Fred Correção
Fred Correção
Fred Correção
Beirão1, Humberto Brant1, Flávio Castro1, Wesley Dáttilo1, Roger Guevara1, G. Wilson
Fernandes1
1
Departamento de Biologia Geral, Universidade Federal de Minas Gerais, Belo
Horizonte, Brazil.
1
Graduate School in Ecology, Conservation and Wildlife Management, Federal
USA.
5
Department of Ecology, University of São Paulo, São Paulo, Brazil.
Abstract
Keywords:
Mountains are important systems for the maintenance of endemic species and
ecological and evolutionary processes (Graham et al. 2014, Noroozi et al. 2018, Silveira
et al. 2018, Rangel et al. 2018). Studies in mountains generally find two patterns of
2016, Longino and Branstetter 2018). Both patterns generate a high spatial β-diversity
among communities (Fernandes et al. 2016, Mota et al. 2018, Peillo et al. 2018), forming
of potentially interacting species that are interconnected by dispersal (Leibold et al. 2004;
Holyoak et al. 2005). There are often large differences in habitat use and movements and
behavior, body size or trophic level (Chase and Bengtsson livro). This variation in
dispersal rates can influence metacommunity structure and the partitioning of diversity.
spatiotemporal scales (Anderson et al. 2011, Silva & Hernández 2018) and it may reflect
two different processes:, species turnover or nestedness among sites (Baselga 2007,
2010). Differences in characteristics among taxa and the use of habtitathabitat use can
the turnover of species species turnover among elevations is generally the main process
that shapes the β-diversity (Mota et al. 2018, Nunes et al. 2016,a,b, Perillo et al. 2017).
and have high rates of endemism (Fujita et al. 2014, Steinbauer et al. 2016, Noroozi et al.
2018).
Tropical mountain ranges in South America are fundamental as cradles
andbirthplace and museums of biodiversity (Rangel et al. 2018). Among them, Espinhaço
Range, with about 1200 km of extension, stands out for its diversity and endemism
(Silveira et al. 2016, 2018). In the highest regions of the Espinhaço RangeChain of the
Epinhaço, above 1000 m. a.s.l. are generally found two main habitats are generally found:,
forest islands, known as capões de mata (Fernandes 2016, Silveira et al. 2018). In these
systems, we will find species that we can be classified classify as habitat use either as
two distinguishable habitats (Chazdon et al. 2011). The distribution of these habitats in
the landscape and the dispersal rates of taxa can determine the distribution of species in
the metacommunity and the processes that structure the β-diversity and to the nestedness
of the metacommunity.
processes responsible for nestedness patterns may be important for conservation planning
because it can help to highlight which factors have the greatest influence on species
et al., 2002). Nevertheless, the underlying causal processes linked to the nestedness
patterns in metacommunities are still poorly understood (Fischer and Lindenmayer 2005,
mais permeável, como o campo rupestre, deve contribuir mais para o aninhamento da
taxonomic group.
Scarabaeidae: Scarabaeinae). These insects are highly abundant, diverse and have a wide
geographical distribution (Davis et al. 2002, Gibb et al. 2017) and fulfil a suite of play
many important ecological roles functions (Folgarait 1998, Nichols et al. 2008, Nunes et
al. 2018). These taxa are dominant components of terrestrial biomes worldwide, are
highly sensitive to habitat gradients and have been used as bioindicators of environmental
conditions (Bonebrake et al. 2010, Nichols et al. 2007, Parr 2010, Spector 2006). In
addition also, a multi-taxa responses study as it can provide valuable information for the
development of effective conservation strategies in tropical mountains systems. Commented [DFV1]: Não entendi muito bem essa parte,
seria bom ver o texto original em português para poder
fazer uma correção mais precisa
Methods
Study area
The study was performed in the southern portion of the Espinhaço mountain range, called
Serra do Cipó, in Minas Gerais, southeast Brazil (19º14’19”S, 43º31’35”W; Fig. 1). The
Espinhaço formation extends almost continuously along 1200 km from northeast to
southeast Brazil, crossing two states (Bahia and Minas Gerais) and acting as a barrier
separating two Brazilian biomes, the Cerrado (Brazilian savanna) in the west and the
Atlantic Forest in the east. This mountain chain hosts high species richness, several
threatened and endemic species (Echternacht et al. 2011; Fernandes 2016; Silveira et al.
2016) in approximately 1% of the Brazilian territory (Bitencourt and Rapini 2013). Above
900 m a.s.l., there is the predominance of a grassland matrix called campo rupestre. The
(Hopper et al. 2016; Silveira et al. 2016). Immersed in the matrix of campo rupestre, there
are the Capões de Mata (hereafter called forest islands), which are natural patches of
forest islands emerging in mountaintops. These forest islands occur mainly above 1200
with the Atlantic Forest domain (Coelho et al. 2018). The occurrence of forest islands is
restricted to erosion valleys without boulders, generally associated with headwaters and
small streams (Coelho et al. 2018). Average annual rainfall ranged between 1300 and
Sampling design
The study was performed in the two main systems found above 1000 m a.s.l. in
the Serra do Cipó, the campo rupestre and forests islands (Figure 1). We sampled 26 sites,
totaling an area of 66.149 km2, of which 11 sites of forests islands and 15 of campo
rupestre, and all sites are at least 200 m apart. Samplings occurred in two seasons in each
system, between 2013 and 2014, in the beginning and ending of rainy and dry seasons.
region of each forest island a set of 20 x 50 m plot, where five pitfall traps were installed,
one in the center of the plot and four at its vertices. Along sites of campo rupestre we
used transects separated by at least 250 m apart, each consisting of five traps separated
by 50 m (25 traps per elevation). Iin each site we placed five soil pitfalls separated by 50
m. Pitfalls were 9 cm deep and 15 cm in diameter, contained 250 ml of a salt and detergent
solution and each trap was left in the field for 48 hours. To quantify the fruit-feeding
butterflies metacommunity we used five Van-Someren Rydon traps baited with fermented
banana with sugar cane juice. We sampled the fruit-feeding butterflies per for five
consecutive days in each season. Finally, to quantify the dung beetle metacommunity we
used in the central region of each forest island a set of 20 x 50 m plot where four pitfall
traps were installed at its vertices to collect dung beetles. Along sites of campo rupestre
sites we used transects separated by at least 250 m, each consisting of three traps separated
by 100 m (nine traps per elevation). Iin each site we placed five soil pitfalls separated
from each other by 50 m. Pitfalls were 9 cm deep and 15 cm in diameter, contained 250
ml of a salt and detergent solution, and were baited with 25 g of fresh human feces. Each
All taxa were collected, preserved and transported to the laboratory where all
captured individuals were identified to the lowest taxonomical level possible. We used
identification guides (Baccaro et al. 2015, BUTTERFLY, Vaz-de-Mello et al. 2011) and
consultation ofconsulted taxonomic experts. Dr. Rodrigo Feitosa for ants, Dr, Andre? for
buterflies and Dr. Fernando Vaz-de-Mello for dung beetles. All sampled material is
institutions where the specialists of each taxa are associated. Ants were deposited kept in
the Universidade Federal do Paraná, butterfly were deposited kept in the XXX and dung
beetles were deposited kept in the Zoological Collection of Universidade Federal de Mato
Grosso. All necessary sampling permits were issued by Instituto Chico Mendes de
Data analyses
First, the species were classified based on their habitat use using the multinomial
classification models proposed by Chazdon et al. (2011). This analysis is based on the
relative abundance of individuals per taxa (ants, butterflies and dung beetles), minimizing
bias due to differences in sampling intensities between habitats and insufficient sampling
within each habitat (Chazdon et al. 2011). Using this analysis per taxa, these models
classifyies species into four groups: campo rupestre specialists (CRS), forest island
specialists (FIS), generalists and too rare to classify. We used a cutoff point of K = 0.5
for a simple majority rule or liberal threshold and a conservative P value of 0.005. We
used a conservative approach because our goal was to analyze the whole community
(Chazdon et al. 2011). We carried out the CLAM analysis using vegan v 2.4-0 (Oksanen
et al. 2015).
β-diversity in two ways: between communities using a taxonomic approach, including all
taxonomic groups and per taxa (ants, buttlerflies and dung beetles) and a second analysis
using a habitat use approach (CRS, FIS, generalists and too rare). In all cases,
nestedness to β-diversity using the Sørensen (βSOR) and Simpson (βSIM) indices (Baselga
partitioned total β-diversity (βSOR) into the component derived from species turnover
(βSIM) and the component derived from species gain and loss or nestedness (βnest). We
performed analysies using taxonomic approach and habitat use approach using the
coefficient) (Baselga & Orme 2012). For each data set, we also calculated pairwise
dissimilarity measures using the function beta.pair from the same R package.
the analysis in two steps. First, we tested if metacommunities of all species and per
taxonomic groups (ants, butterflies and dung beetles) were significantly nested with
respect to habitats (campo rupestre and forests islands) and to species. We estimated
nestedness using the NODF metric (Almeida-Neto et al. 2008) in the software program
ANINHADO (Guimarães and Guimarães 2006). NODF values for the whole
these values. In this case, we computed 1000 simulated networks generated by the Null
Model II (Bascompte et al. 2003). In this null model, the probability of occurrence of a
species is proportional to the number of species for of both sites and insect species. We
or decrease the overall network nestedness, and compared this to our random
contributions of habitats (campos rupestre or forest islands) using all insects or separately
by taxonomic group (ants, butterflies or dung beetles) and after that we explored the
relative contributions of species classification per habitat use (CRS, FIS, generalists or
too rare) to the nested pattern of the network. To test the predictions that the campo
and in each of the three taxonomic groups as response variables and the habitat (campo
rupestre or forest island) as explanatory factor using generalized linear models (GLMs).
contribution to nestedness as response variable and the group taxonomic (ants, buttlerflies
and dung beetles), species classification (CRS, FIS, generalists or too rare) and the
Gaussian distribution, and we checked for the distribution of errors and over-dispersion
in the data. The minimal model was defined as the removal of non-significant variables
(p > 0.05) and when significant differences were observed in variables more than two
analysis to determine differences among levels (Crawley 2013). For residual and contrast
analyses analysis we used the RT4Bio package in R (Reis Jr. et al. 2013). All analyses
Results
We recorded 211 insect species/morphospecies in this landscape, involving 128
ants, 33 butterflies and 50 dung beetles (Table 1). In this metacommunity CLAM
forest island specialists, seven species (3.3 %) as generalists of habitats and 164 species
(77.7 %) were indicated as too rare to classify (Table 1). Of the ant species recorded,
six as generalists of habitats and 101 species were indicated as too rare to classify (Table
1). By the butterflies, CLAM classified only one specie as campo rupestre specialist, one
specie as forest island specialist, 31 species were indicated as too rare to classify and no
species was classified as habitat generalist (Table 1). Finally, for dung beetles CLAM
classified nine species as campo rupestre specialists, eight as forest island specialists,
only one specie as generalist of habitats and 32 species were indicated as too rare to
a taxonomic or a habitat use approach (Table 2). Turnover was the main component of β-
diversity, representing more than 90%, including all taxonomic groups or per taxa (Fig.
2a). When analyzing the metacommunity after habitat use classification, the β-diversity
We found that the metacommunity network was significantly nested, using all
species and per taxonomic groups (p < 0.05, Table 3), indicating that the species
the GLMs was possible to verify that there is no difference in contribution between
habitats to nestedness pattern (p> 0.05) when comparing campo rupestre and forest
islands, using all species and per taxonomic groups (Table 4). To habitat use classification
we found that species generalists that most contributed to metacommunity nested pattern
(p < 0.05, Table 4). We also verified that the taxonomic group does not differ in relation
to contributed to nested (p > 0.05, Fig. 3). Although the number of generalist species
represents only 3.3% of the total, we confirm the prediction that these species contributed
more to nestedenns pattern, when compared to habitat specialists or too rare species. We
also verified that of forest islands specialists (FIS) contributed more to nested pattern,