RESEARCH
PAPER
Latitudinal gradient in niche breadth of
brachyuran crabs
Katherine J. Papacostas1,2* and Amy L. Freestone1,3
1Department of Biology, Temple University,
Philadelphia, PA 19122, USA, 2Moore Center
for Science and Oceans, Conservation
International, Arlington, VA 22202, USA,
3Smithsonian Tropical Research Institute,
Apartado 0843-03092 Balboa, Ancon, Panama
ABSTRACT
Aim Niche breadth has long been hypothesized to decrease at low latitudes and
contribute to global patterns of species diversity. Range size, phylogenetic related-
ness and body size also have hypothesized relationships with both latitude and
niche breadth, which may further affect niche breadth patterns. Existing terrestrial
data are inconclusive and few data exist on latitudinal gradients in niche breadth in
the marine realm. We tested the latitude–niche breadth relationship in a marine
system while exploring the correlations of both variables with range size, and
accounting for relatedness and body size.
Location Global.
Methods We compiled a global dataset on the dietary niche breadth of 39
brachyuran crab species from existing studies and additional analyses on species
collected in Connecticut and Florida, USA and Bocas del Toro, Panama. Estimates
of latitude, range size, clade and body size were obtained for each species. We then
tested for correlations among focal variables and examined the strength of their
relationships with diet breadth.
Results Latitude was the strongest predictor of niche breadth in temperate
species, and the latitude–niche breadth relationship was stronger in larger-bodied
species. The strongest predictor of the niche breadth of tropical species was clade,
with the newest clade having the narrowest diet. Niche breadth was related to range
size for both temperate and tropical species. Tropical species had larger ranges on
average than temperate species.
Main conclusions We found an interesting division in the niche breadth rela-
tionships of temperate and tropical species; diets of temperate species were posi-
tively correlated with latitude, range size and body size, and diets of tropical species
were related to range size and clade. Therefore, only temperate species demon-
strated the predicted positive relationship between niche breadth and latitude,
while evolutionary history was a stronger predictor of niche breadth in tropical
species.
Keywords
Body size, brachyuran crab, diet, gut content analysis, latitude, niche breadth,
phylogenetic relatedness, range size.
*Correspondence: Katherine J. Papacostas,
Department of Biology, Temple University,
Philadelphia, PA 19122, USA.
E-mail: kpapacostas@temple.edu
INTRODUCTION
Niche breadth is defined as the range of environmental condi-
tions and resources that a species can utilize (MacArthur, 1968),
and can range from very broad (i.e. generalist species) to very
narrow (i.e. specialist species). Species interaction strength, such
as competition, may determine niche breadth according to
classic ecological theory; weaker species interactions should
allow for species to evolve a broader niche (Bolnick et al., 2010)
while strong interactions should, over time, drive species to
bs_bs_banner
Global Ecology and Biogeography, (Global Ecol. Biogeogr.) (2015)
© 2015 John Wiley & Sons Ltd DOI: 10.1111/geb.12400
http://wileyonlinelibrary.com/journal/geb 1
utilize fewer resources and show greater specialization, which in
turn should reduce interspecific competition (Bolnick et al.,
2010).
It has long been hypothesized that niche breadth varies with
latitude (MacArthur, 1972). Species diversity peaks in the
tropics (Hillebrand, 2004b) and studies have shown that species
interaction strength varies spatially, being strongest at lower
latitudes (Schemske et al., 2009; Freestone et al., 2011). These
strong species interactions could drive a higher prevalence of
specialist species at lower latitudes, which is a proposed mecha-
nism contributing to the coexistence of tropical species
(Schemske et al., 2009). Despite this, research on the relation-
ship between niche breadth and latitude remains inconclusive.
Terrestrial studies have shown mixed results (Vazquez & Stevens,
2004), with approximately half of existing studies supporting a
latitudinal gradient in niche breadth [e.g. Krasnov et al. (2008)
and Belmaker et al. (2012), who specifically examined richness–
specialization patterns] and half refuting it (e.g. Fiedler, 1998;
Slove & Janz, 2010). In the marine realm, examination of lati-
tudinal gradients in specialization are rare (but see Rohde, 1978;
Sunday et al., 2011) and it is unclear how niche breadth relates
to latitude in marine species.
Other factors may also be related to niche breadth, either
alone or in conjunction with latitude. For instance, the range of
resources a species is able to utilize can influence geographical
range size (Slatyer et al., 2013), with generalist species having
larger ranges and specialist species having narrower ranges.
However, range size may also be correlated with latitude;
Rapoport’s rule predicts that species at higher latitudes have
larger ranges on average than species restricted to lower latitudes
(Stevens, 1989), potentially due to the increased seasonal vari-
ability at higher latitudes selecting for broader climatic toler-
ances (i.e. broader environmental niches; Fernandez & Vrba,
2005). Both latitude and range size thus have a hypothesized
relationship with niche breadth, but may also be related and
could interactively influence interpretations of patterns of niche
breadth. For instance, recent research suggests that a narrow
dietary niche breadth at lower latitudes may be related to the
smaller range size of low-latitude species rather than having a
direct relationship with latitude (Slove & Janz, 2010).
Phylogenetic relatedness may also affect niche breadth
(Barnagaud et al., 2014). Specifically, very closely related species
are expected to be more ecologically similar due to niche con-
servatism (Kerkhoff et al., 2014), thus the niche breadth of such
species may also be similar. Conflicting hypotheses exist for the
distribution of related clades across latitude. The ‘out of the
tropics’ model suggests that more species originate in the tropics
and then spread towards the poles (Jablonski et al., 2006), in
which case newer clades should be more prevalent in the tropics.
In contrast, the ‘tropical niche conservatism’ hypothesis posits
that older, more basal, clades remain closer to the tropics, and
newer derived clades that have overcome the cold winter tem-
perature barrier are more prevalent at higher latitudes (Hawkins
& DeVries, 2009).
Finally, body size has also been proposed to contribute to diet
breadth; large species have been suggested to be more generalist,
and able to utilize a wider variety of prey resources, than smaller
species (Ashmole, 1968; Novotny & Basset, 1999). Support for
this hypothesis is mixed, however, and other more recent studies
of vertebrate and invertebrate taxa suggest that larger-bodied
animals may target larger more energetically profitable prey and
thus may be more specialized (e.g. Costa et al., 2008). Body size
is also predicted by Bergmann’s rule to increase with latitude
(Bergmann, 1848), with lower average temperatures being a
hypothesized driver of greater cell growth (van Voorhies, 1996),
resulting in another potentially confounding factor in the niche
breadth–latitude relationship.
We therefore examined latitudinal trends in niche breadth
using brachyuran crabs as a model system and diet as a measure
of niche breadth, while accounting for variation in phylogenetic
relatedness and body size. We further explored the bivariate
relationships of range size with niche breadth and latitude to aid
in interpreting findings. Due to limited phylogenetic variation
in the brachyuran crab group, and to mixed support for body
size as an important predictor of niche breadth, we hypothesized
that these covariates would have limited importance in under-
standing niche breadth–latitude relationships in brachyuran
crabs. We therefore predicted that dietary niche breadth would
increase with latitude and that species range sizes would be
positively correlated with both dietary niche breadth and
latitude.
METHODS
We conducted a global analysis of brachyuran crab diets by
compiling published data (n = 30 species) and completing diet
analyses on temperate, subtropical and tropical crabs (n = 10
species; one species whose diet had been analysed in numerous
other studies and nine species whose diets had not been previ-
ously analysed). We searched ISI Web of Science for relevant
studies using the broad search string ‘crab’ AND ‘marine’ AND
‘diet’. We then included data from studies that conducted gut
content analyses on a population of crabs of one or more species
in their native range, and reported the average diet composition
of those species as either volumetric proportion (%V) data or
frequency of occurrence (%O) data (a full list of studies
included and species examined can be found in Appendices S1
and S2 in the Supporting Information). Volumetric proportion
(%V) is defined as the estimated volume per prey item in an
individual’s gut, averaged across all the analysed guts of that
species (total n = 100%). Frequency of occurrence (%O) is
defined as the number of times a prey item is found in an
individual’s stomach (Ni; for i = category 1 to n), divided by the
total number of guts analysed, multiplied by 100 (total
n > 100%).
To supplement the data obtained from the literature, gut
content analyses were conducted on animals collected from
Connecticut and Florida, USA (41.320° N, 72.057° W; n = 2
native crab species) and Bocas del Toro, Panama (9.352° N,
82.258° W; n = 8 native crab species). Specimens in each region
were gathered at the same time of day (morning), and during
the late summer/autumn. Specimens were preserved within 2 h
K. J. Papacostas and A. L. Freestone
Global Ecology and Biogeography, © 2015 John Wiley & Sons Ltd2
of collection. For each crab we noted the sex and size, measuring
the carapace width to the nearest millimetre. We then removed
the carapace, extracted the stomach with forceps and measured
the width of the cardiac stomach (posterior ventral edge) to the
nearest millimetre (Griffen & Mosblack, 2011). Using a dissec-
tion microscope, we visually identified gut contents for individ-
uals of each species (4–16 individuals per species, N = 88) to the
lowest possible taxonomic level (Appendix S3). Gut contents
were separated by food type into grid cells on a Petri dish. We
determined the proportional contribution of each food type
from the number of grid cells (or portions of grid cells) that
each occupied. This method provided an estimate of the per-
centage of each food type in the diet (Griffen & Mosblack,
2011). We used these data to calculate both %V and %O of prey
items for each species.
In order to standardize the dataset (total n = 100%), we con-
verted frequency of occurrence to relative frequency (%F) of
prey items (Safi & Kerth, 2004) using the following equation:
% (% ) % .F O O= ( )
=
∑i iin 1
We then calculated Levins’ standardized measure of niche
breadth (Hurlbert, 1978) using %V and %F data for each species
(n = 39) as follows:
B B nA = ( ) ( )– – ,1 1
where BA is Levins’ standardized niche breadth, B is Levins’
measure of niche breadth and n is a constant that reflects the
total number of prey resources used across all species. Levins’
measure of niche breadth (Levins’ B) (Safi & Kerth, 2004) was
calculated as
B Bi
i
n
i
i
n
= ( ) = ( )
= =
∑ ∑1 121 21% % .F or V
Standardizing Levins’ B puts the measure of niche breadth on a
scale from zero to one. Those species with a Levins’ B closer to
zero have a narrower diet and are more specialized, and those
with a Levins’ B closer to one have a broader diet and are more
generalist. When multiple studies examined the same crab
species, we calculated a standardized Levins’ B from the average
%V or %F of prey items across the studies examining that par-
ticular species in order to obtain one value for Levins’ B per crab
species. Most studies had unidentifiable prey items (usually
labelled ‘detritus’ or ‘unidentifiable remains’), which were
excluded from analyses. Additionally, the way in which %V and
%F are inferred from stomach contents differs fundamentally, so
we tested for a relationship between standardized Levins’ B
values calculated from both %V and %F, using our own data.
Calculations of standardized Levins’ B values using %V and %F
data had a strong linear relationship (r2 = 0.86, N = 10,
P < 0.0001), and Levins’ B values using %F data were converted
to Levins’ B values using %V data using the following equation
prior to other analyses:
B BA from %V data A from %F data( ) = × −0 5944 0 0044. . .( )
Our estimates of diet breadth are conservative: brachyuran crabs
are considered a generalist taxon (Griffen & Mosblack, 2011)
and the taxonomic resolution of prey items reported in the
literature is coarse (generally phylum, class or order) due to the
feeding mode of crabs (McGaw & Reiber, 2000). However, visual
gut content analysis was the most commonly used method for
examining crab diet and provides a useful estimate of diet
breadth. Also, prey availability, a factor that can influence diet
breadth (Petraitis, 1979), was unavailable in this dataset. Given
our coarse taxonomic resolution of prey items, however, it is
highly likely that these broad prey categories were available as
food items in all locations (e.g. Gosling, 2003; Horton et al.,
2013), ensuring a conservative and comparable estimate of diet
breadth for all focal species.
After diet breadth had been calculated for each species we
obtained estimates for our focal predictor variables. Most crabs
in our dataset were coastal species representative of all oceans
apart from the Southern, although the majority of temperate
studies were conducted in the east and west Atlantic (Appendix
S2). For each crab species, latitudinal range size and latitudinal
position were estimated primarily through species distribution
data and species occurrence maps found on the World
Register of Marine Species database (WoRMS Editorial Board
2014). We calculated latitudinal range size (hereafter referred to
as range size) by subtracting the degree of latitude of the lowest-
latitude observations of the species in its native range from the
highest-latitude extent of the species in either the Northern or
Southern Hemisphere. If species ranges extended across both
hemispheres, the total distribution of the species in degrees
latitude was used for this estimate of range size (Appendix S2).
We then used these data to determine the midpoint of each
species’ geographic range (hereafter termed latitude). This mid-
point approach is generally accepted for use in biogeographic
studies (e.g. Rohde, 1999; Krasnov et al., 2008; Sunday et al.,
2012) based on the classical ‘abundant centre’ hypothesis which
posits that species abundances peak in the centre of their ranges
(Wulff & Brissenden, 1943; Rohde, 1999; Fenberg &
Rivadeneira, 2011; but see Sagarin & Gaines, 2002; Ruggiero &
Werenkraut, 2007, for limitations of this method). Only one
species lacked distribution information on WoRMS, and for
that species we consulted the literature directly to determine its
latitudinal distribution (Appendix S2). Each crab species was
then categorized by clade; while Brachyura is a monophyletic
group, it can be divided into seven clades of increasing distance
from the ancestral species (Brosing et al., 2007). The 39
species studied fell into three clades: (1) Majoidea (n = 10),
(2) Cancridae/Portunidae/Xanthoidea (n = 17), and (3)
Neobrachyura (n = 12); Majoidea are closest to the ancestral
species (Brosing et al., 2007). Average body sizes for each species
were obtained from the literature (Appendix S2).
We tested for linear relationships between: (1) latitude and
range size, as predicted by Rapoport’s rule (Stevens, 1989), (2)
latitude and clade, and (3) latitude and average body size as
predicted by Bergmann’s rule (Bergmann, 1848) to detect any
potential multicollinearity in the dataset. Approximately half of
our data were for tropical species (n = 18) and the other half for
Latitudinal gradient in crab niche breadth
Global Ecology and Biogeography, © 2015 John Wiley & Sons Ltd 3
temperate species (n = 21), and in these preliminary analyses we
found an interesting division in latitude–range size relationships
between tropical and temperate species (see Appendix S4 and
Results). Tropical species were defined as those whose latitudinal
midpoints fell between 23.5° N and 23.5° S, and temperate
species as those whose latitudinal midpoints fell between 23.5
and 90° N or S. To ensure that this division was not driven
intrinsically by the way latitudinal range size and midpoint were
calculated (since temperate species with large ranges could by
default have a higher latitudinal midpoint, while tropical species
spanning both hemispheres would not), we conducted a
randomization test to compare the slope (i.e. the strength) of
our latitude–range size relationship for temperate species with
the slope that would be observed for the relationship purely by
chance. We generated 10,000 random datasets consisting of 21
high-latitude range limits and 21 low-latitude range limits
between 0° and 90°, including a condition that random range
limits would continue to be generated until all the midpoints of
those ranges were above 23.5° (R 3.2.1; R Core Team, 2014).
From these range limits we calculated the latitudinal range sizes
and latitudinal midpoints for each dataset. We then calculated
the slope of the latitude–range size relationship for each of the
datasets, as well as the 95% confidence interval of these slopes, to
compare with the slope of our own dataset; our observed slope
of 0.535 was well outside the 95% confidence interval (−0.103,
−0.097) of the randomized dataset, supporting that the relation-
ship observed in our dataset was much stronger than expected
by chance. We therefore proceeded to analyse tropical and tem-
perate species separately.
We examined our focal predictors of diet breadth for both
temperate and tropical species by conducting backward stepwise
(Neter et al., 1996) model selection using the corrected Akaike
information criterion (AICc), which is the AIC adjusted for
small sample size (Burnham & Anderson, 2002). We further
calculated Akaike weights (wi) and evidence ratios (wj/wi) for
each model for comparison purposes. The predictor variables
included in the two full models were as follows: latitude, clade,
average body size, body size range of individuals dissected (to
control for sample bias in the dataset), sample size for each
species (to control for sample bias in the dataset), and interac-
tions between latitude and our focal covariates (see Tables 1 &
2). To complement the model selection we also ran linear regres-
sions to assess the relationship between all continuous predictor
variables and diet breadth as well as an ANOVA for the effect of
clade on niche breadth. Range size was not included in the
model selection and the bivariate relationship between diet
breadth and range size was analysed separately since the
expected direction of causality of this relationship is the oppo-
Table 1 Results of model selection for tropical species. Comparison of the best fit model and the four other models, listing the corrected
Akaike information criterion (AICc), the difference from the ‘best fit’ model (ΔAICc), the Akaike weight (wi) and the evidence ratio (wj/wi)
for each model. Main effects are midpoint of latitudinal range (MLR), clade (C), average body size of crabs sampled (ABS), sample size for
each crab species (SS) and body size range of crabs within the sample (BSR).
Model No. of parameters Variables included in model AICc ΔAICc wi wj/wi
M5 (best fit) 1 Main effects: C −53.30 0.00 0.822 1
M4 2 Main effects: C, ABS −50.22 3.08 0.177 4.655
M3 3 Main effects: C, ABS, SS −40.93 12.37 1.70 × 10−3 4.85 × 102
M2 4 Main Effects: MLR, C, ABS, SS 0.18 53.48 2.01 × 10−12 4.10 × 1011
Interactions: MLR × C, MLR × ABS
M1 (full) 6 Main Effects: MLR, C, SS, BSR 34.91 88.21 5.76 × 10−20 1.43 × 1019
Interactions: MLR × C, MLR × ABS
Table 2 Results of model selection for temperate species. Comparison of the best fit model and the four other models, listing the corrected
Akaike information criterion (AICc), the difference from the ‘best fit’ model (ΔAICc), the Akaike weight (wi) and the evidence ratio (wj/wi)
for each model. Main effects are midpoint of latitudinal range (MLR), clade (C), average body size of crabs sampled (ABS), sample size for
each crab species (SS) and body size range of crabs within the sample (BSR).
Model No. of parameters Variables included in model AICc ΔAICc wi wj/wi
M5 (best fit) 4 Main effects: MLR −61.96 0 0.575 1.00
M4 2 Main effects: MLR, ABS −60.93 1.03 0.343 1.67
Interactions: MLR × ABS
M3 5 Main effects: MLR, ABS, SS −58.06 3.9 8.18 × 10−2 7.03
Interactions: MLR × ABS
M2 6 Main Effects: MLR, C, ABS, SS −35.33 26.63 9.48 × 10−7 6.06 × 105
Interactions: MLR × C, MLR × ABS
M1 (full) 7 Main Effects: MLR, RS, C, SS, BSR −24.93 37.03 5.23 × 10−9 1.10 × 108
Interactions: MLR × C, MLR × ABS
K. J. Papacostas and A. L. Freestone
Global Ecology and Biogeography, © 2015 John Wiley & Sons Ltd4
site of the other factors of interest (Ashmole, 1968; MacArthur,
1972; Krasnov et al., 2008; Slatyer et al., 2013). One outlier was
identified in the dataset Callinectes sapidus (average body size of
12.2 cm, size range 0.1–24 cm and sample size of 4117 individ-
uals which were 4, 8 and 40 times greater than those values for all
other crabs, respectively; Appendix S2) and was removed prior
to analyses.
RESULTS
Overall, brachyuran crabs have broad diets characteristic of gen-
eralist species, using prey resources from 25 taxonomic catego-
ries (Appendix S5). Most crabs were found to utilize both plant
and animal resources, with crustaceans and molluscs being the
dominant animal food groups (Appendix S5). Plant and algal
matter were the dominant food items used across species, with
over 70% of the individuals examined consuming some type
of vascular plant and over 60% consuming algal material
(Appendix S5).
Our hypothesis of a positive latitude–range size relationship
was only supported by temperate species, while a positive rela-
tionship between range size and diet breadth was supported
across the dataset. First, the relationship between latitude and
range size differed significantly for tropical and temperate
species (Fig. 1a, b, Appendix S4). We found that tropical species
had large ranges on average that had no relationship with lati-
tude (P = 0.75, r2adj = 0.007, N = 17; Fig. 1a, Appendix S4), while
temperate species had range sizes that increased with latitude as
expected (P = 0.0005, r2adj = 0.46, N = 21; Fig. 1b, Appendix S4).
Range size was positively correlated with diet breadth in both
tropical (Fig. 1c, linear regression: P = 0.028, r2adj = 0.23, N = 17)
and temperate species (Fig. 1d, linear regression: P = 0.0017,
r2adj = 0.38, N = 21).
Our hypothesis that diet breadth would decrease with
increasing latitude was only supported for temperate species.
Latitude was not a strong predictor of diet breadth in tropical
species (P = 0.32; Table 1, Fig. 1e), but was a strong predictor of
diet breadth in temperate species (P = 0.023, r2adj = 0.20, N = 21;
Figure 1 Comparison of the
relationships between latitude, range size
and diet breadth in tropical and
temperate brachyuran crab species.
Linear regressions were used to analyse
relationships between latitude (midpoint
of latitudinal range) and range size
(degrees of latitude) for (a) tropical
species and (b) temperate species, as well
as diet breadth and range size for (c)
tropical species and (d) temperate
species. Both linear regressions and
model selection were used to test the
strength of latitude and diet breadth
relationships for (e) tropical species and
(f) temperate species. Each point is a
brachyuran crab species. Diet breadth
was calculated from the published
literature and from our diet analyses on
crab species collected in Connecticut,
Florida and Panama (n = 21 temperate
species, and 17 tropical species). Positive
relationships were found between
latitude and range size for temperate
species only (panel b). Positive
relationships between diet breadth (as
represented by standardized Levins’ B
values) and range size were found for
both temperate and tropical species
(panels c and d). Latitude was related to
diet breadth for temperate species but
was not retained in the best fit model for
tropical species (panels e and f).
Latitudinal gradient in crab niche breadth
Global Ecology and Biogeography, © 2015 John Wiley & Sons Ltd 5
Table 1), being retained in the top two best-fit models in the
temperate analysis (Table 2, Fig. 1f). Therefore, taken together
with the results from bivariate range size relationships, we found
that higher-latitude temperate species with large ranges had the
broadest diets, while small-ranged, lower-latitude temperate
species had narrower diets.
In addition to latitude predicting the diet breadth of temper-
ate species, the second best fitting model of the temperate
dataset which also had substantial support (P = 0.038,
r2adj = 0.27, N = 21), retained body size and a body size × latitude
interaction term (Table 2, Fig. 2). Latitude had a weak relation-
ship with average body size for temperate species, with slightly
larger individuals being found at higher latitudes and smaller
individuals being found at lower latitudes (linear regression:
P = 0.04, r2adj = 0.16, N = 21; Fig. 2a). There was not a strong
relationship between body size alone and diet breadth (P = 0.24;
Fig. 2b), but the marginally significant interaction term sug-
gested that the relationship between latitude and diet breadth
increased in strength as body size increased (P = 0.08; Fig. 2c).
Although influential, this interaction may be of lesser impor-
tance in predicting diet breadth than latitude, since latitude was
significant in the simple regressions and retained in both best-fit
models (Table 2).
We found no relationship between clade and latitude in our
dataset (ANOVA for temperate species, r2adj = 0.11, N = 21,
F2/18 = 2.35, P = 0.1; ANOVA for tropical species, r2adj = 0.18,
N = 17, F2/14 = 2.73, P = 0.1), but we did find a relationship
between clade and diet breadth for tropical species. Interestingly,
clade was the strongest predictor of diet breadth for tropical
species, being the only factor retained in the best-fit model
(P = 0.0042, r2adj = 0.48, N = 17; Table 1). Those tropical species
belonging to the most evolutionarily recent clade,
Neobrachyura, had considerably narrower diets than the two
older brachyuran clades represented in our dataset (Table 1,
Fig. 3a), while the older tropical clades had diet breadths that
were more comparable to mid/high-latitude temperate species
(Fig. 1d, Fig. 3a). While the trend of the temperate data inter-
estingly mirrors these results for tropical species (Fig. 3a, b),
clade was not supported as a major predictor of diet in temper-
ate species, as the model containing clade was not significant
and had far less strength than the best-fit model (P = 0.77;
Table 2, Fig. 3b).
DISCUSSION
Temperate brachyuran crabs were found to be more specialized
at lower latitudes, but in contrast we found no latitude–niche
breadth relationship for tropical species. Body size and latitude
further had an interactive effect on diet breadth of temperate
species, with larger-bodied species demonstrating a broader diet
at higher latitudes than at lower latitudes. In the tropics, the
strongest single predictor of diet breadth was clade. Our results
suggest that niche breadth patterns may relate to latitudinal
gradients in competition strength (Schemske et al., 2009) and
species richness (Belmaker et al., 2012), both of which are
thought to increase towards lower latitudes. Specialization can
alleviate strong competition by reducing niche overlap, thereby
facilitating the coexistence of species as well as higher species
richness (Dyer et al., 2007) at low temperate latitudes. Species
interactions (Schemske et al., 2009) and associated evolutionary
Figure 2 Relationships among body size, latitude and diet
breadth for temperate crab species. Linear relationships between
(a) average body size and Levins’ B, (b) latitude (midpoint of
latitudinal range) and average body size, with each point
representing a brachyuran crab species and (c) a body
size × latitude interaction plot demonstrating the predicted
strength of the latitude–diet breadth relationship if the smallest
body size (0.75 cm) and largest body size (10.5 cm) are held
constant across latitude; the plot indicates that the strength of the
relationship increases as body size increases. A linear regression
was used to analyse the relationship between latitude and average
body size, while a regression as well as model selection were used
to test the strength of the average body size relationship with diet
breadth, and the relationship of the body size × latitude
interaction with diet breadth.
K. J. Papacostas and A. L. Freestone
Global Ecology and Biogeography, © 2015 John Wiley & Sons Ltd6
selection pressure (Roulin et al., 2009), however, are expected to
be strongest in the tropics. Therefore evolutionary history rather
than latitude may be more predictive of niche breadth in tropi-
cal species.
There is little marine research exploring latitude–niche
breadth relationships; one study found mixed results for host
specificity in parasites across latitude (Rohde, 1978) while two
other studies supported positive relationships between breadth
of thermal tolerance and both latitude and range size (Sunday
et al., 2011, 2012). In these studies, however, other possible
explanatory variables such as body size and phylogenetic rela-
tionships were not examined in conjunction with latitude. By
examining these other important variables, and also by analys-
ing our temperate and tropical data separately, our results high-
light several key mechanisms that may differentially drive large-
scale patterns of niche breadth in tropical and temperate species.
The fact that clade is the major factor associated with diet
breadth in our tropical dataset suggests that evolutionary
history may be a key predictor of niche breadth in tropical
species. In the tropics, species diversity has been found to be
Figure 3 Comparison of the
relationship between clade and diet
breadth for tropical and temperate
species. Majoidea is closest to the
ancestral species (the oldest clade
represented in the dataset) and
Neobrachyura is furthest (the newest
clade). Data are presented as box plots
for (a) tropical species and (b) temperate
species; the lower boundaries of the
boxes represent the second quartile, the
line within the box represents the
median and the upper boundary of the
box indicates the third quartile. The bar
below the box spans the first quartile,
and the bar above the box spans the
fourth quartile. The means are indicated
with individual data points. Samples
sizes for each clade are detailed in the
figure. Different letters above boxes
indicate significant differences between
groups, as determined using Tukey’s
honestly significant difference test.
Latitudinal gradient in crab niche breadth
Global Ecology and Biogeography, © 2015 John Wiley & Sons Ltd 7
greater than at higher latitudes, potentially due to climatic sta-
bility, high diversification rates and/or historical factors (Pianka,
1966; Hillebrand, 2004b; Jablonski et al., 2006; Mittelbach et al.,
2007). New clades in high-diversity systems may therefore
evolve as specialists as a mechanism to coexist and increase the
efficiency of their resource usage and minimize competition
with older, more generalist clades. These results do not neces-
sarily mean that evolutionary history is unimportant at higher
latitudes, since temperate clades showed a similar niche breadth
pattern to tropical clades (i.e. narrowed niche breadth in the
newest clade), although this pattern was not significant.
However, our results are consistent with existing hypotheses
suggesting that there may be differences in the importance of
biotic interactions between temperate and tropical latitudes
over evolutionary time-scales (Mittelbach et al., 2007; Roulin
et al., 2009).
Latitude was the strongest predictor of diet breadth for tem-
perate species, but body size and a latitude–body size interaction
were also found to be important. Body size alone did not
strongly influence diet breadth, but the positive relationship
between latitude and diet breadth was strongest in larger-bodied
species. Although it has been suggested that large-bodied species
are able to consume prey of a broader size range, and thus a
wider diversity of prey, than smaller species (e.g. Diaz, 1994),
some studies have demonstrated that larger-bodied species pref-
erentially target fewer, larger, more energy efficient organisms
(e.g. Costa et al., 2008). Our results may suggest that large-
bodied crabs utilize these alternative foraging strategies depend-
ing on latitude; large-bodied species at high latitudes may be
able to target a wider range of prey due to lower predator diver-
sity (Paine, 1966; Hillebrand, 2004a,b) and reduced competition
(Schemske et al., 2009). As both predator diversity and compe-
tition strength are hypothesized to increase at lower latitudes
(Paine, 1966; Pianka, 1966; Schemske et al., 2009), specialization
may allow more efficient foraging in large-bodied low-
temperate/subtropical species.
Although niche breadth was correlated with range size across
all latitudes, we observed differential support between temper-
ate and tropical species for Rapoport’s rule (predicting a positive
relationship between range size and latitude). The ranges of
tropical species were very large on average and had no relation-
ship with latitude, but the ranges of temperate species had a
significant positive correlation with latitude. Support for
Rapoport’s rule varies in the literature (Stevens, 1989; Rohde,
1999); strong support has been found in studies conducted in
the Northern Hemisphere, but previous global analyses have
found weak patterns (see Ruggiero & Werenkraut, 2007). Exist-
ing marine studies have found little support for the hypothesis
(e.g. Rohde & Heap, 1996; Macpherson, 2003), and it has been
suggested that Rapoport’s rule may be weaker in marine systems
than terrestrial systems (Ruggiero & Werenkraut, 2007), poten-
tially due to variation in propagule dispersal (Byers & Pringle,
2006), larger average scales of connectivity in marine systems
(Carr et al., 2003), less temperature variability than terrestrial
systems (Sunday et al., 2012) or currents having strong influ-
ences on the range sizes of marine species (Gaylord & Gaines,
2000). Our results, however, suggest that large-ranged tropical
species could also obscure the pattern.
Variation in tropical and temperate range patterns may be
driven by differences in the abiotic and biotic factors to which
they are exposed. For instance, the ranges of tropical species are
likely to be restricted by cold boundaries (Sunday et al., 2011,
2012). As climate is less spatially and temporally variable at low
latitudes (Pianka, 1966), and the tropics span a wide geographi-
cal area (47° of latitude), the entire tropics could be within the
thermal tolerance range of a tropical species. Indeed, many of
our tropical species had ranges that spanned the majority of the
tropics (14 out of 17 species had ranges > 45° of latitude) or
extended into subtropical areas where temperature extremes are
still mild. Therefore, although we found that dietary niche
breadth and range size were correlated, breadth of environmen-
tal tolerance may also be important in driving these large range
sizes of tropical species. Temperate species ranges are also set by
cold boundaries, although they experience broader temperature
extremes and are likely to have broader thermal tolerances
(Sunday et al., 2011, 2012). However, their low-latitude range
limits may also be set by biotic boundaries such as competitive
exclusion from tropical species; species interactions are also
determinants of range size (Briers, 2003; Holt & Barfield, 2009)
and low-temperate species may have the smallest ranges, as well
as the smallest niches, because they experience these abiotic and
biotic limitations on both latitudinal range limits.
Overall, we found a stark division in predictors of diet
breadth in temperate and tropical species, with latitude being a
strong predictor for temperate species along with a latitude–
body size interaction, and evolutionary history as the best pre-
dictor for tropical species. We found positive range size
relationships with niche breadth for all species, except that
tropical species in general had larger ranges and broader niches
than expected. The large ranges of tropical species are likely
influenced by large geographical areas within their thermal tol-
erances, and broad tropical niche breadth was shown to be
driven by older, more generalist clades. These results suggest that
competition may be important to varying degrees across lati-
tude; weak at high latitudes, strong enough at low temperate
latitudes to influence niche breadth patterns among species, and
strongest in the tropics where it may influence niche breadth
patterns among clades. Therefore, ecological mechanisms (e.g.
competition strength) may be primarily driving patterns of
niche breadth in the temperate zone, while evolutionary mecha-
nisms (e.g. selection due to competition) may be more predic-
tive of niche breadth patterns in tropical, high-diversity areas.
These potentially differing mechanisms driving niche breadth
may also contribute to large-scale maintenance of diversity in
both temperate and tropical systems.
ACKNOWLEDGEMENTS
We thank C. Clark, J. Hoffman and M. Vaca for assistance with
data entry, the Smithsonian Tropical Research Institute’s Bocas
del Toro Research Station and the Smithsonian Marine Station
for use of their facilities for subtropical and tropical crab collec-
K. J. Papacostas and A. L. Freestone
Global Ecology and Biogeography, © 2015 John Wiley & Sons Ltd8
tions, Z. Mckie-Krisberg and A. Durkin for coding assistance, E.
Cordes, R. Sanders, B. Sewall and P. Petraitis for useful feedback
during the development of the project, and J. Belmaker as well as
three anonymous reviewers. This study was supported by the
Smithsonian/Link Foundation fellowship programme
(Smithsonian Marine Station Contribution at Fort Pierce,
Florida, no. 1002), the Quebec-Labrador Foundation’s Sounds
Conservancy Grant, and the National Science Foundation
through both the Graduate STEM Fellows in K-12 Education
programme under grant no. 0841377 and the Division of Ocean
Sciences under grant no. 1225583. Any opinion, findings, con-
clusions or recommendations expressed in this material are
those of the authors, do not necessarily reflect the position of the
National Science Foundation and no official endorsement
should be inferred.
REFERENCES
Ashmole, N.P. (1968) Body size, prey size, and ecological segre-
gation in five sympatric tropical terns (Aves: Laridae). System-
atic Zoology, 17, 292–304.
Barnagaud, J.-Y., Daniel Kissling, W., Sandel, B., Eiserhardt,
W.L., Sekercioglu, C.H., Enquist, B.J., Tsirogiannis, C. &
Svenning, J.-C. (2014) Ecological traits influence the
phylogenetic structure of bird species co-occurrences world-
wide. Ecology Letters, 17, 811–820.
Belmaker, J., Sekercioglu, C.H. & Jetz, W. (2012) Global patterns
of specialization and coexistence in bird assemblages. Journal
of Biogeography, 39, 193–203.
Bergmann, C. (1848) Uber die Verhaltnisse der
Warmeokonomieder Tiere zu ihrer Grosse. E. A. Huth,
Göttingen.
Bolnick, D.I., Ingram, T., Stutz, W.E., Snowberg, L.K., Lau, O.L.
& Paull, J.S. (2010) Ecological release from interspecific com-
petition leads to decoupled changes in population and indi-
vidual niche width. Proceedings of the Royal Society B:
Biological Sciences, 277, 1789–1797.
Briers, R.A. (2003) Range limits and parasite prevalence in a
freshwater snail. Proceedings of the Royal Society B: Biological
Sciences, 270, S178–S180.
Brosing, A., Richter, S. & Scholtz, G. (2007) Phylogenetic analy-
sis of the Brachyura (Crustacea, Decapoda) based on charac-
ters of the foregut with establishment of a new taxon. Journal
of Zoological Systematics and Evolutionary Research, 45, 20–32.
Burnham, K.P. & Anderson, D.R. (2002) Model selection and
multimodel inference: a practical information-theoretic
approach. Springer-Verlag, New York.
Byers, J.E. & Pringle, J.M. (2006) Going against the flow: reten-
tion, range limits and invasions in advective environments.
Marine Ecology Progress Series, 313, 27–41.
Carr, M.H., Neigel, J.E., Estes, J.A., Andelman, S., Warner, R.R. &
Largier, J.L. (2003) Comparing marine and terrestrial ecosys-
tems: implications for the design of coastal marine reserves.
Ecological Applications, 13, S90–S107.
Costa, G.C., Vitt, L.J., Pianka, E.R., Mesquita, D.O. & Colli, G.R.
(2008) Optimal foraging constrains macroecological patterns:
body size and dietary niche breadth in lizards. Global Ecology
and Biogeography, 17, 670–677.
Diaz, M. (1994) Variability in seed size selection by granivorous
passerines – effects of bird size, bird size variability, and eco-
logical plasticity. Oecologia, 99, 1–6.
Dyer, L.A., Singer, M.S., Lill, J.T., Stireman, J.O., Gentry, G.L.,
Marquis, R.J., Ricklefs, R.E., Greeney, H.F., Wagner, D.L.,
Morais, H.C., Diniz, I.R., Kursar, T.A. & Coley, P.D. (2007)
Host specificity of Lepidoptera in tropical and temperate
forests. Nature, 448, 696–699.
Fenberg, P.B. & Rivadeneira, M.M. (2011) Range limits and
geographic patterns of abundance of the rocky intertidal owl
limpet, Lottia gigantea. Journal of Biogeography, 38, 2286–
2298.
Fernandez, M.H. & Vrba, E.S. (2005) Rapoport effect and
biomic specialization in African mammals: revisiting the cli-
matic variability hypothesis. Journal of Biogeography, 32, 903–
918.
Fiedler, K. (1998) Diet breadth and host plant diversity of
tropical- vs. temperate-zone herbivores: South-East Asian and
west Palaearctic butterflies as a case study. Ecological Entomol-
ogy, 23, 285–297.
Freestone, A.L., Osman, R.W., Ruiz, G.M. & Torchin, M.E.
(2011) Stronger predation in the tropics shapes species
richness patterns in marine communities. Ecology, 92, 983–
993.
Gaylord, B. & Gaines, S.D. (2000) Temperature or transport?
Range limits in marine species mediated solely by flow. The
American Naturalist, 155, 769–789.
Gosling, E. (2003) Ecology of bivalves. Bivalve molluscs: biology,
ecology and culture, pp. 44–78. Wiley-Blackwell, Malden,
MA.
Griffen, B.D. & Mosblack, H. (2011) Predicting diet and con-
sumption rate differences between and within species using
gut ecomorphology. Journal of Animal Ecology, 80, 854–863.
Hawkins, B.A. & DeVries, P.J. (2009) Tropical niche conserva-
tism and the species richness gradient of North American
butterflies. Journal of Biogeography, 36, 1698–1711.
Hillebrand, H. (2004a) Strength, slope and variability of marine
latitudinal gradients. Marine Ecology Progress Series, 273, 251–
267.
Hillebrand, H. (2004b) On the generality of the latitudinal
diversity gradient. The American Naturalist, 163, 192–211.
Holt, R.D. & Barfield, M. (2009) Trophic interactions and range
limits: the diverse roles of predation. Proceedings of the Royal
Society B: Biological Sciences, 276, 1435–1442.
Horton, T., Lowry, J. & De Broyer, C. (2013) World Amphipoda
database. Available at: http://www.marinespecies.org/
amphipoda/ (accessed 2 April 2015).
Hurlbert, S.H. (1978) The measurement of niche overlap and
some relatives. Ecology, 59, 67–77.
Jablonski, D., Roy, K. & Valentine, J.W. (2006) Out of the tropics:
evolutionary dynamics of the latitudinal diversity gradient.
Science, 314, 102–106.
Kerkhoff, A.J., Moriarty, P.E. & Weiser, M.D. (2014) The latitu-
dinal species richness gradient in New World woody
Latitudinal gradient in crab niche breadth
Global Ecology and Biogeography, © 2015 John Wiley & Sons Ltd 9
angiosperms is consistent with the tropical conservatism
hypothesis. Proceedings of the National Academy of Sciences
USA, 111, 8125–8130.
Krasnov, B.R., Shenbrot, G.I., Khokhlova, I.S., Mouillot, D. &
Poulin, R. (2008) Latitudinal gradients in niche breadth:
empirical evidence from haematophagous ectoparasites.
Journal of Biogeography, 35, 592–601.
MacArthur, R.H. (1968) The theory of the niche. Population
biology and evolution (ed. by R.C. Lewontin), pp. 159–176.
Syracuse University Press, New York.
MacArthur, R.H. (1972) Geographical ecology. Princeton Univer-
sity Press, Princeton, NJ.
McGaw, I.J. & Reiber, C.L. (2000) Integrated physiological
responses to feeding in the blue crab Callinectes sapidus.
Journal of Experimental Biology, 203, 359–368.
Macpherson, E. (2003) Species range size distributions for some
marine taxa in the Atlantic Ocean. Effect of latitude and
depth. Biological Journal of the Linnean Society, 80, 437–455.
Mittelbach, G.G., Schemske, D.W., Cornell, H.V. et al. (2007)
Evolution and the latitudinal diversity gradient: speciation,
extinction and biogeography. Ecology Letters, 10, 315–331.
Neter, J., Kutner, M.H., Nachtsheim, C.J. & Wasserman, W.
(1996) Applied linear statistical models, 4th edn. McGraw-Hill,
Boston, MA.
Novotny, V. & Basset, Y. (1999) Body size and host plant spe-
cialization: a relationship from a community of herbivorous
insects on Ficus from Papua New Guinea. Journal of Tropical
Ecology, 15, 315–328.
Paine, R.T. (1966) Food web complexity and species diversity.
The American Naturalist, 100, 65–75.
Petraitis, P.S. (1979) Likelihood measures of niche breadth and
overlap. Ecology, 60, 703–710.
Pianka, E.R. (1966) Latitudinal gradients in species diversity – a
review of concepts. The American Naturalist, 100, 33–46.
R Core Team (2014) R: a language and environment for statistical
computing. R Foundation for Statistical Computing, Vienna,
Austria. Available at: http://www.R-project.org/ (accessed 10
March 2015).
Rohde, K. (1978) Latitudinal differences in host-specificity of
marine Monogenea and Digenea. Marine Biology, 47, 125–
134.
Rohde, K. (1999) Latitudinal gradients in species diversity and
Rapoport’s rule revisited: a review of recent work and what
can parasites teach us about the causes of the gradients?
Ecography, 22, 593–613.
Rohde, K. & Heap, M. (1996) Latitudinal ranges of teleost fish in
the Atlantic and Indo-Pacific Oceans. The American Natural-
ist, 147, 659–665.
Roulin, A., Wink, M. & Salamin, N. (2009) Selection on a
eumelanic ornament is stronger in the tropics than in tem-
perate zones in worldwide-distributed barn owl. Journal of
Evolutionary Biology, 22, 345–354.
Ruggiero, A. & Werenkraut, V. (2007) One-dimensional analyses
of Rapoport’s rule reviewed through meta-analysis. Global
Ecology and Biogeography, 16, 401–414.
Safi, K. & Kerth, G. (2004) A comparative analysis of specializa-
tion and extinction risk in temperate-zone bats. Conservation
Biology, 18, 1293–1303.
Sagarin, R.D. & Gaines, S.D. (2002) The ‘abundant centre’ dis-
tribution: to what extent is it a biogeographical rule? Ecology
Letters, 5, 137–147.
Schemske, D.W., Mittelbach, G.G., Cornell, H.V., Sobel, J.M. &
Roy, K. (2009) Is there a latitudinal gradient in the importance
of biotic interactions? Annual Review of Ecology, Evolution,
and Systematics, 40, 245–269.
Slatyer, R.A., Hirst, M. & Sexton, J.P. (2013) Niche breadth pre-
dicts geographical range size: a general ecological pattern.
Ecology Letters, 16, 1104–1114.
Slove, J. & Janz, N. (2010) Phylogenetic analysis of the
latitude–niche breadth hypothesis in the butterfly subfamily
Nymphalinae. Ecological Entomology, 35, 768–774.
Stevens, G.C. (1989) The latitudinal gradient in geographical
range – how so many species exist in the tropics. The American
Naturalist, 133, 240–256.
Sunday, J.M., Bates, A.E. & Dulvy, N.K. (2011) Global analysis of
thermal tolerance and latitude in ectotherms. Proceedings of
the Royal Society B: Biological Sciences, 278, 1823–1830.
Sunday, J.M., Bates, A.E. & Dulvy, N.K. (2012) Thermal toler-
ance and the global redistribution of animals. Nature Climate
Change, 2, 686–690.
van Voorhies, W.A. (1996) Bergmann size clines: a simple expla-
nation for their occurrence in ectotherms. Evolution, 50,
1259–1264.
Vazquez, D.P. & Stevens, R.D. (2004) The latitudinal gradient in
niche breadth: concepts and evidence. The American Natural-
ist, 164, E1–E19.
WoRMS Editorial Board (2014) World Register of Marine
Species. Available at: http://www.marinespecies.org (accessed
20 September 2014).
Wulff, E.V. & Brissenden, E. (1943) An introduction to historical
plant geography (transl. E. Brissenden, with a foreword by E. D.
Merrill). Chronica Botanica Co., Waltham, MA.
SUPPORTING INFORMATION
Additional Supporting Information may be found in the online
version of this article at the publisher’s web-site:
Appendix S1 References to data sources.
Appendix S2 Full list of species and data included in analyses.
Appendix S3 Diets of crabs collected in Connecticut, Florida
and Panama.
Appendix S4 Relationship between latitude and range size of all
species in the dataset.
Appendix S5 Frequency of prey items consumed across all
species analysed.
K. J. Papacostas and A. L. Freestone
Global Ecology and Biogeography, © 2015 John Wiley & Sons Ltd10
BIOSKETCHES
Katherine Papacostas is interested in both
theoretical and applied ecology as well as conservation.
Her research examines spatial and temporal variation in
marine invasion dynamics, how species interactions
change with latitude and global patterns of marine
resource usage.
Amy Freestone’s research combines community
ecology and macroecology. She is primarily interested
in the impact of species interactions on community
assembly and ecosystem function, and how these
processes structure patterns of species diversity across
spatial and temporal scales.
Editor: Jonathan Belmaker
Latitudinal gradient in crab niche breadth
Global Ecology and Biogeography, © 2015 John Wiley & Sons Ltd 11