Journal of Insect Conservation (2020) 24:17–34

https://doi.org/10.1007/s10841-019-00194-2

ORIGINAL PAPER

Patterns of diversity in a metacommunity of bees and wasps

of relictual mountainous forest fragments
Lucas Neves Perillo1,3 · Newton Pimentel de Ulhôa Barbosa1 · Ricardo R. C. Solar2 ·
Frederico de Siqueira Neves1,2

Received: 31 October 2018 / Accepted: 14 November 2019 / Published online: 20 November 2019
© Springer Nature Switzerland AG 2019

Abstract
Naturally fragmented landscapes provide a suitable opportunity for investigating species dynamics under the sole influence
of habitat fragmentation. Various approaches have used landscape attributes (e.g., patch size and connectivity) to explain
patterns of species diversity and composition. We evaluated the influence that patch and landscape attributes have on bee
and wasp (Hymenoptera: Aculeata) diversity (species richness and abundance) using a natural forest archipelago located in
a campo rupestre (rupestrian grassland) matrix. We also assessed the effects of season on species composition (temporal
β-diversity). Our analysis found higher richness and abundance of bees and wasps in the summer. There was a significant
change in species composition between seasons (species replacement accounts for 88% of β-diversity), with winter com-
munities not representing subsets of summer communities. Evaluation of the relationships between bee and wasp diversity
and landscape attributes (e.g., patch size and isolation, distance between patches and continuous forest distance), only found
a significant relationship for temporal β-diversity, which increases with distance from continuous forest. We propose that
forest islands can be considered transient environments for many species, and so this naturally fragmented metacommunity
depends on continuous forest propagules and its temporal dynamics to sustain diversity among forest islands.

Keywords Spatio-temporal patterns · Atlantic forest · Community structure · Landscape structure · Fragmentation · Campo
rupestre

Introduction proposed several theoretical approaches to understanding the

The contemporary landscape of the Earth is covered by
a mosaic of natural and human-modified environments.
Since the development of the Theory of Island Biogeog-
raphy (MacArthur and Wilson 1967), ecologists have
Electronic supplementary material The online version of this
article (https​://doi.org/10.1007/s1084​1-019-00194​-2) contains
supplementary material, which is available to authorized users.
* Lucas Neves Perillo
lucasnevesperillo@gmail.com.br
1
Laboratório de Ecologia de Insetos, Instituto de
Ciências Biológicas, Universidade Federal de Minas
Gerais, Av. Antônio Carlos, 6627, Belo Horizonte,
Minas Gerais 31270‑901, Brazil
2
Departamento de Genética, Ecologia e Evolução, Instituto de
Ciências Biológicas, Universidade Federal de Minas Gerais,
Belo Horizonte, Brazil
3 of fauna (Cook et al. 2004; Driscoll 2005; Yekwayo et al.
Bocaina Biologia da Conservação, Belo Horizonte, Brazil
dynamics of biodiversity of terrestrial fragmented habitats
(Mendenhall et al. 2014; Patiño et al. 2017). Metapopulation
(Levins 1969) and metacommunity theories (Leibold et al.
2004) brought a spatiotemporal perspective to the subject
(Biswas and Wagner 2012). A comparative meta-analysis of
bird density found patch area to have a significant positive
effect with true oceanic islands, and not for habitat frag-
ments (Brotons et al. 2003), indicating a need for separate
biogeographical theories for true islands and landscape eco-
systems (Mendenhall et al. 2014).
Naturally-fragmented terrestrial landscapes are excep-
tionally valuable for analyzing the potential impacts of
habitat fragmentation on community structure without
the direct influence of human-driven confounding fac-
tors (Driscoll 2005; Haddad et al. 2017). Unlike oceanic
islands, which are surrounded by a matrix that is a barrier
to most terrestrial fauna, forest archipelagos have more
permeable matrices that harbor an important percentage

13
Vol.:(0123456789)

Content courtesy of Springer Nature, terms of use apply. Rights reserved.

18
2016) and provide additional resources to species (Öck-
inger et al. 2012). These matrix-tolerant species can use
both vegetation types (forest and open areas), which allows
constant dispersal events and contributes to the persistence
of species in habitat patches. Such mechanisms could be
linked to metacommunity structure, a concept that offers
different perspectives based on the dispersal characteris-
tics of potentially interacting species (Leibold et al. 2004;
Thompson et al. 2017).
Species can differ in how they respond to fragmentation
and landscape attributes, (e.g., patch characteristics, con-
nectivity and habitat amount) (Haila 2002; Haddad et al.
2017). Many studies have found direct relationships between
patch area and high connectivity in their influence on the
abundance and richness of species of various arthropod
groups (e.g., Tscharntke et al. 2002; Watling and Donnelly
2006; Öckinger et al. 2012; Inclán et al. 2014). There is,
however, no consensus for all invertebrate groups, such as
galling insects (Julião et al. 2004), other herbivorous insects
(Macedo-Reis et al. 2019) and spiders (Nogueira and Pinto-
da-Rocha 2016).
Although anthropically-fragmented areas have been
widely studied (e.g., Cook et al. 2002; Carvalho Guimarães
et al. 2014; Magrach et al. 2014; Rossetti et al. 2017), there
have been few investigations involving natural forest frag-
ments embedded in open matrices, particularly considering
insects (Pereira et al. 2017; Coelho et al. 2018b; da Silva
et al. 2019). A weak species-area relationship was found for
galling species in forest fragments of the Pantanal (Julião
et al. 2004). In the same system, species richness of ants was
found to be positively correlated with herbaceous vegeta-
tion density (Corrêa et al. 2006) and with fragment size and
isolation (Cuissi et al. 2015). And beetle diversity was found
to be related to vegetation complexity (Vieira et al. 2008).
In mountainous natural Atlantic Forest fragments (hereafter
Atlantic Forest islands) within an open vegetation matrix, da
Silva et al. (2019) found patch attributes (e.g., canopy cover,
distance to continuous forest) and landscape attributes (e.g.,
percentage of forest in the landscape, number of patches,
patch shape complexity) to have small general effects on
dung beetle species. Pereira et al. (2017) found commu-
nities of fruit-feeding butterflies to be dependent on local
(i.e., patch attributes) and landscape (i.e., isolation) effects.
Butterfly richness was greater in forest islands with greater
canopy openness and lower understory coverage in the dry
season, while landscape metrics had little effect on richness,
but a strong influence on species composition (Pereira et al.
2017). In spite of being important providers of ecosystem
services, such as pollination (Novais et al. 2016), bees and
wasps of Atlantic Forest islands have been neglected from
such studies. It is unclear whether these species colonize
high-elevation forested habitats and if patch and landscape
attributes underpin the assembly of their communities.
13
Content courtesy of Springer Nature, terms of use apply. Rights reserved.
Journal of Insect Conservation (2020) 24:17–34
Diversity in a given region (γ-diversity) can be parti-
tioned into at least two components—α-diversity, which is
the number of species at the local scale, and β-diversity,
which represents differences in species composition among
sites (Whittaker 1972; Tuomisto 2010). Partitioning
γ-diversity can contribute to conservation strategies by, for
example, determining the number and size of areas that are
needed to protect naturally-fragmented forest diversity. Like-
wise, β-diversity can be decomposed into species replace-
ment (turnover) and species gain/loss (nestedness) (Baselga
2010), with this distinction being particularly important for
elucidating processes underlying biodiversity patterns. Dif-
ferences in habitat use among species determine the forma-
tion and maintenance of β-diversity of tropical mountains,
where the dominant process driving it is species turnover
among elevational belts (Nunes et al. 2016; Perillo et al.
2017; Mota et al. 2018). Turnover can indicate high portions
of endemism, while nestedness can indicate that dispersal
limitation in structuring assemblages (Baselga 2010), which
is a common pattern for tropical insects (Solar et al. 2015;
Perillo et al. 2017; Macedo-Reis et al. 2019). Some tropical
forest insect species undergo significant seasonal changes
due to differences in resource availability over time and due
to drastic changes in weather conditions (Tylianakis et al.
2005; Inclán et al. 2014). Only a subset of summer com-
munities is found active in these forests in the dry and cold
(winter) season (see Neves et al. 2010; da Silva et al. 2019).
We aimed to explore spatiotemporal patterns of bee
and wasp diversity (i.e., richness, abundance and temporal
β-diversity) in Atlantic Forest islands in a mountain complex
in eastern South America (Coelho et al. 2016). We aimed to:
(1) investigate seasonal effects on bee and wasp diversity; (2)
determine if canopy cover and understory density positively
influence species diversity; and (3) investigate whether patch
and landscape attributes (e.g., patch size, level of isolation,
distance between patches and distance to continuous forest)
affect species diversity with higher temporal β-diversity in
small-sized and more isolated patches. We also aimed to
determine if (4) winter communities are subsets of summer
communities.
Materials and methods
Study sites
The study was performed in a natural forest archipelago
situated in the southern Espinhaço Mountain Range, an
important mountainous formation in Brazil due to its high
biodiversity and endemism (Giulietti et al. 1997; Silveira
et al. 2016) (Table S1; Fig. 1a). This mountain range acts
as a humidity barrier, dividing two Brazilian domains—the
Cerrado (Brazilian savanna) and the Atlantic Forest—which
Journal of Insect Conservation (2020) 24:17–34
Fig. 1  a Map showing the 14 natural Atlantic forest islands (green
shapes) selected for sampling of Aculeata wasp and bees in a campo
rupestre matrix at Serra do Cipó, Minas Gerais, Brazil. b Concentric
together host a diverse array of threatened and endemic plant
and animal species (Giulietti et al. 1997; Silveira et al. 2016).
Forest island archipelagos of the southern Espinhaço Range
are most closely associated with the Atlantic Forest domain
(Coelho et al. 2018a), which is located on the east side of
the mountain range. Between the Atlantic Forest and Cer-
rado, and predominantly associated with mountaintops of
the Espinhaço Range (above 900 m asl), are campo rupestre,
grassland mosaics associated with rocky outcrops. Campo
rupestre are Neotropical azonal vegetation complexes (Alves
et al. 2014), that form the matrix of the landscape (i.e., pre-
dominant landcover) and are classified as old, climatically
buffered, infertile landscapes (OCBILs). As such, they have
favored the persistence of old botanical lineages that con-
tinue to diversify, such as the Gondwanan families Eriocaul-
aceae, Velloziaceae and Xyridaceae (Mello-Silva et al. 2011;
Hopper et al. 2016; Silveira et al. 2016).
Mostly immersed in the matrix of campo rupestre are
natural patches of vegetation emerging on mountaintops
(above 1200 m asl) called Capões de Mata, which have a
floristic composition related to Seasonal Semideciduous
Forest—a common Atlantic Forest formation in Southeast
Brazil (Coelho et al. 2016). These relictual forests have an
edaphoclimatic formation, and thus are highly dependent on
climate and soil for their development (Coelho et al. 2016,
Content courtesy of Springer Nature, terms of use apply. Rights reserved.
19
buffer with a 500 m radius (78.5 ha) considering the plot center as
the centroid to classify forest (green) and grassland (orange) land use
classes. (Color figure online)
2018a, b). Physical characteristics of the soil, such as drain-
age, are more relevant than chemical characteristics, such
as nutrient inputs, for the establishment and development
of this plant formation (Coelho et al. 2016). Forest islands
of the Espinhaço Range are directly related to high rates of
relative humidity and nearly constant rainfall (Coelho et al.
2016), caused by the orographic effects of high altitudes
(Barbosa et al. 2010).
Sampling design
A total of 14 forest islands were selected considering preser-
vation status (i.e., exclusion of fragments strongly impacted
by fire or anthropic uses) and dimensions (Figs. 1a, 2;
Table S1). The chosen areas of Capões de Mata vary in
size (1.21–39.89 ha, mean = 8.21 ha, SD = 10.24 ha) and
shape (from circular to elongated limits), and are spaced
at variable distances (0.1–1.5 km). A 20 × 50 m plot was
established in the center of each forest patch, where sam-
pling was performed over 2 years (2014–2015) with two
sampling events in the summer (February 9–16, 2014; Feb-
ruary 21–28, 2015) and two in the winter (August 16–23,
2014; August 22–29, 2015).
A range of methodologies was used during each sampling
period in order to maximize the sampled species diversity.
13
20
Fig. 2  External (a–c) and internal (d) views of Capões de Mata in a campo rupestre matrix at Serra do Cipó. Minas Gerais, Brazil. (c) Sche-
matic view of a 20 × 50 m plot created in each Atlantic forest island
Each Atlantic Forest island received: (1) two Malaise
traps—one soil Malaise trap and one canopy Malaise Win-
dow trap; (2) ten colorless pitfall traps (five on the ground
and five attached to tree trunks); and (3) four Moericke traps
(25 cm diameter yellow pan traps with the container filled
with salty liquid). Moericke and pitfall traps remained in the
field for 48 h (Moericke: 10,752 trap-hours; Pitfall: 26,880
trap-hours), while Malaise traps remained in the field for
7 days (9408 trap-hours each type).
All captured hymenopterans of Aculeata, with the excep-
tion of those of the family Formicidae, were identified to the
lowest taxonomic level possible based on Fernández and
Sharkey (2006) and Silveira et al. (2002). Identifications
were confirmed by specialists of each group whenever pos-
sible. Nevertheless, community analyses were based largely
on morphospecies to enable the use of most of the Aculeata
families on the study, an accepted approach for some hyme-
nopteran families (Derraik et al. 2010; Perillo et al. 2017).
Vegetation structure
Vegetation structure in each Atlantic forest island was
determined using digital images of shrub density and her-
baceous vegetation (understory structure) and the relative
contribution of trunks and branches (canopy structure), by
interpreting black pixels within the selected frame area as
vegetation structures (see Zehm et al. 2003; Nassar et al.
13
Content courtesy of Springer Nature, terms of use apply. Rights reserved.
Journal of Insect Conservation (2020) 24:17–34
2008 for further details). Canopy structure was calculated
using four digital photographs per site (20 × 50 m plot) taken
during each sampling using a digital camera with an 8 mm
fisheye lens (Nassar et al. 2008). The density of understory
vegetation was estimated from 16 photos taken in each site
(20 × 50 m plot) per sampling with a 100 × 100 cm white
screen as a backdrop (Zehm et al. 2003). Canopy photos
were analyzed using Gap Light Analyzer software v.2
(Frazer et al. 1999) while understory photos were processed
using the EBImage package of R (Oles et al. 2012; R Core
Team 2017). In both cases, mean cover and density values
were obtained for each sampling per forest island (for more
details see Pereira et al. 2017).
Landscape variables
Individual study landscapes were defined using concentric
circular sectors with a 500 m radius (78.5 ha), chosen to
encompass each of the 14 forest islands with the plot center
as the centroid (Fig. 1b). The buffer area was chosen because
it is sufficient to encompass common foraging activities of
bees and wasps, which significantly decline in species rich-
ness and abundance at around 300 m from the considered
original location (Osborne et al. 1999; Zurbuchen et al.
2010). Due to buffer overlap at larger spatial scales, effects of
spatial or other types of statistical autocorrelation in the data
were minimized. Although herbaceous phytophysiognomies
Journal of Insect Conservation (2020) 24:17–34
(i.e., microhabitats and vegetational structure) of campo rup-
estre are highly heterogeneous, the land within 500 m buff-
ers was classified only as “forest” or “grassland” habitat; all
tree formations were classified as “forest” and all open lands
were classified as “grasslands”, regardless of the type of land
use. Classifications were mapped for the landscape of each
buffer using high-resolution multispectral images (RapidEye
satellites, ~ 5 m resolution), and the R package randomFor-
est v4.6-12 (R Core Team 2017), a method proposed by
Breiman (2001). The classification model was constructed
using all five Rapideye spectral bands plus the Normalized
Difference Vegetation Index (NDVI) and red edge NDVI.
The landscape attributes were analyzed with the software
Fragstats v4.2.1 (McGarigal et al. 2012), using a fixed edge
depth of 20 m and no sampling strategy. The metrics were
calculated for the 500 m circular sector of each forest island.
Ten Fragstats class variables were selected (considering
non-redundant and most biological meaning for the studied
group) (Table S2), which, in addition to five other metrics
(Table S3), were used as explanatory variables. Of these five
metrics, patch size (original perimeter), distance to continu-
ous forest (considering forest limits) and distance to closest
forest island (considering forest limits) were calculated man-
ually using Google Earth Pro tools, while canopy cover and
understory density were calculated as previously described.
Correlation analyses were run (psych R package) testing all
15 explanatory variables, with those exhibiting high cor-
relation values (Pearson correlation coefficient ≥ 0.7; see
Dormann et al. 2013) being excluded sequentially (Fig. S1).
Statistical analysis
All statistical analyses were performed using R v.3.4.0.
Sampling efficiency was analyzed using total species accu-
mulation curves with non-parametric estimators (Jackknife
1 and Chao). The influences of season and patch variables
(canopy cover and understory density) on species richness
and abundance were determined using the identity of each
forest fragment (14) as a random factor in a Generalized Lin-
ear Mixed Model (GLMM), to account for temporal pseu-
doreplication (Bolker et al. 2009), using the most suitable
error distribution for each model (Crawley 2013). The lme4
package and glmer function were used to perform GLMM
analyses. Each model (richness and abundance) was fitted
with five non-correlated fixed-effect explanatory variables
named ‘season’, ‘canopy cover’, ‘understory density’, ‘sea-
son × canopy cover interaction’ and ‘season x understory
density interaction’. Non-significant (P > 0.05) variables and
interactions were removed in order to simplify the model
until the minimal adequate model was obtained by remov-
ing non-significant variables and comparing the full models
with simplified models, using the native function anova in R.
The function Anova of package car, which allows ANOVA
Content courtesy of Springer Nature, terms of use apply. Rights reserved.
21
in mixed models, was used to deem variables significant or
not. All models were subjected to residual analyses to check
for model fit and error structure suitability (Crawley 2013).
Temporal changes of species between seasons (i.e., dif-
ferences in species composition between seasonal samples)
in each forest island were calculated as the percent contribu-
tion of species turnover or nestedness to total β-diversity for
multiple seasons. The results of this calculation were used
to determine whether species substitution or species loss/
gain influenced the general temporal pattern of β-diversity
(Baselga 2010). The percentage contributions of turnover
(βsim) and nestedness (βnes) were also calculated between
seasons in each sample year [using mean values resultant of
Sørensen (βsor) dissimilarity fraction (βsor = βsim + βnes) (see
Baselga 2012)] to determine if winter communities (assem-
blages with lower species richness) were subsets of summer
communities.
Landscape predictors were tested using generalized linear
models (GLMs), considering accumulated richness, abun-
dance and temporal β-diversity (βsor), and using all non-cor-
related vegetation and landscape attributes as explanatory
variables (Tables S2 and S3 and Fig. S1). Model simplifi-
cation was conducted excluding non-significant variables
and interactions to achieve the most parsimonious model.
The best distribution errors were selected for each model
(Crawley 2013) using the R package RT4Bio with Quasi-
Poisson errors.
Results
A total of 3402 specimens were collected, representing 213
morphospecies and 16 families of Aculeata (Table 1 and
2). Of these, 2878 were collected by Malaise traps; 453 by
Moericke traps; 34 by Malaise/Window traps and 33 by Pit-
fall traps (details in Table 1). There was a high number of
rarely sampled species, comprising 79 singletons (37.01%
of the total) and 29 doubletons (13.61% of the total). Only
three morphospecies were sampled in all forest islands.
Jackknife 1 estimator determined sampling adequacy to be
71.79% (estimated richness: 296.70), while the Chao estima-
tor determined it to be 67.32% (estimated richness: 316.40)
(Fig. S2). A total of 87 (41%) morphospecies were collected
exclusively in the summer and 39 (18%) exclusively in the
winter. The families Embolemidae and Andrenidae were
only collected in the winter, while Colletidae was exclu-
sively collected in the summer, all three of which are fami-
lies of low abundance.
Both the richness (Fig. 3) and abundance of bee and wasp
species varied by season, with greater richness (Chi = 17.83;
d.f. = 1; n = 14; P < 0.001) and abundance (Chi = 6.87;
d.f. = 1; n = 14; P < 0.01) in summer (Table 3). Despite this
result, the winter communities did not represent subsets of
13
22 Journal of Insect Conservation (2020) 24:17–34

Table 1  Bee and wasp (Hymenoptera: Aculeata) morphospecies abundance, capture method and number of forest island occurrences in summer
and winter seasons in Serra do Cipó, Minas Gerais, Brazil
Capture method Summer season Winter season Total abundance

Malaise Malaise trap Moericke trap Pitfall trap Islands with spe- Abundance Islands with spe- Abundance
window cies occurrence cies occurrence

APOIDEA
ANDRENIDAE
ANDRENINAE
Andreninae sp. 01 X 0 0 2 2 2
Andreninae sp. 02 X 0 0 1 1 1
Andreninae sp. 03 X 0 0 1 1 1
APIDAE
APINAE
Apinae sp. 01 X 1 1 1 1 2
Apinae sp. 02 X 1 1 0 0 1
APINI
Apis mellifera Lin- X X X 6 7 8 22 29
naeus, 1758
MELIPONINI
Melipona X X X 5 20 13 77 97
(Melipona)
quadrifasciata
Lepeletier, 1836
Trigona spinipes X X X 7 12 3 6 18
(Fabricius, 1793)
Partamona ailyae X 1 1 2 6 7
Camargo, 1980
Trigona braueri X 3 44 2 12 56
Friese, 1900
XYLOCOPINI
CERATININA
Ceratina (Cerati- X 4 4 0 0 4
nula) sp. 01
Ceratina (Cerati- X 1 1 1 1 2
nula) sp. 02
Ceratina (Cerati- X 2 2 0 0 2
nula) sp. 03
Ceratina (Cerati- X 0 0 1 1 1
nula) sp. 04
Ceratina (Cerati- X 1 1 0 0 1
nula) sp. 05
Ceratina (Cerati- X 1 2 0 0 2
nula) sp. 06
Ceratina (Cerati- X 1 1 0 0 1
nula) sp. 07
Ceratina (Cerati- X X 1 1 1 1 2
nula) sp. 08
Ceratina (Crewella) X 1 1 0 0 1
sp. 01
Ceratina (Crewella) X 1 1 0 0 1
sp. 02
EMPHORINI
Melitoma torquata X 1 1 0 0 1
Roig-Alsina, 2009
MELIPONINI
Scaptotrigona X X 0 0 2 2 2
postica (Latreille,
1807)
Trigona sp. 01 X 1 1 0 0 1
Trigona sp. 02 X 1 1 1 1 2
Paratrigona lineata X 1 1 0 0 1
(Lepeletier, 1836)

13

Content courtesy of Springer Nature, terms of use apply. Rights reserved.

Journal of Insect Conservation (2020) 24:17–34 23

Table 1  (continued)
Capture method Summer season Winter season Total abundance

Malaise Malaise trap Moericke trap Pitfall trap Islands with spe- Abundance Islands with spe- Abundance
window cies occurrence cies occurrence

COLLETIDAE
HYLAEINAE
Hylaeus sp. X X 2 2 0 0 2
HALICTIDAE
HALICTINAE
AUGOCHLORINI
Neocorynura X 1 2 3 3 5
aenigma (Gribodo,
1894)
Augochloropsis X 0 0 1 1 1
sp. 01
Augochloropsis X 0 0 1 1 1
sp. 02
Augochloropsis X 0 0 1 1 1
sp. 03
Augochloropsis X 1 1 0 0 1
hebescens (Smith,
1879)
Augochlorini sp. 01 X 1 1 0 0 1
Augochlorini sp. 02 X 1 1 1 2 3
Augochlorini sp. 03 X 0 0 1 1 1
Augochlorini sp. 04 X X 3 4 1 2 6
Augochlorini sp. 05 X 1 1 3 4 5
Augochlorini sp. 06 X 1 1 0 0 1
HALICTINI
Dialictus sp. 01 X X 3 3 2 6 9
Dialictus sp. 02 X 0 0 1 1 1
Dialictus sp. 03 X 1 1 0 0 1
Dialictus sp. 04 X X 4 9 1 1 10
Dialictus sp. 05 X 0 0 1 1 1
Dialictus sp. 06 X 0 0 1 1 1
Dialictus sp. 07 X 0 0 1 1 1
Dialictus sp. 08 X 1 1 0 0 1
Halictini sp. 01 X 1 1 0 0 1
AMPULICIDAE
AMPULICINAE
AMPULICINI
Ampulex sp. 01 X X 5 5 1 1 6
Ampulex sp. 02 X 1 1 0 0 1
DOLICHURINI
Dolichurus sp. 01 X X 3 4 0 0 4
Dolichurus sp. 02 X 4 6 0 0 6
Dolichurus sp. 03 X 0 0 1 2 2
CRABRONIDAE
CRABRONINAE
BOTH-
YNOSTETHINI
Bothynostethini X 2 5 0 0 5
sp. 01
CRABRONINI
Ectemnius sp. 01 X 2 2 3 3 5
Crabronini sp. 01 X 4 7 5 5 12
Crabronini sp. 02 X X 4 4 7 12 16
Crabronini sp. 03 X 0 0 1 1 1
Crabronini sp. 04 X X 1 1 2 2 3

13

Content courtesy of Springer Nature, terms of use apply. Rights reserved.

24 Journal of Insect Conservation (2020) 24:17–34

Table 1  (continued)
Capture method Summer season Winter season Total abundance

Malaise Malaise trap Moericke trap Pitfall trap Islands with spe- Abundance Islands with spe- Abundance
window cies occurrence cies occurrence

LARRINI
Liris sp. 01 X X 10 34 9 20 54
Liris sp. 02 X 1 1 0 0 1
Gastroserina sp. 01 X 1 1 0 0 1
Gastroserina sp. 02 X 1 1 1 1 2
MISCOPHINI
Miscophini sp. 01 X X 3 6 0 0 6
Miscophini sp. 02 X 1 1 0 0 1
TRYPOXYLINI
Pison sp. 01 X X 8 19 5 12
Trypoxylon (Trypar- X X 1 1 1 1 2
gilum) lactitarse
Saussure, 1867
Trypoxylini sp. 01 X X 7 19 1 1 20
Trypoxylini sp. 02 X 4 4 1 1 5
Trypoxylini sp. 03 X 4 11 2 3 14
Trypoxylini sp. 04 X 5 6 2 2 8
Trypoxylini sp. 05 X 2 2 0 0 2
Trypoxylini sp. 06 X 2 2 1 2 4
Trypoxylini sp. 07 X 1 1 0 0 1
Trypoxylini sp. 08 X 0 0 1 1 1
Trypoxylini sp. 09 X 0 0 1 1 1
PEMPHREDONI-
NAE
PEMPHRE-
DONINI
Pemphredonini X X 3 12 6 15 27
sp. 01
Pemphredonini X X 2 2 3 3 5
sp. 02
Pemphredonini X X 1 2 0 0 2
sp. 03
Pemphredonini X 1 1 0 0 1
sp. 04
SPILOMENINA
Spilomena sp. 01 X 4 6 5 6 12
Spilomena sp. 02 X 1 3 1 1 4
Spilomena sp. 03 X 1 1 2 2 3
Spilomena sp. 04 X X 0 0 2 3 3
PHILANTHINAE
CERCERINI
Eucerceris sp. 01 X 1 2 0 0 2
Eucerceris sp. 02 X 1 1 0 0 1
Eucerceris sp. 03 X 1 1 0 0 1
PHILANTHINI
Thachypus sp. 01 X 0 0 1 1 1
SPHECIDAE
SPHECINAE
AMMOPHILINI
Eremnophila sp. 01 X 1 1 0 0 1
VESPOIDEA
MUTILLIDAE
MUTILLINAE
EPHUTINI
Ephuta sp. 01 X 9 30 2 3 33

13

Content courtesy of Springer Nature, terms of use apply. Rights reserved.

Journal of Insect Conservation (2020) 24:17–34 25

Table 1  (continued)
Capture method Summer season Winter season Total abundance

Malaise Malaise trap Moericke trap Pitfall trap Islands with spe- Abundance Islands with spe- Abundance
window cies occurrence cies occurrence

Ephuta sp. 02 X 2 7 0 0 7
Ephuta sp. 03 X 1 1 0 0 1
Ephuta sp. 04 X 0 0 1 1 1
Ephuta sp. 05 X 1 1 0 0 1
Timulla sp. 01 X 1 1 0 0 1
SPHAEROPTHAL-
MINAE
SPHAEROPTHAL-
MINI
Sphaeropthalmini X 1 1 0 0 1
sp. 01
Sphaeropthalmini X 2 3 1 1 4
sp. 02
Sphaeropthalmini X 0 0 1 1 1
sp. 03
Sphaeropthalmini X 1 1 0 0 1
sp. 04
Sphaeropthalmini X 1 1 0 0 1
sp. 05
Sphaeropthalmini X 0 0 1 1 1
sp. 06
Sphaeropthalmini X 1 1 0 0 1
sp. 07
Xystromutilla sp. 01 X 3 3 1 2 5
Xystromutilla sp. 02 X 1 1 0 0 1
Xystromutilla sp. 03 X 0 0 2 2 2
POMPILIDAE
Pompilidae sp. 01 X 0 0 9 71 71
Pompilidae sp. 02 X X X 11 31 0 0 31
Pompilidae sp. 03 X 1 1 0 0 1
Pompilidae sp. 04 X X 6 18 0 0 18
Pompilidae sp. 05 X 1 1 0 0 1
Pompilidae sp. 06 X X 3 3 1 1 4
Pompilidae sp. 07 X X 3 3 0 0 3
Pompilidae sp. 08 X X 1 1 1 1 2
Pompilidae sp. 09 X X 2 2 3 6 8
Pompilidae sp. 10 X 1 1 1 1 2
Pompilidae sp. 11 X 2 3 0 0 3
Pompilidae sp. 12 X X 1 1 1 2 3
Pompilidae sp. 13 X X 7 12 7 12 24
Pompilidae sp. 14 X X 4 34 4 45 79
Pompilidae sp. 15 X 2 2 0 0 2
Pompilidae sp. 16 X X 5 14 3 5 19
Pompilidae sp. 17 X 0 0 1 1 1
Pompilidae sp. 18 X X X 8 13 2 2 15
Pompilidae sp. 19 X X 6 43 8 23 66
Pompilidae sp. 20 X X X 7 26 2 8 34
Pompilidae sp. 21 X 1 1 0 0 1
Pompilidae sp. 22 X 0 0 1 1 1
Pompilidae sp. 23 X X 4 6 5 6 12
Pompilidae sp. 24 X X 4 7 2 2 9
Pompilidae sp. 25 X X 2 2 0 0 2
Pompilidae sp. 26 X X 6 10 7 11 21
Pompilidae sp. 27 X 7 89 3 13 102
Pompilidae sp. 28 X X 7 50 1 1 51

13

Content courtesy of Springer Nature, terms of use apply. Rights reserved.

26 Journal of Insect Conservation (2020) 24:17–34

Table 1  (continued)
Capture method Summer season Winter season Total abundance

Malaise Malaise trap Moericke trap Pitfall trap Islands with spe- Abundance Islands with spe- Abundance
window cies occurrence cies occurrence

Pompilidae sp. 29 X 1 1 0 0 1
Pompilidae sp. 30 X 1 9 0 0 9
Pompilidae sp. 31 X X 2 3 1 1 4
Pompilidae sp. 32 X X 1 1 1 1 2
Pompilidae sp. 33 X X 3 4 1 2 6
Pompilidae sp. 34 X X 3 4 2 4 8
Pompilidae sp. 35 X 1 1 0 0 1
Pompilidae sp. 36 X X 2 4 0 0 4
Pompilidae sp. 37 X 1 1 0 0 1
Pompilidae sp. 38 X 1 1 0 0 1
Pompilidae sp. 39 X 1 1 0 0 1
Pompilidae sp. 40 X X 1 2 0 0 2
Pompilidae sp. 41 X 1 1 0 0 1
SCOLIIDAE
SCOLIINAE
Campsomeris sp. 01 X X X 6 20 5 10 30
Campsomeris sp. 02 X X 2 4 1 9 13
Campsomeris sp. 03 X 1 1 0 0 1
Campsomeris sp. 04 X 1 1 0 0 1
TIPHIIDAE
ANTHOBOSCINAE
Anthoboscinae X 8 2 10 0 2
sp. 01
MYZININAE
Pterombrus sp. 01 X 1 33 0 0 33
THYNNINAE
ELAPHROPTERINI
Ornepetes sp. 01 X X 0 0 1 8 8
RHAGIGASTERINI
Aelurus sp. 01 X X 5 25 0 0 25
Aelurus sp. 02 X X X 0 12 8 0 12
Aelurus sp. 03 X 6 6 0 0 6
SCOTAENINI
Scotaenini sp. 01 X X X 1 0 0 69 69
Scotaenini sp. 02 X 1 10 0 0 10
Scotaenini sp. 03 X 0 1 3 0 1
Scotaenini sp. 04 X 0 0 1 1 1
Scotaena sp. 01 X 4 0 0 1 1
TIPHIINAE
Tiphia sp. 01 X X X X 12 140 6 3 143
Tiphia sp. 02 X X X 5 83 0 12 95
Tiphia sp. 03 X X 7 7 0 0 7
VESPIDAE
EUMENINAE
Eumeninae sp. 01 X 0 0 0 1 1
POLISTINAE
EPIPONINI
Agelaia multipicta X X X 13 48 11 42 90
(Haliday, 1836)
Agelaia myrme- X 4 20 1 6 26
cophila (Ducke,
1905)
Polybia bifasciata X 1 2 0 0 2
Saussure, 1854

13

Content courtesy of Springer Nature, terms of use apply. Rights reserved.

Journal of Insect Conservation (2020) 24:17–34 27

Table 1  (continued)
Capture method Summer season Winter season Total abundance

Malaise Malaise trap Moericke trap Pitfall trap Islands with spe- Abundance Islands with spe- Abundance
window cies occurrence cies occurrence

Polybia chrysotho- X 1 0 1 1 1
rax (Lichtenstein)
Polybia fastidiosus- X X X X 2 126 1 69 195
cula Saussure,
1854
Polybia lugubris X 4 1 0 0 1
Ducke, 1905
Polybia occidentalis X X 0 5 1 1 6
(Olivier,1791)
Polybia sericea X X 1 6 0 0 6
(Olivier, 1792)
Polybia sp. X 1 1 0 0 1
MISCHOCYT​TAR​
INI
Mischocyttarus X 2 5 0 1 6
rotundicollis
(Cameron,1912)
POLISTINI
Polistes versicolor X 2 1 1 0 1
(Olivier,1791)
Polistes lanio (Fab- X X 0 1 1 1 2
ricius, 1775)
Polistes deceptor X 0 0 1 1 1
Schulz, 1905
CHRYSIDOIDEA
BETHYLIDAE
BETHYLINAE
Bethylinae sp. 01 X X 1 1 5 7 8
EPYRINAE
Epyrinae sp. 01 X X X 14 92 14 99 191
Epyrinae sp. 02 X X 9 47 4 9 56
Epyrinae sp. 03 X 0 0 1 1 1
Epyrinae sp. 04 X X X 3 3 2 3 6
Epyrinae sp. 05 X 0 0 1 1 1
Epyrinae sp. 06 X 1 1 0 0 1
PRISTOCERINAE
Pristocerinae sp. 01 X 4 16 7 14 30
Pristocerinae sp. 02 X X 4 4 3 5 9
Pristocerinae sp. 03 X 2 5 2 3 8
Pristocerinae sp. 04 X 6 6 0 0 6
Pristocerinae sp. 05 X X X X 14 244 14 87 331
Pristocerinae sp. 06 X X 13 167 14 317 484
Pristocerinae sp. 07 X 4 10 5 25 35
Pristocerinae sp. 08 X 2 2 0 0 2
Pristocerinae sp. 09 X 2 3 0 0 3
Pristocerinae sp. 10 X 2 2 0 0 2
Pristocerinae sp. 11 X 5 14 11 72 86
Pristocerinae sp. 12 X X 3 5 5 11 16
Pristocerinae sp. 13 X 1 1 2 2 3
Pristocerinae sp. 14 X X 1 1 1 1 2
Pristocerinae sp. 15 X 0 0 1 1 1
CHRYSIDIDAE
AMISEGINAE
Amisega sp. 01 X 1 1 0 0 1
CHRYSIDINAE
Pleurochrysis sp. 01 X 1 1 3 4 5

13

Content courtesy of Springer Nature, terms of use apply. Rights reserved.

28 Journal of Insect Conservation (2020) 24:17–34

Table 1  (continued)
Capture method Summer season Winter season Total abundance

Malaise Malaise trap Moericke trap Pitfall trap Islands with spe- Abundance Islands with spe- Abundance
window cies occurrence cies occurrence

Caenochrysis sp. 01 X X 2 4 2 3 7
Ipsiura sp. 01 X 2 2 0 0 2
DRYINIDAE
Dryinidae sp. X 1 1 0 0 1
ANTEONINAE
Deinodryinus sp. 01 X X 0 0 2 3 3
DRYININAE
Dryinus sp. 01 X 1 5 0 0 5
Dryinus sp. 02 X 3 3 2 2 5
Dryinus sp. 03 X X 2 4 5 5 9
Dryinus sp. 04 X X 3 5 0 0 5
Thaumatodryinus X 1 1 3 5 6
sp. 01
Thaumatodryinus X 1 1 0 0 1
sp. 02
GONATOPODINAE
Neodryinus sp. 01 X X 1 1 5 8 9
EMBOLEMIDAE
Embolemus sp. 01 X 0 0 2 4 4
Embolemus sp. 02 X 0 0 1 1 1
Total 1980 1422 3402

Table 2  Richness and Family Summer Winter Total

abundance for each Aculeata
(Hymenoptera) family collected Richness Abundance Richness Abundance Richness Abundance
in summer and winter in a
Brazilian forest archipelago CHRYSIDOIDEA
Bethylidae 19 624 17 658 22 1282
Chrysididae 4 8 2 7 4 15
Dryinidae 8 21 5 19 9 40
Embolemidae 0 0 2 9 2 9
VESPOIDEA
Mutillidae 12 51 7 11 16 62
Pompilidae 38 407 23 221 41 628
Scoliidae 4 26 2 19 4 45
Tiphiidae 10 319 6 94 14 413
Vespidae 11 216 9 123 14 339
APOIDEA
Andrenidae 0 0 3 4 3 4
Apidae 20 104 11 130 22 234
Colletidae 1 2 0 0 1 2
Crabronidae 30 169 23 99 35 268
Halictidae 12 26 14 27 20 53
Sphecidae 5 7 2 1 6 8
Total 174 1980 126 1422 213 3402

summer communities, with a low contribution of nested- (understory vegetation density and canopy vegetation
ness between seasons (nestedness summer/winter for first cover) affected species richness, but abundance was nega-
sample year: 0.22%; nestedness summer/winter for second tively influenced by understory density during the summer
sample year: 0.39%). None of the local variables measured (Chi = 4.13; d.f. = 1; n = 14; P < 0.05) (Fig. 4; Table 3). This

13

Content courtesy of Springer Nature, terms of use apply. Rights reserved.

Journal of Insect Conservation (2020) 24:17–34 29

Fig. 3  Bee and wasp species richness in summer and winter seasons in Atlantic Forest islands. Bars represent the standard error with the mean
value (diamond)

Table 3  Results of minimum Dependent variable Coefficients of fixed effects

adequate models (GLMMs),
showing the effects of seasonal Parameters Estimate Std. error z-value Pr (> |z|)
variables on bee and wasp
species richness and abundance Richness Intercept 3.112 0.079 39.29 < 0.001
Season − 0.322 0.076 − 4.22 < 0.001
Abundance Intercept 4.5237 0.2264 19.983 < 0.001
Season − 0.3319 0.1266 − 2.621 0.009
Understory density − 1.0149 0.4997 − 2.031 0.042

Adjusted to a negative binomial error distribution

relationship, however, was strongly influenced by two points
belonging to one specific forest island. Therefore, despite
this result deserves further investigation, we suggest caution
analyzing this result in our data.
We found high absolute values of temporal dissimilar-
ity (temporal β-diversity) for the forest islands, to which
species replacement (turnover) contributed relatively more
than did nestedness. β-diversity represented 73% of total
diversity between sampling periods with 88% of this being
represented by species temporal turnover (βsim). Temporal
β-diversity (βsor) for each forest island, used as a dependent
variable, was directly related only to distance to continuous
forest (Deviance = 0.03; d.f. = 12; Pseudo R-Squared = 0.43;
P = 0.04), with it increasing with increasing distance (DCF
variable: Table S2 and Fig. 5). Distance to continuous forest
was not correlated with any other variable (Fig. S1). Besides
distance to continuous forest, there were no significant
effects of the landscape attributes on richness, abundance
or temporal β-diversity.
Discussion
In general, of the patch and landscape attributes considered
in this study, only distance to continuous forest seemed
to be an important spatial factor driving changes in the

13

Content courtesy of Springer Nature, terms of use apply. Rights reserved.

30 Journal of Insect Conservation (2020) 24:17–34

Fig. 4  Species abundance in summer (black dots) and winter (gray triangles) seasons in Atlantic forest islands with an understory density inter-
action

Fig. 5  Relationship between bee and wasp species temporal β-diversity (βsor) (percentage) in each Atlantic forest island and its distance from
continuous forest (km)

13

Content courtesy of Springer Nature, terms of use apply. Rights reserved.

Journal of Insect Conservation (2020) 24:17–34
composition of Aculeata species in natural forest fragments.
This spatial pattern indicates that local species diversity (i.e.,
species that occupy each forest island) is limited by disper-
sal (Leibold et al. 2004; Jamoneau et al. 2012), a pattern
also observed among beetles in naturally fragmented areas
(Driscoll 2008).
We found seasonal variation in bee and wasp distribution,
a pattern also found for butterflies (Pereira et al. 2017), ants
(Soares et al. 2013) and dung beetles (da Silva et al. 2019)
in Atlantic Forest islands. This pattern could be explained
by temporal variation in the availability of food for adults,
such as nectar or pollen resources (Inclán et al. 2014) and/
or changes in temperature (Oliver et al. 2016). It also could
be controlled by competition-colonization trade-offs, with
species competing for resources and inducing colonization
of empty habitats at the regional scale (Van Nouhuys and
Hanski 2002), as indicated by the metacommunity “patch
dynamics” paradigm (Leibold et al. 2004). This temporal
variation may also be ruled by trade-offs between food avail-
ability, which is abundant in summer, and mobility for forag-
ing scarce resources in winter. Insect seasonality in seasonal
tropical systems is determined mostly by variation in rain
and humidity (Wolda 1988; Abrahamczyk et al. 2011), but in
some cases temperature can be the main variable responsible
for seasonal patterns (Silva et al. 2011). The environment is
always humid at high altitudes in the Espinhaço Mountain
Range due to the nebular condensation of humid air (Coelho
et al. 2016), but temperature varies seasonally (Long-Term
Ecological Research Program Project—PELD Campos Rup-
estres da Serra do Cipó; unpublished data). Hence, increases
in bee and wasp richness can be determined by elevated
rainfall and temperature regimes in the summer.
We also found that patch attributes, such as canopy
openness and understory density, may not influence species
richness. The only interaction our data revealed was that
between seasonality and understory density, with a nega-
tive relationship with bee and wasp abundance in summer.
However, researchers have found a contrasting response for
ants, with higher richness and diversity in patches with more
plant cover (Corrêa et al. 2006; Oliver et al. 2016). In the
same system as the present study, the abundance of frugivo-
rous butterflies was not affected by understory coverage in
either season, but there was higher richness in patches with
greater understory coverage during the summer and an oppo-
site pattern (higher richness in lower understory coverage)
in the winter (Pereira et al. 2017). A covered understory
seemed to hinder the movement by hymenopteran parasi-
toids (Dall’Oglio et al. 2016), while floral resources were
more abundant inside forests when light passed through
vegetation (Walters and Stiles 1996). Flowering in altitudi-
nal grasslands is concentrated in the summer (Rocha et al.
2016), a season when matrix-tolerant bee and wasp species
constantly uses these open areas to forage. The lack of a
Content courtesy of Springer Nature, terms of use apply. Rights reserved.
31
relationship between understory cover and bee and wasp
abundance in the winter suggests that most species use for-
est islands as refuges and as main foraging areas during this
period of harsh climatic conditions and reduced resources
in open matrix areas (Rocha et al. 2016), regardless of the
understory density of forest islands.
The high temporal species replacement found in the stud-
ied Atlantic Forest islands has been documented in simi-
lar environments as the most important component of dis-
similarity in species over time (Inclán et al. 2014; Oliver
et al. 2016). Species turnover can also be correlated with
geographic distance (Misiewicz et al. 2014; Perillo et al.
2017) and its positive relationship with distance from a
continuous Atlantic Forest remnant shows its importance
to species distribution dynamics, and indicates a spatially
explicit model (Leibold et al. 2004). Forest islands can be
considered transient environments for many species (Vieira
et al. 2008), where their presence can be considered a sample
of the dispersers coming from continuous forest and even
from open surrounding areas, which can explain the high
rates of rare species (singletons and doubletons; 50.62%).
Similar rates of rare species have been commonly observed
in tropical arthropod studies (Coddington et al. 2009; Bas-
set et al. 2012), and especially in high-elevation tropical
mountains (Perillo et al. 2017), although these rates should
diminish with increased sampling effort due to the prob-
ability of capturing more propagules (i.e., specimens that
randomly disperse from a source area). Our community
analyses were based on morphospecies sorting, which is an
accepted approach for some hymenopteran families (Derraik
et al. 2010), especially for assessing patterns of richness and
changes in species composition (Oliver and Beattie 1996).
Most of the landscape attributes of the present study had
no significant effects on bee and wasp species, and area,
isolation and spatial connectivity could not alone explain
species diversity (such as predicted by Prugh et al. 2008;
Haddad et al. 2017). The species-area relationship does
not represent a consensus, with studies recording positive
(Benedick et al. 2006; Watling and Donnelly 2006; Öckinger
et al. 2012; Cuissi et al. 2015), no (Julião et al. 2004; Vieira
et al. 2008; Banks-Leite et al. 2012; Nogueira and Pinto-da-
Rocha 2016), and even negative relationships (Lion et al.
2016) with species diversity. The forest-matrix boundary is
probably not an efficient barrier for most bee and wasp spe-
cies (Tscharntke et al. 2012), and matrices can also harbor
an important percentage of fauna (Cook et al. 2004; Driscoll
2005; Yekwayo et al. 2016), masking the influence of con-
nectivity or patch size, for example. Moreover, long foraging
distances can be traversed by a small percentage of individu-
als (Zurbuchen et al. 2010), but this can be insufficient to
prevent genetic differentiation between distant forest frag-
ments (Suni et al. 2014). Landscape structure probably has
a particular influence on bees and wasps depending on the
13
32
species and their different attributes. Future investment in
determining which species are specialists to a specific envi-
ronment (i.e., forest-dependent or matrix-tolerant species)
can bring new insights to analyses of hymenopteran commu-
nities. For example, only forest-dependent dung beetles were
found to be influenced by patch and landscape attributes in
the same study system (da Silva et al. 2019).
Therefore, landscape and patch attributes could not
explain the pattern of species diversity, especially species
richness, observed in the studied campo rupestre forest
island archipelago. The distribution of bee and wasp spe-
cies is consistent with the “patch dynamics” metacommu-
nity paradigm (Leibold et al. 2004; Jamoneau et al. 2012),
by which each patch may be occupied or unoccupied and
local species diversity is limited by dispersal and species
competitive ability (Leibold et al. 2004; Biswas and Wag-
ner 2012). This provides a valid explanation of spatial and
temporal dynamics, driven by habitat heterogeneity and
environmental filters, with forest fragments connected by
dispersal among islands.
Acknowledgements We would like to thank the following students
from the Insect Ecology Lab for assistance with field work: Luiz
Eduardo Macedo Reis, Flávio Siqueira de Castro, Daniela Melo, Luiz
Fernando Ferreira, Geanne Pereira, Humberto Brant and Heron Hilário,
and we highlight the invaluable help of Caio Silveira Marques and
Thaís Silva Tavares. We extend our gratitude to G. Wilson Fernandes
for providing climatic data; Reserva Vellozia and GSG for logistical
support; José Eustáquio dos Santos Júnior and Rogério Lopes for speci-
men identification; and Francisco Carvalho Diniz and Pedro Giovâni
da Silva for text revision. Dr. Marcel Coelho provided very impor-
tant advice regarding the project and provided insightful comments
regarding the text. This study was supported by Fundação de Amparo
à Pesquisa do Estado de Minas Gerais (FAPEMIG), Conselho Nacional
de Desenvolvimento Científico e Tecnológico (Long-Term Ecological
Research PELD-CRSC-17) and Coordenação de Aperfeiçoamento de
Pessoal de Nível Superior.
Compliance with ethical standards
Conflict of interest The authors declare that they have no conflict of
interest.
References
Abrahamczyk S, Kluge J, Gareca Y et al (2011) The influence of cli-
matic seasonality on the diversity of different tropical pollina-
tor groups. PLoS ONE 6:e27115. https​://doi.org/10.1371/journ​
al.pone.00271​15
Alves RJV, Silva NG, Oliveira JA, Medeiros D (2014) Circumscribing
campo rupestre: megadiverse Brazilian rocky montane savanas.
Braz J Biol 74:355–362. https:​ //doi.org/10.1590/1519-6984.23212​
Banks-Leite C, Ewers RM, Metzger JP (2012) Unraveling the driv-
ers of community dissimilarity and species extinction in
fragmented landscapes. Ecology 93:2560–2569. https​: //doi.
org/10.1890/11-2054.1
Barbosa O, Marquet PA, Bacigalupe LD et al (2010) Interac-
tions among patch area, forest structure and water fluxes in a
13
Content courtesy of Springer Nature, terms of use apply. Rights reserved.
Journal of Insect Conservation (2020) 24:17–34
fog-inundated forest ecosystem in semi-arid Chile. Funct Ecol
24:909–917. https​://doi.org/10.1111/j.1365-2435.2010.01697​.x
Baselga A (2010) Partitioning the turnover and nestedness components
of beta diversity. Glob Ecol Biogeogr 19:134–143. https​://doi.org
/10.1111/j.1466-8238.2009.00490​.x
Baselga A (2012) The relationship between species replace-
ment, dissimilarity derived from nestedness, and nestedness.
Glob Ecol Biogeogr 21:1223–1232. https ​ : //doi.org/10.111
1/j.1466-8238.2011.00756​.x
Basset Y, Cizek L, Cuenoud P et al (2012) Arthropod diversity in a
tropical forest. Science 338:1481–1484. https​://doi.org/10.1126/
scien​ce.12267​27
Benedick S, Hill JK, Mustaffa N et al (2006) Impacts of rain forest
fragmentation on butterflies in northern Borneo: species richness,
turnover and the value of small fragments. J Appl Ecol 43:967–
977. https​://doi.org/10.1111/j.1365-2664.2006.01209​.x
Biswas SR, Wagner HH (2012) Landscape contrast: a solution to hid-
den assumptions in the metacommunity concept? Landsc Ecol
27:621–631. https​://doi.org/10.1007/s1098​0-012-9732-5
Bolker BM, Brooks ME, Clark CJ et al (2009) Generalized linear
mixed models: a practical guide for ecology and evolution. Trends
Ecol Evol 24:127–135. https:​ //doi.org/10.1016/j.tree.2008.10.008
Breiman L (2001) Random forests. Mach Learn 45:5–32. https​://doi.
org/10.1023/A:10109​33404​324
Brotons L, Mönkkönen M, Martin JL (2003) Are fragments islands?
Landscape context and density-area relationships in boreal forest
birds. Am Nat 162:343–357. https​://doi.org/10.1086/37688​7
Carvalho Guimarães CD, Viana JPR, Cornelissen T (2014) A meta-
analysis of the effects of fragmentation on herbivorous insects.
Environ Entomol 43:537–545. https​://doi.org/10.1603/EN131​90
Coddington JA, Agnarsson I, Miller JA et al (2009) Undersampling
bias: the null hypothesis for singleton species in tropical arthro-
pod surveys. J Anim Ecol 78:573–584. https​://doi.org/10.111
1/j.1365-2656.2009.01525​.x
Coelho MS, Fernandes GW, Pacheco P et al (2016) Archipelago of
montane forests surrounded by rupestrian grasslands: new insights
and perspectives. In: Fernandes GW (ed) Ecology and conser-
vation of mountaintop grasslands in Brazil. Springer, Cham, pp
129–156
Coelho MS, Carlos PP, Pinto VD et al (2018a) Connection between
tree functional traits and environmental parameters in an archi-
pelago of montane forests surrounded by rupestrian grasslands.
Flora Morphol Distrib Funct Ecol Plants 238:51–59. https​://doi.
org/10.1016/j.flora​.2017.04.003
Coelho MS, Neves FS, Perillo LN et al (2018b) Forest archipelagos:
a natural model of metacommunity under the threat of fire. Flora
Morphol Distrib Funct Ecol Plants 238:244–249. https​://doi.
org/10.1016/j.flora​.2017.03.013
Cook WM, Lane KT, Foster BL, Holt RD (2002) Island theory, matrix
effects and species richness patterns in habitat fragments. Ecol
Lett 5:619–623. https​://doi.org/10.1046/j.1461-0248.2002.00366​
.x
Cook WM, Anderson RM, Schweiger EW (2004) Is the matrix really
inhospitable? Vole runway distribution in an experimentally
fragmented landscape. Oikos 104:5–14. https​://doi.org/10.111
1/j.0030-1299.2004.12761​.x
Corrêa MM, Fernandes GW, Leal IR (2006) Diversidade de formigas
epigéicas (Hymenoptera: Formicidae) em capões do pantanal sul
matogrossense: relações entre riqueza de espécies e complexi-
dade estrutural da área. Neotrop Entomol 35:724–730. https:​ //doi.
org/10.1590/S1519​-566X2​00600​06000​02
Crawley MJ (2013) The R book, 2nd edn. Wiley, Chichester
Cuissi RG, Lasmar CJ, Moretti TS et al (2015) Ant community in natu-
ral fragments of the Brazilian wetland: species–area relation and
isolation. J Insect Conserv 19:531–537. https​://doi.org/10.1007/
s1084​1-015-9774-5
Journal of Insect Conservation (2020) 24:17–34
da Silva PG, Nunes CA, Ferreira LF et al (2019) Patch and landscape
effects on forest-dependent dung beetles are masked by matrix-
tolerant dung beetles in a mountaintop rainforest archipelago.
Sci Total Environ 651:1321–1331. https​: //doi.org/10.1016/j.
scito​tenv.2018.09.195
Dall’Oglio OT, Ribeiro RC, De Souza RF et al (2016) Can the under-
story affect the hymenoptera parasitoids in a eucalyptus plan-
tation? PLoS ONE 11:e0151165. https​://doi.org/10.1371/journ​
al.pone.01511​65
Derraik JGB, Early JW, Closs GP, Dickinson KJM (2010) Morphos-
pecies and taxonomic species comparison for hymenoptera. J
Insect Sci 10:1–7. https​://doi.org/10.1673/031.010.10801​
Dormann CF, Elith J, Bacher S et al (2013) Collinearity: a review
of methods to deal with it and a simulation study evaluating
their performance. Ecography 36:027–046. https​://doi.org/10.
1111/j.1600-0587.2012.07348​.x
Driscoll DA (2005) Is the matrix a sea? Habitat specificity in a natu-
rally fragmented landscape. Ecol Entomol 30:8–16. https​://doi.
org/10.1111/j.0307-6946.2005.00666​.x
Driscoll DA (2008) The frequency of metapopulations, metacom-
munities and nestedness in a fragmented landscape. Oikos
117:297–309. https​://doi.org/10.1111/j.2007.0030-1299.16202​
.x
Fernández F, Sharkey MJ (2006) Introducción a los hymenoptera de
la región neotropical. Editora Guadalupe Ltda, Bogotá
Frazer GW, Canham CD, Lertzman KP (1999) Gap light analyzer
(GLA), version 2.0: imaging software to extract canopy struc-
ture and gap light transmission indices from true-colour fisheye
photographs, users manual and program documentation. ©1999
Simon Fraser Univ. Burn. Br. Columbia, Inst. Ecosyst. Stud.
Millbrook, New York
Giulietti AM, Pirani JR, Harley RM (1997) Espinhaço range
region—Eastern Brazil. In: Davis SD, Heywood VH, Herrera-
MacBryde O et al (eds) Centres of plant diversity: a guide and
strategy for their conservation -, vol 3. The Americas. WWF/
IUCN Publications Unit, Cambridge, pp 397–404
Haddad NM, Holt RD, Fletcher RJ et al (2017) Connecting mod-
els, data, and concepts to understand fragmentation’s ecosys-
tem-wide effects. Ecography 40:1–8. https​://doi.org/10.1111/
ecog.02974​
Haila Y (2002) A conceptual genealogy of fragmentation research:
from island biogeography to landscape ecology. Ecol Appl
12:321–334
Hopper SD, Silveira FAO, Fiedler PL (2016) Biodiversity hotspots
and Ocbil theory. Plant Soil 403:167–216. https:​ //doi.org/10.1007/
s1110​4-015-2764-2
Inclán DJ, Cerretti P, Marini L (2014) Interactive effects of area and
connectivity on the diversity of tachinid parasitoids in highly
fragmented landscapes. Landsc Ecol 29:879–889. https​://doi.
org/10.1007/s1098​0-014-0024-0
Jamoneau A, Chabrerie O, Closset-Kopp D, Decocq G (2012) Frag-
mentation alters beta-diversity patterns of habitat specialists
within forest metacommunities. Ecography 35:124–133. https​://
doi.org/10.1111/j.1600-0587.2011.06900​.x
Julião GR, Amaral MEC, Fernandes GW, Oliveira EG (2004) Edge
effect and species–area relationships in the gall-forming insect
fauna of natural forest patches in the Brazilian Pantanal. Biodivers
Conserv 13:2055–2066. https​://doi.org/10.1023/B:BIOC.00000​
40006​.81958​.f2
Leibold MA, Holyoak M, Mouquet N et al (2004) The metacommunity
concept: a framework for multi-scale community ecology. Ecol
Lett 7:601–613. https​://doi.org/10.1111/j.1461-0248.2004.00608​
.x
Levins R (1969) Some demographic and genetic consequences of envi-
ronmental heterogeneity for biological control. Bull Entomol Soc
Am 15:237–240. https​://doi.org/10.1093/besa/15.3.237
Content courtesy of Springer Nature, terms of use apply. Rights reserved.
33
Lion MB, Garda AA, Santana DJ, Fonseca CR (2016) The conserva-
tion value of small fragments for atlantic forest reptiles. Bio-
tropica 48:265–275. https​://doi.org/10.1111/btp.12277​
MacArthur RH, Wilson OE (1967) The theory of island biogeogra-
phy. Princeton University Press, Princeton
Macedo-Reis LE, Quesada M, de Siqueira Neves F (2019) Forest
cover drives insect guild diversity at different landscape scales
in tropical dry forests. For Ecol Manage 443:36–42. https​://doi.
org/10.1016/j.forec​o.2019.04.007
Magrach A, Laurance WF, Larrinaga AR, Santamaria L (2014)
Meta-analysis of the effects of forest fragmentation on inter-
specific interactions. Conserv Biol 28:1342–1348. https​://doi.
org/10.1111/cobi.12304​
McGarigal K, Cushman SA, Ene E (2012) FRAGSTATS v4: spatial
pattern analysis program for categorical and continuous maps
Mello-Silva R, Santos DYAC, Salatino MLF et al (2011) Five
vicarious genera from Gondwana: the Velloziaceae as shown
by molecules and morphology. Ann Bot 108:87–102. https​://
doi.org/10.1093/aob/mcr10​7
Mendenhall CD, Karp DS, Meyer CFJ et al (2014) Predicting biodi-
versity change and averting collapse in agricultural landscapes.
Nature 509:213–217. https​://doi.org/10.1038/natur​e1313​9
Misiewicz TM, Kraichak E, Rasmussen C (2014) Distance and
habitat drive fine scale stingless bee (Hymenoptera: Apidae)
community turnover across naturally heterogeneous forests in
the Western Amazon. Sociobiology 61:407–414. https​ ://doi.
org/10.13102​/socio​biolo​gy.v61i4​.407-414
Mota GS, Luz GR, Mota NM et al (2018) Changes in species com-
position, vegetation structure, and life forms along an altitudi-
nal gradient of rupestrian grasslands in south-eastern Brazil.
Flora Morphol Distrib Funct Ecol Plants 238:32–42. https​://
doi.org/10.1016/j.flora​.2017.03.010
Nassar JM, Rodríguez JP, Sánchez-Azofeifa A et al (2008) Manual
of methods: human, ecological and biophysical dimensions of
tropical dry forests. Gráficas Lauki C.A, Caracas
Neves FS, Oliveira VHF, Espírito-Santo MM et al (2010) Succes-
sional and seasonal changes in a community of dung beetles
(Coleoptera: Scarabaeinae) in a Brazilian tropical dry forest.
Braz J Nat Conserv 8:160–164. https​://doi.org/10.4322/natco​
n.00802​009
Nogueira AA, Pinto-da-Rocha R (2016) The effects of habitat size
and quality on the orb-weaving spider guild (Arachnida: Araneae)
in an atlantic forest fragmented landscape. J Arachnol 44:36–45.
https​://doi.org/10.1636/P15-19.1
Novais SMA, Nunes CA, Santos NB et al (2016) Effects of a possible
pollinator crisis on food crop production in Brazil. PLoS ONE
11:1–12. https​://doi.org/10.1371/journ​al.pone.01672​92
Nunes CA, Braga RF, Figueira JEC et al (2016) Dung beetles along
a tropical altitudinal gradient: environmental filtering on taxo-
nomic and functional diversity. PLoS ONE 11:e0157442. https​://
doi.org/10.1371/journ​al.pone.01574​42
Öckinger E, Lindborg R, Sjödin NE, Bommarco R (2012) Land-
scape matrix modifies richness of plants and insects in grass-
land fragments. Ecography 35:259–267. https​://doi.org/10.111
1/j.1600-0587.2011.06870​.x
Oles A, Pau G, Smith M, et al (2012) EBImage: 4.10.1., community
ecology package. R package version
Oliver I, Beattie AJ (1996) Invertebrate morphospecies as surrogates
for species: a case study. Conserv Biol 10:99–109. https​://doi.org
/10.1046/j.1523-1739.1996.10010​099.x
Oliver I, Dorrough J, Doherty H, Andrew NR (2016) Additive and
synergistic effects of land cover, land use and climate on insect
biodiversity. Landsc Ecol 31:2415–2431. https​://doi.org/10.1007/
s1098​0-016-0411-9
Osborne JL, Clark SJ, Morris RJ et al (1999) A landscape-scale
study of bumble bee foraging range and constancy, using
13
34
harmonic radar. J Appl Ecol 36:519–533. https​://doi.org/10.104
6/j.1365-2664.1999.00428​.x
Patiño J, Whittaker RJ, Borges PAV et al (2017) A roadmap for island
biology: 50 fundamental questions after 50 years of the theory
of island biogeography. J Biogeogr 44:963–983. https​://doi.
org/10.1111/jbi.12986​
Pereira GCN, Coelho MS, Beirão MV et al (2017) Diversity of fruit-
feeding butterflies in a mountain archipelago of rainforest. PLoS
ONE 12:e0180007. https:​ //doi.org/10.1371/journa​ l.pone.018000​ 7
Perillo LN, Neves FS, Antonini Y, Martins RP (2017) Compositional
changes in bee and wasp communities along neotropical moun-
tain altitudinal gradient. PLoS ONE 12:e0182054. https​://doi.
org/10.1371/journ​al.pone.01820​54
Prugh LR, Hodges KE, Sinclair ARE, Brashares JS (2008) Effect of
habitat area and isolation on fragmented animal populations. Proc
Natl Acad Sci USA 105:20770–20775. https​://doi.org/10.1073/
pnas.08060​80105​
R Core Team (2017) R: a language and environment for statistical
computing. R Foundation for Statistical Computing, Vienna
Rocha NMWB, Carstensen DW, Fernandes GW et al (2016) Phenol-
ogy patterns across a rupestrian grassland altitudinal gradient. In:
Fernandes GW (ed) Ecology and conservation of mountaintop
grasslands in Brazil. Springer, Cham, pp 275–289
Rossetti MR, Tscharntke T, Aguilar R, Batáry P (2017) Responses
of insect herbivores and herbivory to habitat fragmentation: a
hierarchical meta-analysis. Ecol Lett 20:264–272. https​://doi.
org/10.1111/ele.12723​
Silva NAP, Frizzas MR, Oliveira CM (2011) Seasonality in insect
abundance in the “Cerrado” of Goiás State, Brazil. Rev Bras
Entomol 55:79–87. https​://doi.org/10.1590/S0085​-56262​01100​
01000​13
Silveira FA, Melo GAR, Almeida EAB (2002) Abelhas Brasileiras:
sistemática e identificação. Belo Horizonte
Silveira FAO, Negreiros D, Barbosa NPU et al (2016) Ecology and
evolution of plant diversity in the endangered campo rupestre: a
neglected conservation priority. Plant Soil 403:129–152. https​://
doi.org/10.1007/s1110​4-015-2637-8
Soares SA, Suarez YR, Fernandes WD et al (2013) Temporal variation
in the composition of ant assemblages (Hymenoptera, Formicidae)
on trees in the Pantanal floodplain, Mato Grosso do Sul, Bra-
zil. Rev Bras Entomol 57:84–90. https​://doi.org/10.1590/S0085​
-56262​01300​01000​13
Solar RRC, Barlow J, Ferreira J et al (2015) How pervasive is biotic
homogenization in human-modified tropical forest landscapes?
Ecol Lett 18:1108–1118. https​://doi.org/10.1111/ele.12494​
Suni SS, Bronstein JL, Brosi BJ (2014) Spatio-temporal genetic
structure of a tropical bee species suggests high dispersal over
a fragmented landscape. Biotropica 46:202–209. https​://doi.
org/10.1111/btp.12084​
Thompson PL, Rayfield B, Gonzalez A (2017) Loss of habitat and con-
nectivity erodes species diversity, ecosystem functioning, and sta-
bility in metacommunity networks. Ecography 40:98–108. https​
://doi.org/10.1111/ecog.02558​
13
Content courtesy of Springer Nature, terms of use apply. Rights reserved.
Journal of Insect Conservation (2020) 24:17–34
Tscharntke T, Steffan-Dewenter I, Kruess A, Thies C (2002) Char-
acteristics of insect populations on habitat fragments: a
mini review. Ecol Res 17:229–239. https ​ : //doi.org/10.104
6/j.1440-1703.2002.00482​.x
Tscharntke T, Tylianakis JM, Rand TA et al (2012) Landscape mod-
eration of biodiversity patterns and processes: eight hypoth-
eses. Biol Rev 87:661–685. https​: //doi.org/10.1111/j.1469-
185X.2011.00216​.x
Tuomisto H (2010) A diversity of beta diversities: Straightening up a
concept gone awry. Part 1. Defining beta diversity as a function
of alpha and gamma diversity. Ecography 33:2–22. https:​ //doi.org
/10.1111/j.1600-0587.2009.05880​.x
Tylianakis JM, Klein AM, Tscharntke T (2005) Spatiotemporal varia-
tion in the effects of a tropical habitat gradient on Hymenoptera
diversity. Ecology 86:3296–3302. https:​ //doi.org/10.1890/05-0371
Van Nouhuys S, Hanski I (2002) Colonization rates and distances
of a host butterfly and two specific parasitoids in a fragmented
landscape. J Anim Ecol 71:639–650. https​: //doi.org/10.104
6/j.1365-2656.2002.00627​.x
Vieira L, Lopes FS, Fernandes WD, Raizer J (2008) Comunidade de
carabidae (Coleoptera) em manchas florestais no pantanal, Mato
Grosso do Sul, Brasil. Iheringia Série Zool 98:317–324. https​://
doi.org/10.1590/S0073​-47212​00800​03000​05
Walters BB, Stiles EW (1996) Effect of canopy gaps and flower patch
size on pollinator visitation of Impatiens capensis. Bull Torrey
Bot Club 123:184–188. https​://doi.org/10.2307/29967​93
Watling JI, Donnelly MA (2006) Fragments as islands: a synthesis of
faunal responses to habitat patchiness. Conserv Biol 20:1016–
1025. https​://doi.org/10.1111/j.1523-1739.2006.00482​.x
Whittaker RH (1972) Evolution and measurement of species diversity.
Taxon 21:213–251. https​://doi.org/10.2307/12181​90
Wolda H (1988) Insect seasonality: why? Annu Rev Ecol Syst 19:1–18.
https​://doi.org/10.1146/annur​ev.es.19.11018​8.00024​5
Yekwayo I, Pryke JS, Roets F, Samways MJ (2016) Surrounding
vegetation matters for arthropods of small, natural patches of
indigenous forest. Insect Conserv Divers 9:224–235. https​://doi.
org/10.1111/icad.12160​
Zehm A, Nobis M, Schwabe A (2003) Multiparameter analysis of
vertical vegetation structure based on digital image processing.
Flora Morphol Distrib Funct Ecol Plants 198:142–160. https​://
doi.org/10.1078/0367-2530-00086​
Zurbuchen A, Landert L, Klaiber J et al (2010) Maximum foraging
ranges in solitary bees: only few individuals have the capability
to cover long foraging distances. Biol Conserv 143:669–676. https​
://doi.org/10.1016/j.bioco​n.2009.12.003
Publisher’s Note Springer Nature remains neutral with regard to
jurisdictional claims in published maps and institutional affiliations.
Terms and Conditions
Springer Nature journal content, brought to you courtesy of Springer Nature Customer Service Center GmbH (“Springer Nature”).
Springer Nature supports a reasonable amount of sharing of research papers by authors, subscribers and authorised users (“Users”), for small-
scale personal, non-commercial use provided that all copyright, trade and service marks and other proprietary notices are maintained. By
accessing, sharing, receiving or otherwise using the Springer Nature journal content you agree to these terms of use (“Terms”). For these
purposes, Springer Nature considers academic use (by researchers and students) to be non-commercial.
These Terms are supplementary and will apply in addition to any applicable website terms and conditions, a relevant site licence or a personal
subscription. These Terms will prevail over any conflict or ambiguity with regards to the relevant terms, a site licence or a personal subscription
(to the extent of the conflict or ambiguity only). For Creative Commons-licensed articles, the terms of the Creative Commons license used will
apply.
We collect and use personal data to provide access to the Springer Nature journal content. We may also use these personal data internally within
ResearchGate and Springer Nature and as agreed share it, in an anonymised way, for purposes of tracking, analysis and reporting. We will not
otherwise disclose your personal data outside the ResearchGate or the Springer Nature group of companies unless we have your permission as
detailed in the Privacy Policy.
While Users may use the Springer Nature journal content for small scale, personal non-commercial use, it is important to note that Users may
not:

1. use such content for the purpose of providing other users with access on a regular or large scale basis or as a means to circumvent access
control;
2. use such content where to do so would be considered a criminal or statutory offence in any jurisdiction, or gives rise to civil liability, or is
otherwise unlawful;
3. falsely or misleadingly imply or suggest endorsement, approval , sponsorship, or association unless explicitly agreed to by Springer Nature in
writing;
4. use bots or other automated methods to access the content or redirect messages
5. override any security feature or exclusionary protocol; or
6. share the content in order to create substitute for Springer Nature products or services or a systematic database of Springer Nature journal
content.
In line with the restriction against commercial use, Springer Nature does not permit the creation of a product or service that creates revenue,
royalties, rent or income from our content or its inclusion as part of a paid for service or for other commercial gain. Springer Nature journal
content cannot be used for inter-library loans and librarians may not upload Springer Nature journal content on a large scale into their, or any
other, institutional repository.
These terms of use are reviewed regularly and may be amended at any time. Springer Nature is not obligated to publish any information or
content on this website and may remove it or features or functionality at our sole discretion, at any time with or without notice. Springer Nature
may revoke this licence to you at any time and remove access to any copies of the Springer Nature journal content which have been saved.
To the fullest extent permitted by law, Springer Nature makes no warranties, representations or guarantees to Users, either express or implied
with respect to the Springer nature journal content and all parties disclaim and waive any implied warranties or warranties imposed by law,
including merchantability or fitness for any particular purpose.
Please note that these rights do not automatically extend to content, data or other material published by Springer Nature that may be licensed
from third parties.
If you would like to use or distribute our Springer Nature journal content to a wider audience or on a regular basis or in any other manner not
expressly permitted by these Terms, please contact Springer Nature at

onlineservice@springernature.com