Universidade de São Paulo
Biblioteca Digital da Produção Intelectual - BDPI
Departamento de Ecologia - IB/BIE
Artigos e Materiais de Revistas Científicas - IB/BIE
2014-04-16
Island isolation reduces genetic diversity and
connectivity but does not significantly elevate
diploid male production in a neotropical orchid
bee
Conservation Genetics, Dordrecht, v.15, n.5,p.1123-1135, 2014
http://www.producao.usp.br/handle/BDPI/44764
Downloaded from: Biblioteca Digital da Produção Intelectual - BDPI, Universidade de São Paulo
Conserv Genet (2014) 15:1123–1135
DOI 10.1007/s10592-014-0605-0
RESEARCH ARTICLE
Island isolation reduces genetic diversity and connectivity but does
not significantly elevate diploid male production in a neotropical
orchid bee
Samuel Boff • Antonella Soro • Robert J. Paxton
Isabel Alves-dos-Santos
•
Received: 3 December 2013 / Accepted: 29 March 2014 / Published online: 16 April 2014
Ó Springer Science+Business Media Dordrecht 2014
Abstract There is concern about the worldwide decline
of bees, in which genetic factors may play a role. As
populations of these haplodiploid insects suffer habitat
fragmentation and subsequent isolation, theory predicts an
increase in inbreeding and a concomitant increase in
inviable or sterile diploid males, a product of reduced
allelic diversity at the sex determining locus, which could
lead to a diploid male extinction vortex. To test this idea,
we genotyped 1,245 males of one orchid bee, Euglossa
cordata, a species with low diploid male production on the
mainland. We genotyped bees from the Brazilian mainland
and three islands using five highly variable microsatellite
Electronic supplementary material The online version of this
article (doi:10.1007/s10592-014-0605-0) contains supplementary
material, which is available to authorized users.
S. Boff
Departamento de Biologia, Faculdade de Filosofia Ciências e
Letras de Ribeirão Preto, Universidade de São Paulo, Av.
Bandeirantes, 3900, Bloco 2, Ribeirão Prêto, SP 14040-901,
Brazil
S. Boff (&) A. Soro R. J. Paxton
General Zoology, Institute for Biology,
Martin-Luther-University Halle-Wittenberg,
Hoher Weg 8, 06120 Halle (Saale), Germany
e-mail: [email protected]
R. J. Paxton
German Centre for Integrative Biodiversity Research (iDiv),
Halle-Jena-Leipzig, Deutscher Platz 5e, 04103 Leipzig,
Germany
I. Alves-dos-Santos
Departamento de Ecologia, Instituto de Biociências,
Universidade de São Paulo, Rua do Matão 321,
trav 14. Cidade Universitária, São Paulo, SP 05508-900, Brazil
loci. Allelic richness was lowest on the most remote island
38 km from the mainland and, though the degree of differentiation across localities was modest (global FST =
0
0.034; global GST ¼ 0:153, both P \ 0.001) and isolation
by distance was weak (Mantel test, r = 0.614, P = 0.056),
sea was revealed to be a significant barrier to inferred
historic gene flow (partial Mantel test of distance over sea,
r = 0.831, P = 0.003). Only seven males were diploid
(mean diploid male production, DMP, 0.6 %). Though the
proportion of diploid males was highest on the most remote
island (1.3 %), differences in DMP across study areas were
statistically non-significant. Thus island isolation leads to
reduced genetic variation at putatively neutral microsatellite loci, but E. cordata nevertheless seems to maintain
allelic diversity at the sex locus, possibly because of sufficient gene flow, or because it is a locus under balancing
selection, or because of the joint action of these two evolutionary forces: migration and selection. These and other
bee species may be able to maintain sufficient variability at
the sex locus to avoid entering the DM extinction vortex,
even on relatively isolated islands or habitat fragments.
Keywords Complementary sex determination Euglossini Euglossa cordata Gene flow Genetic
structure Microsatellite
Introduction
Habitat fragmentation and loss are among the main factors
responsible for observed declines in biodiversity (Dobson
et al. 2006; Cardinale et al. 2012). As populations decrease
in size, genetic factors may play a role in accelerating
extinction; examples include inbreeding depression and
123
1124
reduced adaptability from loss of genetic diversity (Brook
et al. 2002; Frankham 1995, 2003; O’Grady et al. 2008;
Spielman et al. 2004). Concern over the role of genetics in
conservation has prompted a plethora of population genetic
studies to assess the genetic diversity of populations of
species increasingly exposed to habitat fragmentation and
loss (e.g. Allendorf and Luikart 2007; Jordan and Snell
2008; Darvill et al. 2010; Davis et al. 2010).
Islands surrounded by water (in contrast to habitat
islands) deserve special attention as an example of naturally fragmented areas. Island populations are more susceptible to extinction than mainland populations, whether
threatened by alien species or habitat loss (Frankham 1995,
1998; Gerlach 2008; Dennis et al. 2008; New 2008), and
may be more sensitive to genetic factors (bottlenecks and
genetic drift) causing population decline as consequence of
loss of genetic variation and inbreeding than mainland
populations because of small effective population size, Ne,
and reduced immigration (Frankham 1997). Loss of genetic
variation means that populations may be less equipped to
evolve in response to environmental change whilst
inbreeding is associated with inbreeding depression in
normally outbred populations (Frankham et al. 2004). It is
in acknowledgement of these factors that the IUCN has
recognised genetic diversity as one level of biodiversity
deserving conservation (McNeely et al. 1990). True islands
represent a suitable system to study the impact of fragmentation on populations because, for terrestrial organisms, islands have no surrounding habitat to assist
migrants. Furthermore, island populations surrounded by
water conform more closely to the ‘island model’ of Sewall
Wright and therefore population genetic data can be more
directly interpreted within this consistent body of theoretical work (e.g. Wright 1965, 1969; Frankham 1997, 1998;
Frankham et al. 2004). Bees on islands have often served as
useful subjects for studying the effects of isolation on
genetic diversity and differentiation with respect to the
‘island model’ (e.g. Darvill et al. 2006; Davis et al. 2010).
Bees are ecologically important because of the ecosystem service they provide as pollinators of wild and cultivated plants (Kearns et al. 1998), including ca. 70 % of
world crops (Klein et al. 2007; Gallai et al. 2009). Yet there
is growing evidence for their decline worldwide (Brown
and Paxton 2009). The extent to which the loss of genetic
diversity of bee populations impacts on their survival
(Zayed et al. 2004, Zayed 2009) is an open question. On
the one side, haplodiploids organism such as bees could
have an advantage with regards to small population size in
that their genetic load could be purged through exposure of
deleterious alleles in the haploid sex. On the other side,
their system of sex determination predisposes them to the
production of inviable or sterile so-called diploid males
though inbreeding (e.g. Cook 1993; Cook and Crozier
123
Conserv Genet (2014) 15:1123–1135
1995). Like the majority of Hymenoptera, bees are haplodiploid with a sex determination system termed arrhenotokous parthenogenesis (arrhenotoky) (Crozier and Pamilo
1996). Females are usually produced from fertilized eggs
while males are produced from unfertilized eggs. At the
genic level, sex in the honey bee (Apis mellifera) is
determined by a single locus, the sex determination locus
(Beye et al. 2003; csd or sex locus), and so-called single
locus complementary sex determination (sl-CSD) is considered widespread in the Hymenoptera, including bees
(Cook 1993; van Wilgenburg et al. 2006; Harpur et al.
2013), though not in the Hymenoptera Parasitica (Heimpel
and de Boer 2008; Verhulst et al. 2010), where multilocus
CSD is also found (de Boer et al. 2008). In the case of
species with sl-CSD, diploid individuals heterozygous at
the sex locus develop into females and haploid individuals
into males. Diploid male production (DMP) can therefore
be considered a negative consequence of inbreeding
(Zayed and Packer 2005). Yet it is currently unclear what
the impact of habitat fragmentation is in reducing genetic
diversity and promoting inbreeding and DMP in these
insects.
Diploid males therefore increase the genetic load in bee
populations (and other haplodiploids with sl-CSD) brought
about through inbreeding and reduced allelic diversity at
the sex locus. In some haplodiploid species, DMP may
skew the sex ratio towards males (Cowan and Stahlhut
2004; Elias et al. 2010), theoretically restricting population
growth (Stouthamer et al. 1992) and further lowering
genetic diversity. As the number of alleles at the sex locus
decreases, DMP may increase through a higher probability
of matched matings (mating between partners sharing the
same allele at the sex locus), further reducing both allelic
diversity at the sex locus and population size, leading to a
diploid male (DM) extinction vortex (Zayed and Packer
2005). Gene flow and the influx of new alleles at the sex
locus as well as a high rate of reproduction, a fixed sex
ratio and lack of mate choice can theoretically rescue a
population entering the DM extinction vortex (Hein et al.
2009).
Diploid males are detected through genetic analysis as
males heterozygous at one or more selectively neutral
nuclear markers. They have already been described in over
60 Hymenopteran species (van Wilgenburg et al. 2006;
Heimpel and de Boer 2008), including several bee species
(e.g. Paxton et al. 2000, Zayed and Packer 2001; Darvill
et al. 2006; Tavares et al. 2010; Whidden and Owen 2011),
and including orchid bees (Euglossini: Takahashi et al.
2008; Souza et al. 2010; Freiria et al. 2012). Euglossine
bees (Hymenoptera, Apidae, Euglossini) are endemic to the
Neotropics, where they are important pollinators of many
plant families, particularly orchids, to which males are
attracted (Jansen 1971, Dressler 1982, Roubik and Hanson
Conserv Genet (2014) 15:1123–1135
2004). The use of synthetic scents similar to those produced by orchids can be used to sample male euglossines,
and this facilitates their study (e.g. Cordeiro et al. 2013).
In populations of euglossine bees, the frequency of
diploid males has been found to be high in studies
employing allozyme markers (Roubik et al. 1996; Zayed
et al. 2004; López-Uribe et al. 2007) yet those employing
microsatellites suggest low frequencies of DMs (Cerântola
et al. 2010; Souza et al. 2010; Zimmermann et al. 2011;
Freiria et al. 2012). According to Souza et al. (2010), allozyme-based genotyping for the detection of diploid males
may be flawed because of allele mis-scoring, possibly due
to protein instability. DNA is more stable and therefore
microsatellite genotyping more reliable for the detection of
diploid males (Schlötterer 2004).
Recent genetic (microsatellite) analyses of mainland
populations of orchid bees, including our study species
Euglossa cordata, suggest that DMP is very low and
inferred historic gene flow very high, with little population
genetic differentiation over 10’s to 100’s of km (Souza
et al. 2010; Zimmermann et al. 2011); in one study, E.
cordata showed weak, albeit significant, genetic differentiation across 310 km of continental São Paulo state (Souza
et al. 2010) and, in another study in the same state, there
was a lack of differentiation across [500 km (Cerântola
et al. 2010). But it is unclear whether low DMP in E.
cordata is because suitable habitat is widespread, i.e.
mainland populations are not fragmented, or because
individuals are highly mobile and their populations are
large, facilitating dispersal and population connectivity, or
because the sex locus is under balancing selection, facilitating the maintenance of allelic diversity. Another open
question is how genetic diversity at neutral genetic markers
varies with DMP; island populations of the bumble bee
Bombus muscorum show reduced genetic variation and
contain diploid males (Darvill et al. 2006), suggesting that
DMP may be correlated with a population’s genetic
variability.
Here we studied populations of E. cordata, an orchid
bee species that is broadly distributed and abundant in
southeastern Brazil, to investigate the impact of island
isolation on genetic variation and DMP. Islands are ideal
for this purpose as the intervening water is unsuitable for
these insects, so fragmentation is easily defined, and
because other bee species (a bumble bee) are thought to
exhibit elevated DMP on islands (Darvill et al. 2006). We
tested the following hypotheses: (1) genetic diversity is
lower in island populations and decreases with geographic
distance of an island from the mainland; (2) population
structure is higher in island populations and increases with
geographic distance from the mainland; and (3) the frequency of diploid males is higher on smaller and more
isolated islands.
1125
Materials and methods
Study areas
The study was carried out on three islands (Ilhabela, Ilha de
Búzios and Ilha da Vitória, part of the Ilhabela State Park,
and hereafter termed Ilhabela, Búzios and Vitória) and at
the adjacent mainland (São Sebastião) of the north coast of
the state of São Paulo, Brazil (Fig. 1; Tables 1, S1). The
three islands originated as hilltops that became surrounded
by water through rising sea levels after the last period of
glaciation during the late Quaternary-Pleistocene, approximately 12–20,000 years ago (Suguio and Matin Suguio
and Martin 1987, Ângelo 1989). Islands are separated from
the mainland by 2-38 km and from each other by 7 km
(Ilhabela to Búzios) to 11 km (Búzios to Vitória) of open
water (Fig. 1). The other island in the vicinity, Ilha Anchieta, is large (828 Ha), only 600 m from the mainland,
and much further from any of our study islands (minimum
20 km; see Fig. 1).
Sampling
A total of 1,305 male E. cordata was sampled across the 4
land masses (the three islands and the mainland) comprising 14 sampling locations (Fig. 1) between 2007 and
2010 (Tables 1, S1). Sampling was carried out on sunny
days, from 8:00 to 15:00; E. cordata males were collected
in flight when attracted to the following chemical baits:
amyl acetate, phenyl alcohol, ethyl acetate, benzyl benzoate, methyl benzoate, beta-ionone, methyl butyrate, methyl
cinnamate, cineole, eucalyptol, eugenol, myrcene, methyl
salicylate and vanillin.
Sample processing
Among the Euglossini, only males are reported to be
attracted to aromatic compounds (Roubik and Hanson
2004). All sampled individuals were nevertheless sexed as
males on the basis of their greatly expanded and convex
rear tibiae, an obvious morphological and sexually
dimorphic character. Three legs were removed from the
same side of all collected specimens, preserved in 99 %
ethanol, and stored at -20 °C. The rest of the specimens
were prepared for dry preservation, identified and deposited in the Entomological Collection Paulo Nogueira Neto
(Coleção Entomológica Paulo Nogueira Neto—CEPANN)
of the Bee Laboratory at the Institute of Biosciences,
Universidade de São Paulo in Brazil.
DNA was extracted from the legs in 95 ll of 5 %
ChelexÒ-100 (Walsh et al. 1991) with the addition of 5 ll
per sample of proteinase K (10 mg/ml) and by processing
samples in a thermocycler with the following program: 1 h
123
1126
Conserv Genet (2014) 15:1123–1135
Fig. 1 The four study areas on the northern coast of São Paulo,
southeastern Brazil, where E. cordata males were collected: mainland
(São Sebastião; localities 1–3) and three islands: Ilhabela (localities
4–12), Ilha de Búzios (locality 13), and Ilha da Vitória (locality 14).
See Supplementary Table S1 for collection dates with sample sizes
Table 1 Details of the 14
localities nested within four
study areas in São Paulo, Brazil,
from which 1,305 E. cordata
males were collected; N,
number of males sampled per
area; Ng, number of males
successfully genotyped per area;
L2n, locality (1–14) and date
(month and year) when a diploid
male was sampled
Study area
Coordinates
Island area
(ha)
Localities
N/Ng
São Sebastião
(mainland)
45°240 W,
23°450 S
Mainland
1–3
280/263
Ilhabela
45°210 W,
23°470 S
33593
L2n locality/
date
1/Mar 2009
2/Jan 2010
2/Aug 2010
4–12
612/580
4/Mar 2008
17/Nov 2009
24/Nov 2009
0
Ilha de Búzios
45°08 W,
23°480 S
755
13
328/322
Ilha da Vitória
45°010 W,
23°450 S
221
14
85/80
14/Jan 2009
1305/
1245
7 diploid males
Total
at 55 °C, 15 min at 99 °C, 1 min at 37 °C. DNA extracts
were thereafter frozen at -20 °C till required for PCR.
PCR and analysis of amplified fragments
We genotyped males at five microsatellite loci (Egc17,
Egc18, Egc24, Egc26 and Egc51) developed by Souza
123
and the footnote to Supplementary Table S1 for names of localities 1–
14. Filled circles represent the midpoint between localities 1 and 3
(mainland) and the geographic centre of Ilhabela, used in isolation by
distance analyses
et al. (2007) for E. cordata. PCRs were set up in 12.5 ll
reaction volumes comprising 19 PCR Master Mix (Promega), 0.5 U Taq DNA polymerase (Promega), 400 lM
each dNTP, 3 mM MgCl2, 400 nM of each primer, 1 ll
DNA extract (ca. 10 ng DNA) and made up to 12.5 ll with
ddH2O. PCR cycling temperatures followed Souza et al.
(2007) for E. cordata. Forward primers were end-labelled
Conserv Genet (2014) 15:1123–1135
with fluorophores and multiplexed as in Souza et al.
(2010). PCR products were resolved in a MegaBace 1000
DNA autosequencer and trace files of fragments were
inspected by eye then sized using the program Fragment
Profile, version 2.2. Samples that did not amplify or whose
fragments were ambiguous were re-amplified. Of the 1,305
individuals, 60 did not amplify at any locus, 1,245 individuals amplified at one or more loci and 96 % of 1,245
individuals amplified at C3 loci; the total rate of nonamplification of alleles was 6.2 %.
Genotyping error rates were estimated by re-genotyping
ca. 7.5 % of individuals (n = 96) following the protocol
used for the first genotyping. Genotyping error was calculated as the number of mismatches between duplicated
genotypes divided by the total number of re-genotyped
individuals (Bonin et al. 2004). In the case of a mis-match
of a homozygote, results of the first PCR were employed.
Linkage across loci was checked using Genepop after
diploidising the dataset (Raymond and Rousset 1995).
Genetic diversity
Measures of genetic diversity were calculated on our entire
dataset of 1,245 genotyped male individuals using Microsatellite Analyser (MSA 4.05; Dieringer and Schlötterer
2003). Samples were pooled into the mainland (São Sebastião) and the three islands (Ilhabela, Búzios and Vitória).
Though the mainland and largest island close to the mainland (Ilhabela) comprised multiple sampling localities, differences in allelic frequencies between localities within each
land mass and across temporal samples within a locality
were small and, for Ilhabela, largely non-significant (see
genetic differentiation below). Pooling data into the four
land masses also allowed a direct comparison with DMP, for
which the data are pooled into the same four land masses.
We used allelic richness and expected heterozygosity
(estimated with 1,000 permutations) as genetic diversity
indices to account for differences in sample size across sites.
To test for differences in allelic richness (rarefied to a
common sample size of 48, the minimum number of sampled individuals for each locus and land mass) and expected
heterozygosity while accounting for differences in variability among loci (e.g. Petit et al. 2005), we used analyses of
covariance (ANCOVA), where the covariates were the mean
allelic richness and the mean expected heterozygosity per
locus across the entire study, respectively. ANCOVA’s were
performed using SPSS (IBM SPSS Statistics v19, 2010). We
also checked for differences among sites in private alleles,
those unique to one of the four study areas, using the software GenAlEx 6.41 (Peakall and Smouse 2006). Differences
in the mean number of private alleles between study areas
were tested for significance using ANOVA in SPSS (IBM
SPSS Statistics v19, 2010).
1127
Frankham’s (1998) effective inbreeding coefficient (Fe)
for each island population was estimated as:
Fe ¼ 1 HIs =HM ;
where HIs and HM are the expected heterozygosity for
island and mainland populations, respectively (Frankham
1998).
Genetic differentiation
We also used MSA 4.05 (Dieringer and Schlötterer 2003)
to compute pairwise and global genetic differentiation, FST
(Weir and Cockerham 1984), with Bonferroni adjustment
for multiple testing. We calculated Hedrick’s (2005) pair0
wise GST in Genodive (Meirmans and Van Tienderen
0
2004) and global GST in MSA 4.05 (Dieringer and
Schlötterer 2003) to take into account the high variability
of genetic markers like microsatellites by rescaling FST to
lie between 0 and 1. Significance for all estimators of
genetic differentiation was derived by permuting across
loci and, for each locus, across pairs of study sites, using an
exact test with 1,000 permutations, and not assuming
Hardy–Weinberg equilibrium because we did not have
female (diploid) data with which to check mating structure.
For calculation of genetic differentiation, we pooled
data across localities within a land mass (within the
mainland or within an island) so as to achieve sample sizes
[60 haploid individuals and thereby obtain a more accurate assessment of differentiation independent of sampling
variance. To do so, within any one locality we pooled data
from different months that were temporally adjacent and
not significantly differentiated. Then, within a land mass
(i.e. within the mainland or within Ilhabela), we pooled
data from different localities that were not significantly
differentiated, resulting in two statistically differentiated
mainland populations (São Sebasitão A and São Sebastião
B), Ilhabela (all localities combined), Búzios and Vitória
(see Supplementary Table S1 for further details) and a
dataset of n = 1,029 males. Though this still resulted in
one small population sample, that of Vitória with n = 31
males (Supplementary Table S1), results from analyses in
which we combined all temporal samples from Vitória
(total n = 79 males) were qualitatively similar (Supplementary Table S3).
Landscape genetics
To test whether allelic frequencies varied according to a
pattern of isolation by distance (IBD), we performed
Mantel tests and partial Mantel tests as implemented in the
program XLSTAT with 10,000 randomizations using the
dataset analyzed for genetic differentiation, in which
samples had been pooled into five study sites. We tested for
123
1128
Conserv Genet (2014) 15:1123–1135
a significant correlation between a matrix of pairwise
genetic distances (transformed to FST/1 - FST) and the
logarithm of geographic (Euclidean) distance (Rousset
1997) among all pairs of localities (n = 5 sites). For site
São Sebasitão A comprising data from two localities, we
used their geographic midpoint for the calculation of geographic distance to other populations whilst for population
Ilhabela, comprising data from nine localities, we used the
island’s geographic centre (Fig. 1). To test whether sea
acted as barrier to gene flow, we used a partial Mantel test
in which a matrix of pairwise genetic distances (FST/1 - FST)
was correlated with a matrix of distance over sea (minimum distance across water between pairs of sampling
localities), while controlling for geographic distance using
a third matrix of log-transformed Euclidean distances as a
covariate. The use of sophisticated resistance surfaces
would be more helpful to estimate the resistance of open
water to dispersal than the one used here (implicitly, the
distance over water was considered directly and linearly
proportional to cost) but would require additional, independent sites (islands), which do not exist in this study
system.
Diploid males
equilibrium (HWE). Though we did not analyse females
from our study system with which to estimate mating
structure, E. cordata females from continental populations
were found not to deviate from HWE (Souza et al. 2007,
2010), and moderate levels of inbreeding having only a
weak effect on Phet (Paxton et al. 2000). Therefore error in
our assumption of HWE will have little impact on Phet. Phet
was estimated separately for the mainland and each island,
as for genetic diversity, so as to include all males genetically analysed; using the reduced dataset of n = 1,029
males, as for the analysis of population differentiation and
landscape genetics, would have led to the exclusion of two
of the seven diploid males (Supplementary Table S1).
For our entire dataset of 1,305 males, 60 failed to
amplify at any locus. In addition, a further 27, 74, 83, 66
and 136 failed to amplify at loci Egc17, Egc18, Egc24,
Egc26 and Egc51 respectively. Null alleles might account
for some or all of these missing data. To explicitly take
into account both the possibility of null alleles reducing
Phet and the smaller sample size due to lack of amplification of a clearly scorable allele, we also estimated
Phet_null, where:
!,
L
N
X
Y
2
pj :
Phet null ¼ 1 ðxi Þ þ ð1 inull Þ:inull
j¼1
Diploid males were classified as those individuals heterozygous at one or more locus (e.g. Cowan and Stahlhut
2004) whereas males with only one allele at every locus
were classified as haploid. To confirm ploidy, all putative
diploid males were re-amplified using single-plex PCRs to
confirm heterozygosity. Genotypes of diploid males could
in all cases be unambiguously scored at one or more loci as
possessing two clear peaks (alleles) of similar intensity
(height) in electropherograms. Only one diploid male was
scored as heterozygous at alleles differing by only one
repeat motif (Supplementary Table S2), though we found it
easy to score this locus (Egc24) so we feel its two alleles
are unlikely to have been confused with so-called stutter
arising from slippage during PCR (e.g. Schlötterer 1998).
The power of detecting diploid males, those heterozygous at one or more microsatellite loci, depends on the
number and heterozygosity of the loci used. We calculated
the detection power of our loci, Phet, as the probability of
an individual being heterozygous at one or more loci with
the formula:
!
L
N
X
Y
Phet ¼ 1 ðx2i Þ ;
j¼1
i1
where xi is the frequency of allele i, summation is across
N alleles of a locus and multiplication is across L loci. This
assumes that the population was at Hardy–Weinberg
123
i1
and where xi are allele frequencies revised by taking
account of null alleles, where inull is the frequency of a
putative null allele at a locus, which is calculated by
assuming that all individuals lacking a scorable allele at a
locus were diploid homozygotes for a putative null allele,
i.e. inull is calculated as the square root of the frequency of
individuals lacking a scorable PCR product at a locus, and
where pj is the number of individuals that actually amplified at locus j in a sampling site divided by the total number
of individuals analysed at that sampling site for which at
least one locus was scored. This sets an upper estimate for
inull and therefore also a lower bound for Phet_null
accounting for null alleles and a reduced sample size.
Phet_null was also estimated for the mainland and each
island separately.
We then calculated the frequency of diploid males as the
number of males heterozygous at one or more microsatellite locus (and therefore diploid), divided by the total
number of genotyped males. The binomial 95 % confidence interval (2 tailed) of the proportion of diploid males
was calculated in J.C. Pezzullo’s Interactive Stats javascript (http://statpages.org/confint.html) according to Souza
et al. (2010). We used a Fisher exact test to test for significant differences in the frequency of diploid males
across the mainland (São Sebastião) and the three islands
(Ilhabela, Búzios and Vitória).
Conserv Genet (2014) 15:1123–1135
1129
Table 2 Genetic diversity in 1,245 E. cordata males at five microsatellite loci on the mainland (SSeb, São Sebastião) and three islands (Ibela,
Ilhabela; Buz, Ilha de Búzios; Vit, Ilha da Vitória)
Area
N
SSeb
263
Ibela
580
Buz
322
Vit
80
All
1,245
Egc17
Egc18
Egc24
Egc26
Egc51
MGD (s.d.)
P
PA
He
0.860
0.774
0.672
0.810
0.614
0.76 (0.09)
Na
13
14
11
11
7
11.2 (2.7)
AR
10.71
10.05
6.22
9.24
5.38
8.8 (2.4)
He
0.863
0.795
0.699
0.835
0.544
0.76 (0.13)
Na
15
23
13
12
10
14.6 (5.0)
AR
10.52
10.72
5.86
9.15
5.80
9.4 (2.7)
He
Na
0.866
11
0.783
14
0.665
8
0.796
9
0.591
6
0.74 (0.11)
9.6 (3.0)
AR
9.97
9.76
5.15
8.25
4.96
7.9 (2.5)
BC
He
0.842
0.743
0.645
0.804
0.665
0.74 (0.08)
C
Na
10
9
4
9
6
7.6 (2.5)
AR
9.55
8.65
3.99
8.30
5.67
7.5 (2.4)
He
0.860
0.774
0.672
0.810
0.614
–
Na
17
25
16
12
10
16 (5.8)
6
AB
18
A
0
0
24
N number of genotyped males; He expected heterozygosity, Na number of alleles, AR allelic richness, MGD mean genetic diversity (mean across
loci, standard deviation in brackets), P significance of difference in AR across study localities by ANCOVA, where different letters represent
P \ 0.05; and number of private alleles (PA), number of alleles unique to that study area
Results
The average genotyping error rate for the five microsatellite loci used in this study was \3 % and loci were generally not out of linkage disequilibrium (P [ 0.05 for
almost all pairwise comparisons). The exceptions were loci
Egc18–Egc24 at location 2 and Egc18–Egc51 at location 3,
likely an effect of small sample size.
Genetic diversity
All five loci were highly variable, with an average of 16
alleles per locus (Table 2). ANCOVA controlling for variation in variability across loci revealed that allelic richness
was significantly different among the four land masses
(ANCOVA F = 5.342, df = 3, 16, P = 0.010), though
expected heterozygosity was not (ANCOVA F = 1.136,
df = 3, 16, P = 0.365). The most isolated island, Vitória,
exhibited the lowest allelic richness, whereas Ilhabela, the
largest island close to the mainland, had the highest allelic
richness, followed by the mainland (São Sebastião)
(Table 2).
Private alleles were present at low frequency, varying
between 0.010 and 0. The highest number of private alleles
was detected on Ilhabela (18 private alleles), followed by
the mainland, São Sebastião (six private alleles), a significant difference (F = 5.88, df = 1, 8, P = 0.042; see
Table 2). Private alleles were not found in Búzios and
Vitória, supporting the view that they represent more
isolated populations containing a subset of the alleles found
on the large island and the mainland.
Frankham’s (1998) effective inbreeding coefficient was
low for Ilhabela (0.012) and increased for Búzio (0.022)
and Vitória (0.023). These results again support the view of
that isolated islands harboured reduced genetic diversity.
Genetic differentiation
Overall, the degree of differentiation was modest; global
0
FST was 0.034 (P = 0.0001) and GST was 0.153
(P = 0.0006). Among the five study sites, nine of ten pairs
were significantly genetically differentiated (Table 3), with
differentiation greatest between the most distant island,
Vitória, and all other sites for FST. For GST, pairwise
comparisons gave a similar pattern as for FST, though
differentiation between the island of Búzios and the
mainland (São Sebastião B) was marginally greater than
between Vitória and two other sites (Table 3). Pairwise FST
0
and GST were highly correlated (Mantel test, r = 0.991,
P \ 0.0001) and results based on both measures were
qualitatively identical, so we henceforth present results
based on analysis of FST alone.
There was a positive relationship between genetic differentiation and log10 of Euclidean distance among pairs of
sites, though the relationship was of borderline significance
(Mantel test, r = 0.614; P = 0.056; see Fig. 2). Sea,
however, seemed to act as a barrier to gene flow; differentiation between pairs of sites versus distance across sea
123
1130
Conserv Genet (2014) 15:1123–1135
0
Table 3 Pairwise genetic differentiation of E. cordata as FST (lower left diagonal) and GST (upper right diagonal) among five sampled areas on
the mainland (São Sebastião A and B) and islands (Ilhabela, Ilha da Búzios, Ilha da Vitória), comprising 1,029 genotyped males
Area (N males)
São Sebastião A
São Sebastião B
Ilhabela
Ilha da Búzios
Ilha da Vitória
Pairwise genetic differentiation
São Sebastião A (151)
–
0.114
0.004
0.142
0.347
São Sebastião B (87)
0.035
–
0.056
0.211
0.174
Ilhabela (443)
0.003
0.019
–
0.143
0.291
Ilha de Búzios (322)
0.038
0.057
0.037
–
0.191
Ilha da Vitória (31)
0.100
0.059
0.083
0.060
–
See text and Supplementary Table S1 for details of how localities on the same land mass were pooled
For FST, values in bold are significant (P \ 0.05) after Bonferroni adjustment of significance
whilst taking into account log10 of Euclidean distance was
highly significant (partial Mantel test for effect of distance
over sea, r = 0.831, P = 0.003).
Diploid males
Males detected as diploids were confirmed as heterozygous
at one or more loci by repeating PCRs on these individuals
at all their heterozygous loci, giving the same two alleles
that differed by [1 repeat motifs for six of the seven diploid males at one or more loci (Supplementary Table S2).
The power of detecting diploid males was very high for the
mainland and all islands, even when taking into account
null alleles (Phet [ 0.99, Table 4), and making redundant
the need to adjust DMP for genetic non-detection.
Out of 1,245 genotyped males, seven were diploid
(DMP is 0.6 %); five were heterozygous at one microsatellite locus, one was heterozygous at two loci and the other
was heterozygous at three loci (Supplementary Table S2).
The highest frequency of diploid males (1.3 %) was on
Vitória, the smallest and most distant island from the
mainland (Table 4; Fig. 1). However, differences in the
proportion of diploid males among the four study areas
were non-significant (pairwise Fisher exact tests, all n.s.).
We note that the upper 95 % confidence interval attached
to the estimate of DMP for Vitória was 6.7 %, double that
of any other site. A larger sample size (greater statistical
power) would be needed for this most distant site from the
mainland to confirm whether or not its DMP was elevated
slightly beyond that of other sites, though we can already
reject the hypothesis that is elevated by [7 %.
Table 4 Diploid males in E. cordata across the four study areas
(mainland or SSeb, São Sebastião; Ibela, Ilhabela; Buz, Ilha de
Búzios; Vit, Ilha da Vitória)
Area
Fig. 2 Isolation by distance relationship in E. cordata as genetic
distance (FST/1 - FST) upon log10 geographical distance (in metres)
for five sampled areas (total n = 1,029 males, data in Supplementary
Table S1). The least squares best-fit line is shown for ease of
interpretation (Mantel test, r = 0.614, P = 0.056). The filled circle
represents the only distance not over sea, between São Sebastião A
and São Sebastião B)
123
TDM/TGM
Phet
Phet_null
FDM
95 % CIs of FDM
SSeb
3/263
0.9993
0.9963
0.011
0.002–0.032
Ibela
3/580
0.9993
0.9926
0.005
0.001–0.014
Buz
0/322
0.9991
0.9977
0.000
0.000–0.011
Vit
1/80
0.9988
0.9972
0.013
0.000–0.067
Total
7/1245
0.9991
0.9959
0.006
0.002–0.011
TDM total number of diploid males, TGM total number of successfully genotyped males, Phet probability of detection of a diploid male,
Phet_null probability of detecting a diploid males per locality adjusted
for putative null alleles, FDM frequency of diploid males, 95 % CIs
95 % binomial confidence intervals
Conserv Genet (2014) 15:1123–1135
Discussion
With respect to our original hypotheses, we found that
populations of the orchid bee E. cordata on two small
islands [20 km distant from the mainland exhibited
reduced genetic variation and significant differentiation
from the mainland or a large, near-shore (ca. 2 km from the
mainland) island, yet DMP was extremely low and did not
vary significantly across study areas. Clearly, variability at
neutral loci diminished with island isolation whereas the
sex locus, a locus under balancing selection, appeared not
to be impacted.
Genetic diversity and differentiation on islands
First we address the reduced genetic diversity we detected
on islands distant from the mainland. According to island
biogeography theory, the probability of species colonization and persistence are related to distance from source
population and island size, respectively (MacArthur and
Wilson 1967). The patterns of genetic diversity we observe
are consistent with theoretical expectations of island biogeography (Jansen 1981). Distance from a source population and island area also determine patterns of genetic
diversity on islands; the more isolated and smaller is an
island, the genetically more depauperate it is at neutral
markers, and genetic variation is lost in inverse relation to
population size (Frankham 1995, 1997). Our finding suggests that, though E. cordata can migrate to two distant
islands, genetic differentiation is impacted by the isolation
of Búzios and particularly Vitória. In addition, the small
surface area of these two islands (see Table 1) may be
associated with small Ne and greater genetic drift, leading
to loss of genetic diversity on these islands, which results in
lower standing genetic variation at drift-migration-mutation equilibrium. Given that our study system comprises
three islands that covary in isolation and size, we are
unable to separate the relative importance of island isolation or island size in reducing the genetic diversity of
populations of E. cordata.
In relation to genetic differentiation, our data for E.
cordata suggest that 11 km of sea (the distance between
the most distant island, Vitória, and the next most distant,
Búzios) represent a significant, if not insurmountable,
barrier to inferred gene flow, a greater resistance (sensu
Spear et al. 2010) than terrestrial habitat. For the widespread sweat bee Halictus rubicundus, sea ([20 km of the
Irish Sea) also appears to represent a barrier to inferred
gene flow (Soro et al. 2010); populations on Ireland were
significantly differentiated from populations on Britain,
even after controlling for Euclidean distance, and those on
Ireland showed reduced genetic diversity. A similar pattern
has also been detected in the primitively eusocial bumble
1131
bee Bombus hortorum (Goulson et al. 2011) on the Western
Isles of Scotland; sea represented a barrier to inferred gene
flow, and genetic diversity was lower on islands than the
mainland. For the rare bumble bee B. muscorum in the
same island system of Scotland, significant differentiation
was observed at distances [10 km (Darvill et al. 2006).
High genetic differentiation has also been detected between
island and mainland populations in the solitary bee Colletes
floralis (Davis et al. 2010), studied in the same Scottish
island system. Nevertheless Davis et al. (2010) found no
evidence that natural geographic discontinuity represented
by the sea, i.e. between mainland and islands, constituted a
barrier to gene flow beyond isolation by distance, and
genetic diversity was no different on islands versus the
mainland. Despite differences between studies, we can
nevertheless conclude that open water represents a barrier
to the movement of E. cordata and probably many other
bee species.
An important follow-on question is, ‘How much water
represents a barrier to movement?’ Females of other orchid
bees (Eufriesea surinamensis and Eulaema polychroma)
have been recorded flying over 2.5–5 km of open water
(Jansen 1971). For E. cordata in our study system, 2 km
(mainland-Ilhabela) seems to represent a weak barrier to
movement but 7 km (Ilhabela-Búzio) to 11 km (BúzioVitória) of open water seems to represent a more significant barrier. Whether these distances can be generalized to
other orchid bee species in the Neotropics is not known.
Gene flow in E. cordata could be primarily mediated by
females or males. Though Euglossa orchid bee females are
generally philopatric (Garófalo 1985, Zimmermann et al.
2009), they are able to fly long distances (Jansen 1971) and
promote gene flow. Interestingly, E. cordata has more than
one generation per year (Garófalo 1985), with generations
asynchronous across localities. Males produced in one
locality may be attracted to neighboring ones (Paxton
2005) because they are able to detect scents over several
km (Dressler 1982; Ayasse et al. 2001) and, as a consequence, promote occasional long distance dispersal.
The high genetic variability and lack of genetic differentiation of E. cordata on the mainland and the near-shore
island Ilhabela that we detected is more typical of mainland
populations of orchid bees (Zimmermann et al. 2011; Suni
and Brosi 2011), including other studies of E. cordata
(Cerântola et al. 2010; Souza et al. 2010; Rocha Filho et al.
2013). Cerântola et al. (2010), for example, found effectively zero differentiation for E. cordata across [500 km
of mainland São Paulo, Brazil, whereas Souza et al. (2010)
found weak though significant differentiation for E. cordata across 310 km of São Paulo. Lack of differentiation
may be indicative of large population size, high dispersal
capacity, or both. However, we caution against generalization across all 200 plus orchid bee species as some vary
123
1132
markedly in their population genetic structure (Suni and
Brosi 2011; Rocha Filho et al. 2013).
Diploid males
Across the study, we estimated a low frequency of diploid
males in E. cordata. One possibility is that diploid males
were less attracted to odour baits that haploid males and so
we underestimated their frequency. We have no evidence
for or against this idea, though note that we did capture
seven diploid males at baits, so clearly E. cordata males are
viable and are seemingly attracted to baits. Our estimates
and interpretation of DMP also assume Hardy–Weinberg
equilibrium and sl-CSD. We did not sample E. cordata
females (diploid individuals heterozygous at the sex locus)
with which to estimate inbreeding because they are not
attracted to odour baits. But Souza et al. (2007) and Cerântola et al. (2010) found inbreeding coefficients in
mainland E. cordata not to differ from zero, suggesting the
species mates randomly. In addition, our estimates of DMP
are robust to slight deviations from random mating (Paxton
et al. 2000). Null alleles are also unlikely to explain the
lack of diploid males we observed in our study. With
respect to sex determination, Cerântola et al. (2010) and
this study both found diploid males in E. cordata, indicating that this species possesses CSD. Patterns of DMP
suggest that many bee species probably possess sl-CSD (as
opposed to mulitlocus CSD; Harpur et al. 2013; cf. Paxton
et al. 2000), though other haplodiploid Hymenoptera possess multilocus CSD, such as the parasitoid wasp Cotesia
vestalis (de Boer et al. 2008). If E. cordata possesses
multilocus CSD, this might help explain its low DMP.
We now interpret DMP in E. cordata in terms of
diversity at the sex locus (or loci), a locus under balancing
selection. Small population size leads to genetic erosion
and an increased probability of inbreeding (Frankham
1998, Frankham et al. 2011). Although gene flow can
negate the negative effects of inbreeding depression in
small populations (Frankham 1998), our results show that
gene flow is insufficient to maintain genetic diversity at
neutral loci in E. cordata on small and distant islands.
Given the apparent lack of DMP on islands distant from the
mainland, yet a decrease in genetic diversity at neutral loci,
we suggest that balancing selection, complemented by
occasional immigration of novel sex alleles, may largely
suffice to maintain allelic diversity at the sex locus in E.
cordata on islands distant from the mainland. Differentiating the relative importance of these two evolutionary
forces will give a better understanding of when a population with sl-CDS is likely to enter a DM extinction vortex.
The mainland-island system studied by us nevertheless
gives empirical support to the simulation-based study of
Hein et al. (2009), showing the viability of an isolated
123
Conserv Genet (2014) 15:1123–1135
population of a species with sl-CSD is not immediately
threatened by extinction if it is even modestly connected by
gene flow to other populations, if its rate of reproduction is
sufficiently large, its sex ratio is fixed and it does not
practice mate choice.
Mate or gamete selection could also act to reduce DMP
and would mean that DMP is a poor proxy for allelic
diversity at the sex locus. The bumble bee Bombus terristris exhibits inbreeding avoidance through kin discrimination in pre-copulatory mate choice (Whitehorn et al. 2009).
Avoidance of mating with kin would theoretically reduce
DMP in this and other bee species such as E. cordata.
Many Euglossa females mate only once (Zimmermann
et al. 2009), though it is not known whether they avoid
mating with males carrying one of their own sex alleles.
Gamete selection has also been proposed to explain low
DMP in an inbred bee (Paxton et al. 2000), but there is
again no empirical support for the concept of sex allele
recognition by gametes.
The low DMP in E. cordata supports a growing body of
work that has found low DMP on mainland America in a
range of orchid bee species (Souza et al. 2010; Zimmermann et al. 2011). For example, Souza et al. (2010)
detected \1 % DMP using microsatellites among 1,010
individuals of 27 orchid bee species from several areas on
mainland America, including E. cordata. Only eight diploid males were detected by microsatellite analysis among
353 individuals of Eufriesea violacea from Brazilian forest
fragments (Freiria et al. 2012). The frequency of DMs was
even smaller or null for other studies in Brazil and Mexico
(Cerântola et al. 2010, Zimmermann et al. 2011). Rocha
Filho et al. (2013) reported zero diploid males in several
euglossine species, including E. cordata, on Anchieta, an
island close to our study sites (Fig. 1), but located at a
distance from the mainland (600 m) smaller than the
known flight capacity of several Euglossini bees (Jansen
1971, 1981; Dressler 1982).
Theoretically, isolation can impact both genetic diversity and genetic differentiation of terrestrial populations,
regardless of whether that isolation represents habitat
fragmentation into habitat islands on the mainland or actual
islands separated by water. Here we addressed the effect of
isolation on the genetic variability of neutral loci and the
sex locus in potentially small populations of Euglossini in a
system of islands located at increasing distance from the
mainland. It prompts the important question of whether the
results we obtained from a 10 to 20,000 year old system
within a protected area, which is probably in mutation–
migration–drift equilibrium, can be extrapolated to systems
where fragmentation is the sudden outcome of pressure
from human activities, such as the rapidly increasing
intensification of agriculture typical of countries, like
Brazil, undergoing fast economic expansion. What are the
Conserv Genet (2014) 15:1123–1135
consequences for a small population of being suddenly
disconnected from its main source of genetic variation?
Will it have enough genetic variability at loci under
directional selection in addition to those under balancing
selection, or enough time to respond to changing selective
pressures, or will it enter a DM extinction vortex (Zayed
and Packer 2005; cf. Hein et al. 2009)? These are important
questions for which studies with orchid bees are warranted
on the effects of drift and isolation on the genetic variability of selectively neutral loci and those under selection
in an increasingly fragmented landscape.
Acknowledgments We thank Carlos Eduardo Pinto, Guaraci Duran
Cordeiro, Morgana Sazan, Paulo César Fernandes and Tiago Caetano
for their help in the field, André Nemésio and Leo Rocha Filho for
species identification, Flávio Oliveira Francisco, Leandro Rodrigues
Santiago, Paulo Gonçalves and Rita Radzeviciute for help in the lab,
and Maria Cristina Arias, Tomás Murray and the referees for many
very helpful comments that helped improve the manuscript. Petra
Liebe kindly processed samples in a MegaBace in the Molecular
Ecology Lab at the Martin Luther University Halle-Wittenberg, for
which we thank her and Robin Moritz. The Fundação de Amparo à
Pesquisa do Estado de São Paulo (FAPESP) supported our project.
The Coordenação de Aperfeiçoamento de Pessoal de Nı́vel Superior
(CAPES) and the Deutscher Akademischer Austausch Dienst
(DAAD) funded the first author. SISBIO (Sistema de Informação e
Autorização em Biodiversidade) provided a collecting permit
(Nr.10453-1) and IBAMA (Insituto Brasileiro do Meio Ambiente e
dos Recursos Naturais Renováveis) provided a biological sample
export permit (Nr.10BR005548/DF).
References
Allendorf FW, Luikart G (2007) Conservation and the genetics of
populations. Blackwell, Malden
Ângelo S (1989) Ilhas do Litoral Paulista, Série documentos.
Secretaria do Meio Ambiente (SMA), São Paulo
Ayasse M, Paxton RJ, Tengö J (2001) Mating behavior and chemical
communication in the Order Hymenoptera. Annu Rev Entomol
46:31–78
Beye M, Hasselmann M, Fondrk MK, Page RE, Omholt SW (2003)
The gene csd is the primary signal for sexual development in the
honeybee and encodes an SR-type protein. Cell 114:419–429
Bonin A, Bellemain E, Bronken Eidesen P, Pompanon F, Brochmann
C, Taberlet P (2004) How to track and assess genotyping errors
in population genetics studies. Mol Ecol 13:3261–3273
Brook BW, Tonkyn DW, O‘Grady JJ, Frankham R (2002) Contribution of inbreeding to extinction risk in threatened species.
Conserv Ecol 6:16. http://www.consecol.org/vol6/iss1/art16.
Accessed 27 Jan 2013
Brown MJF, Paxton RJ (2009) The conservation of bees: a global
perspective. Apidologie 40:410–416
Cardinale BJ, Duffy E, Gonzalez A, Hooper DU, Perrings C, Venail
P, Narwani A, Mace GM, Tilman D, Wardle DA, Kinzig AP,
Daily GC, Loreau M, Grace JB, Lagauderie A, Srivastava DS,
Naeem S (2012) Biodiversity loss and its impact on humanity.
Nature 486:59–67
Cerântola NCM, Oi CA, Cervini M, Del Lama MA (2010) Genetic
differentiation of urban populations of Euglossa cordata from
the state of São Paulo, Brazil. Apidologie 42:214–222
1133
Cook JM (1993) Sex determination in the Hymenoptera: a review of
models and evidence. Heredity 71:421–435
Cook JM, Crozier RH (1995) Sex determination and population
biology in the Hymenoptera. Trends Ecol Evol 10:281–286
Cordeiro GD, Boff S, Caetano TA, Fernandes PC, Alves-dos-Santos I
(2013) Euglossine bees (Apidae) in Atlantic forest areas of São
Paulo State, southeastern Brazil. Apidologie 44:254–267
Cowan DP, Stahlhut JK (2004) Functionally reproductive diploid and
haploid males in an inbreeding hymenopteran with complementary sex determination. PNAS 101:10374–10379
Crozier RH, Pamilo P (1996) Sex allocation and kin selection. Evolution
of social insect colonies. Oxford University Press, Oxford
Darvill B, Ellis JS, Lye GC, Goulson D (2006) Population structure
and inbreeding in a rare and declining bumblebee, Bombus
muscorum (Hymenoptera: Apidae). Mol Ecol 15:601–611
Darvill B, O’Connor S, Lye GC, Waters J, Lepais O, Goulson D
(2010) Cryptic differences in dispersal lead to differential
sensitivity to habitat fragmentation in two bumblebee species.
Mol Ecol 19:53–63
Davis ES, Murray TE, Fitzpatrick Ú, Brown MJF, Paxton RJ (2010)
Landscape effects on extremely fragmented populations of a rare
solitary bee, Colletes floralis. Mol Ecol 19:4922–4935
de Boer JG, Ode PJ, Rendahl AK, Vet LEM, Whitfield JB, Heimpel
GE (2008) Experimental support for multiple-locus complementary sex determination in the parasitoid Cotesia vestalis. Genetics
180:1525–1535
Dennis RLH, Dapporto L, Shreeve TG, John E, Coutsis JG, Kudrna
O, Saarinen K, Ryrholm N, (Bob) Williams WR (2008)
Butterflies of European islands: the implications of the geography and ecology of rarity and endemicity for conservation.
J Insect Conserv 12:205–236
Dieringer D, Schlötterer C (2003) Microsatellite Analyser (MSA
4.05): a platform independent analysis tool for large microsatellite data sets. Mol Ecol Notes 3:167–169
Dobson A, Lodge D, Alder J, Cumming GS, Keymer J, McGlade J,
Mooney H, Risak JA, Sala O, Wolters V, Wall D, Winfree R,
Xenopoulos MA (2006) Habitat loss, trophic collapse, and the
decline of ecosystem service. Ecology 87:1915–1924
Dressler RL (1982) Biology of the orchid bees (Euglossini). Annu
Rev Ecol Syst 13:373–394
Elias J, Dorn S, Mazzi D (2010) No evidence for increased extinction
proneness with decreasing effective population size in a
parasitoid with complementary sex determination and fertile
diploid males. BMC Evol Biol 10:366
Frankham R (1995) Conservation genetics. Annu Rev Genet
29:305–327
Frankham R (1997) Do island population have less genetic variation
than mainland populations? Heredity 78:311–327
Frankham R (1998) Inbreeding and extinction: island populations.
Conserv Biol 12:665–675
Frankham R (2003) Genetics and conservation biology. C R Biologies
326:S22–S29
Frankham R, Ballou JD, Briscoe DA (2004) A primer of conservation
genetics. Cambridge University Press, Cambridge
Frankham R, Ballou JD, Eldridge MDB, Robert CL, Ralls K, Dudash
MR, Fenster CB (2011) Predicting the probability of outbreeding
depression. Conserv Biol 25:465–475
Freiria GA, Bombarda Ruim J, Souza RF, Sofia SH (2012) Population
structure and genetic diversity of the orchid bee Eufriesea
violacea (Hymenoptera, Apidae, Euglossini) from Atlantic
Forest remnants in southern and southeastern Brazil. Apidologie
43:392–402
Gallai N, Salles J-M, Settele J, Vaissière BE (2009) Economic
valuation of the vulnerability of world agriculture confronted to
pollinator decline. Ecol Econ 68:810–821
123
1134
Garófalo CA (1985) Social structure of Euglossa cordata nests
(Hymenoptera: Apidae: Euglossini). Entomol Gen 11:77–83
Gerlach J (2008) Preliminary conservation status and needs of an
oceanic island fauna: the case of Seychelles insects. J Insect
Conserv 12:293–305
Goulson D, Kaden JC, Lepais O, Lye GC, Darvill B (2011)
Population structure, dispersal and colonization history of the
garden bumblebee Bombus hortorum in the Western Isles of
Scotland. Conserv Genet 12:867–879
Harpur BA, Sobhani M, Zayed A (2013) A review of the
consequences of complementary sex determination and diploid
male production on mating failures in the Hymenoptera.
Entomol Exp Appl 146:156–164
Hedrick PW (2005) A standardized genetic differentiation measure.
Evolution 59:1633–1638
Heimpel GE, de Boer JG (2008) Sex determination in the Hymenoptera. Annu Rev Entomol 53:209–230
Hein S, Poethke HJ, Dorn S (2009) What stops the ‘diploid male
vortex’? A simulation study for species with single locus
complementary sex determination. Ecol Model 220:1663–1669
IBM Corporation (2010) IBM SPSS Statistics for Windows, Version
19.0., Released 2010. IBM Corporation, Armonk
Jansen DH (1971) Euglossine bees as long-distance pollinators of
tropical plants. Science 171:203–205
Jansen DH (1981) Reduction in Euglossine bee species richness on
Isla del Cano, a Costa Rican offshore island. Biotropica
13:238–240
Jordan MA, Snell HL (2008) Historical fragmentation of islands and
genetic drift in populations of Galápagos lava lizards (Microlophus albemarlensis complex). Mol Ecol 17:1224–1237
Kearns CA, Inouye DW, Waser NM (1998) Endangered mutualisms:
the conservation of plant-pollinator interactions. Annu Rev Ecol
Evol Syst 29:83–112
Klein AM, Vaissière BE, Cane JH, Steffan-Dewenter I, Cunningham
SA, Kremen C, Tscharntke T (2007) Importance of pollinators in
changing landscapes for world crops. Proc R Soc B 274:303–313
López-Uribe MM, Almanza MT, Ordóñez M (2007) Diploid male
frequencies in Colombian population of Euglossini bees.
Biotropica 39:660–662
MacArthur RH, Wilson EO (1967) The theory of island biogeography. Princeton University Press, Princeton
McNeely JA, Miller KR, Reid WV, Mittermeier RA, Werner TB
(1990) Conserving the World’s biological diversity. IUCN,
World Resources Institute, Conservation International, WWFUS and the World Bank, Washington
Meirmans PG, Van Tienderen PH (2004) GENOTYPE and GENODIVE: two programs for the analysis of genetic diversity of
asexual organisms. Mol Ecol Notes 4:792–794
New TR (2008) Insect conservation on islands: setting the scene and
defining the needs. J Insect Conserv 12:197–204
O’Grady JJ, Reed DH, Brook BW, Frankham R (2008) Extinction risk
scales better to generations than years. Anim Conserv 11:442–451
Paxton RJ (2005) Male mating behavior and mating systems of bees:
an overview. Apidologie 36:145–156
Paxton RJ, Thorén PA, Gyllenstrand N (2000) Microsatellite DNA
analysis reveals low diploid male production in a communal bee
with inbreeding. Biol J Linnean Soc 69:483–502
Peakall R, Smouse PE (2006) GenAlEx 6: genetic analysis in Excel.
Population genetic software for teaching and research. Mol Ecol
Notes 6:288–295
Petit RJ, Deguilloux M-F, Chat J, Grivet D, Garnier-Géré P,
Vendramin GG (2005) Standardizing for microsatellite length
in comparisons of genetic diversity. Mol Ecol 14:885–890
Raymond M, Rousset F (1995) GENEPOP (V. 1.2): population
genetics software for exact tests and ecumenicism. J Hered
86:248–249
123
Conserv Genet (2014) 15:1123–1135
Rocha Filho LC, Cerântola NCM, Garófalo CA, Del Lama MA
(2013) Genetic differentiation of the Euglossini (Hymenoptera,
Apidae) population on a mainland coastal plain and an insular in
Southeastern Brazil. Genetica 141:65–74
Roubik DW, Hanson PE (2004) Orchid bees of tropical America:
biology and field guide. Heredia, Costa Rica
Roubik DW, Weight LA, Bonilla MA (1996) Population genetics,
diploid males, and limits to social evolution of Euglossine bees.
Evolution 50:931–935
Rousset F (1997) Genetic differentiation and estimation of gene flow from
F-statistics under isolation by distance. Genetics 145:1219–1228
Schlötterer C (1998) Microsatellites. In: Hoelzel AR (ed) Molecular
genetic analysis of populations. A practical approach. Oxford
University Press, Oxford, pp 237–261
Schlötterer C (2004) The evolution of molecular markers—just a
matter of fashion? Nat Rev Genet 5:63–69
Soro A, Field J, Bridge C, Cardinal SC, Paxton RJ (2010) Genetic
differentiation across the social transition in a socially polymorphic sweat bee, Halictus rubicundus. Mol Ecol 19:3351–3363
Souza RO, Cervini MA, Del Lama MA, Paxton RJ (2007) Microsatellite loci for euglossine bees (Hymenoptera: Apidae). Mol
Ecol Notes 7:1352–1356
Souza RO, Del Lama MA, Cervini M, Mortari N, Eltz T, Zimmermann Y, Bach C, Brosi BJ, Suni S, Quezada-Euán JJG, Paxton
RJ (2010) Conservation genetics of neotropical pollinators
revisited: microsatellite analysis suggests that diploid males are
rare in orchid bees. Evolution 64:3318–3326
Spear SF, Balkenhol N, Fortin M-J, McRae BH, Scribner KIM (2010)
Use of resistance surfaces for landscape genetic studies:
considerations for parameterization and analysis. Mol Ecol
19:3576–3591
Spielman D, Brook BW, Frankham R (2004) Most species are not
driven to extinction before genetic factors impact them. PNAS
101:15261–15264
Stouthamer R, Luck RF, Werren JH (1992) Genetics of sex
determination and improvement of biological control using
parasitoids. Environ Entomol 21:427–435
Suguio K, Martin L (1987) Classificação da costa e evolução
geológica das planı́cies litorâneas do sudeste e sul do Brasil. In:
Anáis do Simpósio sobre ecossistemas da Costa Sul e Sudeste
Brasileira: sı́ntese dos conhecimentos, pp 1–2
Suni SS, Brosi BJ (2011) Population genetics of orchid bees in a
fragmented tropical landscape. Conserv Genet. doi:10.1007/
s10592-011-0284-z
Takahashi J, Ayabe T, Mitsuhata M, Shimizu I, Ono M (2008)
Diploid male production in a rare locally distributed bumblebee,
Bombus florigelus (Hymenoptera, Apidae). Insect Soc 55:43–50
Tavares MG, Carvalho CR, Soares FAF (2010) Genome size variation
in Melipona species (Hymenoptera: Apidae) and sub-grouping
by their DNA content. Apidologie 41:636–642
Van Wilgenburg E, Driesses G, Beukboom LW (2006) Single locus
complementary sex determination in Hymenoptera: an ‘‘unintelligent’’ design? Front Zool. doi:10.1186/1742-9994-3-1
Verhulst EC, Beukeboom LW, van de Zande L (2010) Maternal
control of haplodiploid sex determination in the wasp Nasonia.
Science 328:620–623
Walsh PS, Metzeger DA, Higuchi R (1991) Chelex 100 as a medium
for simple extraction of DNA for PCR-based typing from
forensic material. Biotechniques 10:506–513
Weir BS, Cockerham CC (1984) Estimating F-statistics for the
analysis of population structure. Evolution 38:1358–1370
Whidden TL, Owen RE (2011) Frequencies of diploid male in natural
population of three North American bumble bee (Bombus)
species (Hymenoptera: Apidae). Genetics 104:83–87
Whitehorn PR, Tinsley MC, Goulson D (2009) Kin recognition and
inbreeding reluctance in bumblebees. Apidologie 40:627–633
Conserv Genet (2014) 15:1123–1135
Wright S (1965) The interpretation of population structure by Fstatistics with special regard to systems of mating. Evolution
19:395–420
Wright S (1969) Evolution and the genetics of populations. The
theory of gene frequencies, vol 2. University of Chicago Press,
Chicago
Zayed A (2009) Bee genetics and conservation. Apidologie 40:
237–262
Zayed A, Packer L (2001) High levels of diploid male production in a
primitively eusocial bee (Hymenoptera: Halictidae). Heredity
87:631–636
Zayed A, Packer L (2005) Complementary sex determination
substantially increases extinction proneness of haplodiploid
population. PNAS 102:10742–10746
1135
Zayed A, Roubik DW, Packer L (2004) Use of diploid male
frequency data as an indicator of pollinator decline. Proc R Soc
Lond B 271:9–12
Zimmermann Y, Roubik DW, Quezada-Euan JJG, Paxton RJ, Eltz T
(2009) Single mating in orchid bees (Euglossa, Apinae):
implications for mate choice and social evolution. Insect Soc
56:241–249
Zimmermann Y, Schorkopf DLP, Moritz RFA, Pemberton RW,
Quezada-Euan JJG, Eltz T (2011) Population genetic structure of
orchid bees (Euglossini) in anthropogenically altered landscapes.
Conserv Genet 12:1183–1194
123