ArticlePDF Available

Relative importance of phenotypic trait matching and species' abundances in determining plant - Avian seed dispersal interactions in a small insular community


Abstract and Figures

Network theory has provided a general way to understand mutualistic plant-animal interactions at the community level. Still, the mechanisms responsible for interaction patterns remain controversial. In this study we use a combination of statistical models and probability matrices to evaluate the relative importance of species morphological and nutritional (phenotypic) traits and species abundance in determining interactions between fleshy-fruited plants and birds that disperse their seeds. Models included variables associated with species abundance, a suite of variables associated with phenotypic traits (fruit diameter, bird bill width, fruit nutrient compounds), and the species identity of the avian disperser. Results show that both phenotypic traits and species abundance are important determinants of pairwise interactions. However, when considered separately, fruit diameter and bill width were more important in determining seed dispersal interactions. The effect of fruit compounds was less substantial and only important when considered together with abundance-related variables and/or the factor 'animal species'. Published by Oxford University Press on behalf of the Annals of Botany Company.
Content may be subject to copyright.
Research Article
SPECIAL ISSUE: Island Plant Biology: Celebrating Carlquist’s
Relative importance of phenotypic trait matching and
species’ abundances in determining plant avian seed
dispersal interactions in a small insular community
´n Gonza
´lez-Castro1,2, 3*, Suann Yang2,4, Manuel Nogales1and Toma
´s A. Carlo2
Island Ecology and Evolution Research Group (CSIC-IPNA), C/Astrofı
´sico Francisco Sa
´nchez n83, 38206, La Laguna, Tenerife,
Canary Islands, Spain
Department of Biology, Pennsylvania State University, 208 Mueller Laboratory, University Park, PA 16802, USA
Present address: Instituto de Ciencia Innovacio
´n Tecnologı
´a y Saberes Universidad Nacional de Chimborazo, Avenida Antonio Jose
Sucre, Riobamba, Ecuador
Present address: Biology Department, Presbyterian College, 503 South Broad Street, Clinton, SC 29325, USA
Received: 30 October 2014; Accepted: 10 February 2015; Published: 5 March 2015
Guest Editor: Donald Drake
Citation: Gonza
´lez-Castro A, Yang S, Nogales M, Carlo TA. 2015. Relative importance of phenotypic trait matching and species’
abundances in determining plant–avian seed dispersal interactions in a small insular community. AoB PLANTS 7: plv017; doi:10.1093/
Abstract. Network theory has provided a general way to understand mutualistic plant– animal interactions at the
community level. However, the mechanisms responsible for interaction patterns remain controversial. In this studywe
use a combination of statistical models and probability matrices to evaluate the relative importance of species mor-
phological and nutritional (phenotypic) traits and species abundance in determining interactions between fleshy-
fruited plants and birds that disperse their seeds. The models included variables associated with species abundance,
a suite of variables associated with phenotypic traits (fruit diameter, bird bill width, fruit nutrient compounds), and the
species identity of the avian disperser. Results show that both phenotypic traits and species abundance are important
determinants of pairwise interactions. However, when considered separately, fruit diameter and bill width were more
important in determining seed dispersal interactions. The effect of fruit compounds was less substantial and only
important when considered together with abundance-related variables and/or the factor ‘animal species’.
Keywords: Dispersal; frugivory; mutualistic networks; oceanic islands; probability matrices.
A ubiquitous mutualistic plant animal interaction is
that between fleshy-fruited plants and the fruit-eating
animals that disperse their seeds (Jordano 2000). Seed
dispersal interactions are complex because they involve
multiple species of animals and plants, the temporal
and spatial variability of such interactions (Yang et al.
2014) and the influence of frugivore behaviour and physi-
ology, as well as the chemistry of fruits (Jordano 2000;Carlo
*Corresponding author’s e-mail address:
Published by Oxford University Press on behalf of the Annals of Botany Company.
This is an Open Access article distributed under the terms of the Creative Commons Attribution License (
licenses/by/4.0/), which permitsunrestricted reuse,distribution, and reproduction inany medium,provided the originalwork is properly cited.
AoB PLANTS &The Authors 2015 1
by guest on April 22, 2015 from
and Yang 2011). Network theory has emerged as a useful
tool to deal with such complexity and to search for organ-
izational and coevolutionary patterns in community-wide
plant–frugivore interactions (e.g. Bascompte et al. 2003,
2006;Jordano et al. 2003;Bascompte and Jordano 2007;
Rezende et al. 2007;Gonza
´lez et al. 2010;Mello et al.
2011;Aizen et al. 2012;Stouffer et al. 2012). However, the
mechanisms responsible for interaction patterns in such
networks (e.g. nestedness, modularity, interaction asym-
metry, degree distribution) remain unclear.
Two hypotheses are available to explain how mutualistic
interactions influence the structure of mutualistic net-
works. The first is the neutrality hypothesis (so-called
abundance hypothesis), which states that observed pat-
terns within a community are due torandom species inter-
actions. According to neutrality, probabilities of observing
a plant– disperser interaction chiefly depend on the abun-
dance of species. For example, observing both common
and rare frugivores feeding on common fruiting plant spe-
cies is more likely than on rare ones. This implies that
abundance will be positively correlated with the level of
generalization in the mutualistic interactions, i.e. highly
abundant species could artificially appear as generalists
that are highly connected in the mutualistic network,
and rare species as more specialized (e.g. Dupont et al.
´zquez 2005;Va
´zquez et al. 2007;Schleuning
et al. 2011).
On the other hand, the phenotypic traits hypothesis
postulates that interaction patterns result from morpho-
logical, physiological, behavioural or evolutionary con-
straints that condition interaction probabilities between
potential mutualistic partners (Jordano et al. 2003;
Rezende et al. 2007;Santamarı
´a and Rodrı
2007;Dupont and Olesen 2009;Mello et al. 2011;Olesen
et al. 2011). Among phenotypic traits, the most commonly
used in analyses of seed dispersal networks are the
disperser bill width and fruit diameter (i.e. this will deter-
mine whether or not a seed can be swallowed and
dispersed), as well as accessibility restrictions by frugi-
vores (e.g. Rezende et al. 2007;Olesen et al. 2011;Burns
2013). However, although it has been shown that the
chemical compounds of fruit can be important in deter-
mining frugivory and seed dispersal interactions (Jordano
2000 and references therein), such traits have not been
used previously in network analyses. In this study we
incorporate, for the first time, fruit nutritional compounds
into the analysis of a frugivory network.
Some studies have demonstrated that both mechanisms
(abundance and phenotypic traits) can work hand-in-hand
to shape network structure (e.g. Stang et al. 2006;
Bascompte and Jordano 2007). In this vein, Verdu
Valiente-Banuet (2011) demonstrated that facilitative
interactions between plants were better explained by a
combination of abundance and phylogenetic relationships
than by these variables separately. Still, both hypotheses
are not necessarily mutually exclusive and ecologists are
beginning to examine their relative importance. Because
interaction networks can be presented as adjacency
matrices, Va
´zquez et al. (2009a)proposed using probabil-
ity matrices (derived from species abundance and
their spatial–temporal overlap) to assess the relative
importance of abundance and phenotypic traits to deter-
mine the observed patterns of mutualistic interactions.
This approach is useful to predict aggregate network
parameters, but not effective in predicting pairwise
interactions (Va
´zquez et al. 2009a).
To improve this approach, here we model pairwise in-
teractions between fleshy-fruited plants and their avian
dispersers, as a response to species’ phenotypic traits,
as well as to the species’ abundance. Then we use inter-
action frequencies predicted by the best statistical model
to build a probability matrix as proposed by Va
´zquez et al.
(2009a)to assess the ability of that model to predict ag-
gregate network parameters. Solving these questions will
help us to understand evolutionary and ecological forces
driving the assemblage of interactions in mutualistic
communities (Bascompte and Jordano 2007).
Study area
The study was carried out in Los Adernos, a Mediterranean
scrubland habitat site located in the northwest region of
Tenerife (Canary Islands, UTM: 28R 317523 E/3138253 N,
220 m above sea level (a.s.l.)). The climate is Mediterra-
nean, with mean annual rainfall ranging between 200
and 400 mm and mean temperature between 16 and
19 8C. The fleshy-fruited plant community is mainly com-
posed of Asparagus plocamoides Webb ex Svent., Jasmi-
num odoratissimum L., Rubia fruticosa Aiton, Rhamnus
crenulata Aiton and Heberdenia excelsa (Aiton) Banks ex
DC. There are fouravian disperser species: Sylvia atricapilla,
S. melanocephala,Turdus merula and Erithacus rubecula.
The study was conducted in two sampling periods encom-
passing two whole years: from June 2008 through May
2009, and from January 2010 through December 2010.
Seed dispersal interactions
In order to characterize the set of interactions between
fleshy-fruited plants and avian dispersers, we focussed
on seeds recovered from the faeces of birds captured
with mist nets that were opened from dawn until dusk
23 days per month. We computed mist-netting effort
by multiplying the mist-net length by the number of
hours they were operated. Faecal samples were analysed
with a dissecting scope for seeds, which were counted
2AoB PLANTS &The Authors 2015
´lez-Castro et al. — Species traits, abundance and seed dispersal networks on oceanic islands
by guest on April 22, 2015 from
and identified to species level. In order to take into ac-
count interspecific differences on the number of seeds
produced per fruit, we divided the number of seeds dis-
persed by the mean number of seeds produced per fruit.
Doing so gave us a better estimation of the number of
times a given disperser visit a plant species for fruits.
With these data we constructed an interaction matrix
based on the interaction frequency between fleshy-fruited
plants and avian dispersers [see Supporting Information].
Network theory has been usually applied for large and
complex communities, whereas the community in this
study is small (four animal and nine plant species).
Small communities, however, are less prone to sample
bias than large ones (Blu
¨thgen 2010), and the reliability
of studies will be greater when the more accurate is the
sampling of interactions. We used an accumulation
curve to prove the robustness of our sampled interac-
tions, with a curve slope lower than 0.03 after all our 54
mist-netting sessions [see Supporting Information].
Explanatory variables
We considered eight explanatory variables associated
with phenotypic traits (six variables) and abundance
hypotheses (two variables) in order to explain interaction
frequency between plants and animals (i.e. number of
dispersed seeds). Explanatory variables for phenotypic
traits hypothesis were fruit organic compounds (fibre,
lipids, sugars and proteins), the identity of bird species
and size overlap between fruit diameter and bill
width of birds (hereafter size overlap). Although a wide
variety of fruit chemical compounds may influence the
choice of fruits by birds (Jordano 2000 and references
therein), we selected sugars, fibre, proteins and lipids
based on a study on Mediterranean avian-dispersed fruits
(Herrera 1987). We decided to include the factor ‘animal
species’ involved in each plant animal interaction
because species identity is important to predict animal
interaction patterns (Carlo et al. 2003;Carnicer et al. 2009).
The two explanatory variables used to test for the
abundance hypothesis were the product of abundance
of interacting species (hereafter abundance) and tem-
poral overlap of species phenophase length (time length
which plants display fruits and bird species are present at
the study site). Although phenophase length and hence
temporal overlap are, to some extent, species-specific
traits, they can also be considered as metrics of abun-
dance because a species can be abundant either by pro-
ducing high fruit densities and/or by being available over
long time periods (Va
´zquez et al. 2009a). Moreover, the
phenophase length of fruiting plants can also be affected
by external factors to the plant such as weather condi-
tions or the depletion of fruit crops.
Fruit nutrient compounds. Chemical analyses of fruits
were performed by Canagrosa Laboratories (http://www. Amount of compounds was calculated
as percentage of dry mass by different methods: Kjeldahl
method for proteins, gravimetric plus digestion with
acid-detergent solution for fibre and Soxhlet extraction
with hexane for lipids. The amount of sugars was
calculated based on the remaining organic material
following the equation:
sugars =NFES ×100
100 RH ,(1)
where RH is the relative humidity of the sample and NFES
(nitrogen-free extractive substances) is calculated as
NFES =100 RH proteins lipids fibre.(2)
Animal species. We accounted for the animal species
identity as a factor to explain the interaction frequency.
Although quantification of animal traits exists for the
species studied here (Herrera 1984;Jordano 1987), our
system is unfortunately too small (36 interactions)
to support models with these additional explanatory
variables. We decided to give priority to use fruit nutrient
compounds because animal traits have been previously
used in some extent on network analyses, whereas fruit
nutrients have not yet.
Size overlap. To account for individual variability of fruit
and bill size, we decided to use the range (mean+SD) in
both fruit diameter and bird gape width instead of just
comparing their mean values. For each pair of species,
we calculated the percentage of range of fruit diameter
that was equal or smaller than the maximum value of
the bird gape width. For example, if the diameter of a
fruit species ranges from 7.0 to 8.0 mm (variation
range ¼1.0 mm), then the resulting overlap with the
gape of a bird that ranges between 9.0 and 10.0 mm
would be 100 %. However, for a bird with a gape width of
6.0– 7.5 mm, the size overlap is 50 %, because only half of
the fruit variation range (0.5 out of 1 mm) could be
‘swallowed’ by this second bird. For those interactions
with pairs of species which size overlap was 0 % we
arbitrarily establish a size overlap of 1 ×10
Abundance and temporal overlap. The variable
abundance is the product of abundance of the interacting
species. To assess fruit abundance we used 20 plots of 5 m
randomly placed. We visited every plot monthly and
estimated the number of fruits per metre square for every
plant species (visual counting method, Chapman et al.
AoB PLANTS &The Authors 2015 3
´lez-Castro et al. — Species traits, abundance and seed dispersal networks on oceanic islands
by guest on April 22, 2015 from
1992). We estimated the cumulative abundance after the
two study years and then calculated the relative fruit
abundance for every plant species as the percentage of
fruits of each species from the total community-wide fruit
crop. Bird abundance was estimated by using a simple
mixed effect regression analysis for every 100 h of
sampling: [Individuals ×m
¼2.15 +4.117 ×(100 ×C);
P¼0.001; N¼152], where Cis the number of captured
birds per unit of effort. To build this regression we used
unpublished data (A.G.-C.) from the same study area.
Censuses were performed twice per month. Therefore, we
considered the date as a random effect factor to avoid an
effect of temporal pseudo-replication. To ensure bird
detectability in censuses, a band 25 m wide was surveyed,
where all individuals (seen or heard) were counted. We
consider that all disperser birds had equivalent capture
probability in mist nets because shrubs mostly dominate
the study site, and frugivorous birds have same movement
patterns, between shrubs, where mist nets where placed.
Every 15 days we recorded the presence of species of
fruits and birds to obtain the length of species pheno-
phase. The temporal overlap is defined as the percentage
of days with respect to the whole study period (i.e. 730
days) that pairs of species coincided in the study area.
For example, if a plant species fruited for 60 days, and a
bird species was present in the study area for 60 days, but
fruit and bird species coincided for only 30 days, then they
had 4.11 % of temporal overlap (30 out of 370 days).
Modelling interaction frequency between pairs
of species
We modelled the log-transformation of interaction fre-
quency (estimated as explained above) as a response to
the explanatory variables by using a generalized least
squares (GLS) model in the nlme package (Pinheiro et al.
2009) implemented in R 2.11 (R Development Core Team
2015). The variance was not homogeneous and changed
with the predictor ‘size overlap’, thus the GLS model
allowed us to work with a normal error distribution and
taking into account the variance structure by using the
function ‘varFixed’, implemented in the nlme package
(Zuur et al. 2009).
As our study system was small, with only 36 plant
frugivore interactions, we had to build different models
with different subsets of explanatory variables to avoid
model over fitting (sensu Burnham and Anderson 2002).
Therefore, we used different combinations of phenotypic
traits and abundance variables in different models [see
Supporting Information].Bydoingso,eachstatistical
model included only a maximum of five explanatory vari-
ables (including main effects of variables and/or their
statistical). In models that we could not include all fruit
compounds together, we separated them into two differ-
ent sets: one included ‘non-energetic’ compounds (fibre
and proteins) and the other included ‘highly energetic’
compounds (sugars and lipids).
We ranked models according to the AIC value and com-
puted the Akaike’s weight as an estimation of the prob-
ability of a given model to be the best candidate model
explaining the observed interactions (Burnham and
Anderson 2002). To evaluate the importance of different
explanatory variables we used a multi-model inference,
based on the sum of Akaike’s weight of each model
where each explanatory variable appeared (Burnham
and Anderson 2002). Different variables appeared in
a very different number of models (e.g. size overlap
appeared in three models, whereas ‘animal species’
appeared in 14 models), based on the natural history of
the fruiting plants and animals rather than their statistic-
al importance. For such a reason, we averaged the sum of
Akaike’s weight by dividing it by the total number of mod-
els where each variable appeared.
Prediction of aggregate network parameters
With the interaction frequency predicted by the best stat-
istical model (that showed the lowest AIC value) we cre-
ated an expected interaction matrix. Subsequently, we
normalized this matrix by dividing their elements by
their total number of predicted interactions to obtain a
probability matrix. With this probability matrix, we used
simulations based on the approach of Va
´zquez et al.
(2009a)to assess the capacity of our best model to pre-
dict different aggregate network parameters: connec-
tance (proportion of realized interactions respect to
total cells in the interaction matrix), interaction evenness
(which is a Shannon index proposed by Tylianakis et al.
2007; the higher the index, the more evenly distributed
are interactions in the matrix), nestedness (the degree
to which specialists interact with proper subsets of the
species that generalists interact with) and interaction
asymmetry for fruits and for dispersers (Va
´zquez et al.
2007,2009a). We performed 1000 randomizations for
the model (i.e. the probability matrix), calculated the
mean and 95 % confidence interval of each parameter
and assessed if the observed value for each parameter
in the interaction matrix recorded at Los Adernos fell
within such confident interval.
We had problems in simulating nestedness values
with the temperature algorithm proposed by Rodrı
´s and Santamarı
´a (2006),perhapsduetosmall
size of our interaction matrix. Therefore, we used the
Nestedness metric based on Overlap and Decreasing Fill
(NODF; Almeida-Neto et al. 2008) and ‘weighted nested-
ness’ (Galeano et al. 2009) algorithms. All the analyses
were run with the R-code provided by Va
´zquez et al.
4AoB PLANTS &The Authors 2015
´lez-Castro et al. — Species traits, abundance and seed dispersal networks on oceanic islands
by guest on April 22, 2015 from
(2009a)but with modifications to include NODF and
weighted nestedness measures as implemented in the
bipartite (Dormann et al. 2009) package of R statistical
software (R Development Core Team 2015).
Statistical models
We created 24 statistical models, containing each ex-
planatory variable separately, as well as different combi-
nations of them [see Supporting Information]. The best
model to explain plantbird interactions combined both
species phenotypic traits and species abundance. This
model (AIC ¼215.00) included the factor ‘animal species’
and variables related with species matching (size overlap
between fruit diameter and bird bill width, temporal over-
lap of species phenophase length and the product of spe-
cies abundance). The second best model (AIC ¼24.57)
included some fruit compounds (fibre and proteins), the
factor ‘animal species’ and both abundance-related vari-
ables (temporal overlap and species abundance) [see Sup-
porting Information]. The null model (that with only the
intercept) was better than three models based only on
species abundance, size overlap and ‘size overlap ×ani-
mal species’, respectively [see Supporting Information].
Considering the averaged sum of Akaike’s weight, the
most important variable was size overlap between fruit
diameter and bird bill width, followed by species temporal
overlap, species abundance and the factor ‘animal spe-
cies’ (Table 1). However, the difference between temporal
overlap and species abundance was not significant (the
order of magnitude of that difference was 1 ×10
According to the Akaike’s weight, fruit nutrient com-
pounds were less important variables (Table 1). However,
models that combined both fruit compounds with the
identity of animal species and/or variables related with
abundance were among the best fitted [see Supporting
In general, birds dispersed more frequently plant spe-
cies with a lower amount of sugars and lipids in their
fruits, with the exception of T. merula,whichtendto
select those fruits with higher sugar content (Fig. 1).
Birds, with the exception again of T. merula, more often
dispersed fruits with higher protein content. Curiously,
fibre-rich fruits, which are of low digestibility, also had,
in general, a high dispersal frequency (Fig. 1). Interaction
frequency was also higher for pairs of species with a
higher product of their abundance and higher temporal
overlap, with the exception of S. melanocephala (Fig. 2).
With respect to aggregate network parameters, the
probability matrix based on the best statistical model
was able to predict only nestedness, based on both
‘NODF’ and ‘weighted nestedness’ algorithms (Table 2).
Our study shows that the best way to understand pairwise
interactions of the plant– frugivore network in the scrub-
lands of Tenerife is using both the phenotypic traits
and the abundance of species. Previous studies in other
sites and mutualistic interactions (e.g. pollination)
have reached similar conclusions (Stang et al. 2006;
Bascompte and Jordano 2007;Verdu
´and Valiente-
Banuet 2011), but our study stands out in showing that
matching of two phenotypic traits (fruit diameter and
bird bill width) is a stronger determinant of mutualistic in-
teractions than species’ abundances (Table 1). Our best
model was able to predict only one network parameter,
nestedness (NODF and ‘weighted nestedness’), which is
in contrast to previous studies able to predict more para-
meters (Va
´zquez et al. 2009a;Verdu
´and Valiente-Banuet
2011). However, it is important because nestedness has
been proposed to be an important structural feature de-
termining species coexistence and diffuse coevolution
(Bascompte et al. 2003). Therefore, our results support
that nestedness may be strongly influenced by both fruit-
bill matching, as well as by species abundance as previ-
ously proposed (e.g. Va
´zquez 2005;Rezende et al. 2007).
Previous studies have demonstrated that an important
trait related to seed dispersal frequency is the overlap be-
tween fruit diameter and bird gape width (e.g. Rezende
et al. 2007;Olesen et al. 2011),whichisinaccordwith
our best model. Increasing the size overlap between the
Table 1. Relative importance of each explanatory variable to
determine fruit–avian disperser interactions in Los Adernos
(Northwest of Tenerife Island). Variables are ranked from most
important to less important according to the averaged sum of
Akaike’s weight, w
(Burnham and Anderson 2002). The higher the
value, the more important the explanatory variable is. As the
number of models in which each variable appears is established by
our knowledge of fruit and bird natural history, and not because of
statistical reasons, we decided to use the averaged sum of
Akaike’s weight.
Explanatory variable Number of models
in which variable
sum of w
Size overlap 3 0.333333333
Temporal overlap 7 0.143707147
Species abundance 7 0.143707147
Animal species 14 0.072078903
Fibre 5 0.001181644
Proteins 5 0.001181644
Sugars 5 0.000568973
Lipids 5 0.000568973
AoB PLANTS &The Authors 2015 5
´lez-Castro et al. — Species traits, abundance and seed dispersal networks on oceanic islands
by guest on April 22, 2015 from
bill widths and fruit diameters increases the probability of
a successful seed dispersal interaction between a bird
plant species pair (e.g. Pratt and Stiles 1985;Wheelwright
1985;Jordano 1987;Levey 1987). Size restriction could
explain why in this community, some small bird species
depended heavily on smaller fruits of low digestibility
(e.g. S. melanocephalaR. crenulata) than on more prof-
itable but larger fruits (e.g. Tamus edulis, with a 97.87 %
of sugars). This size restriction could also explain why the
smallest passerine (S. melanocephala) has very few inter-
actions with the large-fruited H. excelsa (Fig. 2), despite
these two species having a high temporal overlap and
high abundances.
Although the importance of fruit chemistry in mediat-
ing plantfrugivore interactions has been amply demon-
strated (Jordano 2000), this study is the first to include
them as part of structural analyses of a mutualistic
network. We found some relationship between fruit
nutrient amount and interaction frequency (Fig. 1). How-
ever, fruit compounds were weak predictors of fruitbird
interaction frequency when considered independently
(Table 1; and see low values of R
in Fig. 1). On the
other hand, six out of the seven best models included
combinations of fruit compounds with the identity of ani-
mal species and/or species abundance [see Supporting
Information]. This suggests that importance of fruit
compounds in determining fruit bird interactions should
be considered in a global context of additional ecological
In general, small birds dispersed plant species with
fruits of low digestibility and profitability (i.e. low content
of sugars and lipids and high content of fibre) more fre-
quently, whereas T. merula tended to show the opposite
pattern (Fig. 1). This result makes sense if we consider
that small disperser birds of the thermophilous scrub-
lands (E. rubecula,S. atricapilla and S. melanocephala)
are characterized by having fruit-dominated diets and
short gut-passage times (see Herrera 1984 for same
bird species). Thus, when birds with short gut-passage
times consume fruits of low digestibility, they would
need to increase the rates of fruit intake to maintain
their energy and nutrient assimilation balance (Barboza
et al. 2009), whereas birds with long gut-passage times
(T. merula in our case) would not need to increase the in-
take rate because fruit pulp remains longer in the gut
(Barboza et al. 2009).
Our results also confirm the abundance hypothesis be-
cause of the positive relationship between the frequency
of seed dispersal and the product of species abundances
and their temporal overlap, with the above-mentioned
exception of the S. melanocephala–H. excelsa interaction
(Fig. 2). According to the averaged sum of Akaike’s
Figure 1. Relationship between the content of different fruit nutrient compounds and interaction frequency in Los Adernos. The figure shows the
relationship with different avian disperser species. For each species (different lines) the fit (R
) is shown.
6AoB PLANTS &The Authors 2015
´lez-Castro et al. — Species traits, abundance and seed dispersal networks on oceanic islands
by guest on April 22, 2015 from
Figure 2. Relationship of species phenophase temporal overlap, and product of species abundance with interaction frequency in Los Adernos.
The figure shows the relationship with different avian disperser species. For each species (different lines), the fit (R
) is shown.
Table 2. Observed and simulated (mean and 95 % confident interval) values of six network parameters. Network parameters with observed
value that coincide with confidence interval of simulation are in bold.
Network parameter Observed value Mean value Lower limit Upper limit
Connectance 0.7777778 0.9170556 0.8611111 0.9444444
Nestedness (NODF) 41.6666667 30.0674603 16.6666667 48.8095238
Weighted nestedness 0.3467869 0.1300832 20.1311452 0.4306871
Interaction evenness 0.8341048 0.8738392 0.8471696 0.899341
Interaction asymmetry for birds 20.2723253 20.1723619 20.2142811 20.1448604
Interaction asymmetry for plants 0.1602762 0.148106 0.1422437 0.15625
AoB PLANTS &The Authors 2015 7
´lez-Castro et al. — Species traits, abundance and seed dispersal networks on oceanic islands
by guest on April 22, 2015 from
weights, temporal overlap and species abundance were
more important than fruit chemical compounds (Table 1).
Thus, as the product of species abundances and/or their
temporal overlap increases, the more likely it is that species
will interact with each other. According to previous studies
(e.g. Dupont et al. 2003;Va
´zquez 2005;Va
´zquez et al. 2007;
Schleuning et al. 2011), species abundance (as number of
individuals) should be sufficient by itself to explain fruit–
bird interactions. However, like for phenotypic traits, a
greater importance of species abundance (and temporal
overlap) emerged from four models where abundance-
related variables appear jointly with other phenotypic traits
[see Supporting Information].
Although the importance of temporal overlap and spe-
cies abundance was not significantly different (Table 1),
the model solely based on temporal overlap of species’ phe-
nophase fit better than the model only based on the prod-
uct of species’ abundance [see Supporting Information]
(Fig. 2). This finding is in accordance with previous studies
on mutualistic networks (Olesen et al. 2008;Gonza
Castro et al.2012b). We have to note that although we con-
sider temporal overlap as an abundance-related variable,
species’ phenophase length is, in some extent, a species-
specific trait (Olesen et al. 2008;Gonza
´lez-Castro et al.
2012b). Therefore, the importance of species’ temporal
overlap for plant– disperser interactions might be consid-
ered as an influence of phylogeny of species.
We can think of at least two explanations for the rela-
tively weak effect of species abundance on plant
disperser interactions in our community, when compared
with the effect of temporal overlap. One could be due to
the small size of this community (four animal and nine
plant species). Larger communities invariably have a
higher potential number of interactions, which makes
them more difficult to sample appropriately (Blu
2010). In contrast, in small communities sampling
community-wide interactions is more precise. For
example, comparing two closely related Mediterranean
habitats, Gonza
´lez-Castro et al. (2012a)found that the
effect of abundance on interaction asymmetry was
lower in the small-sized community than in the large
one. Another possibility is that the method used to esti-
mate species abundances affects the outcome of the
models (Va
´zquez et al.2009b). Previous studies have
used interaction frequency as a measure of abundance
(e.g. Va
´zquez et al. 2007;Schleuning et al. 2011). Thus,
there is an obvious lack of independence between the
response (interactions) and the predictor variable (abun-
dance). But, in this study, we measured species abun-
dance independently of the animalplant interaction
using captures and censuses. In this sense, our abun-
dance estimates are uncorrelated to our interaction
data, and thus more appropriate than those used by
previous studies. We suggest that future studies should
use independent estimators of species’ abundances
when trying to assess the effects of abundance on inter-
action patterns.
The first six models in our ranking included the factor
‘animal species’ [see Supporting Information]. The inclu-
sion of this factor, and/or its interaction term with fruit
compounds or abundance-related variables, generally im-
proved the equivalent models in which ‘animal species’
was excluded [see Supporting Information]. Importance
of interaction terms between fruit compounds and ‘animal
species’ reveals the relevance of some animal phenotypic
traits in determining fruit-disperser interactions, and sug-
gests that different animal species might respond to fruit
nutrients in different ways (Jordano 2000). Therefore, our
results support the assertion of Jordano (2000) that: ‘the
profitability of a given fruit should be examined in the con-
text of an interaction with a particular frugivore species’.
The model including the statistical interaction ‘animal
species ×species abundance’ slightly improves the fit of
the model with respect tothe model only based on species
abundance (AIC ¼375.124 and 376.489, respectively)
[see Supporting Information] (see also different bird re-
sponses in Fig. 2). This result is consistent with interspecific
differences in the capacity of birds to respond to changing
fruit abundances (Carnicer et al. 2009). This species-
specific response of birds to fruit abundance makes it
more difficult for abundance to determine community-
wide interactions by itself.
Although fruit-bill size overlap seems to be the most
important variable when considered independently,
both species abundance and phenotypic traits were
important in determining fruit bird interactions. The
small community size (36 interactions) constrained us
to use different subsets of explanatory variables in differ-
ent models. However, species abundance and phenotypic
traits are inseparable in a community. Therefore, the
approach we used will be useful to examine more diverse
communities, but using more realistic models (i.e. includ-
ing all explanatory variables together). Although obtain-
ing detailed phenotypic data in larger communities is
challenging, it will allow a better understanding of eco-
logical organization and coevolutionary processes shap-
ing mutualistic plantanimal communities.
Sources of Funding
A.G.-C. benefited from a JAE-PRE fellowship from the
Consejo Superior de Investigaciones Cientı
´ficas (Spain).
The National Science Foundation under a grant awarded
in 2008 funded S.Y. This work was financed by the
8AoB PLANTS &The Authors 2015
´lez-Castro et al. — Species traits, abundance and seed dispersal networks on oceanic islands
by guest on April 22, 2015 from
Spanish Ministry of Science and Education project
(CGL2007-61165/BOS) and is also framed within project
GGL2010-18759/BOS, supported by FEDER funds from
the European Union.
Contributions by the Authors
T.A.C. and A.G.-C. conceived the idea. M.N. and A.G.-C. car-
ried out the fieldwork, S.Y. and A.G.-C. analysed the data.
S.Y., M.N., T.A.C. and A.G.-C. wrote the paper.
Conflict of Interest Statement
None declared.
We thank Benito Pe
´rez, Daniel Gonza
´lez, Yurena Gavila
David Padilla and Juan Carlos Illera for their help during
fieldwork. We also thank Diego P. Va
´zquez for useful com-
ments on an early version of the manuscript. CANAGROSA
Laboratories performed chemical analyses of fruits. The
Cabildo (insular council) of Tenerife, Parque Rural de
Teno and the landowner Teobaldo Me
´ndez provided per-
mission to work at the study site. The Biology Department
and the Eberly College of Science at Penn State University
provided support to T.A.C., S.Y. and A.G.-C. during the
Spring and Summer of 2011.
Supporting Information
The following additional information is available in the
online version of this article
Table S1. Interaction frequency between plants
(columns) and animals (rows) recorded at the study site.
Figure S1. Accumulative curve of interactions recorded
at the study site against the sampling effort (mist-netting
Tabl e S 2. Candidate statistical models, ranked according
to their AIC value.
Literature Cited
Aizen MA, Sabatino M, Tylianakis JM. 2012. Specialization and rarity
predict nonrandom loss of interactions from mutualist networks.
Science 335:1486–1489.
Almeida-Neto M, Guimara
˜es P, Guimara
˜es PR, Loyola RD, Ulrich W.
2008. A consistent metric for nestedness analysis in ecological
systems: reconciling concept and measurement. Oikos 117:
Barboza PS, Parker KL, Hume ID. 2009. Integrative wildlife nutrition.
Berlin: Springer.
Bascompte J, Jordano P. 2007. Plant-animal mutualistic networks:
the architecture of biodiversity. Annual Review of Ecology, Evolu-
tion, and Systematics 38:567–593.
Bascompte J, Jordano P, Melian CJ, Olesen JM. 2003. The nested
assembly of plant-animal mutualistic networks. Proceedings of
the National Academy of Sciences of the USA 100:9383–9387.
Bascompte J, Jordano P, Olesen JM. 2006. Asymmetric coevolution-
ary networks facilitate biodiversity maintenance. Science 312:
¨thgen N. 2010. Why network analysis is often disconnected from
community ecology: a critique and an ecologist’s guide. Basic
and Applied Ecology 11:185– 195.
Burnham KP, Anderson DR. 2002. Model selection and multi-model infer-
ence: a practical information-theoretic approach.NewYork:Springer.
Burns KC. 2013. What causes size coupling in fruit– frugivore
interaction webs? Ecology 94:295– 300.
Carlo TA, Yang S. 2011. Network models of frugivory and seed
dispersal: challenges and opportunities. Acta Oecologica 37:
Carlo TA, Collazo JA, Groom MJ. 2003. Avian fruit preferences across a
Puerto Rican forested landscape: pattern consistency and impli-
cations for seed removal. Oecologia 134:119– 131.
Carnicer J, Jordano P, Melia
´n CJ. 2009. The temporal dynamics of re-
source use by frugivorous birds: a network approach. Ecology 90:
Chapman CA, Chapman LJ, Wangham R, Hunt K, Gebo D, Gardner L.
1992. Estimators of fruit abundance of tropical trees. Biotropica
Dormann CF, Fru
¨nd J, Blu
¨thgen N, Gruber B. 2009. Indices, graphs
and null models: analyzing bipartite ecological networks.The
Open Ecology Journal 2:7–24.
Dupont YL, Olesen JM. 2009. Ecological modules and roles of species
in heathland plant insect flower visitor networks. Journal of
Animal Ecology 78:346– 353.
Dupont YL, Hansen DM, Olesen JM. 2003. Structure of a plant–
flower-visitor network in the high-altitude sub-alpine desert of
Tenerife, Canary Islands. Ecography 26:301–310.
Galeano J, Pastor JM, Iriondo JM. 2009. Weighted-Interaction Nest-
edness Estimator (WINE): a new estimator to calculate over
frequency matrices. Environmental Modelling and Software 24:
1342– 1346.
´lez AMM, Dalsgaard B, Olesen JM. 2010. Centrality measures
and the importance of generalist species in pollination networks.
Ecological Complexity 7:36– 43.
´lez-Castro A, Traveset A, Nogales M. 2012a.Seeddispersal
interactions in the Mediterranean Region: contrasting patterns
between islands and mainland. Journal of Biogeography 39:
´lez-Castro A, Yang S, Nogales M, Carlo TA. 2012b. What deter-
mines the temporal changes of species degree and strength in
an oceanic island plant-disperser network? PLoS One 7:e41385.
Herrera CM. 1984. Adaptation to frugivory of Mediterranean avian
seed dispersers. Ecology 65:609– 617.
Herrera CM. 1987. Vertebrate-dispersed plants of the Iberian Penin-
sula: a study of fruit characteristics. Ecological Monographs 57:
Jordano P. 1987. Frugivory, external morphology and digestive
system in Mediterranean sylviid warblers Sylvia spp. Ibis 129:
Jordano P. 2000. Fruits and frugivory. In: Fenner M, ed. Seeds: the
ecology of regeneration in plant communities. Wallingford: CAB
International, 125–165.
AoB PLANTS &The Authors 2015 9
´lez-Castro et al. — Species traits, abundance and seed dispersal networks on oceanic islands
by guest on April 22, 2015 from
Jordano P, Bascompte J, Olesen JM. 2003. Invariant properties in
coevolutionary networks of plant-animal interactions. Ecology
Letters 6:69– 81.
Levey DJ. 1987. Seed size and fruit-handling techniques of avian
frugivores. The American Naturalist 129:471– 485.
˜es PR, Kalko EK, Jordano P, de
Aguiar MA. 2011. The modularity of seed dispersal: differences
in structure and robustness between bat– and bird fruit
networks. Oecologia 167:131–140.
Olesen JM, Bascompte J, Elberling H, Jordano P. 2008. Temporal
dynamics in a pollination network. Ecology 89:1573–1582.
Olesen JM, Bascompte J, Dupont YL, Elberling H, Rasmussen C,
Jordano P. 2011. Missing and forbidden links in mutualistic
networks. Proceedings of the Royal Society B: Biological Sciences
Pinheiro J, Bates D, DebRoy S, Sarkar D, The R Core Team. 2009. nlme:
linear and nonlinear mixed effects models. R package version
Pratt TK, Stiles EW. 1985. The influence of fruit size and structure on
composition of frugivore assemblages in New Guinea. Biotropica
R Development Core Team. 2015. R: a language and environment for
statistical computing. Vienna: R Foundation for Statistical
Rezende EL, Jordano P, Bascompte J. 2007. Effects of phenotypic
complementarity and phylogeny on the nested structure of
mutualistic networks. Oikos 116:1919 1929.
calculate the nestedness temperature of presence-absence
matrices. Journal of Biogeography 33:924–935.
´a L, Rodrı
´s MA. 2007. Linkage rules for plant-
pollinator networks: trait complementarity or exploitation
barriers? PLoS Biology 5:e31.
Schleuning M, Blu
¨thgen N, Flo
¨rchinger M, Braun J, Schaefer HM,
¨hning-Gaese K. 2011. Specialization and interaction strength
in a tropical plant-frugivore network differ among forest strata.
Ecology 92:26– 36.
Stang M, Klinkhamer PGL, van der Meijden E. 2006. Size constraints
and flower abundance determine the number of interactions in a
plant-flower visitor web. Oikos 112:111–121.
Stouffer DB, Sales-Pardo M, Sirer MI, Bascompte J. 2012. Evolutionary
conservation of species’ roles in food webs. Science 335:1489–1492.
Tylianakis JM, Tscharntke T, Lewis OT. 2007. Habitat modification
alters the structure of tropical host –parasitoid food webs. Nature
´zquez DP. 2005. Degree distribution in plant-animal mutualistic
networks: forbidden links or random interactions? Oikos 108:
´zquez DP, Melia
¨thgen N, Krasnov BR,
Poulin R. 2007. Species abundance and asymmetric interaction
strength in ecological networks. Oikos 116:1120–1127.
´zquez DP, Chacoff NP, Cagnolo L. 2009a. Evaluating multiple deter-
minants of the structure of plant-animal mutualistic networks.
Ecology 90:2039– 2046.
´zquez DP, Blu
¨thgen N, Cagnolo L, Chacoff NP. 2009b. Uniting pat-
tern and process in plant-animal mutualistic networks: a review.
Annals of Botany 103:1445 –1457.
´M, Valiente-Banuet A. 2011. The relative contribution of
abundance and phylogeny to the structure of plant facilitation
networks. Oikos 120:1351–1356.
Wheelwright NT. 1985. Fruit-size, gape width, and the diets of
fruit-eating birds. Ecology 66:808–818.
Yang S, Albert R, Carlo TA. 2014. Transience and constancy of inter-
actions in a plant-frugivore network. Ecosphere 4:147.
Zuur AF, Ieno EN, Walker NJ, Savaliev AA, Smith GM. 2009. Mixed
effects modelsand extensions in ecology with R.NewYork:Springer.
10 AoB PLANTS &The Authors 2015
´lez-Castro et al. — Species traits, abundance and seed dispersal networks on oceanic islands
by guest on April 22, 2015 from
... Examples of such characteristics are nutritional content (Gautier-Hion et al., 1985;Jordano, 2000;Cazetta et al., 2012;Blendinger et al., 2015), fruit size, seed quantity, and seed size (Wheelwright, 1985), and detectability as determined by their color (Van der Pijl, 1982;Schaefer and Schmidt, 2004;Schmidt et al., 2004;Cazetta et al., 2012;Duan et al., 2014;Ordano et al., 2017). In addition to characteristics of fruit availability, the organization of these bird-plant frugivory interactions can be affected by the relationship between a bird's bill and fruit morphology (González-Castro et al., 2015;Bender et al., 2018), the spatiotemporal overlap of species (Ramos-Robles et al., 2016), and the relative abundance of fruits and birds (González-Castro et al., 2015). These variables represent neutral-and niche-based processes that may operate in structuring frugivory networks (Machado-de-Souza et al., 2019). ...
... Examples of such characteristics are nutritional content (Gautier-Hion et al., 1985;Jordano, 2000;Cazetta et al., 2012;Blendinger et al., 2015), fruit size, seed quantity, and seed size (Wheelwright, 1985), and detectability as determined by their color (Van der Pijl, 1982;Schaefer and Schmidt, 2004;Schmidt et al., 2004;Cazetta et al., 2012;Duan et al., 2014;Ordano et al., 2017). In addition to characteristics of fruit availability, the organization of these bird-plant frugivory interactions can be affected by the relationship between a bird's bill and fruit morphology (González-Castro et al., 2015;Bender et al., 2018), the spatiotemporal overlap of species (Ramos-Robles et al., 2016), and the relative abundance of fruits and birds (González-Castro et al., 2015). These variables represent neutral-and niche-based processes that may operate in structuring frugivory networks (Machado-de-Souza et al., 2019). ...
... For months in which we actually recorded an interaction, but the phenology data did not match such pair of species, we assigned an arbitrary value of 1 × 10 −8 . This would allow us to run the analyses avoiding biases in our results (González-Castro et al., 2015;Gonzalez and Loiselle, 2016). ...
Full-text available
Frugivory interactions between birds and fruit-bearing plants are shaped by the abundance of its interacting species, their temporal overlap, the matching of their morphologies, as well as fruit and seed characteristics. Our study evaluates the role of seven factors of fruits and plants in determining the frequency of whole-fruit consumption by birds. We studied the frugivory network of a Neotropical periurban park in Xalapa, Veracruz, Mexico, and quantified relative abundance and phenology of birds and fruit, as well as fruit morphology, chromatic and achromatic contrast, and nutritional content. Using a maximum likelihood approach, we compared the observed interaction network with 62 single- and multiple-variable probabilistic models. Our network is composed of 11 plants and 17 birds involved in 81 frugivory interactions. This network is nested, modular, and relatively specialized. However, the frequency of pairwise interactions is not explained by the variables examined in our probabilistic models and found the null model has the best performance. This indicates that no single predictor or combination of them is better at explaining the observed frequency of pairwise interactions than the null model. The subsequent four top-ranking models, with ΔAIC values < 100, are single-variable ones: carbohydrate content, lipid content, chromatic contrast, and morphology. Two- and three-variable models show the poorest fit to observed data. The lack of a deterministic pattern does not support any of our predictions nor neutral- or niche-based processes shaping the observed pattern of fruit consumption in our interaction network. It may also mean that fruit consumption by birds in this periurban park is a random process. Although our study failed to find a pattern, our work exemplifies how investigations done in urban settings, poor in species and interactions, can help us understand the role of disturbance in the organization of frugivory networks and the processes governing their structure.
... width of frugivorous birds and diameter of fruits) is among the most important niche factors (Burns, 2013;Dehling et al., 2014;González-Castro et al., 2015). Although both niche and neutral processes may influence the structure of mutualistic interactions in any community, the challenge is to unravel in which circumstances one kind of factor predominates over the other, or how much of the patterns are explained by each factor. ...
... The chemical constituents of fruits may influence fruit choice by birds, being related to the degree of frugivory of birds (i.e., a proxy for the importance of fruits in the diet), but fruit chemistry is a neglected component of studies investigating trait matching among plants and their fruit eaters (Pizo et al., 2021). González-Castro et al. (2015) documented a positive effect of fruit nutrient content and interaction frequency among four frugivorous birds and their food plants in a scrubland habitat in Tenerife (Canary Islands), but Sebastián-González et al. (2017) report fruit chemistry to be unimportant in determining interaction rules in a species-rich community formed by birds, mammals, a fish, and a reptile in the Brazilian Pantanal. Thus, there is conflicting evidence about the effects of fruit nutrients on bird-plant interaction networks (but see Pizo et al., 2021). ...
... However, as other studies demonstrate (González-Castro et al., 2015;Sebastián-González et al., 2017), species abundance, though important, is not enough to predict interaction frequencies. Niche factors, particularly the mass (pulp-to-seed) ratio and the nutrient profile of fruits, here represented by lipid content, also play a measurable role. ...
Neutral and niche factors influence the structure of frugivory and seed dispersal networks. While the former refers to the abundance of interacting species, niche factors refer to traits that mediate interactions between species (e.g., morphology). The challenge is to unravel in which circumstances one kind of factor predominates over the other, or how much variation is explained by each factor. We investigated the relative contributions of abundance and trait matching (considering fruit's size, pulp mass to seed mass ratio, and lipid content, and bird's degree of frugivory, body mass, and gape width) as drivers of the frequency of interactions between frugivorous birds and fleshy fruits in an area of Cerrado second‐growth vegetation in Brazil. We expected that the abundance of species would be the most important factor due to the predominance of common small‐bodied generalist bird species and small‐seeded fleshy‐fruited plants. Species abundance was indeed an important factor explaining interactions, although in limited fashion. This is because niche factors also helped to explain bird–plant interaction frequencies in the community, particularly the mass ratio of fruits and, to a lesser extent, their lipid content through its interaction with abundance and its negative correlation with mass ratio. In addition to the importance of both bird and plant abundances, these results underscore the role played by plant functional traits in maintaining community function and ecosystem services even in habitats dominated by common generalist birds and small‐seeded plants. Abstract in Portuguese is available with online material. Fatores neutros e fatores relacionados ao nicho ecológico influenciam a estrutura das redes de frugivoria e dispersão de sementes. Enquanto os primeiros se referem à abundância das espécies que interagem entre si, os fatores relacionados ao nicho se referem a características das espécies que medeiam as interações (por exemplo, morfologia). É importante investigar em que circunstâncias um tipo de fator predomina sobre o outro, ou quanto da variação nas interações é explicada por cada fator. Nós investigamos as contribuições relativas da abundância e das características de frutos carnosos e aves frugívoras (o tamanho, a proporção de polpa, o teor de lipídios dos frutos e o grau de frugivoria, massa corporal e largura do bico das aves) como determinantes da frequência de interações entre eles em uma área de vegetação secundária de Cerrado no Brasil. Esperávamos que a abundância das espécies fosse o fator mais importante devido à predominância de espécies comuns de aves generalistas de pequeno tamanho e de plantas com sementes pequenas. A abundância foi de fato um fator importante para explicar a frequência das interações, embora de forma limitada. Isso porque fatores de nicho também ajudaram a explicar as frequências de interação ave‐planta na comunidade, particularmente a proporção de polpa dos frutos e, em menor medida, seu conteúdo lipídico por meio de sua interação com a abundância e sua correlação negativa com a proporção de polpa. Esses resultados ressaltam a importância das características funcionais das plantas para manter a função de dispersão de sementes na comunidade e os serviços do ecossistema, mesmo em habitats dominados por aves generalistas comuns e plantas com sementes pequenas, como tipicamente acontece em ambientes alterados. We investigated the relative contributions of abundance and trait‐matching as drivers of the frequency of interactions between frugivorous birds and fleshy fruits in an area of Cerrado second‐growth vegetation in Brazil to found that, in addition to the importance of both bird and plant abundances, plant functional traitsalso play a role even in habitats dominated by common generalist birds and small‐seeded plants.
... Frugivores vary in size, foraging behavior, movements and digestive physiology (Wheelwright 1985, Jordano 2000, Levey and Martínez del Rio 2001, Morales et al. 2013, González-Castro et al. 2015. Similarly, fruits vary in size, color displays, shape, and nutritional rewards, all of which influence fruit selection and seed dispersal by frugivores. ...
... The effects of frugivore abundance on the dispersal of relatively rare plant species have a quantitative effect on the seed rain, but the effects of frugivore richness are of a qualitative nature (Carlo 2005, García and Martínez 2012, González-Castro et al. 2015. A higher diversity of frugivores increases the chances of interactions with plants based on behavioral intra and interspecific facilitation mechanisms that take place among foraging frugivores (Saracco et al. 2004, Schleuning et al. 2015. ...
... Frugivore richness increased the magnitude of the equalizing effect and the antiapostatic dispersal probably through complementary effects among frugivore species (Schleuning et al. 2015) since the identity of the frugivore has an important role in determining patterns of fruit choice and the dispersal of rare seeds (González-Castro et al. 2015, Morán-López et al. 2020. Frugivore diversity affects the structure of seed dispersal networks and can increase the functional complementarity of species in a community (Petchey and Gaston 2002, Bastolla et al. 2009, Sebastián-González et al. 2015, Peña et al. 2020) even in low-specialized networks with high redundancy and low complexity as in plant-frugivore networks in tropical fragmented landscapes such as our study sites (Menke et al. 2012, Schleuning et al. 2012, Emer et al. 2018. ...
The diversity of tropical forests is strongly shaped by mutualistic interactions involving plants and frugivores that disperse their seeds. However, it is little known how decreases in the diversity of frugivores can affect seed dispersal patterns, plant community composition and species' coexistence in tropical forest landscapes. Here, we investigated the effects of bird frugivore diversity on seed dispersal of rare plant species and on the magnitude of equalizing effects on the seed rain in open areas within 12 fragmented landscapes in the Brazilian Atlantic Forest. We monitored the production of bird‐dispersed seeds and bird abundance in forest fragments, and sampled the seed rain and the activity of birds attracted to experimental tree nuclei established in neighboring pastures. The activity of frugivores in tree nuclei was positively correlated with the diversity of birds recorded in nearby forest fragments, and the seed rain diversity increased with frugivore activity. The proportion of seeds dispersed more frequently than expected by chance in tree nuclei increased linearly with the species' richness of birds. The richness and abundance of active frugivores in deforested areas promoted a seed rain with evenness and diversity up to five times greater than the seed pool available in forest fragments due to the proportional increase in the dispersal of rare plant species and a concomitant proportional decrease in the dispersal of dominant fruiting plants. Furthermore, every additional bird species detected in a site was associated with a 10% increase in the equalizing effect on dispersed seeds' relative abundance. Our results show that the aggregated behavior of avian frugivore communities on deforested areas results in higher species richness in the seed rain of plant communities and underscore the urgency to reduce bird species' loss and the simplification of their communities in tropical landscapes.
... But a small bird species that is a pulp-robber of a large-seeded fruit can, at the same time, be a highly effective disperser of plants with small-seeded fruits (Carlo et al. 2003). Contingencies like this illustrate the importance of trait variance and matching between the phenotypes of birds and fruits (González-Castro et al. 2015;Machado-de-Souza et al. 2019;González-Varo et al. 2022). In fact, morphological, behavioral, and physiological traits, as well as the relative abundance of bird species, can be more important than overall diet composition in determining the role of bird species as seed dispersers (e.g., Carlo et al. 2003). ...
... The traits of birds and fruits set boundaries to frugivory interactions, and how often and where such interactions occur (González-Castro et al. 2015). For birds, important traits include body morphology and size, while relevant plant traits include fruit and seed size, the architecture of infructescence, nutrients, secondary metabolites, and phenology. ...
The mutualistic interactions between frugivorous birds and their food plants are a staple of Neotropical ecosystems. Here, we review the major concepts and principles that govern interactions between Neotropical birds and plants. Morphological traits of birds such as beak size, morphology, and gut physiology interact with fruit traits such as the size of seeds and fruits, and the nutritional and chemical properties of fruit pulp to determine the identity of the dominant community-wide frugivory interactions that take place throughout Neotropical ecosystems. We discuss tradeoffs that occur between fruit handling capacity and seed dispersal efficiency of birds, as well as tradeoffs between the digestion of complex vs. simple nutrient classes, and the overall dietary composition and foraging strategies of birds and the implications to community processes. We also comment on the importance of Neotropical avian frugivores in shaping dynamics of forest regeneration and on the conservation status of the major families of Neotropical frugivores and conclude by identifying promising venues for future research in the field of frugivory and seed dispersal.
... finding and interacting with plants ( González-Castro et al. 2015;Vizentin-Bugoni et al. 2021), we formulated two contrasting expectations: (i) we expected a modular and non-nested network if traits of individuals and their spatial contexts vary, forming discrete subsets within the population that could attract distinct subsets of seed dispersers or; (ii) we expected a nested and non-modular network if traits and spatial contexts of individuals vary continuously forming a gradient, in which less connected and more specialized individuals would interact with a subset of bird species that interact with the more connected and less specialized individuals. ...
... This pattern is consistent with other studies on intraspecific variation of animal-dispersed plants (Crestani et al. 2019;Miguel et al. 2018). The emergence of such modules may occur due to differences in habitat use among seed dispersers and variation in their relative abundance, both of which are known to influence the probability of encounter of mutualistic partners and shape plant-animal networks (Vázquez et al. 2007;Krishna et al. 2008;González-Castro et al. 2015). Habitat specialist dispersers may have been important for the emergence of modules, because they only interact with individual plants located within their preferential habitats. ...
Full-text available
While network analyses have stimulated a renewed interest in understanding patterns and drivers of specialization within communities, few studies have explored specialization within populations. Thus, in plant populations, causes and consequences of individual variation in their interactions with mutualistic animals remain poorly understood. Studying a Brazilian pepper (Schinus terebinthifolia) population, we measured the extent of individual variation in interactions with seed dispersers and tested whether connectivity (number of seed dispersers) and specialization (exclusiveness of partners) are associated with phenotypic and phenological traits of individuals and their spatial context. We found that: (i) individuals varied broadly in their connectivity and specialization on seed dispersers; (ii) phenotypic traits and spatial context matter more than fruiting duration in determining how many and how exclusive are seed dispersers of an individual; (iii) the individual-based network was nested and indicated that the less connected individuals were shorter, occurred in neighborhoods with fewer fruits, and tended to interact with a subset of the partners of more generalist individuals which, in turn, were taller and inserted in higher fruit density neighborhoods; (iv) modularity indicated the existence of subsets of individuals that interacted disproportionately with distinct groups of partners, which may occur due to differences in bird habitat use across the landscape. Our study underlines a remarkable interindividual variation that is overlooked when interactions are compiled to describe species-level interactions. Traits and spatial contexts that define variation among individuals may have important implications not only for fitness but also for sampling and description of interactions at species level.
... A VIF value > 10 is generally considered to indicate multicollinearity (Queen et al., 2002). In the case of bird morphometric variables, those related to the bill (depth, width, and length) had VIFs >10; we therefore retained bill width, as this trait can indicate whether a seed can be swallowed and dispersed (González-Castro et al., 2015). The same procedure was carried out for the fruit variables, which resulted in the elimination of equatorial diameter. ...
... These results also suggest the importance of bird morphological traits in the establishment of mutualistic relationships with plants exhibiting an ornithocorous dispersal syndrome. Relationships between specific fruits and frugivorous birds based on size traits are usually more evident in phylogenetically distant groups (e.g., González-Castro et al., 2015;Lord, 2004), so morphological variations may be less noticeable in more closely-related groups (Rezende et al., 2007). For example, frugivorous bird species with body mass differences exceeding 100 g interact with groups of plants with greater variation in their fruit sizes (Buitrón-Jurado and Sanz, 2020;García, 2016;Rezende et al., 2007). ...
Fruit size and shape are key traits that affect the interactions between plants and their seed dispersers within communities. Yet, the co-adaptive morphological links between fruits and bird seed dispersers within highly biodiverse Neotropical dry forests are still poorly understood. In this study, we evaluated the relationship between the morphometric fruit traits of Bursera, a group of dominant trees in tropical dry forests, and the morphometric body traits of flycatcher birds of the Tyrannidae family, their main seed dispersers. We found that the largest bird species primarily consumed large- and medium-sized fruits, while smaller birds targeted plants with small- and medium-sized fruits. Migratory flycatchers were the core species involved in Bursera fruit removal, except in Bursera species fruiting during the rainy season, when these birds are not present. Also, small-sized fruits with an oblong shape tended to be removed more frequently by small-sized birds, while large-oblong fruits tended to be more consumed by larger birds. According to body traits, bill width, tarsus, and wing length had significant effects on fruit removal. Wider-billed Tyrannus, Myiarchus, and Myiodynastes species removed almost half of the fruits. The Myiarchus genus showed the longest tarsus and removed up to 58% of all fruits. The Tyrannus species, which rely on flight to obtain food, showed the greatest wing length. Our findings suggest that morphological traits of Bursera fruits such as shape and size may be fitted to interact with the body traits of their bird dispersers. Studies on seasonally dynamic bird-plant interactions are crucial for a better understanding of the ecological and evolutionary drivers in Neotropical dry forests.
... Nevertheless, we acknowledge that our trait-based simulation approach has limitations. First, the simulations of plantfrugivore interactions could be refined further by accounting for species abundances along the elevational gradient and by considering additional plant traits, such as the nutritional content of fruits(González-Castro et al., 2015). Furthermore, the ongoing debate on mechanisms and possible niche shapes describing the probability of seed-dispersal interactions(Burns, 2013;González-Varo, Onrubia, et al., 2021) leaves room for future improvements of the traitmatching model, although previous work has shown that different types of niche shapes can yield similar findings in simulation studies of seed-dispersal distances ...
Full-text available
Climate change causes shifts in species ranges globally. Terrestrial plant species often lag behind temperature shifts, and it is unclear to what extent animal‐dispersed plants can track climate change. Here, we estimate the ability of bird‐dispersed plant species to track future temperature change on a tropical mountain. Tropical elevational gradient (500–3500 m.a.s.l.) in the Manú biosphere reserve, Peru. From 1960–1990 to 2061–2080. Fleshy‐fruited plants and avian frugivores. Using simulations based on the functional traits of avian frugivores and fruiting plants, we quantified the number of long‐distance dispersal (LDD) events that woody plant species would require to track projected temperature shifts on a tropical mountain by the year 2070 under different greenhouse gas emission scenarios [representative concentration pathway (RCP) 2.6, 4.5 and 8.5]. We applied this approach to 343 bird‐dispersed woody plant species. Our simulations revealed that bird‐dispersed plants differed in their climate‐tracking ability, with large‐fruited and canopy plants exhibiting a higher climate‐tracking ability. Our simulations also suggested that even under scenarios of strong and intermediate mitigation of greenhouse gas emissions (RCP 2.6 and 4.5), sufficient upslope dispersal would require several LDD events by 2070, which is unlikely for the majority of woody plant species. Furthermore, the ability of plant species to track future changes in temperature increased in simulations with a low degree of trait matching between plants and birds, suggesting that plants in generalized seed‐dispersal systems might be more resilient to climate change. Our study illustrates how the functional traits of plants and animals can inform predictive models of species dispersal and range shifts under climate change and suggests that the biodiversity of tropical mountain ecosystems is highly vulnerable to future warming. The increasing availability of functional trait data for plants and animals globally will allow parameterization of similar models for many other seed‐dispersal systems.
... Therefore, fruit traits are subjected to an array of evolutionary forces imposed by mutualistic and agonistic interactions (Jordano, 1995;Cipollini and Levey, 1997;Mack, 2000). Fruit trait combinations that have emerged from complex evolutionary pathways currently affect the ability of seed dispersers to interact with them (González-Castro et al., 2015;Blendinger et al., 2016;Dehling et al., 2016). In addition, foraging preferences of dispersers according to their handling skills or digestive capabilities can lead to a differential use of fruit trait combinations (Valenta and Nevo, 2020;Rojas et al., 2021). ...
Background and Aims Fruit traits and their interrelations can affect foraging choices by frugivores, and hence, the probability of mutualistic interactions. Certain combinations of fruit traits that determine the interaction with specific seed dispersers are known as dispersal syndromes. The dispersal syndrome hypothesis (DSH) states that seed dispersers influence the combination of fruit traits found in fruits. Therefore, fruit traits can predict the type of dispersers with which plant species interact. Here, we analysed whether fruit traits’ relationships can be explained by DSH. To do so, we estimated the interrelation between morphological, chemical and display groups of fruit traits. In addition, we tested the importance of each trait-group defining seed dispersal syndromes. Methods Using phylogenetically corrected fruit traits’ data and fruit-seed disperser networks, we tested the relationships among morphological, chemical and display fruit traits with Pearson’s correlations and phenotypic integration indices. Then, we used perMANOVA to test if the fruit traits involved in the analysis supported seed dispersers’ functional types. Key results Morphological traits showed strong intra-group relationships, contrasting to chemical and display traits whose intra-group trait relationships were weak or null. Accordingly, only the morphological group of traits supported three broad seed disperser functional types (birds, terrestrial mammals and bats), consistently with the DSH. Conclusions Altogether, our results give some support to the DSH. Here, the three groups of traits interacted in different ways with seed dispersers’ biology. Broad functional types of seed dispersers would adjust fruit consumption to anatomical limitations imposed by fruit morphology. Once this anatomic filter is surpassed, seed dispersers use almost all the range of variation in chemical and display fruit traits. This suggests that the effect of seed dispersers on fruit traits is modulated by hierarchical decisions. First, morphological constraints define which interactions can actually occur; subsequently, display and composition determine fruit preferences.
... Another factor that may have caused functional clustering in addition to fire and flooding may have been a restricted pool of dispersers due to flooding, for example reducing zoochory in which dispersal by birds is important. The dispersal and frugivory network has a strong relationship with plant traits so that changes in this interaction can restrict the number and identity of the dispersing agents (González-Castro et al., 2015). It is also possible that the characters chosen here were insufficient to detect patterns of environmental filtering in typical cerrado. ...
In central Brazil, there are strong gradients and discontinuities in vegetation structure and composition between the forests of southern Amazonia and the open savannas of South America's Cerrado. These transitions are often controlled by disturbance processes, and the ability of vegetation to respond to climatic and environmental changes may depend on the regeneration traits of the different floras present. In this study we aim to assess the regeneration traits of tree communities of the Amazon-Cerrado transition and to understand how they differ among and within the markedly different vegetation types. We sampled 39 one-hectare long-term monitoring plots that include typical cerrado (TC = 10), cerradão (CD = 3), gallery forests (GF = 3), floodplain forests (FF = 6), seasonal and open rainforests (SF = 17). The regeneration traits assessed included dispersal syndrome (zoochory, anemochory, and autochory), fruit consistency (dry and fleshy), number of seeds per fruit, and diaspore dimensions (width and length). We found differences among the vegetation types, in all regeneration traits. These tended to be aggregate by vegetation structure, being similar for cerrado and cerradão species, and similar for SF, FF and GF (more forested vegetation). Vegetation types did not differ in functional diversity, however, while regeneration traits among seasonal and open rainforests were well-dispersed, in floodplain forests they were more clustered. Tree species depend substantially on fauna (zoochoric species between 42 and 86% in vegetation) for the dispersal across all habitats. By considering regeneration traits in the study of tree recruitment and establishment, we will increase our understanding about the dynamics of tree communities in neotropical forests and savannas.
... Structural properties of interaction networks, such as nestedness and modularity, are influenced by species abundance, species traits and resource properties which has been studied in various types of interactions (Ballarin et al., 2019;Calatayud et al., 2017;Chamberlain & Holland, 2009;Donatti et al., 2011;Gonz alez-Castro et al., 2015;Sazatornil et al., 2016;Sfair et al., 2018;Stang et al., 2006;Tur et al., 2015;Vizentin-Bugoni et al., 2016). However, the factors underlying the structural properties of the relationship between hermit crabs and gastropod-shells that was recently evaluated through the network approach remain unvisited (see Rodrigues et al., 2020 andda Silva et al., 2020). ...
The relationship between hermit crabs (Paguroidea) and gastropod shells has attracted the attention of researchers toward a network approach. These networks seem to present recurrent patterns that are often modular, i.e., a pattern in which the network can be divided into compartments that represent species interacting more with species belonging to its own compartment than with species composing another compartment. A modular interaction network occurs when species tend to interact more intensively within subsets of species (modules) than with species outside of it. Since modularity is a characteristic in which different hermit crabs interact with different subsets of shells, we proposed that these patterns should be a reflection of different morphometric traits, as well as that species specialization (measure which indicates if a given species is specialist or generalist in resource use) could be related to such traits. Three different hermit crab-shell networks in which the modularity and species specialization have already been determined were chosen. The animals were sampled in three different regions with the same sampling effort in the same type of substrate. After sampling, the animals were taken to the laboratory where they were identified and measured. A PERMANOVA was used for the hermit crab morphometric traits from each region using each compartment as grouping variable. In order to test if morphometric traits influence species specialization (d 0), linear mixed models were created and selected through Akaike's Information Criterion. Our data show that hermit crabs presented different morpho-metric traits in each module they occupied. Also, d 0 was influenced by morphomet-ric traits; thus, hermit crabs with different sizes need different types of shells, which reflect in different specialization levels.
Full-text available
Plant-animal mutualistic interactions such as frugivory and seed dispersal display great variation in time due to fluctuations in fruit abundance, animal abundance, and behavior. In particular, some species participate in interactions with other species only transiently, while other species are active for longer periods of time. Species with a longer period of activity are able to interact with more species, and thus engage in constant participation in an interaction network. Species with high constancy would thus be expected to help maintain the biodiversity of a community; however, the manner in which constant species link to their partners may be critical to species coexistence. Because species that interact with many partners concurrently could create more competition compared to those species that interact sequentially with many partners, evaluating the concurrence in an interaction network sheds light on how the network can maintain biodiversity. In this study, we investigate how phenological patterns of fruit production and frugivore presence affect the temporal variation of a plant-frugivore network, and focus on the manner in which high degree species collect their interactions over time. We found a clear separation of activity periods: most species appeared only briefly and participated in relatively few interactions, or showed activity for longer time periods and participated in more interactions. Species that were active for longer time periods often shifted interactions, resulting in a sequential collection of their partners in time, rather than concurrence. For the seed dispersal mutualism in particular, sequential accumulation of partners may allow plant species more opportunities to disperse their seeds compared to concurrence. We suggest that for temporally and spatially heterogeneous landscapes, sequential accumulation of partners would serve to reduce competition and facilitate coexistence of species.
Full-text available
In a mid-elevational rainforest in New Guinea, size and structure of fruits influenced feeding visits by various species of frugivorous birds: large (diam. > 12 mm) structurally unprotected fruits were taken mostly by fruit-pigeons and bowerbirds, structurally-protected fruits were taken mostly by birds of paradise, and small, structurally-unprotected fruits were taken by nearly all species. Fruit-pigeons as compared with birds of paradise have short, thin bills; small feet; and fewer behaviors for reaching and handling fruits. Birds of paradise use complex food-handling techniques for removing fruits in capsules and other protective structures that are similar to techniques used to capture insects hidden in bark and dried foliage. Absence of large structurally-unprotected fruits in the diets of birds of paradise is not explained entirely as an effect of fruit size. Specialization for frugivory in the entirely frugivorous fruit-pigeons has resulted in exploitation of specific fruit types (large and small structurally-unprotected fruits) rather than all fruit types. Most fruiting plants were visited by a subset of frugivores, as influenced by fruit size and structure. Attracting specific feeding assemblages might be regarded as an adaptation of the plant for enhancing particular patterns of seed dispersal. However, other selective pressures also operate on the evolution of fruit size and structure.
Nutrition spans a wide range of mechanisms from acquisition of food to digestion, absorption and retention of energy substrates, water and other nutrients. Nutritional principles have been applied to improving individual health, athletic performance and longevity of humans and of their companion animals, and to maximizing agricultural efficiency by manipulating reproduction or growth of tissues such as muscle, hair or milk in livestock. Comparative nutrition borrows from these tra- tional approaches by applying similar techniques to studies of ecology and physiology of wildlife. Comparative approaches to nutrition integrate several levels of organization because the acquisition and flow of energy and nutrients connect individuals to populations, populations to communities, and communities to ecosystems. Integrative Wildlife Nutrition connects behavioral, morphological and biochemical traits of animals to the life history of species and thus the dynamics of populations. An integrated approach to nutrition provides a practical framework for understanding the interactions between food resources and wildlife popu- tions and for managing the harvest of abundant species and the conservation of threatened populations. This book is for students and professionals in animal physiology and ecology, conservation biology and wildlife management. It is based on our lectures, dem- strations and practical classes taught in the USA, Canada and Australia over the last three decades. Instructors can use Integrative Wildlife Nutrition as a text in wildlife and conservation biology programs, and as a reference source for related courses in wildlife ecology.
IN 41 bird species of mediterranean scrublands of S Spain, seed dispersers (feeding on whole fruits and voiding seeds unharmed) are indistinguishable from nonfrugivores and fruit predators (feeding on pulp or seeds and not performing dispersal) in the ratio of gizzard mass, liver mass, and intestine length to body mass, but differ significantly in bill morphology and average gut passage time (GPT). Seed dispersers tend to be flatter and broader billed than other groups, and have a wider mouth relative to bill width. GPT of seed dispersers is signficantly shorter than that of other groups. Morphological correlates of seed dispersers suggest that adaptations for insectivory serve as preadaptations for feeding on whole fruits. Shorter GPT's, in contrast, appear to be an adaptation indispensable to sustained, intense frugivory. seasonal frugivory imposes limits on commitments to permanent structural modifications, and more subtle functional adaptations of seed dispersers to plants are as relevant as the more conspicuous structural adaptations reported for year-round frugivores.-from Author
Limitations of linear regression applied on ecological data. - Things are not always linear additive modelling. - Dealing with hetergeneity. - Mixed modelling for nested data. - Violation of independence - temporal data. - Violation of independence spatial data. - Generalised linear modelling and generalised additive modelling. - Generalised estimation equations. - GLMM and GAMM. - Estimating trends for Antarctic birds in relation to climate change. - Large-scale impacts of land-use change in a Scottish farming catchment. - Negative binomial GAM and GAMM to analyse amphibian road killings. - Additive mixed modelling applied on deep-sea plagic bioluminescent organisms. - Additive mixed modelling applied on phyoplankton time series data. - Mixed modelling applied on American Fouldbrood affecting honey bees larvae. - Three-way nested data for age determination techniques applied to small cetaceans. - GLMM applied on the spatial distribution of koalas in a fragmented landscape. - GEE and GLMM applied on binomial Badger activity data.