![Proportion of dispersed seeds [volume of intact to total seeds on the interval (0 – 1), mu ] modeled as a function of fi sh size (SL — standard length). Pantanal: (a) Piaractus mesopotamicus . Amazon: (b) Brycon amazonicus , and (c) Myloplus torquatus .](https://www.researchgate.net/profile/Joisiane-Araujo/publication/279636223/figure/fig3/AS:284594382360578@1444863996863/Proportion-of-dispersed-seeds-volume-of-intact-to-total-seeds-on-the-interval-0-1_Q320.jpg)
Content uploaded by Joisiane K. Mendes Araujo
Author content
All content in this area was uploaded by Joisiane K. Mendes Araujo on Mar 16, 2018
Content may be subject to copyright.
Overfishing disrupts an ancient mutualism between frugivorous fishes
and plants in Neotropical wetlands
Sandra Bibiana Correa
a
, Joisiane K. Araujo
b
, Jerry M.F. Penha
b
,CatiaNunesdaCunha
b
,
Pablo R. Stevenson
c
,JillT.Anderson
a,
⁎
a
Department of Genetics, Odum School of Ecology, University of Georgia, 120 Green St., Athens, GA 30602, USA
b
Instituto de Biociências, Universidade Federal de Mato Grosso, Ave. Fernando Correia 2367, Cuiabá, MT, Brazil
c
Departamento de Ciencias Biológicas, Universidad de los Andes, Carrera 1 No. 18A-12,Bogotá, Colombia
abstractarticle info
Article history:
Received 7 May 2015
Received in revised form 13 June 2015
Accepted 14 June 2015
Available online xxxx
Keywords:
Seed dispersal
Defaunation
Flooded forest
Flooded savannah
Amazon
Pantanal
Defaunation is disrupting plant–animal interactions worldwide. The overhunting of frugivores disrupts seed
dispersal and diminishesplant regeneration, yet investigations of frugivoreoverexploitation neglect anancient
guild: fruit-eating fish. For nearly five decades, Neotropical frugivorous fishes have been intensively harvested.
These fishing activities have reduced population sizes of some species by up to 90% and have likely altered
populations to younger, smaller individuals. Here we evaluate potential ecological consequences of overfishing
frugivores for seed dispersal and recruitment dynamics. We analyzed dietary data from seven fruit-eating fish
species in Amazonian and Pantanal wetlands to test the hypothesis that seed dispersal effectiveness increases
with fish size within and across species. Relative to small individuals, larger fish dispersed large numbers of
seeds of a higher diversity of plants and a greater range of seed sizes. For some seed species, dispersal by larger
fish augmented germination success, relative to seeds dispersed by smaller fishes. Large Piaractus mesopotamicus
in the Pantanal disperse seeds of 27% more speciesthan fishes under the minimum size limit for this fishery.Our
results indicate that the ongoing overexploitation of multiple frugivorous fish species coulddepress the quantity
and diversity of seeds dispersed, as well as the quality of seed dispersal in wetland habitats that extend over 15%
of the area of South America.
© 2015 Elsevier Ltd. All rights reserved.
1. Introduction
The disruption of pollination and seed dispersal mutualisms directly
affects plant reproductive success and threatens biodiversity (Aslan
et al., 2013; Dirzo et al., 2014; Valiente-Banuet et al., 2015). The vast
majority of plants rely on animals to disperse their genes and progeny
(Jordano, 2000; Ollerton et al., 2011). Forexample, vertebrate frugivores
disperse the seeds of 70–95% of woody plant species in tropical forests
(e.g., Jordano, 2000). Seed dispersal sets the initial template of plant dis-
tribution in the landscape, shapes the pool of interacting plant species,
and potentially facilitates species coexistence through competition/
colonization tradeoffs (Howe and Smallwood, 1982; Seidler and
Plotkin, 2006).
Frugivorous vertebrates are overhunted in tropical forests across the
globe, which reduces plant recruitment, alters plant species composi-
tion, diminishes biodiversity, and causes numerous indirect changes to
communities (Caughlin et al., 2015; Markl et al., 2012; Poulsen et al.,
2013). As hunters prefer big prey, overharvesting is particularly
problematic for large-seeded plant species, which require seed dispersal
by sizable frugivores (Caughlin et al., 2015; Effiom et al., 2014). Local
extinction of large frugivores can even induce rapid evolutionary changes
in seed size (Galetti et al., 2013). A burgeoning body of work evaluates the
ecological repercussions of hunting frugivores (Aslan et al., 2013), but
neglects the largest clade of vertebrates: fish.
Worldwide, over 275 species of fish consume fruits and disperse
seeds; of these, at least 150 inhabit South American wetlands (Horn
et al., 2011), where they disperse seeds of at least 566 plant species
from 82 families (Correa et al., 2015). Neotropical wetlands extend
across eight countries and the three largest South American river basins
(Amazon, Orinoco, and Paraná–Paraguay), occupying at least 15% of
the continent (Junk and Piedade, 2010). Fruiting is synchronized with
the annual flood, lasting up to seven months, and many fruits and
seeds exhibit adaptations for dispersal by water (hydrochory) or fish
(ichthyochory; Ferreira et al., 2010; Kubitzki and Ziburski, 1994).
During lengthy flooded seasons, fishes spend ~87% of their time in
floodplain habitats (Anderson et al., 2011) where seeds can germinate
after floodwaters recede (Ferreira et al., 2010).
Globally, selective harvesting of fish concentrates on large individ-
uals, inducing changes in population structure by favoring the survival
of smaller fishes that reproduce earlier (Allan et al., 2005; Palkovacs,
Biological Conservation 191 (2015) 159–167
⁎Corresponding author: Tel.: +1 706 542 0853.
E-mail address: jta24@uga.edu (J.T. Anderson).
http://dx.doi.org/10.1016/j.biocon.2015.06.019
0006-3207/© 2015 Elsevier Ltd. All rights reserved.
Contents lists available at ScienceDirect
Biological Conservation
journal homepage: www.elsevier.com/locate/bioc
2011). A fishery-induced reduction in body size of frugivorous species
would have profound effects on plant regeneration if larger fishes are
better seed dispersers than smaller fishes (Anderson et al., 2011;
Correa et al., 2015; Galetti et al., 2008; Kubitzki and Ziburski, 1994).
Large Neotropical frugivorous fishes are prized in commercial, subsis-
tence, and recreational fisheries, and are threatened by ongoing overex-
ploitation (Isaac and Ruffino, 1996; Peixer et al., 2007). Frugivorous
fishes are heavily consumed in the Amazon, Orinocoand Pantanal, con-
tributingto food security andeconomic growth (Barthem and Goulding,
2007; Mateus et al., 2004; Rodriguez et al., 2007). Intensive commercial
fisheries developed in the 1970s across the Amazon and Pantanal with
the introduction of nylon gill nets and large capacity freezer rooms
(Mateus et al., 2004; Santos and Santos, 2005). By the early 1990s,
stocks of Colossoma macropomum in the Central Amazon were already
overexploited (Isaac and Ruffino, 1996) and commercial and recrea-
tional fishing activities have reduced population sizes of Piaractus
mesopotamicus in the Pantanal by 90% (Albuquerque et al., 2012;
Peixer et al., 2007). In addition to population declines, overfishing likely
has shifted body size of these species to smaller individuals that repro-
duce earlier (Palkovacs, 2011), however the lack of basin-wide continu-
ous historic fishery statistics (Mateus et al., 2004; Ruffino, 2008)hinders
our ability to assess changes in population structure of overexploited
populations. Striking population declines in other large species such as
catfishes have shifted fishing pressure to under-exploited stocks of
large frugivorous fishes in remote locations (Agudelo et al., 2012;
Albuquerque et al., 2012; Mateus et al., 2004; Reinert and Winter,
2002). Small- and medium-sized frugivorous fish species are also inten-
sively exploited in the Amazon and Pantanal (Mateus et al., 2004;
Santos and Santos, 2005). These species are heavily consumed by river-
ine people; therefore, a large portion of the capture is not accounted for
in commercial fishery statistics, leading to an underestimation of the
real fishing pressure (Castello et al., 2013).
Here, we examine the ecological consequences of overfishing in two
expansive and diverse Neotropical wetlands: the Pantanal and the
Amazon. In both regions, we predict that the largest fishes disperse
more seeds of a greater diversity of plant species, and that large-
bodied fish species are among the primary vectors of dispersal for
plant species with big seeds (e.g., Stevenson, 2011). Additionally, we
examine size-dependent shifts in seed predation, testing the prediction
that larger fishes depredate fewer seeds than smaller individuals. In a
preliminary analysis, Correa et al. (2015) found that smaller fishes had
a greater probability of destroying seeds than larger fishes across
three species of Brycon, presumably because big fishes are more adept
at swallowing fleshy fruits entire. Finally, we tested whether seed
germination success increases with fish size in a greenhouse experiment.
Overexploitation could fundamentally alter fish–fruit interactions
from high quality seed dispersal by large fishes to greater rates of seed
predation by smaller fishes.
2. Materials AND Methods
2.1. Study areas
2.1.1. Brazilian Pantanal
Spanning over 160 000 km
2
, the Pantanal is one of the largest
tropical wetlands in the world (Junk and Nunes da Cunha, 2005). We
conducted our study in the SESC Pantanal Private Natural Heritage
Reserve (1063 km
2
;16°30′51″S, 56°22′38″W; Fig. 1b) in the northern
Pantanal. The Pantanal Reserve constitutes a mosaic of seasonally
flooded savannas with small patches of semi-deciduous mono-
dominant forests of cambará (Vochysia divergens, Vochysiaceae) and
gallery forests. Tree diversity in forests is low (up to 10 species with
dbh N10 cm per 0.1 ha; Arieira et al., 2011)andtheflooding regime is
moderate (2 m depth for 5 months annually, Junk and Nunes da
Cunha, 2005).
2.1.2. Colombian Amazon
The lowlands of the Amazon River and its tributaries contain vast
expanses of seasonally flooded forests. Amazonian floodplains cover
approximately 250000 km
2
, constituting the largest wetland on earth
(Junk, 1993). We sampled the lower Caquetá River (1°16′32″S, 69°43′
50″W; Fig. 1a), where the floodplain supports a continuous, relatively
undisturbed evergreen forest with high tree diversity (up to 54species
with dbh N10 cm per 0.1 ha; Duivenvoorden, 1996) that floods up to
9mdepthfor6monthsannually(Rodriguez, 1991).
2.2. Fish sampling
To sample frugivorous fishes of varying sizes, we baited hooks of
different sizes with ripe local fruits, as well as with dough made of
cooked manioc flour and artificial fruit flavors, and fished with pole
and line. In the Amazon, hooks were also suspended from vegetation
in the flooded forest and monitored hourly (Correa and Winemiller,
2014), which is not feasible in the Pantanal due to abundant Caiman
populations. Our sampling selectively captures fruit-eating fish species
(Correa and Winemiller, 2014). Immediately after removal from the
hook, we euthanized fishes with Tricaine methanesulfonate (MS-222).
We recorded species identity, standard length (SL—length from the tip
of the snout to the posterior end of the hypural bones, excluding the
caudal fin), weight, and mouth gape of the fish, and collected stomach
and intestinal samples by dissection. This research complies with
animal use guidelines (AUP# 2194-100789-011314).
In the Pantanal, we sampled fishes in four habitats (river channels,
savannas, mono-dominant and gallery forests) during the flooding
season of 2014 (February–May) for a total of n= 374 individuals of
four species (Serrasalmidae: P. mesopotamicus,n=217,14–59 cm SL;
Mylossoma duriventre,n= 50, 11–21 cm SL; Myloplus tiete,n=40,
11–16 cm SL. Characidae: Brycon hilarii,n=67, 11–36 cm SL; Fig. 2).
In the Amazon, we sampled fishes in three floodplain forest habitats
(river channels, islands within the river channel, and floodplain
channels) during the peak flooding (June–August 2014) for a total of
n= 285 individuals of four species(Serrasalmidae: Myloplus torquatus,
n=69,12–28 cm SL; M. duriventre,n=34,15–27 cm SL. Characidae:
Brycon amazonicus,n= 52, 23–35 cm SL; Brycon melanopterus,n=
130, 12–27 cm SL; Fig. 2). These sample sizes exclude fishes with
empty digestive tracts (n= 20 in the Pantanal and n= 6 in the Ama-
zon). Fish nomenclature follows Reis et al. (2003).
2.3. Diet estimation
We separated stomach (excluding bait) and intestinal contents into
food categories (intact seeds, masticated seeds, and fruit pulp), calculat-
ed the volume of each category by water displacement, and identified
and quantified intact seeds. Intact seeds pass through the digestive
system undamaged, and are likely viable, whereas masticated seeds
are destroyed during consumption and digestion. We measured length
and width of at least 3 intact seeds per taxa when possible.
To assess the number of species fruiting during the flooded season,
we established 100 m long × 10 m wide transects in all sampled habitat
types: savanna (n= 3), mono-dominant (n= 3) and gallery forest
(n= 4) in the Pantanal; and floodplain forest parallel (n=5)andper-
pendicular to the rivers' edge (n= 3) and along floodplain channels
(n= 2) in the Amazon. Biweekly, we recorded the number of plants
per species with ripe and immature fruits. Botanical vouchers were de-
posited at the Herbarium of the Universidade Federal de Mato Grosso,
Cuiabá, Brazil and the Colombian Amazon Herbarium (COAH),
Colombia. Plant nomenclature follows Angiosperm Phylogeny Group
and Tropicos (http://www.tropicos.org).
We tested the hypothesis that seed dispersal effectiveness increases
with fish size within and across fish species using four metrics of
dispersal quality and quantity: (1) proportion of intact to total seed
matter (the probability of dispersing rather than destroying seeds),
160 S.B. Correa et al. / Biological Conservation 191 (2015) 159–167
(2) number of seeds, (3) seed species richness, and (4) seed size. We
modeled these response variables as a function of individual fish body
size (SL), fish species, and the size by species interaction. When the
interaction was non-significant, we removed it. We analyzed these
four response variables separately because of differentstatistical distri-
butions, and sites separately because of differences in fish and plant
species compositions between the Pantanal and Amazon. We adjusted
probability values for multiple tests of the same dataset (Pantanal vs.
Amazon) using the Benjamin and Hochberg (1995) correction. Models
were implemented in R (version 3.1.2).
To test if the probability of dispersing seeds increases with fish size,
we calculated the proportion of seeds that were intact in the digestive
tract of each fish relative to total seed matter (intact seed volume:
intact + masticated seed volume), which is directly related to the
probability of seed predation (seed predation = 1 −seed dispersal).
Fishes with entirely masticated seeds had values of 0, whereas fishes
with entirely intact seeds had values of 1. We limited this analysis to
fishes with seed matter (intact ormasticated) in their digestive contents
(n=325fishes of 4 species inthe Pantanal, and 261 fishes of 4 species
in the Amazon). We implemented zero-one inflated beta models in
GAMLSS (version 4.3-1 Rigby and Stasinopoulos, 2005) with the BEINF
family to test the proportion of intact seeds as a function of fish size
(SL), fish species, and the interaction. Zero-one inflated beta regression
analyzes proportions as a mixture of Bernoulli and beta distributions,
and is the only approach currently capable of accommodating propor-
tional data that include values of 0 and 1, i.e., on the interval [0,1].
These models simultaneously estimate 3 parameters: (1) the probabili-
ty that this proportion has a valueof 0 (nu); (2) the expected value for
the beta component (values between 0 and 1, mu); and (3) the proba-
bility that this proportion has a value of 1 (tau). If larger fishes are better
seed dispersers, we expect negative relationships between fish size and
nu (the probability that digestive contents contained 0 intact seeds),
and positive relationships between fish size and mu (proportions
between 0 and 1) and tau (probability that digestive contents contained
only intact seeds).
We then restricted the datasets to individuals with intact seeds in
their digestive contents to assess the quantity and quality of dispersal
among fishes serving as seed dispersers (n=229fishes of 4 species in
the Pantanal, and 157 fishes of 4 species in the Amazon). The number
of intact seeds in fish digestive tracts was modeled with a quasipoisson
regression, which fitted the data better than a Poisson model given
overdispersion in the dataset. We analyzed seed species richness with
Fig. 1. Map of the study sites. (a) RíoCaquetá, Amazonas,Colombia, Northwestern Amazonia. Samplingwas conducted betweenthe confluence with the Río Mirití (A) and the floodplain
forest below the Córdoba rapids (B). (b) Reserva Particular do Patrimônio Natural (RPPN) SESC Pantanal, Mato Grosso, Brazil, Northern Pantanal. Sampling was conducted in Riozinho
(C) and the Rio Cuiabá (D). Dark green areas in (b) are monodominant forests of Vochysia divergens. Satellite images were retrieved from ESRI. (For interpretation of the references to
color in this figure legend, the reader is referred to the web version of this article.)
161S.B. Correa et al. / Biological Conservation 191 (2015) 159–167
a Poisson regression of the number of seed species present in individual
gut contents. Finally, since gape size constrains the size of the seeds that
a frugivore can consume (Wheelwright, 1985), we hypothesized that
large fishes are unique vectors of dispersal for the biggest seeds. We
calculated the size (length × width) of the largest intact seed in the
digestive tract of each fish to estimate the upper bound of seed sizes
that a fish can consume. Mouth gape was tightly correlated with body
length (R
2
=0.89,F
1,526
=4092,pb0.0001) across our 7 fish species;
therefore, we used fish size as a proxy for gape size to test the hypothesis
that the size of dispersed seeds increases with fish size. Since large fishes
can swallow both small and large seeds, heterogeneity in seed size
increased with fish size in our datasets; therefore, we conducted a
generalized least squares regression in which variance in seed size is
proportional to fish size (Zuur et al., 2009) in the R package nlme
(version 3.1-118, Pinheiro et al., 2014).
2.4. Germination experiment
To test whether fishes enhance germination, we conducted a
greenhouse experiment comparing seeds removed from fish digestive
tracts vs. local control seeds without fruit pulp. Removing the pulp
from control seeds tested if (1) fish consumption increases germination
success via scarification, in which case gut-processed seeds should have
a greater probability of germinating than control seeds or (2) fish
consumption influences germination simply by eliminating fruit pulp,
in which case gut-processed seeds should have equivalent germination
success as control seeds (Traveset et al., 2008). We randomly planted
individual seeds of the most abundant species in fish diets at each
site (Pantanal: 11 species, total n= 1660; Amazon: 5 species, total
n= 746; see Appendix A for additional details) into pots with 40, 60
or 80 mL of soil, depending on seed size. Pots were filled with commer-
cial potting soil for forest plants (Pantanal) or local forest soil that was
sieved to remove seeds and large debris (Amazon). Trays were housed
in a screened greenhouse (50% shade) at each field site. We monitored
seeds for germination daily (Pantanal: 8 months, Amazon: 6 months)
and watered as needed.
We conducted a logistic regression to test the effects of seed species,
treatment (de-pulped control seeds vs. gut treatment by fish) and
their interaction on germination success in separate analyses for
the Pantanal and Amazon datasets (Proc Logistic, SAS ver. 9.3). Owing
to quasi-separation of data points, we used a penalized likelihood
method (Firth, 1993). A second logistic regression examined
whether germination success of seeds processed by fish increased
with fish size.
Fig. 2. Variation in body size across species in local assemblages of frugivorous fish species in the Pantanal (Mt—Myloplus tiete,Md—Mylossoma duriventre,Bh—Brycon hilarii, and
Pm—Piaractus mesopotamicus) and Amazon (Md—Mylossoma duriventre,Mtq—Myloplus torquatus,Bm—Brycon melanopterus, and Ba—Brycon amazonicus). Photographs represent the
largest adult individuals per species captured during the study, except for Piaractus mesopotamicus (largest individual: 59 cm SL) and Mylossoma duriventre in the Amazon (largest
individual: 27 cm SL). SL—standard length. Photos by S.B. Correa and J.K. Araujo.
162 S.B. Correa et al. / Biological Conservation 191 (2015) 159–167
3. Results
We found 42938 intact seeds of 53 species in the digestive tracts
of 229 individual fishes in the Pantanal and 150624 intact seeds of
70 species in the digestive tracts of 163 fishes in the Amazon (see
Appendix B for additional details). Fishes dispersed 52% and 14% of
the fleshy-fruited plant species recorded during vegetation transects
in the Pantanal and Amazon, respectively (Appendix C). However,
several seed species found intact in fish guts were not observed in
the vegetation transects (Pantanal: 11 species, Amazon: 54 species;
Appendix B–C), suggesting that these highly mobile fishes may have
consumed fruits in distant wetlands (Anderson et al., 2011). A single
fish carried up to 3554 (mean seed length and width =
5.5 mm × Proc Logistic, SAS 4.8 mm; fish size: 40.5 cm SL) and 8386
(mean seed length and width = 4.4 mm × 2.0 mm; fish size: 23 cm
SL) seeds in its digestive tract, in the Pantanal and Amazon, respectively.
Fish-dispersed seeds ranged in size from 0.8 mm × 0.6 mm (Banara
arguta, Flacourtiaceae) to 40.7 mm × 19.6 mm (Couepia uiti,
Chrysobalanaceae) in the Pantanal, and 0.6 mm × 0.5 mm (Ficus sp.
Moraceae) to 27.1 mm × 11.9 mm (Pouteria sp., Sapotaceaceae) in the
Amazon. In the Pantanal and Amazon, fishes dispersed 10 and 12 large-
seeded species (N10 mm wide, Stevenson, 2011), respectively.
3.1. Seed dispersal effectiveness
Concordant with predictions, the probability of dispersing intact
seeds increased with body size (see Appendix D for full model results).
We observed the expected negative relationship between fish size and
the probability a fish destroyed all consumed seeds (nu) fortwo species:
a 1 cm increase in fish standard length decreased the odds of complete
seed mastication (nu)by6.6%forP. mesopotamicus in the Pantanal (95%
CI: 2.6, 10.3; padjusted for multiple tests = 0.008) and by 28.0% for
B. melanopterus in the Amazon (95% CI: 9.1, 43.0; adjusted p=0.028).
Furthermore, we observed the expected positive relationship between
fish size and mu [volume of intact to total seeds on the interval (0,1)]
for three species: a 1 cm increase in fish standard length increased the
odds of seed dispersal (mu)by2.7%forP. mesopotamicus in the Pantanal
(95% CI: 0.74%, 4.77%; adjusted p= 0.031; Fig. 3a) and by 58.0%
for B. amazonicus in the Amazon (95% CI: 11.3%, 125.4%; adjusted
p=0.035, Fig. 3b). M. torquatus followed a similar overall pattern
(Fig. 3c), but the p-value was non-significant after correctionfor multi-
ple tests (raw p= 0.034, adjusted p=0.081).
The abundance, species richness and size of dispersed seeds in-
creased with fish size in both wetlands. In the Pantanal, larger fishes dis-
persed more seeds (Fig. 4a, Appendix E) irrespective of fish species
(SL × species, F
3,221
. = 0.65, p= 0.59). In the Amazon, a 1 cm increase
in standard length increased the number of intact seeds by 1.18 (95%
CI: 1.02, 1.35; p= 0.026) for M. torquatus (Fig. 4b), but there was no re-
lationship between fish size and intact seed abundance for
B. amazonicus,B. melanopterus and M. duriventre (SL × species,
F
3,149
= 4.76, adjusted p= 0.021). In both wetlands, the species rich-
ness of dispersed seeds increased with fish size (Fig. 4c, d; Appendix
E), independent of fish species (SL × species, Pantanal: F
3,221
= 0.11,
p=0.95;Amazon:F
1,149
=1.79,p= 0.15), as did the size of dispersed
seeds (Fig. 4e, f; Appendix E; SL × species, Pantanal: F
3,221
=0.71,p=
0.55; Amazon: F
3,149
= 0.85, p=0.47).
In the Pantanal, the effect of treatment (de-pulped control seeds vs.
gut treatment by fish) on germination success differed across seed spe-
cies (seed species × treatment: χ
2
=32.96,p= 0.048; Appendix F). For
8 of 11 seed species, there was no difference in germination success be-
tween seed treatments. For B. arguta (Flacourtiaceae), control seeds
were more likely to germinate than those dispersed by B. hilarii.For
Duroia duckei (Rubiaceae) and Mouriri guianensis (Melastomataceae),
control seeds had higher germination success than those dispersed
by P. mesopotamicus, but not those dispersed by other fish species
(Appendix F). When we analyzed the effect of fish size on germination
success of fish-ingested seeds, we found that larger fishes increased
the odds of germination for four out of 11 seed species (Fig. 5,fish
size × seed species: χ
2
= 29.95, p= 0.001; Appendix G–H), and there
was no relationship between germination success and fish size for the
remaining 7 species.
In the Amazon, germination success differed between seed species
(χ
2
= 23.04, p= 0.0001), but did not differ between seed treatment
(gut passage by various fish species vs. control seeds: χ
2
= 5.07,
p= 0.17). Nor did we find evidence for a seed species by treatment
interaction (χ
2
=4.79,p= 0.44; Appendix F); therefore, in the
Amazon, fishes neither enhance nor depress seed germination success
relative to seeds removed from fruit pulp manually. However, when
we assessed the effect of fish size on germination of fish-ingested
Fig. 3. Proportion of dispersedseeds [volume of intact to total seeds on the interval (0–1),
mu] modeled as a function of fish size (SL—standard length). Pantanal: (a) Piaractus
mesopotamicus. Amazon: (b) Brycon amazonicus, and (c) Myloplus torquatus.
163S.B. Correa et al. / Biological Conservation 191 (2015) 159–167
seeds, we found that a 1 cm increase in standard length enhanced the
odds of germination by 16.7% (95% CI: 4–31%, χ
2
= 6.9, p= 0.009;
Fig. 6). This effect held across seed species, as we found no significant
interaction between fish size and seed species (fish size × seed species:
χ
2
=5.92,p=0.21).
4. Discussion
Seed dispersal effectiveness increased with fish body size in two
distinct regions with different flora, fauna and landscape features. At
any given time, fishes hadaccess to multiple species of fruits of variable
morphologies (Appendix B). In addition, the species composition of
fruits changed across the season in both regions; fruits available to fish-
es captured early in the season were different from those available to
fishes captured late in the season (Correa et al., in prep.). Despite the
variability in the diets of individual fish sampled in different habitats
at different times, we found remarkably consistent results in the
Pantanaland Amazon: relative to small individuals, bigger fishes masti-
cated a lower proportion of the seeds they consume, instead dispersing
larger numbers of intact seeds of a greater plant diversity and of a wider
range of seed-sizes.
Furthermore, big fishes enhanced germination success across
multiple seed species. Frugivores influence the probability of seed
germination by removing the pulp and/or scarifying the seed coat
Fig. 4. Abundance, diversity and maximum size of intact seeds dispersed by fishes in two Neotropical wetlands(Pantanal: a, c, e; Amazon: b, d, f) modeled as a function of fish size
(SL—standard length). pvalues were corrected for multiple comparisons per data set. Black bars represent the range of body sizes per species of individuals included in the analyses.
Species codes follow those in Fig. 2.
164 S.B. Correa et al. / Biological Conservation 191 (2015) 159–167
(Traveset et al., 2008). In our study, ingestion by fishes did not augment
or inhibit germination for most seed species relative to manually de-
pulped seeds, suggesting that fish consumption influences germination
simply by eliminating fruit pulp. However, germination success of
multiple seed species increased with fish size. This positive relationship
between germination success and fish size could be related to longer
retention time in larger intestinal tracts, as intestinal length increases
with fish body size in all of our focal fish species except M. tiete
(R
2
=0.91,F
13,444
=341.4,pb0.0001). Laboratory experiments feeding
seeds of various species to fishes of different body sizes could help elu-
cidate possible physical or chemical changes to seed coats in response to
longer retention times (Pollux, 2011; Traveset et al., 2008).
Effective seed dispersal enhances seedling recruitment and directly
contributes to plant community structure (Jordano, 2000; Wang and
Smith, 2002). Most studies assessing seed dispersal effectiveness focus
on birds and mammals (Schupp et al., 2010) as these are the classic
model systems for endozoochorous seed dispersal (Fleming and Kress,
2013). Only relatively recently have empiricists turned their attention
to quantifying seed dispersal effectiveness of frugivorous fishes.
Previous studies of three of the largest frugivorous fish species in the
Neotropics found that big P. mesopotamicus disperse a greater number
of intact seeds of a common palm in the Pantanal (Galetti et al., 2008),
that the volume of intact seeds increased with body size for
C. macropomum and P. brachypomus in the Peruvian Amazon, and
passage through guts of adult C. macropomum accelerated seed germi-
nation for a common pioneer tree in Amazonian flooded forests
(Anderson et al., 2009). Here, we demonstrate that these patterns,
observed in a handful of population-level studies, can scale to local
communities of frugivorous fishes containing small- and large-sized
Fig. 5. Germination probabilitymodeled as a function of fish size(SL—standardlength) for four plantspecies dispersedby fishes in the Pantanal: (a) Banaraarguta,(b)Cayaponiapodantha,
(c) Eugenia inundata, and (d) Passiflora cf. edulis.
Fig. 6. Germinationprobability modeledas a function of fish size (SL—standard length)for
five plant species dispersed by fishes in the Amazon. The five species are shown in one
panel because the fish size by seed species interaction was non-significant.
165S.B. Correa et al. / Biological Conservation 191 (2015) 159–167
species. Within and across small- to large-sized fish species, large
individuals are better seed dispersers.
4.1. Consequences of overharvesting frugivorous fishes
Big frugivores are key components of seed dispersal networks in
tropical wetlands because of their consumption of large numbers of
fruits, their long seed-retention times, extensive movement patterns,
and unique ability to disperse large-seeded species (Anderson et al.,
2011; Galetti et al., 2008; Kitamura, 2011; Stevenson et al., 2014;
Woodward et al., 2005; Wotton and Kelly, 2012). Given that large fish
species are the main target of commercial fishing operations (Allan
et al., 2005), overexploitation of large frugivorous fishes likely has pro-
found implications for the recruitment of large-seeded canopy species
and the maintenance of diversity in wetland forests. For example, in
our study, large P. mesopotamicus in the Pantanal disperse 27% more
seed species than individuals under the minimum size limit for this fish-
ery (total length = 45 cm). Simulation models suggest that the loss of
large-mammal seed dispersers from Thai forests depressed survival
and growth atmultiple life stages of a canopy tree species, and reduced
population viability (Caughlin et al., 2015). Population declines of forest
elephants in the Congo eliminated seedling recruitment for 12 species
of elephant-dependent large seeded-trees due to increased predation
of undispersed seeds (Beaune et al., 2013). Finally, in western Amazonia,
intact forests with large-bodied primates had higher seedling diversity
and better recruitment of medium- and large-seeded canopy species
relative to overhunted sites (Stevenson, 2011).
In our study, the probability of seed predation decreased with fish
size for P. mesopotamicus,B. melanopterus and B. amazonicus, and we
observed a non-significant trend in that direction for M. torquatus.
Such size-dependent shifts in seed predation could be difficult to reverse
if commercial fisheries induce evolutionary changes in frugivorous fish
populations to smaller individuals that mature early (Palkovacs, 2011).
Overfishing could change the nature of fruit–fish relationships from
primarily mutualistic to increasingly antagonistic with negative
consequences for plant community structure and diversity. Trees in
Amazonian flooded forest typically do not form seed banks, as their
seeds germinate shortly after the water recedes (Ferreira et al., 2010).
Plant communities with negligible seed banks are particularly susceptible
to increased seed predation, which diminishes seedling recruitment
(Vaz Ferreira et al., 2011).
Our study also demonstrates that larger fishes arecapable of dispers-
ing seeds of a broader range of sizes than smaller fishes (Fig. 4e, f). Gape
size limits the size of seeds a frugivore can consume and disperse
(Wheelwright, 1985), which means that large primates and birds
are typically the primary vectors of dispersal for large-seeded plant
species in terrestrial systems (Kitamura, 2011; Stevenson, 2011).
P. mesopotamicus is the biggest frugivorous fishes in the Pantanal. It
can achieve more than five times the length of reproductive adults of
M. tiete (15 cm SL), the smallest frugivorous species therein (Fig. 2).
The size of the largest seed (M. guianensis: 7.6 mm × 5.6 mm) dispersed
by M. tiete, is merely 8% of the size of the largest seed (C. uiti:
40.7 mm × 19.6 mm) dispersed by P. mesopotamicus. Within assem-
blages of frugivorous fishes, adults of the largest fish species disperse
suites of seed species that are unavailable to smaller individuals of
those fish species and to adults of smaller fish species.
Mutualistic interactions between frugivorous fishes and tropical
plants date back to the Late Cretaceous (~ 70 Ma) in South America
(Correa et al., 2015; Thompson et al., 2014), which coincides with the
radiation of angiosperms (Berendse and Scheffer, 2009), and pre-dates
most bird- and mammal–fruit interactions (Correa et al., 2015).
Knowledge of how seed dispersal services of fishes compare with
those of non-aquatic frugivores in wetlands is rather limited, although
observations from one fish species in the Pantanal suggest that fishes
disperse a different suite of species (Donatti et al., 2011). If indeed,
fishes, birds and mammals disperse different seed species, the loss of
one or several key frugivorous fishes may not be mitigated by the
remaining frugivore groups (e.g., Effiom et al., 2014).
4.2. Conclusions
Through a comprehensive multi-system, multi-species study of the
role of fish size on seed dispersal, we documented that the quality and
quantity of seed dispersal increases with body size within and across
fish species. Larger individuals disperse a greater number of viable
seeds of a higher diversity of plants, and are unique vectors of dispersal
for big seeds. As fishes of varying size classes differ in their seed dispers-
al services, we predict that small fishes will not be able to maintain seed
dispersal in wetlands if populations of large fishes continue to decline.
Future empirical studies on recruitment dynamics and spatial aggrega-
tion of fish-dispersed seed species in areas with intact fish populations
vs. overfished areas should be conducted to test the consequences of
seed dispersal limitation for plant population viability (e.g., Caughlin
et al., 2015). Overexploitation of fruit-eating fishes could disrupt an an-
cient mutualism, potentially leading to declines in plant recruitment
success and diversity in wetlands that rely on fruit-eating fishes. To
conserve wetland plant communities, we need to protect not just
the habitat, but the interactions that occur within that habitat. Plant
species composition could change dramatically in floodplain forests if
overfishing removes the fishes that provide the greatest seed dispersal
quantity and quality. Although minimum legal capture sizes have
been established to preserve reproductive potential, this management
strategy does not consider the role of these fishes in wetland plant
regeneration. A scientifically robust management plan that preserves
larger individuals, such as maximum size thresholds (e.g., Pierce,
2010), can help safeguard a key ecosystem process.
Acknowledgments
We thank Jessika Sanabria, Tafnys Hadassa, Érika de-Faria, Pedro
de-Anunciaçao, Ademar, Rodrigo Brandāo, Lázaro Ramos, Anderson
Alvarenga, Ivo Brandāo, Oilton de-Moraes, Márcio Correa, Jhon
Patarroyo, Margarita Roa, Jarvis Rodriguez and Benedicto Neira for field
assistance. We thank Fernando Barbosa, Helio Ferreira, Norida Canchala
and Dairon Cárdenas for plant identifications and Ivonne Vargas for
seed identifications. We thank Tainá F.D. Rodrigues (Universidade Feder-
al de Mato Grosso) for preparing the map. We thank Seth Wenger,
Thomas Pendergast, Mauro Galetti, and two anonymous reviewers for
valuable comments on a previous draft. Permits for this research were
granted by the Chico Mendes Institute of Biodiversity Conservation,
Brazil (License# 42085-1) and the National Authority of Environmental
Licenses, Colombia (Act# 1177 to Universidad de los Andes). We thank
the Eppley Family Foundation, SESC Pantanal, the University of South
Carolina, and the University of Georgia for support of this research.
Equipment donated by IdeaWild (to S.B.C.) was used in this research.
Appendix A. Supplementary data
Supplementary data to this article can be found online at http://dx.
doi.org/10.1016/j.biocon.2015.06.019.
References
Agudelo, E., Bonilla, C.A., Gomez, G.A., Salvino, H., Trujillo, D.L., 2012. Evolución de las
longitudes corporales para la pesquería comercial de bagres en la amazonía
colombiana.Rev.Colomb.Amazón.5,176–195.
Albuquerque, S.P., Catella, A.C., Campos, F.L.d.R., Santos, D.C.d., 2012. Sistema de Controle
da Pesca de Mato Grosso do Sul SCPESCA/MS 17-2010. EMBRAPA, Corumbá, MS.
Allan, J.D., Abell, R., Hogan, Z., Revenga, C., Taylor, B.W., Welcomme, R.L., Winemiller, K.,
2005. Overfishing of inland waters. Bioscience 55, 1041–1051.
Anderson, J.T., Saldaña-Rojas, J., Flecker, A.S., 2009. High-quality seed dispersal by
fruit-eating fishes in Amazonian floodplain habitats. Oecologia 161, 279–290.
Anderson,J.T., Nuttle, T., Saldaña-Rojas, J.S., Pendergast,T.H., Flecker, A.S.,2011. Extremely
long-distance seed dispersal by an overfished Amazonian frugivore. Proc. R. Soc.
Lond. Ser. B Biol. Sci. 278, 3329–3335.
166 S.B. Correa et al. / Biological Conservation 191 (2015) 159–167
Arieira, J., Karssenberg, D., de Jong, S.M., Addink, E.A., Couto, E.G., Nunes da Cunha, C.,
Skøien, J.O., 2011. Integrating field sampling, geostatistics and remote sensing to
map wetland vegetation in the Pantanal, Brazil. Biogeosciences 8, 667–686.
Aslan, C., Zavaleta, E., Tershy, E.S., Croll, B., D., 2013. Mutualism disruption threatens
global plant biodiversity: a systematic review. PLoS ONE 8 (e66993).
Barthem, R., Goulding, M., 2007. An Unexpected Ecosystem. The Amazon as Revealed by
Fisheries. Asociación para la Conservación de la Cuenca Amazónica (ACCA), Missouri
Botanical Garden Press, Lima.
Beaune, D., Fruth, B., Bollache, L., Hohmann, G., Bretagnolle, F., 2013. Doom of the
elephant-dependenttrees in a Congo tropical forest. For.Ecol. Manag. 295, 109–117.
Benjamin, Y., Hochberg, Y., 1995. Controlling the false discovery rate: a practical and
powerful approach to multiple testing. J. R. Stat. Soc. A. Stat. Soc. 57, 289–300.
Berendse, F., Scheffer, M., 2009. The angiosperm radiation revisited, an ecological
explanation for Darwin's ‘abominable mystery’. Ecol. Lett. 12, 872-885.
Castello, L., McGrath,D.G., Arantes, C.C.,Almeida, O.T., 2013. Accounting forheterogeneity
in small-scale fisheries management: the Amazon case. Mar. Policy 38, 557–565.
Caughlin,T.T., Ferguson, J.M., Lichstein, J.W., Zuidema, P.A.,Bunyavejchewin, S., Levey,D.J.,
2015. Loss of animal seed dispersal increases extinction risk in a tropical tree species
due to pervasive negative density dependence across life stages. Proc. R. Soc. Lond.
Ser. B Biol. Sci. 282, 20142095.
Correa, S.B., Winemiller, K.O., 2014. Niche partitioning among frugivorous fishes in
response to fluctuating resources in the Amazonian floodplain forest. Ecology 95,
210–224.
Correa, S.B., Costa-Pereira, R.,Fleming, T.H., Goulding, M., Anderson, J.T.,2015. Neotropical
fruit–fish interactions: eco-evolutionary dynamics and conservation. Biol. Rev. Camb.
Philos. Soc. http://dx.doi.org/10.1111/brv.12153.
Dirzo, R., Young, H.S., Galetti, M., Ceballos, G., Isaac, N.J., Collen, B., 2014. Defaunation in
the Anthropocene. Science 345, 401–406.
Donatti, C.I., Guimaraes, P.R., Galetti, M., Pizo, M.A., Marquitti, F.M.D., Dirzo, R., 2011.
Analysis of a hyper-diverse seed dispersal network: modularity and underlying
mechanisms. Ecol. Lett. 14, 773–781.
Duivenvoorden, J.F., 1996. Patterns of tree species richness in rain forests of the middle
Caquetá area, Colombia, NW Amazonia. Biotropica 28, 142–158.
Effiom, E.O., Birkhofer, K., Smith, H.G., Olsson, O., 2014. Changes of community
composition at multiple trophic levels due to hunting in Nigerian tropical forests.
Ecography 37, 367–377.
Ferreira, C.S., Piedade, M.T.F., Oliveira-Wittmann, A.d., Franco, A.C., 2010. Plant
reproduction in the Central Amazonian floodplains: challenges and adaptations.
AoB Plants http://dx.doi.org/10.1093/aobpla/plq009.
Firth, D., 1993. Bias reduction of maximum likelihood estimates. Biometrika 80, 27–38.
Fleming, T.H., Kress, W.J., 2013. The Ornaments of Life. Coevolution and Conservation in
the Tropics. The University of Chicago Press, Chicago.
Galetti, M., Donatti, C.I., Pizo, M.A., Giacomini, H.C., 2008. Big fish are the best: seed
dispersal of Bactris glaucescens by the Pacu fish (Piaractus meso potamicus)inthe
Pantanal, Brazil. Biotropica 40, 386–389.
Galetti, M., Guevara, R., Côrtes, M.C., Fadini, R., Von Matter, S., Leite, A.B., Labecca, F.,
Ribeiro, T., Carvalho, C.S., Collevatti, R.G., Pires, M.M., Guimarães Jr., P.R., Brancalion,
P.H., Ribeiro, M.C., Jordano, P., 2013. Functional extinction of birds drives rapid
evolutionary changes in seed size. Science 340, 1086–1090.
Horn, M.H., Correa, S.B., Parolin, P., Pollux, B.J.A., Anderson, J.T., Lucas, C., Widmann, P.,
Tiju, A., Galetti, M., Goulding, M., 2011. Seed dispersal by fishes in tropical and tem-
perate fresh waters: the growing evidence. Acta Oecol. 37, 561–577.
Howe, H.F., Smallwood, J., 1982. Ecology of seed dispersal. Annu. Rev. Ecol. Syst. 13,
201–228.
Isaac, V.J., Ruffino, M.L., 1996. Population dynamics of tambaqui, Colossoma macropomum
Cuvier, in the lower Amazon, Brazil.Fish. Manag. Ecol. 106, 110–127.
Jordano, P., 2000. Fruits and frugivory. In: Fenner, M. (Ed.), Seeds: The Ecology of
Regeneration in Plant Communities, 2nd edition CABI International, Wallingford,
UK, pp. 125–166.
Junk, W.J., 1993. Wetlands of tropical South America. In: Whigham, D.F., Dykyjová, D.,
Hejný, S. (Eds.), Wetlands of the World: Inventory, Ecology and Management Vol. I.
Springer, The Netherlands, pp. 679–739.
Junk, W.J., Nunes da Cunha, C., 2005. Pantanal: a large South American wetland at a
crossroads. Ecol. Eng. 24, 391–401.
Junk, W.J., Piedade, M.T.F., 2010. An introduction to South American wetland forests:
distribution, definitions and general characterization. In: Junk, W.J., Piedade, M.T.F.,
Wittmann, F., Schöngart, J., Parolin, P. (Eds.), Central Amazonian Floodplain Forests:
Ecophysiology, Biodiversity, and Sustainable Management. Springer, New York,
pp. 4–25.
Kitamura, S., 2011. Frugivory and seed dispersal by hornbills (Bucerotidae) in tropical
forests. Acta Oecol. 37, 531–541.
Kubitzki, K., Ziburski, A., 1994. Seed dispersal in flood-plain forests of Amazonia.
Biotropica 26, 30–43.
Markl, J.S., Schleuning, M., Forget, P.M., Jordano, P., Lambert, J.E., Traveset, A., Wright, S.J.,
Böhning-Gaese, K., 2012. Meta-analysis of the effects of human disturbance on seed
dispersal by animals. Conserv. Biol. 26, 1072–1081.
Mateus, L.A.d.F., Penha, J.M.F., Petrere, M., 2004. Fishing resources in the rio Cuiabá Basin,
Pantanal do Mato Grosso, Brazil. Neotropical Ichthyol. 2, 217–227.
Ollerton, J., Winfree, R., Tarrant, S., 2011. How many flowering plants are pollinated by
animals? Oikos 120, 321–326.
Palkovacs, E.P., 2011. The overfishing debate: an eco-evolutionary perspective. Trends
Ecol. Evol. 26, 616–617.
Peixer, J., Catella, A.C., Petrere-Júnior, M., 2007. Yield per recruit of the Pacu Piaractus
mesopotamicus (Holmberg, 1887) in the Pantanal of Mato Grosso do Sul, Brazil.
Braz. J. Biol. 67, 561–567.
Pierce, R.B., 2010. Long-term evaluations of length limit regulations for Northern Pike in
Minnesota. N. Am. J. Fish Manag. 30, 412–432.
Pinheiro, J., Bates, D., DebRoy, S., Sarkar, D., 2014. R Development Core Team (2014).
nlme: linear and nonlinear mixed effects models. R Package Version 3.1-118.
R Foundation for Statistical Computing, Vienna.
Pollux, B.J.A., 2011. The experimental study of seed dispersal by fish (ichthyochory).
Freshw.Biol.56,197–212.
Poulsen, J.R., Clark, C.J., Palmer, T.M., 2013. Ecological erosion of anAfrotropical forest and
potential consequences for tree recruitment and forest biomass. Biol. Conserv. 163,
122–130.
Reinert, T.R., Winter, K.A., 2002. Sustainability of harvested Pacu (Colossoma
macropomum) populations in the northeastern Bolivian Amazon. Conserv. Biol. 16,
1344–1351.
Reis, R.E., Kullander, S.O., Ferraris, C.J.J. (Eds.), 2003. Check List of the Freshwater Fishes of
South and Central America. EDIPUCRS, Porto Alegre.
Rigby, R.A., Stasinopoulos, D.M., 2005. Generalized additive models for location, scale and
shape (with discussion). J. R. Stat. Soc. Ser. C Appl. Stat. 54 (part 3), 507–554.
Rodriguez, C.A., 1991. Commercial Fisheries in the Lower Caqueta River. Tropenbos-
Colombia, Bogotá, D.C.
Rodriguez, M.A., Winemiller, K.O., Lewis Jr., W.M., Taphorn Baechle, D.C., 2007. The
freshwater habitats, fishes, and fisheries of the Orinoco River Basin. Aquat. Ecosyst.
Health Manag. 10, 140–152.
Ruffino, M.L., 2008. Sistema integrado de estatística pesqueira para a Amazônia. Pan-Am.
J. Aquat. Sci. 3, 193–204.
Santos, G.M.d., Santos, A.C.M.d., 2005. Sustentabilidade da pesca na Amazônia. Estud.
Avançados 19, 165–182.
Schupp, E.W., Jordano, P., Gomez, J.M., 2010. Seed dispersal effectiveness revisited:
a conceptual review. New Phytol. 188, 333–353.
Seidler, T., Plotkin, J.B., 2006. Seed dispersal and spatial pattern in tropical trees. Public
Libr. Sci. Biol. 4, e344.
Stevenson, P.R., 2011. The abundance of large Ateline monkeys is positively associated
with the diversity of plants regenerating in Neotropical forests. Biotropica 43,
512–519.
Stevenson, P.R., Link, A., Onshuus, A., Quiroz, A.J., Velasco, M., 2014. Estimation of seed
shadows generated by Andean woolly monkeys (Lagothrix lagothricha lugens). Int.
J. Primatol. 35, 1021–1036.
Thompson, A.W., Betancur-R, R., López-Fernández, H., Ortí, G., 2014. A time-calibrated,
multi-locus phylogeny of piranhas, pacus, and allies (Characiformes: Serrasalmidae)
and a comparison of species tree methods. Mol. Phylogenet. Evol. 81, 242–257.
Traveset, A., Rodríguez-Pérez, J., Pías, B., 2008. Seed trait changes in dispersers' guts and
consequences for germination and seedling growth. Ecology 89, 95–106.
Valiente-Banuet, A., Aizen, M.A., Alcántara, J.M., Arroyo, J., Cocucci, A., Galetti, M., García,
M.B., García, D., Gómez, J.M., Jordano, P., Mendel, R., Navarro, L., Obeso, J.R., Oviedo,
R., Ramírez, N., Rey, P.J., Traveset, A., Verdú, M., Zamora, R., 2015. Beyond species loss:
the extinction of ecological interactions in a changing world. Funct. Ecol. 29, 299–307.
Vaz Ferreira, A., Bruna, E.M., Vasconcelos, H.L., 2011. Seed predators limit plant
recruitment in Neotropical savannas. Oikos 120, 1013–1022.
Wang, B.C., Smith, T.B., 2002. Closing the seed dispersal loop. Trends Ecol. Evol. 17,
379–386.
Wheelwright, N.T., 1985.Fruit size, gape width, andthe diets of fruit-eating birds. Ecology
66, 808–818.
Woodward, G., Ebenman, B., Emmerson, M., Montoya, J.M., Olesen, J.M., Valido, A.,
Warren, P.H., 2005. Body size in ecological networks. Trends Ecol. Evol. 20, 402–409.
Wotton, D.M., Kelly, D., 2012. Do larger frugivores move seeds further? Body size, seed
dispersal distance, and a case study of a large, sedentary pigeon. J. Biogeogr. 39,
1973–1983.
Zuur, A.F., Ieno, E.N., Walker, N.J., Saveliev, A.A., Smith, G.M., 2009. Mixed Effects Models
and Extensions in Ecology with R. Springer, NY.
167S.B. Correa et al. / Biological Conservation 191 (2015) 159–167
