggesting individuals favoured
the use of protected mangrove habitat. In Sharkland, fast growth
was also found to be associated with higher mortality probabilities,
however, no association was found between exploration of the
FIGURE 4 Foraging habitat (on a
spectrum from high- risk exposed seagrass
to low- risk protected mangrove as
designated δ13C) as a function of captive
exploration personality in (A) North Sound
and (B) Sharkland. Solid lines represent
significant linear regressions
DHELLEMM ES Et aL .
novel open- field, distance swam from shore and growth rate. It was
proposed that a personality driven growth mortality trade- off may
arise because foraging in open habitats (e.g. seagrass) can be more
productive, but also more dangerous (Dhellemmes, Finger, Smukall,
et al., 2020). This suggestion is substantiated by our results here
and further corroborated by Hussey et al. (2017) whereby high δ13C
values (i.e. seagrass foraging) predicted high growth rate in Bimini's
juvenile lemon sharks. Dhellemmes, Finger, Smukall, et al. (2020) of-
fered two potential explanations for the absence of a link between
personality and life history in Sharkland: (a) The trade- off between
growth and mortality was mediated via a different personality trait
in this subpopulation or (b) the environmental conditions did not al-
ways favour the association between personality and life histor y.
Here, we found that the link between personality and forag-
ing behaviour was unstable among years and subpopulations, and
that predator abundance was a main driver of the trait association.
This suggests that the obser ved lack of relationship between traits
in Sharkland when all years were pooled together is due to yearly
fluctuations in predator abundance. Predation has often been hy-
pothesized as a driver of the association between personality traits
(i.e. behavioural syndromes) with high predation leading to stronger
associations (e.g. Bell, 2004; Dingemanse et al., 20 07). Here, preda-
tion not only influenced the strength of the trait association, but also
its direction. In North Sound, where predator abundance was lower,
more explorative sharks had δ13C values representative of offshore
seagrass foraging, according to our expectations. This indicates that
lower predator abundance reduced the cost of offshore foraging,
leading individuals to behave in accordance with their personality
score measured in captivity. When predator abundance was high,
however, sharks did not forage according to expectations from their
captive test (i.e. explorative individuals foraged more in safer man-
grove habitat, see Figure 5, Sharkland, 2015). One explanation for
this observation could relate to the fact that a captive personalit y
test such as the novel open- field assay provides a safe environment
where food is provided ad libitum (to avoid hunger biases, Biro &
Booth, 2009). In the absence of ecological pressures present in the
natural environment, animals may behave in accordance with their
personality phenot ype. Behaviour, however, is plastic and expected
to fluctuate according to various ex ternal factors (Rodríguez- Prieto
et al., 2011). For instance, low sociabilit y was linked to high disper-
sal in mosquitofish Gambusia affinis, but this trait correlation was
negated under high predator abundance (Cote et al., 2013). For
Sharkland lemon sharks, predator abundance, could be expected to
dampen sharks' risky foraging behaviour, given a reduction in forag-
ing effort in response to perceived predation risk has been shown
in numerous studies (Ferrari et al., 2009). The amount of food con-
sumed by reef fishes, for example, was shown to drastically reduce
when presented with predator decoys (Catano et al., 2016). This ar-
gument is contradicted by our finding that δ13C values in Sharkland
indicate a comparatively higher use of seagrass habitats than in
North Sound. However, this result could be influenced by variability
in the distribution and density of mangrove and seagrass habitats
between North Sound and Sharkland.
In contrast, the behaviour of less explorative sharks (in captive
trials) was influenced by predator abundance in Sharkland in an un-
expected way: when predator abundance was high, less exploratory
individuals in captivity foraged predominantly in more dangerous
exposed seagrass habitat. This result could suggest that food re-
source and/or space availabilit y in the mangrove habitat in Sharkland
alone is insufficient to support the requirements of the whole sub-
population. Forced high density of juvenile lemon sharks in the man-
grove habitat due to predator presence could increase intraspecific
competition, with more exploratory personality types potentially
dominating and expelling lower risk- taking sharks forcing them to
adopt a new foraging specialization (i.e. switching from a mangrove
to seagrass dominated diet).
FIGURE 5 Foraging habitat (on a spectrum from high- risk
exposed seagrass to low- risk protected mangrove as designated
by δ13C) as a function of exploration score for each year and each
subpopulation. The Spearman's rho for each subsample is given at
the bottom right of the plots. Where correlations were significant,
we present the linear regression of foraging habitat on exploration
score as a solid line. (Significance: .p = 0.05, *p < 0.05, **p < 0.01)
DHELLEMM ES Et aL .
The fact that less exploratory individuals foraged in more dan-
gerous seagrass habitats might also partly be due to state- dependent
processes (i.e. driven by the internal state of individuals). Green sea
turtles Chelonia mydas i n Shark Bay wer e fo und to sh if t for aging hab -
itat according to their body condition in the presence of their main
predator, tiger sharks Galeocerdo cuvier (Heit hau s et al., 20 07 ). Wh en
predation risk was high (i.e. tiger sharks were abundant), turtles in
poor body condition foraged in high risk, but profitable habitats,
whil e tur t le s in go od bo dy cond iti on pref erred safe r, but less pro duc-
tive habitats. For juvenile lemon sharks in Bimini, high exploration
score and seagrass foraging have both been linked to higher growth
rates (Dhellemmes, Finger, Smukall, et al., 2020; Hussey et al., 2017)
suggesting that explorative individuals may have better body condi-
tion than less explorative sharks. This state- dependent explanation
seems highly plausible given turtles with good body condition pre-
ferred high- risk habitat under low predator abundance, identical to
the behaviour of lemon sharks in this study (Heithaus et al., 2007).
An increasing body of literature suggests that observed be-
havioural correlations among individuals might not be representative
of what is happening at the within individual level (‘individual gam-
bit’ Brommer, 2013; Niemelä & Dingemanse, 2018). To avoid making
the assumption that among- individual correlations are representa-
tive of within individual correlations the use of multivariate models
is advised which allow for the decomposition of variances within
and between individuals (Niemelä & Dingemanse, 2018). Such sta-
tistical tools require large sample sizes and multiple measurements,
which was not possible due to logistical constraints (e.g. population
size, difficulty of captures, duration of the captive tests) inherent to
studying long- lived and naturally low abundance large species. While
we took the ‘individual gambit’ we argue that our study provides an
important step in understanding how natural conditions (e.g. com-
petition and predation) shape the covariance between personality
and ecologically relevant behaviours and their associated impact
on life history. Furthermore, we acknowledge that foraging habitat
might be highly plastic and therefore our single measurement of δ13C
may not accurately represent foraging specialization. Previous work
based on multiple measurements per individual at the same study
site, however, would suggest our data provides a reasonable proxy
(Hussey et al., 2017). We also accept that recent evidence suggests
that certain tissues may retain maternal isotopic influence for peri-
ods >1 year, which could bias interpretation of juvenile shark forag-
ing habitat (Niella et al., 2021). We state that the issue of isotopic
turnover rate is species dependent (i.e. related to growth rate and
physiolog y) and we remain confident that isotope values in fin tis-
sue of juvenile lemon sharks are reliable (Hussey et al., 2017; Rangel
et al., 2020). While we were only able to record our measures of for-
aging specialization and personality annually, future studies should,
when possible, repeatedly test the relationship between foraging
specialization and personality in fast changing environmental condi-
tions to overcome this caveat. One potential solution to address this
challenge is to use tissues with faster isotopic turnover rates than
fin: for instance the isotopic turnover rate of plasma occurs over
a few weeks, allowing for more frequent sampling (as opposed to
~1 year, Matich et al., 2015). However, the downfall to this approach
would be the increased number of captures, and therefore stress,
needed for repeated samples. Therefore, animal welfare may pre-
clude this approach for some species.
To conclude, we found that the association between personal-
ity measured in captivity and use of high- versus low- risk foraging
habitats was principally regulated by one main environmental fac-
tor: predator abundance. This result provides a potential explanation
FIGURE 6 Effect sizes as a function
of (A) predator abundance and (B)
intraspecific competition (both centred at
the between subpopulation level). Solid
lines represent significant effects in the
final model. The colours represent the
different subpopulations, and each point
is a different year. The points are scaled
according to sample size (a larger point
indicates a larger sample size)
FIGURE 7 Effect sizes as a function of predator abundance
(centred at the within subpopulation level). Solid lines represent
significant effects in the final model. The colours represent the
different subpopulations, and the points are scaled according to
sample size (a larger point indicates larger sample size)
DHELLEMM ES Et aL .
regarding why disentangling the association between life- history
traits and personalit y has been complex in terms of conflicting re-
sults repor ted to date: the association between captive personalit y
and wild traits is plastic and is regulated by relevant ecological pres-
sures. We argue in this case that the study of consistent individual
differences in behaviour and their ecological consequences would
benefit from approaches that account for variability in relevant eco-
logical pressures. Multi- population, multi- year studies in wild animals
where ecological conditions can be monitored will in this case be an
important addition to the field, along with highly controlled studies
in captivity where conditions can be manipulated experimentally.
First, we thank S. Isaac the manager of the Movement and Trophic
Ecology Laboratory of the University of Windsor (Integrative
Biology) for preparing the fin samples and undertaking the stable
isotope analysis. We thank the staff and volunteers at the Bimini
Biological Field Station for their help in collection of these valuable
data. We also thank the Save Our Seas Foundation (SOS367) and the
Elsa Neuman Stipendium des Landes for their continued support of
this project as well as the NSERC Discovery Funds awarded to NEH.
Last, we thank our dear friend H. Chowdhurry for his invaluable con-
tribution to this study.
CONFLICT OF INTEREST
The authors declare that they have no conflict of interest.
F.D., T.L.G., J.K. and N.E.H. conceived the ideas and designed meth-
odology; F.D., M.J.S. and T.L.G. collected the data; F.D. and N.E.H.
analysed the samples; F.D. and M.J.S. analysed the data; F.D., M.J.S.
and N.E.H. led the writing of the manuscript. All authors contributed
critically to the drafts and gave final approval for publication.
DATA AVA ILAB ILITY STATE MEN T
Data available from the Dryad Digital Repository https://doi.
org/10.5061/dryad.rr4xg xd8f (Dhellemmes et al., 2021).
Félicie Dhellemmes https://orcid.org/0000-0002-2043-4653
Matthew J. Smukall https://orcid.org/0000-0003-3790-3644
Tristan L. Guttridge https://orcid.org/0000-0002-6706-1452
Jens Krause https://orcid.org/0000-0002-1289-2857
Nigel E. Hussey https://orcid.org/0000-0002-9050-6077
Araújo, M. S., Bolnick, D. I., & Layman, C. A. (2011). The ecological causes
of individual specialisation. Ecology Letters, 14, 948– 958. https://doi.
org /10.1111/j .1461- 0248 .2011.01662.x
Araújo, M. S., Guimarães, P. R., Svanbäck, R., Pinheiro, A., Guimarães,
P., dos Reis, S. F., & Bolnick, D. I. (2008). Network analysis re-
veals contrasting effects of intraspecific competition on individ-
ual vs. population diets. Ecology, 89(7) , 1981– 19 93. https ://doi .
Archard, G. A., & Braithwaite, V. A. (2010). The importance of wild pop-
ulations in studies of animal temperament. Journal of Zoology, 281,
149– 160. https://doi.org/10.1111/j.1469- 7998.2010.00714.x
Bedoya- Perez, M. A., C arthey, A. J. R., Mella, V. S. A., McAr thur, C., &
Banks, P. B. (2013). A practical guide to avoid giving up on giving- up
densities. Behavioral Ecology and Sociobiology, 67(10), 1541– 1553.
h t t p s : // d o i . o r g / 1 0 . 1 0 0 7 / s 0 0 2 6 5 - 0 1 3 - 1 6 0 9 - 3
Bell, A. M. (2004). Behavioural differences between individu-
als and two populations of stickleback (Gasterosteus aculea-
tus). Journal of Evolutionary Biology, 18(2), 464– 473. https://doi.
org /10.1111/j .1420 - 9101 .2 004.0 0817.x
Binder, T. R., Wilson, A. D. M., Wilson, S. M., Suski, C. D., Godin, J.-
G.- J., & Cooke, S. J. (2016). Is there a pace- of- life syndrome linking
boldness and metabolic capacity for locomotion in bluegill sunfish?
Animal Behaviour, 121, 175– 183. https://doi.org/10.1016/j.anbeh
Biro, P. A., & Booth, D. J. (2009). Extreme boldness precedes starvation
mortality in six- lined trumpeter (Pelates sexlineatus). Hydrobiologia,
635, 3 9 5 – 3 9 8 . h t t p s : // d o i . o r g / 1 0 . 1 0 0 7 / s 1 0 7 5 0 - 0 0 9 - 9 9 0 2 - x
Bolnick , D. I., Yang, L . H., Fordyce, J. A ., Davis, J. M., & Svanbäck,
R. (2002). Measuring individual- level resource specialization.
Ecology, 83(10), 2936– 2941. https://doi.org/10.1890/0012-
Boucher, N. P., Derocher, A. E., & Richardson, E. S. (2020). Spatial and
tempor al variabili ty in ringed seal (Pusa hispida) stable isotopes in the
Beaufort Sea. Ecology and Evolution, 10(10), 4178– 4192. ht tps://doi.
Brommer, J. E. (2013). On between- individual and residual (co)variances
in the stud y of anim al pe rso na lit y: Ar e yo u will in g to tak e the ‘in di vid-
ual gambit’? Behavioral Ecology and Sociobiology, 67(6), 1027– 1032.
h t t p s : // d o i . o r g / 1 0 . 1 0 0 7 / s 0 0 2 6 5 - 0 1 3 - 1 5 2 7 - 4
Carneiro, A. P. B., Bonnet- Lebrun, A. S., Manica, A., Staniland, I. J., &
Phillips, R. A. (2017). Methods for detecting and quantifying indi-
vidual specialisation in movement and foraging strategies of marine
predators. Marine Ecology Progress Series, 578, 151– 166. https://doi.
Catano, L. B., Rojas, M. C ., Malossi, R. J., Peters, J. R., Heithaus, M. R.,
Fourqurean, J. W., & Burkepile, D. E. (2016). Reefscapes of fear:
Predation risk and reef hetero- geneity interact to shape herbivore
foraging behaviour. Journal of Animal Ecology, 85(1 ), 146 – 156 . htt ps: //
doi .org/10.1111/1365 - 2656.124 40
Chapman, D. D., Babcock, E. A., Gruber, S. H., Dibat tist a, J. D., Franks, B.
R., Kessel, S. T., Guttridge, T., Pikitch, E. K., & Feldheim, K. A. (20 09).
Long- term natal site- fidelity by immature lemon sharks (Negaprion
brevirostris) at a subtropical island. Molecular Ecology, 18, 3500– 3507.
https ://doi.or g/10.1111/ j.1365- 294X.200 9.04289.x
Cherel, Y., & Hobson, K. A. (2007). Geographical variation in carbon sta-
ble isotope signatures of marine predators: A tool to investigate their
foraging areas in the Southern Ocean. Marine Ecology Progress Series,
329, 281– 287. https://doi.org/10.3354/meps3 29281
Cortés, E., & Gruber, S. (1990). Diet, feeding habits and estimates of daily
ration of young lemon sharks, negaprion brevirostris (Poey). Copeia,
1990 (1), 204– 218. https://doi.org/10.2307/14 45836
Cote, J., Fogart y, S., Tymen, B., Sih, A., & Brodin, T. (2013). Personality-
dependent dispersal cancelled under predation risk. Proceedings of
the Royal Society B: Biological Sciences, 280(1773), 20132349. https://
doi.org/10.1098/rspb. 2013 .2349
Dammhahn, M., Dingemanse , N. J., Niemel ä, P. T., & Réale, D. (2018). Pace-
of- life syndromes: A framework for the adaptive integration of be-
haviour, physiolog y and life history. Behavioral Ecology and Sociobiology,
72( 3 ) , 6 2 – 7 0 . h t t p s : / / d o i . o r g / 1 0 . 1 0 0 7 / s 0 0 2 6 5 - 0 1 8 - 2 4 7 3 - y
Dhellemmes, F., Finger, J.- S., Laskowski, K. L., Guttridge, T. L., & Krause,
J. (2020). Comparing behavioural syndromes across time and eco-
logical conditions in a free- ranging predator. Animal Behaviour, 162,
23– 33. https://doi.org/10.1016/j.anbeh av.2020.01.009
DHELLEMM ES Et aL .
Dhellemmes, F., Finger, J.- S., Smukall, M. J., Gruber, S. H., Guttridge,
T. L., Laskowski, K. L., & Krause, J. (2020). Personality- driven life
histor y trade- offs differ in two subpopulations of free- ranging
predators. Journal of Animal Ecology, 89(8), 1– 13. https://doi.
org /10.1111/1365- 2656.13283
Dhellemmes, F., Smukall, M. J., Guttridge, T. L., Krause, J., & Hussey,
N. E. (2021). Data from: Predator abundance drives the associa-
tion between exploratory personality and foraging habitat risk in
a wild marine meso- predator. Dryad Digital Repository, h t t p s : //d oi.
DiBattista, J. D., Feldheim, K. A., Gruber, S. H., & Hendry, A. P.
(2007). When bigger is not better: Selection against large
size, high condition and fast growth in juvenile lemon sharks.
Journal of Evolutionary Biology, 20(514), 201– 212. https://doi.
org /10.1111/j .1420 - 9101 .2 006.01210. x
Dingemanse, N. J., & Wright, J. (2020). Criteria for acceptable studies
of animal personalit y and behavioural syndromes. Ethology, 126 (9),
865– 869. https://doi.org/10.1111/eth.13082
Dingemanse, N. J., Wright, J., Kazem, A. J. N., Thomas, D. K.,
Hickling, R., & Dawnay, N. (2007). Behavioural syndromes dif-
fer predictably between 12 populations of three- spined stick-
leback. Journal of Animal Ecology, 76 (6), 1128– 1138. https://doi.
org /10.1111/j .1365- 2656 .2007.012 84 .x
Feldheim, K. A ., Gruber, S. H., & Ashley, M. V. (2002). The breeding biol-
ogy of lemon shark s at a tropical nursery lagoon. Proceedings of the
Royal Societ y of London. Serie s B: Biological Sciences, 269(1501), 1655–
Ferrari, M. C. O., Sih, A., & Chivers, D. P. (2009). The paradox of risk allo-
cation: A review and prospectus. Animal Behaviour, 78(3), 579– 585.
Finger, J.- S., Dhellemmes, F., Guttridge, T. L., Kurvers, R. H. J. M. J. M.,
Gruber, S. H., & Krause, J. (2016). Rate of movement of juvenile lemon
sharks, Negaprion brevirostris in a novel open field, are we measuring
activity or reaction to novelty? Animal Behaviour, 116 (116), 75– 82.
Finger, J.- S., Guttridge, T. L ., Wilson, A. D. M., Gruber, S. H., & Krause,
J. (2018). Are some sharks more social than others? Short- and
long- term consistencies in the social behavior of juvenile lemon
sharks. Behavioral Ecology and Sociobiology, 72(1), 17. ht tp s://doi .
o r g / 1 0 . 1 0 0 7 / s 0 0 2 6 5 - 0 1 7 - 2 4 3 1 - 0
Gruber, S. H., de Marignac, J. R . C., & Hoenig, J. M. (2001). Sur vival of ju-
venile lemon sharks at Bimini, Bahamas, estimated by mark- depletion
experiments. Transactions of the American Fisheries Society, 130 (3),
376– 384. https://doi.org/10.1577/1548- 8659(2001)130
Guenther, A. (2018). Life- history trade- offs: Are they linked to person-
ality in a precocial mammal (Cavia aperea)? Biology Letters, 14(4).
Guttridge, T. L., Gruber, S. H., Franks, B. R ., Kessel, S. T., Gledhill, K. S.,
Uphill, J., Krause, J., & Sims, D. W. (2012). Deep danger: Intra- specific
predation risk influences habitat use and aggregation formation of
juvenile lemon sharks Negaprion brevirostris. Marine Ecology Progress
Series, 445, 279– 291. https://doi.org/10.3354/meps0 9423
Guttridge, T. L., Van Zinnicq Bergmann, M. P. M., Bolte, C., Howey, L.
A., Finger, J. S., Kessel, S. T., Brooks, J. L., Winram, W., Bond, M.
E., Jordan, L. K. B., Cashman, R. C ., Tolentino, E. R., Grubbs, R. D.,
& Gruber, S. H. (2017). Philopatry and regional connec tivit y of the
great hammerhead shark, Sphyrna mo karran in the U. S. and Bahama s.
Frontiers in Marine Science, 4, 3– 18. https://doi.org/10.3389/
Hadfield, J. D. (2010). MCMC methods for multi- response generalized
linear mixed models: The MCMCglmm R package. Journal of Sta tistical
Software, 33 (2), 1– 22.
Hansell, A . C., Kessel, S. T., Brewster, L. R., Cadrin, S. X., Gruber, S. H.,
Skomal, G. B., & Guttridge, T. L. (2018). Local indicators of abun-
dance and demographic s for the coast al shark assemblage of Bimini,
Bahamas. Fisheries Research, 197(Oc tober), 34– 44. https://doi.
Harris, S. M ., Descamps, S., Sneddon, L. U., Bertrand, P., Chastel, O., &
Patrick, S. C. (2020). Personality predicts foraging site fidelit y and
trip repeatability in a marine predator. Journal of Animal Ecology,
89(1), 68– 79. https://doi.org/10.1111/1365- 2656.13106
Heithaus, M. R., Frid, A., Wirsing, A. J., Dill, L. M., Fourqurean, J.
W., Burkholder, D., Thomson, J., & Bejder, L. (2007). State-
dependent risk- taking by green sea turtles mediates top-
down effec ts of tiger shark intimidation in a marine ecosys-
tem. Journal of Animal Ecology, 76 (5), 837– 844. ht tps://doi.
org /10.1111/j .1365- 2656 .2007.01260.x
Hoffmayer, E. R., & Parsons, G. R. (2006). Food habits of three shark
species from the Mississippi sound in the Northern Gulf of Mexico.
Southeastern Naturalist, 2(2), 271– 280. https://doi.org/10.1656/1528-
Hussey, N. E., DiBattist a, J. D., Moore, J. W., Ward, E. J., Fisk, A. T.,
Kessel, S., Guttridge, T. L., Feldheim, K. A., Franks, B. R., Gruber, S.
H., Weideli, O. C., & Chapman, D. D. (2017). Risky business for a ju-
venile marine predator? Testing the influence of foraging strategies
on size and growth rate under natural conditions. Proceedings of the
Royal Societ y B: Biological Sciences, 284(1852), 20170166. ht tps ://doi.
Kalinkat, G. (2014). Bringing animal personalit y research into the food
web arena. Journal of Animal Ecology, 83(6), 1245– 1247. https://doi.
org /10.1111/1365- 2656.12 28 4
Kinney, M. J., Hussey, N. E., Fisk, A. T., Tobin, A. J., & Simpfendorfer,
C. A. (2011). Communal or competitive? Stable isotope analysis pro-
vides evidence of resource partitioning within a communal shark
nurser y. Marine Ecology Progress Series, 439, 263– 276. https://doi.
Lajeunesse, M. J., & Forbes, M. R. (2003). Variable repor ting and
quantitative reviews: A comparison of three meta- analytical
techniques. Ecology Letters, 6(5), 448– 454. https://doi.
Li, Y., Zhang, Y., Hussey, N. E., & Dai, X. (2016). Urea and lipid extraction
treatment effects on δ15N and δ13C values in pelagic sharks. Rapid
Communications in Mass Spectrometry, 30(1), 1– 8. https://doi.
Mangel, M., & Stamps, J. A. (2011). Trade- offs between growth and
mortality and the maintenance of individual variation in grow th.
Evolutionary Ecolog y Research, 3(1), 583– 593.
Matich, P., Kiszka, J. J., Heithaus, M. R., Mourier, J., & Planes, S. (2015).
Short- term shifts of stable isotope (δ13C, δ15N) values in juvenile
sharks within nursery areas suggest rapid shift s in energ y path-
ways. Journal of Experimental Marine Biolog y and Ecology, 465, 83– 91.
Moirón, M., Laskowski, K. L ., & Niemelä, P. T. (2020). Individual differ-
ences in behaviour explain variation in survival: A meta- analysis.
Ecology Letters, 23(2), 399– 408. https://doi.org/10.1111/ele.13438
Morrissey, J. F., & Gruber, S. H. (1993). Home range of juvenile lemon
sharks, Negaprion brevirostris. Copeia, 1993(2), 425– 434. https://doi.
org /10. 23 07/1447141
Newman, S. P., Handy, R. D., & Gruber, S. H. (2007). Spatial and temporal
variations in mangrove and seagrass faunal communities at Bimini,
Bahamas. Bulletin of Marine Science, 80(3), 529– 553.
Niella, Y., Raoult, V., Gaston, T., Peddemors, V. M., Harcourt, R., &
Smoothey, A. F. (2021). Overcoming multi- year impacts of maternal
isotope signatures using multi- tracers and fast turnover tissues in ju-
venile sharks. Chemosphere, 269, 129393. https://doi.org/10.1016/j.
Niemelä, P. T., & Dingemanse, N. J. (2018). On the usage of single mea-
surements in behavioural ecology research on individual differences.
Animal Behaviour, 145, 99– 105. https://doi.org/10.1016/j.anbeh
DHELLEMM ES Et aL .
Olin, J. A ., Hussey, N. E., Fritt s, M., Heupel, M. R., Simpfendorfer, C. A.,
Poulakis, G. R., & Fisk, A. T. (2011). Maternal meddling in neonatal
sharks: Implications for interpreting stable isotopes in young ani-
mals. Rapid Communications in Mass Spectrometry, 25(8), 1008– 1016.
https://doi.org/10.10 02 /rcm.4946
Perals, D., Griffin, A. S., Bartomeus, I., & Sol, D. (2017). Revisiting the
open- field test: What does it really tell us about animal personal-
it y? Animal Behaviour, 123, 69– 79. https://doi.org/10.1016/j.anbeh
Polis, G. A. (1984). Age structure component of niche width and intra-
specific resource partitioning: Can age groups function as ecolog-
ical species? The American Naturalist, 123(4), 541– 564. https://doi.
Powell, E. C., & Taylor, L . A. (2017). Specialists and generalists coexis t
within a population of spider- hunting mud dauber wasps. Behavioral
Ecology, 28(3), 890– 898. https://doi.org/10.1093/behec o/arx050
Pyke, G. H. (1984). Optimal foraging theory: A critical review. Annual
Review of Ecology and Sys tematics, 15(1), 523– 575. ht tps://doi.
o r g / 1 0 . 1 0 1 6 / B 9 7 8 - 0 - 4 4 4 - 6 3 7 6 8 - 0 . 0 0 0 2 6 - 3
R Core Team. (2017). A language and environment for statistical comput-
ing. R Foundation for Statistical Computing. Retrieved from https://
w w w . r - p r o j e c t . o r g / % 0 A
Rangel, B., Hussey, N., Niella, Y., Martinelli, L., Gomes, A., & Moreira, R.
(2020). Neonatal nutritional strategy of a viviparous elasmobranch
with extremely low reproductive output. Marine Ecolog y Progress
Series, 638, 107– 121. https://doi.org/10.3354/meps1 3261
Ratcliffe, N., Adlard, S., Stowasser, G., & McGill, R. (2018). Dietar y diver-
genc e is ass ocia te d wit h in cre as ed intr a- spe cifi c compe tit io n in a ma-
rine predator. Scientific Reports, 8(1), 6827. https://doi.org/10.1038/
s 4 1 5 9 8 - 0 1 8 - 2 5 3 1 8 - 7
Réale, D., Garant, D., Humphries, M. M., Bergeron, P., Careau, V., &
Montiglio, P.- O. (2010). Personality and the emergence of the
pace- of- life syndrome concept at the population level. Philosophical
Transactions of t he Royal Society of Lon don. Series B, B iological Scie nces,
365, 4051– 4063. https://doi.org/10.1098/rstb.2010.0208
Réale, D., Reader, S. M., Sol, D., McDougall, P. T., & Dingemanse,
N. J. (2007). Integrating animal temperament within ecology
and evolution. Biological Reviews, 82(2), 291– 318. https://doi.
org /10.1111/j .1469- 185X .20 07.00010. x
Rodríguez- Prieto, I., Martín, J., & Fernández- Juricic, E. (2011). Individual
variation in behavioural plasticity: Direct and indirect effects of
boldness, exploration and sociability on habituation to predators
in lizards. Proceedings of the Royal Society B: Biological Sciences, 278,
266– 273. https://doi.org/10.1098/rspb.2010.1194
Royauté, R ., Berdal, M. A., Garrison, C. R., & Dochtermann, N. A. (2018).
Paceless life? A meta- analysis of the pace- of- life syndrome hy-
pothesis. Behavioral Ecology and Sociobiology, 72(3), 64. https://doi.
o r g / 1 0 . 1 0 0 7 / s 0 0 2 6 5 - 0 1 8 - 2 4 7 2 - z
Sheppard, C. E., Inger, R., McDonald, R. A., Barker, S., Jackson, A. L.,
Thompson, F. J., Vitikainen, E. I. K., Cant, M. A., & Marshall, H. H.
(2018). Intragroup competition predicts individual foraging special-
isation in a group- living mammal. Ecology Letters, 21(5), 665– 673.
https ://doi.or g/10.1111/e le.12933
Shine, R. (1989). Ecological causes for the evolution of sexual dimor-
phism: A review of the evidence. Quarterly Review of Biology, 64(4),
419– 461. https://doi.org/10.1086/416458
Singer, M. S., Clark, R. E., Johnson, E. R., Lichter- Marck, I. H., Mooney,
K. A., & Whitney, K. D. (2019). Dietary specialization is conditionally
associated with increased ant predation risk in a temperate forest
caterpillar community. Ecolog y and Evolution, 9(21), 12099– 12112.
Spiegel, O., Leu, S. T., Bull, C. M., & Sih, A. (2017). What's your
move? Movement as a link between personality and spatial
dynamics in animal populations. Ecology Letters, 20, 3– 18. https://
doi .org/10.1111/ele .12708
Stamps, J. A. (2007). Growth- mortality tradeoffs and ‘personal-
ity traits’ in animals. Ecology Letters, 10(5), 355– 363. https://doi.
org /10.1111/j .1461- 0248 .2007.01034.x
Svanbäck , R., & Bolnick, D. I. (2007). Intraspecific competition drives
increased resource use diversity within a natural population.
Proceedi ngs of the Royal Society B: B iological Sciences, 274 (1611), 839–
Tinoco, B. A., Graham, C. H., Aguilar, J. M., & Schleuning, M. (2017).
Effects of hummingbird morpholog y on specialization in pollina-
tion networks vary with resource availability. O ikos, 126(1), 52– 60.
https ://doi.or g/10.1111/oik.02998
Toscano, B. J., Gownaris, N. J., Heerhartz, S. M., & Monaco, C. J. (2016).
Personality, foraging behavior and specialization: Integrating behav-
ioral and food web ecology at the individual level. Oecologia, 182(1),
5 5 – 6 9 . h t t p s : / / d o i . o r g / 1 0 . 1 0 0 7 / s 0 0 4 4 2 - 0 1 6 - 3 6 4 8 - 8
Tuma, R. E. (1976). An investigation of the feeding habits of the bull shark,
Carcharhinus leucas, in the Lake Nicaragua- Rio San Juan System.
Investigations of the Ichthyofauna of Nicaraguan Lakes, 39, 533– 538.
van de Pol, M., & Wright, J. (2009). A simple method for distinguish-
ing within- versus bet ween- subject effects using mixed models.
Animal Behaviour, 77(3), 753– 758. ht tp s://doi.org /10.1016/ j.anbeh
Viechtbauer, W. (2010). Conducting meta- analyses in R with the meta-
for package. Journal of Statistical Sof tware. 36(3), 1– 4 8. https://doi.
org/10.1103/Phys R evB.91.121108
Vorenberg, M. M. (1962). Cannibalistic tendencies of lemon and bull
sharks. Copeia, 1962 (2), 455– 456. https://doi.org/10.2307/1440934
Wakefield, E. D., Cleasby, I. R ., Bearhop, S., Bodey, T. W., Davies,
R. D., Miller, P. I., Newton, J., Votier, S. C ., & Hamer, K. C. (2015).
Long- term individual foraging site fidelity – Why some gannets
don't change their spots. Ecology, 96(11), 3058– 3074. https://doi.
Wetherbee, B. M., Gruber, S. H., & Cortés, E. (1990). Diet, feeding habits,
digestion, and consumption in Sharks, with special references to the
Lemon Shark, Negaprion brevirostris. In H. L . J. Pratt, S. H. Gruber, &
T. Taniuchi (Eds.), Elasmobranchs as living re sources: Advance s in the
biology, ecology, systematics , and the status of the fisheries (pp. 29– 46).
U.S. Department of Commerce, National Oceanic and Atmospheric
Administration, National Marine Fisheries Service.
Wolf, M., & Weissing, F. J. (2012). Animal personalities: Consequences
for ecology and evolution. Trends in Ecology & Evolution, 27(8), 452–
Additional supporting information may be found online in the
Supporting Information section.
How to cite this article: Dhellemmes, F., Smukall, M. J.,
Guttridge, T. L., Krause, J., & Hussey, N. E. (2021). Predator
abundance drives the association between explorator y
personality and foraging habitat risk in a wild marine
meso- predator. Functional Ecology, 00, 1– 13. h t t p s : // d o i .
org /10.1111/13 65- 24 35.13874