Content uploaded by Frédéric Normand
Author content
All content in this area was uploaded by Frédéric Normand on Sep 04, 2014
Content may be subject to copyright.
ORIGINAL PAPER
Invasive mango blossom gall midge, Procontarinia
mangiferae (Felt) (Diptera: Cecidomyiidae) in Reunion
Island: ecological plasticity, permanent and structured
populations
P. Amouroux
•
F. Normand
•
S. Nibouche
•
H. Delatte
Received: 14 June 2012 / Accepted: 29 December 2012
Ó Springer Science+Business Media Dordrecht 2013
Abstract Mango blossom gall midge, Procontarinia
mangiferae (=Erosomyia mangiferae Felt), is an
invasive pest that causes economic damage world-
wide. The objectives of our study were to highlight the
genetic and ecological abilities of this monophagous
gall midge to invade new habitats and to evaluate its
genetic structure on an isolated island. This study,
carried out in subtropical Reunion Island, is based on
data from population dynamics surveys and from
molecular analyses (mitochondrial DNA and micro-
satellites). Using 11 microsatellite loci and an exten-
sive sampling of 27 populations at 17 sites, we tested
the genetic differentiation between populations sam-
pled on different mango organs and cultivars at
different seasons and under different climatic and
cultural environments. We checked for the existence
of a seasonal bottleneck. Our results showed that a
single species, P. mangiferae, was present all year
round with no genetic bottleneck at any of the sites
sampled, regardless of the climatic and cultural
conditions, and that it fed on inflorescences and young
leaves. These characteristics showed the ecological
plasticity of P. mangiferae, despite its low genetic
diversity and, consequently, the invasive potential of
this species. Populations in Reunion Island are struc-
tured into two clusters in sympatry and present in
different proportions at each site. One cluster was
more frequently found in the cultivated mango area.
This work provides insights into the relationships
between gall midges and tree host plants in a
subtropical agro-ecosystem, as well as into the role
of the population genetic structure in the establishment
process of a monophagous invasive cecid fly.
Keywords Ecological genetics Invasive pest
species Insects Agriculture Cecidomyiidae
Introduction
Successful invasions are characterized by the match
between the invader and its new habitat, which may
involve a combination of several strategies (Facon
et al. 2006). Genetic structure and life history traits
contribute to the success of invaders. Genetic diver-
sity, provided by large founding populations or
multiple introductions, allows local adaptation in the
Electronic supplementary material The online version of
this article (doi:10.1007/s10530-012-0400-0) contains
supplementary material, which is available to authorized users.
P. Amouroux (&) F. Normand
CIRAD, UPR HortSys, 97455 Saint-Pierre, La Re
´
union,
France
e-mail: paul.amouroux@cirad.fr
P. Amouroux
Universite
´
de la Re
´
union, UMR PVBMT,
97410 Saint-Pierre, La Re
´
union, France
S. Nibouche H. Delatte
CIRAD, UMR PVBMT, 97410 Saint-Pierre, La Re
´
union,
France
123
Biol Invasions
DOI 10.1007/s10530-012-0400-0
new habitat (Dlugosch and Parker 2008; Kolbe et al.
2004; Suarez and Tsutsui 2008). Insects provide
numerous models for studying biological and ecolog-
ical factors that influence the ability of some species to
become established and spread within new environ-
ments (Boubou et al. 2011; Perdereau et al. 2011).
Following the introduction of a species into a new
environment, the invaders are able to evolve rapidly
into the new region in isolation from other individuals
of the species (Launey et al. 2010; Lee 2002). Rapid
loss of genetic diversity in the derived founder
population can occur under the combined effect of
genetic bottleneck and genetic drift (Nei et al. 1975).
Classical population genetics invokes multiple factors
to explain population genetic structure. First, geo-
graphical features can represent barriers to migration
and/or population expansion and generate patterns that
structure the genetic diversity (Launey et al. 2010).
Second, for phytophageous insects, genetic changes
may be required for some host plant shifts or host
range expansions (Tabashnik 1983). The population
genetic structure can then be modified by life history
traits such as dispersal capacity (Phillips et al. 2006),
mating system, which is a genetic link between
generations, population size fluctuation, which
enhances genetic drift during small-sized generation,
and the presence of diapausing individuals, like a seed
bank in the soil, which are expected to offset the
effects of drift (Castillo 1994). Lastly, endosymbiont
bacteria such as Wolbachia spp. can manipulate the
post-zygotic reproduction of their host and induce host
speciation (Sun et al. 2011). The mango blossom gall
midge, Procontarinia mangiferae (Felt), is a world-
wide invasive monophageous pest that was recently
introduced in several countries and is, therefore, a
good model for studying its population genetic
structure with regard to these factors.
Cecidomyiidae, commonly referred to as gall mid-
ges, are one of the largest families of nematocerous
Diptera, with more than 6,024 described species, and
many more undescribed and unknown species world-
wide (Gagne
´
2010). Among these species, 45 are
immigrant (Gagne
´
2010) and many of them are pests of
crops, ornamental plants and forest trees worldwide
(Dorchin 2008). Annual crop pests are the most
frequently studied species and include the Hessian
fly, Mayetiola destructor (Say), a wheat gall midge that
was the first invasive insect to cause economic havoc in
the USA (Stuart et al. 2008). The life cycle of
Cecidomyiidae is closely associated with the phenol-
ogy of their host plant, and gall midges have several
adaptive synchronization strategies including diapause
(Uechi and Yukawa 2006; Yukawa 2000). They are
most often restricted to one organ of a single host plant
(Gagne
´
1989; Jones et al. 1983) and are less often
oligophagous within a single plant genus (Gagne
´
2004).
The genus Procontarinia
contains 15 species, all of
which are associated with mango, Mangifera indica L.
(Anacardiaceae) (Gagne
´
2010). Since invasion suc-
cess has often been linked to diet breadth, herbivores
that feed on numerous plant species have the highest
probability of finding a suitable host plant (Ward and
Masters 2007). However, despite their strict associa-
tion with mango, the mango blossom gall midge,
P. mangiferae (Felt), and the mango leaf gall midge,
P. matteiana Kieffer and Cecconi, are found world-
wide. P. mangiferae is considered to be indigenous to
India and invasive in Thailand, Mauritius, Reunion
Island, Iran, the West Indies and Brazil (CAB 2004;
Gagne
´
2010). This species was described several times
and consequently it has several synonyms: Erosomyia
mangiferae Felt 1911, Mangodiplosis mangiferae
Tavares 1918, Rhabdophaga mangiferae Mani 1938
and Erosomyia indica Grover 1965 (Gagne
´
and
Medina 2004). This species is a serious pest that
reduces potential fruit yield by more than 70 %
(Etienne 1977; Pen
˜
a et al. 1998; Whitwell 1993). It
is currently the major pest of mango flowers in
Reunion Island (Amouroux and Normand 2010). Each
female, with a life span of two or three days, can lay up
to 150 eggs on mango inflorescences. After hatching,
larvae penetrate the mango organ, leave the organ at
the final instar, fall to the ground and bury themselves
in the soil. Adult gall midges emerge from larvae
buried in the soil and induce serious outbreaks during
mango flowering, from the equatorial to the subtrop-
ical regions of India (Prasad 1971). The cycle lasts
from between 14 and 25 days (Pezhman and Askari
2004). Larvae enter into diapause in tropical regions at
the end of the flowering period, but permanent
populations exist in the equatorial regions where
mango flowering may occur all year round. The biotic
and/or abiotic conditions that induce diapause are
unknown (Prasad and Grover 1974).
Procontarinia mangiferae is the only Procontarinia
species able to feed on different organs of mango,
inflorescences and young leaves, inducing different
types of galls (Online Resource 1). Other Procontarinia
P. Amouroux et al.
123
species are restricted either to leaves or to fruits. Raman
et al. (2009) hypothesized that this characteristic could
be indicative of subtle radiation within this species.
Indeed, host shifts frequently result in host races
(Carletto et al. 2009), cryptic species (Blair et al.
2005) or sibling species (Malausa et al. 2007) in the case
of phytophagous insects. Gall midges also provide
examples of radiation. Cook et al. (2011) reported
cryptic species of Dasineura oxycoccana (Johnson)
between two host plant species of the genus Vaccinium.
Some clades of the genus Asphondylia have radiated
from their host plant Larrea tridentate (Sesse and Moc.
ex DC.) Coville by shifting to new organs and by
changing their phenology (Joy and Crespi 2007).
Stireman et al. (2008) also observed that four cryptic
lineages of the gall midge Asteromyia carbonifera
(Osten Sacken) existed on goldenrod leaves without
phenological isolation. However, the existence of two
lineages associated with either inflorescences or leaves
has never been tested for P. mangiferae.
Mango was introduced in Reunion Island in 1770
from India (Le Bellec and Renard 1997). Etienne and
Roura (1974) mentioned the presence of P. mangife-
rae in 1974 for the first time. However, its date of
introduction is unknown. Hugon (1979) confirmed the
presence on the island of P. matteiana on leaves and
P. mangiferae on inflorescences, and observed two
other types of gall on leaves and stems without iden-
tifying the species, suggesting that other gall midges
could be present in Reunion Island.
The objective of our study was to highlight the
genetic and ecological abilities of this monophagous
gall midge to invade new habitats, and to evaluate its
genetic structure on an isolated island. Field popula-
tion dynamics surveys and molecular analyses were
performed in order to find answers to the following
five questions. Is P. mangiferae the unique blossom
gall midge species present on mango in Reunion? Is
P. mangiferae present only in the mango production
area, or in all areas where mango is present? Is
P. mangiferae present seasonally during the mango
flowering period, or is it permanently present thanks
to its ability to feed on young leaves? How can
P. mangiferae genetic diversity be characterized and
are the populations genetically structured according to
biotic (mango organs, cultivars) or abiotic (seasons or
sites) factors? Does P. mangiferae have specific traits
linked to invasiveness?
Materials and methods
Ecological context
Reunion Island is located near the Tropic of Capricorn
in the Indian Ocean (21°06
0
S, 55°29
0
E). It is a sub-
tropical island with a diversity of climates due to a
wide range of elevations with high mountains culmi-
nating at more than 3,000 m a.s.l. (Fig. 1), and to trade
winds that induce contrasted rainfalls between the
windward east coast and the leeward west coast. At
low altitudes (\ 1,000 m a.s.l.), annual rainfall ranges
from 500 to 1,500 mm on the west coast and from
2,000 to 5,000 mm on the east coast (Raunet 1991).
The cool months, from July to September, are
generally dry. Their monthly mean minimum and
maximum temperatures at sea level are 17.7 and
25.8 °C, respectively. The hot and rainy months, from
January to March, have monthly mean minimum and
maximum temperatures at sea level of 22.6 and 30 °C,
respectively (Raunet 1991). Commercial mango
orchards expanded quickly from less than 50 ha in
1970 to more than 300 ha today on the west and
southwest coasts suitable for mango cultivation
(Vincenot and Normand 2009). However, mango trees
are present all over the island, up to 1,000 m a.s.l., in
the backyards of homes and along roadsides. Mango
flowering occurs from July to October. Vegetative
growth begins sporadically during flowering and fruit
growth and flushes after harvest, from January to
April. A vegetative rest period occurs from April to
June.
Field sampling for genetic studies
Twenty-seven populations of 28–35 final instar larvae
were collected at 17 sites across Reunion Island
(Table 1, Fig. 1). Each site was sampled one to four
times during the winter 2009 (June to October) and the
summer 2011 (February) (Table 1). Sampling was
carried out in commercial orchards (cultivars ‘Cog-
shall’ and ‘Jose
´
’) or on isolated trees in backyards and
along roadsides (unknown cultivar). Ten trees were
randomly chosen in commercial orchards, and three to
ten trees in non-cultivated areas. Three to ten larvae per
tree were randomly sampled from infested inflores-
cences or young leaves, depending on the site and the
season. Larvae were stored in 96° ethanol at -20 °C.
Ecological plasticity, permanent and structured populations
123
Population dynamics survey
Quantitative surveys were conducted in one commer-
cial mango orchard (LYJ, cultivar ‘Jose
´
’) located on
the dry west coast, the main mango production area
(Fig. 1; Table 1). Populations of larvae that fell from
mango trees were quantified with funnel-shaped traps
(diameter: 28 cm) filled with water and placed on the
ground under susceptible organs: inflorescences or
elongating vegetative growth units (portion of stem
bearing 10–15 leaves). One trap was placed under each
of 20 randomly selected trees. The number and type of
susceptible organs present vertically above the trap
were recorded. From June to October 2010 (flowering
period), the number of larvae per trap was recorded
twice a week. Trapping was carried out for one week
each month from November 2010 to June 2011
(vegetative growth period), under five to ten trees
bearing susceptible organs. Qualitative surveys were
conducted on two other orchards, the first one in the
dry western area (CPEA, cultivar ‘Cogshall’) and the
second one on the southern area (BP, cultivar ‘Cog-
shall’). The presence of P. mangiferae damage on
young leaves and on inflorescences was recorded each
month on all the trees with inflorescences or young
leaves in all three orchards.
Species identification: morphological criteria
and mitochondrial DNA sequencing
To obtain adult gall midges for morphological iden-
tification, final instar larvae were trapped from inflo-
rescences as described above at the CPEA orchard
(cultivar: ‘Cogshall’, season: winter), and from young
leaves (season: summer) and inflorescences (season:
winter) at the LYJ orchard (cultivar: ‘Jose
´
’). These
larvae were placed in sterile sand at 20 °C under a
constant photoperiod (LD 12:12). Just after emer-
gence, ten males and ten females were stored in 70°
ethanol. These samples were prepared according to the
method developed by Pierre (2011), and were identi-
fied using the description of Felt (1918) and Gagne
´
(1994).
For mitochondrial DNA analysis, a fragment of the
cytochrome oxidase subunit 1 (COI) gene (Yukawa
et al. 2003) was sequenced for 51 P. mangiferae
individuals. These individuals were chosen from
among those collected for microsatellite genotyping
Fig. 1 Location of
sampling sites of
P. mangiferae on Reunion
Island. See Table 1 for
detailed information about
the population codes.
Altitude is indicated by
shades of gray. Dotted lines
delineate the three main
geographical and ecological
areas
P. Amouroux et al.
123
in order to encompass a large diversity of conditions.
They were taken from 15 populations collected at 13
sites: 26 individuals collected in mango orchards (20
from the cultivar ‘Cogshall’ and six from the cultivar
‘Jose
´
’) and 25 collected on wild trees; 40 individuals
collected on inflorescences and 11 on leaves; 43
individuals collected in winter and eight in summer;
26 individuals collected in the western area, 16 in the
eastern area and nine in the southern area; and 19
individuals assigned by
STRUCTURE to cluster A and 20
to cluster B (see Results). We also sequenced the COI
gene for three P. matteiana adults that emerged from
the leaves of two mango trees collected in the LYC
orchard. The primers used for the amplification were
as follows: forward, 5
0
-GGATCACCTGATATAGC
ATTCCC-3
0
(COIS) and reverse, 5
0
-CCCGGTAAAA
TTAAAATATAAACTTC-3
0
(COIA) (Funk et al.
1995). The PCR was performed in a final volume of
25 lL, including 2 lL of DNA, 0.5 lM of each
primer, 0.5 of Taq-DNA-polymerase, 1.5 lLofMgCl
2
(25 mM), 2 lLofDNTP(2.5mM),5lLofbuffer
59 (Promega, Colorless GoTaq
Ò
Flexi Buffer) and
completed with HPLC water. We used the following
procedure for amplification: denaturation at 94 °C
Table 1 Characteristics of the 28 sites where P. mangiferae
populations were sampled: site name, population code (Fig. 1),
area (E: Eastern, S: Southern, W: Western), altitude, environ-
ment around sample sites, cultivar (unknown in natural area),
organ (I: Inflorescence, L: Leaf), season (W: Winter; S:
Summer), sampling month and number (N) of individuals
genotyped
Site name Population code Area Altitude (m a.s.l.) Environment Cultivar Organ Season Month N
Bois Blanc ANS E 115 Natural Unknown I W 08 31*
Bassin plat BP1 S 145 Cultivated Cogshall I W 07 31*
BP2 I W 09 32
Palmiste rouge CIL S 975 Natural Unknown I W 09 32*
Lyce
´
e Agricole CPEA1 W 135 Cultivated Cogshall I W 08 31*
CPEA2 I W 09 32
CPEA3 I W 10 30*
CPEA4 L S 02 31
La Possession DLP W 40 Cultivated Jose
´
I W 08 32
Etang-Sale
´
ESH S 210 Cultivated Jose
´
I W 07 32
Ouaki HOU S 90 Cultivated Jose
´
I W 08 32
La Saline LAS W 485 Natural Unknown I W 09 31*
Saint-Gilles LYC1 W 70 Cultivated Cogshall I W 07 30
LYC2 I W 09 32*
LYC3 L S 02 34
Saint-Gilles LYJ1 W 60 Cultivated Jose
´
I W 07 32*
LYJ2 I W 10 30
Mafate MAF – 1,080 Natural Unknown I W 10 5*
Mare longue MAR E 10 Natural Unknown I W 09 30*
Saint-Leu SLE W 250 Cultivated Cogshall I W 08 34
Sainte-Suzanne SSU1 E 5 Natural Unknown I W 08 28*
SSU2 L W 08 32*
SSU3 L S 02 31
Takamaka TAK E 240 Natural Unknown I W 08 30*
Tour des roches TDR W 1 Cultivated Cogshall I W 08 32*
Tremblet TRE E 40 Natural Unknown I W 09 31
Piton Defaud ZPD1 W 115 Cultivated Jose
´
I W 07 35*
ZPD2 I W 09 30
The asterisks in the last column indicate the sites where individuals were chosen for phylogenetic analyses. At Mafate, the five
individuals sampled were only used for phylogenetic analyses
Ecological plasticity, permanent and structured populations
123
for 5 min, 35 cycles consisting of 60 s of denatur-
ation at 94 °C, 50 s of hybridization at 54 °C, and
120 s of elongation at 72 °C. PCR products were
sent to Macrogen, Inc. (Seoul, Korea) for purifica-
tion and sequencing. The nucleotide sequence data
reported in this paper were deposited in the GenBank
nucleotide sequence database with the accession
numbers JQ823184–JQ823234 for P. mangiferae
and JQ823235–JQ823237 for P. matteiana.
Sequences of the COI gene of two other mango gall
midge species available from GenBank were used as
outgroups in this analysis: P. mangicola (Shi) (acces-
sion numbers AB438110 to AB438112) and P. pustu-
lata (accession numbers FJ820163 to FJ820172). The
analyses were performed on 370 bp fragments shared
by the four species (sizes corresponding to fragments
after alignment). First, sequences were aligned in
FASTA format using
MEGA and CLUSTALW (Tamura
et al. 2007). Sequences were then opened in
PAUP
4.0b10 (Swofford 1998) and the NJ tree was built using
the JC69 model of evolution. Then, 56 models of
evolution were tested. Results of this analysis were
tested with
MODELTEST 3.8 software (Posada and
Crandall 1998) on the Web (http://darwin.uvigo.es/
software/modeltest_server.html). For each data parti-
tion, the best-fitting model of nucleotide evolution was
determined using the Akaike Information Criterion as
implemented in
MODELTEST 3.8. In maximum likelihood
searches, values of parameters for rate matrices, site
heterogeneity and invariance were estimated by
MOD-
ELTEST
on a neighbor-joining tree. Likelihood settings
from the best-fit model were TVM ? I, corresponding
to the parameter Lset = 6 on the
MRBAYES analysis.
MRBAYES 3.1 was used for Bayesian phylogenetic
inference (Ronquist and Huelsenbeck 2003). In all the
analyses, four Markov chains were allowed to sample
parameters and tree topologies every 1,000 generations
for 4,000,000 generations. At the end of each run, the
average standard deviation of split frequencies was
below 0.007. Visual inspection of sampled parameter
values with
TRACER (Tracer 2003–2006, MCMC Trace
File Analyser, A. Rambaut and A. Drummond, Uni-
versity of Oxford, UK) showed that values typically
stabilized after 15,000 generations; the first 100,000
generations were discarded as a conservative burn-in.
For each Bayesian analysis,
TRACER was also used to
check that the autocorrelation between samples was
limited enough to produce effective sample sizes that
were greater than 800 for each parameter. Runs of each
analysis performed with
MRBAYES converged with
PSRF values at 1. Trees were visualised with
FIGTREE
(FigTree, 2006–2009, A. Rambaut).
Microsatellites analysis
Whole genome DNA of the 848 larvae was extracted
according to a protocol modified from Delatte et al.
(2005). Ninety-six larvae were individualized in a
96-well microplate. Each well was filled with 25 lLof
buffer (10 mM Tris HCl, 50 mM KCl, 0.45 % Tween
20, 0.45 % Nonidet P40 and proteinase K to a final
concentration of 500 lg/mL). The microplate was
hermetically sealed and placed in an oven at 65 °C
overnight. The following day, the microplate was
briefly centrifuged and 35 lL of HPLC water was
added to each well. Following the protocol and the
microsatellite loci described by Amouroux et al. (2012),
each individual was genotyped at 11 microsatellite loci:
PmCIRB12, PmCIRB10, PmCIRB6, PmCIRC12,
PmCIRC62, PmCIRD12, PmCIRD8, PmCIRE11,
PmCIRE12, PmCIRE5, PmCIRF5 (Online Resource
2). Electrophoretic analyses were conducted on an
automated ABI Prism 3100 Genetic Analyzer (Applied
Biosystems). Alleles were scored using
GENEMAPPER
version 2.5 software (Applied Biosystems).
The diversity of microsatellite loci within each
population was estimated using the observed hetero-
zygosity (H
obs
) and Nei’s (1987) unbiased expected
heterozygosity (H
n.b.
) computed using GENETIX 4.05
(Belkhir et al. 1996). Significance of linkage disequi-
librium between pairs of loci was tested with the
Fisher’s test in
GENEPOP 4.0 (Rousset 2008). Single and
multilocus fixation indices (F
is
) were estimated
according to Weir and Cockerham (1984). Deviations
from the Hardy–Weinberg equilibrium (HWE) were
tested using a two-tailed Fisher’s exact test in
GENEPOP.
We corrected for multiple testing using the Bonferroni
correction. The
MICROCHECKER program version 2.2
(Van Oosterhout et al. 2004) was used to determine
whether the departures from HWE were due to the
presence of null alleles or to genotyping errors.
Genetic discontinuities between samples from mango
organs (inflorescences vs. leaves), mango cultivars
(‘Jose
´
’, ‘Cogshall’, unknown), areas (western, south-
ern, eastern), and seasons (summer vs. winter) were
quantified through hierarchical analysis of molecular
variance (AMOVA) with nonparametric permutation
procedures using
ARLEQUIN software, version 3.5
P. Amouroux et al.
123
(Excoffier and Lischer 2010). The BOTTLENECK pro-
gram, version 1.2 (Cornuet and Luikart 1996), was
used to detect heterozygote excess for individual
populations, considering a two-phase model (TPM) of
microsatellite mutation, a 70 % stepwise-mutation
model (SMM) and a 30 % infinite allele model (IAM),
and 1,000 replications.
FREENA (Chapuis and Estoup
2007) was used to estimate null allele frequencies, to
calculate pairwise estimators of F
ST
(Weir 1996) and
to apply the ENA correction. The corrected estimator
is designated as F
ST
{ENA}
. The genotypic differentiation
between pairs of populations was tested using a
Fisher’s exact test in
GENEPOP, corrected with the
Bonferroni correction for multiple testing.
Levels of population admixture were quantified
using three Bayesian clustering procedures as imple-
mented in
STRUCTURE 2.3 (Pritchard et al. 2000), TESS
2.3 (Chen et al. 2007) and BAPS 5.2 (Corander et al.
2008). With
STRUCTURE, the admixture model was used
to analyze K = 1–20 clusters, with ten runs for each
K, 1,000,000 replicates, and a 100,000 replicates burn-
in. Default values were used for all other parameters.
The number of clusters was determined by the Evanno
et al. (2005) method. A correspondence analysis
(COA) was performed with
GENETIX 4.05 to visualize
the localization of the clusters identified by
STRUCTURE
along the two first axes that describe the genetic
diversity of the whole population. For each single
population, the allelic diversity (mean number of
alleles per locus) within each cluster was also
estimated. The proportion of individuals assigned to
a cluster by
STRUCTURE was compared between areas
using multiple comparison procedures for binomial
generalized linear models developed by Bretz et al.
(2011). These analyses were conducted with R soft-
ware, version 2.13.1 (R Development Core Team
2011), with the ‘multcomp’ library. Using
TESS,we
initially determined the number of clusters K with the
BYM admixture model, 20 runs per K ranging from 2
to 10, 10,000 sweeps and 1,000 burn-ins. Ten
‘‘dummy’’ individuals were added to represent the
upper mountains of Reunion Island (without mango)
where no population could be sampled (Durand et al.
2009). The number of clusters was determined using
the maximum likelihood. Then, for the best K, we used
50 independent runs with 50,000 sweeps and 10,000
burn-ins.
CLUMPP v.1.1 (Jakobsson and Rosenberg
2007) was used to average the assignment scores over
the ten best runs for
STRUCTURE and over the five best
runs for
TESS. The graphics of assignment scores were
generated by
DISTRUCT (Rosenberg 2004). With BAPS,
we estimated clusters using the spatial clustering of
groups for population mixture analysis. Ten repeti-
tions were computed for each K
max
= 3, 5, 10, 15, 20.
According to the logarithm of marginal likelihood,
100 repetitions of the best number of clusters were
performed to determine the best partition.
Results
Species identification
Based on morphological characteristics, adults col-
lected at two sites, two seasons and two organs were all
identified as P. mangiferae (E. Pierre, pers. comm.).
Phylogeny analysis, based on COI mtDNA of 51
P. mangiferae specimens and 16 other individuals from
three Procontarinia species, clearly distinguished the
four species (Online Resource 3). P. mangiferae
sampled in Reunion Island showed a homogeneous
sequence of the COI mitochondrial gene, except for
four individuals that exhibited a sequential variation of
one nucleotide (different for each individual).
Population dynamics survey
The qualitative survey of mango blossom gall midges
in three commercial orchards showed the presence of
damage by P. mangiferae all year round. Monthly
variations of the population density of P. mangiferae
in the LYJ orchard are presented in Fig. 2. Larvae
were trapped both from inflorescences during flower-
ing and from young leaves during vegetative growth.
Mango flowering, from June to October, corresponded
to the highest population density. The monthly larval
density ranged from 0.5 to 9.8 larvae trapped per
inflorescence per day. During vegetative growth, from
November to May, the monthly larval density ranged
from 0.1 to 1.3 larvae trapped per vegetative growth
unit per day. Fresh galls were then limited to some
scattered young leaves and the number of larvae per
organ was the lowest.
Within-population genetic variability
The analysis of 11 microsatellite loci in 27 populations
of P. mangiferae (n = 848) revealed a low genetic
Ecological plasticity, permanent and structured populations
123
variability, with a mean number of alleles per locus
ranging from 2.09 to 3.18, and an observed heterozy-
gosity (H
n.b.
) ranging from 0.106 to 0.228 (Table 2).
Multilocus estimates of F
is
ranged from -0.103 to
0.347 and showed significant heterozygote deficien-
cies in nine out of 27 populations. The average
proportion of missing data per locus ranged from 0 at
locus PmCIRE5 to 0.042 (SD = 0.063) at locus
PmCIRE11, while the average proportion of missing
data per population ranged from 0 in eight populations
to 0.086 (SD = 0.108) in the SLE population. The
average estimated proportion of null alleles was low
(0.065; SD = 0.019). Exact tests for genotypic dis-
equilibrium resulted in six significant values
(P \ 10
-4
) for the PmCIRE12 and PmCIRF5 loci in
three populations (LYC2, MAR and ZPD1), for
PmCIRC12 and PmCIRB12 in one population
(HOU), and for PmCIRD8 and PmCIRF5 in one
population (BP1).
The largest proportion of the overall molecular
variance was due to variation within populations
(Table 3, range: 98.9–99.3 %, P \ 0.001). The effect
of the area (western, eastern, southern) was significant
but the percentage of explanation of the variance was
only 0.26 %. The effects of cultivars, organs and
seasons were not significant. The various tests of
heterozygosity deficit proposed in
BOTTLENECK (Sign
test and Wilcoxon test, TPM model) suggested a
significant deficit of heterozygosity with the SMM
model in 11 populations (ten populations sampled in
winter and one population in summer), indicating that
these populations were in expansion. No significant
heterozygosity excess was detected in the 27 popula-
tions. The low differences between H
n.b.
and H
obs
confirmed the absence of important bottleneck during
the annual population dynamics. According to the
absence of significant season and organ effects in the
AMOVA, we pooled the individuals sampled at
different dates within each site into a single popula-
tion per site. Significant genotypic differentiation
between these 17 site populations was detected
(P \ 3.7 9 10
-4
, with the Bonferroni correction) in
42 of the 136 pairwise comparisons, with F
ST
{ENA}
values ranging from -0.003 to 0.083 (Table 4).
Twenty-nine of these 42 significant pairwise differ-
entiations concerned a population from the eastern
area versus a population from the western area. No
pairwise comparison was significant among popula-
tions within the western area or among populations
within the southern area.
Bayesian clustering
The three software programs,
STRUCTURE, TESS and
BAPS, used to assess the most probable number of
genetically different populations, converged to the
same result. Two clusters, referred to as A and B, were
identified. The individual assignations with
STRUC-
TURE
, for two and three clusters, are presented in the
Fig. 3. The individual assignations for three clusters
did not show sub-structure. The results of admixture in
STRUCTURE for two clusters showed that individuals
were highly assigned to a cluster. Seventy-one out of
848 individuals were assigned at less than 70 % to a
cluster. However, individuals of each cluster were
present in different proportions at all the sites (Fig. 4).
Multiple comparisons for generalized linear models
showed that individuals assigned to the cluster A were
significantly (P \ 0.016) more present in the western
area (mean frequency = 0.41) than in the eastern
(mean frequency = 0.27) and the southern areas
(mean frequency = 0.3). Furthermore, the genetic
COA allowed graphical representation of the major
axes of genetic differentiation within the sample using
2010−06 2010−09 2010−12 2011−03 2011−06
log10 (number of larvae per organ per day)
0.0 0.2 0.4 0.6 0.8 1.0 1.2
**
Ve
g
etative
g
rowth
Flowering
Fig. 2 Population dynamics of the final instar larvae of
P. mangiferae from June 2010 to July 2011 at the LYJ orchard
in Saint-Gilles. Means and standard deviations were log
10
-
transformed. Asterisks indicate observations of damage without
trapped gall midges. Below the x-axis, the mango cycle is
described by gray and black lines, for flowering and vegetative
growth, respectively
P. Amouroux et al.
123
the clustering identified by STRUCTURE (Fig. 5). The
first two axes of the COA represented 100 % of the
entire genetic variation. Significant differences of the
mean number of alleles per locus were observed
between clusters. Individuals belonging to the cluster
A had significantly (P \ 0.01) more alleles per locus
(2.47 ± 0.59) than individuals belonging to the cluster
B (1.89 ± 0.34).
The results of admixture in
TESS showed a contrast
between eastern populations, with more than 60 %
of the individuals belonging to cluster B and no
individual belonging to cluster A, and western popu-
lations, with the individuals belonging to cluster A or
B and numerous low assigned individuals (individual
assignation to a cluster \ 0.7). The highest posterior
probability with
BAPS was obtained for a structure with
two clusters, one that grouped the populations from the
western and southern areas from MAR to DLP, except
for HOU, and the second one that grouped the
populations from the eastern area (SSU, TRE, ANS,
TAK). Populations from MAR, HOU, CIL and BP had
an absolute value of the logarithm of the marginal
Table 2 Mean (standard deviation) value of parameters characterizing the genetic variability of 27 populations of P. mangiferae on
Reunion Island (location code according to Table 1), based on the analysis of 11 polymorphic microsatellite loci
Population
code
Average number
of allele per
locus
H
n.b.
H
obs
F
is
(FSTAT) Proportion of
missing data
Mean estimated
proportion of null
allele
ANS 2.45 0.187 (0.221) 0.135 0.034 0.000 0.058 (0.065)
BP1 2.36 0.209 (0.249) 0.153 0.048 0.000 0.049 (0.107)
BP2 2.27 0.227 (0.221) 0.173 0.034 0.000 0.053 (0.084)
CIL 2.36 0.197 (0.239) 0.154 -0.005 0.020 (0.025) 0.048 (0.085)
CPEA1 2.91 0.296 (0.216) 0.194 0.347* 0.015 (0.022) 0.094 (0.098)
CPEA2 2.82 0.273 (0.245) 0.193 0.296* 0.009 (0.015) 0.086 (0.080)
CPEA3 2.91 0.263 (0.208) 0.209 0.209* 0.018 (0.026) 0.065 (0.084)
CPEA4 2.91 0.247 (0.232) 0.203 0.183 0.012 (0.017) 0.045 (0.064)
DLP 2.82 0.274 (0.232) 0.228 0.005 0.009 (0.015) 0.074 (0.083)
ESH 2.36 0.207 (0.239) 0.137 0.103 0.003 (0.009) 0.075 (0.084)
HOU 2.09 0.208 (0.219) 0.199 -0.145 0.000 0.042 (0.071)
LAS 2.82 0.279 (0.251) 0.221 0.041 0.026 (0.040) 0.078 (0.095)
LYC1 2.91 0.282 (0.239) 0.188 0.336* 0.000 0.096 (0.111)
LYC2 2.45 0.243 (0.236) 0.187 0.237* 0.028 (0.033) 0.077 (0.078)
LYC3 3.18 0.308 (0.238) 0.227 0.267* 0.083 (0.098) 0.084 (0.082)
LYJ1 2.91 0.241 (0.237) 0.184 0.241* 0.003 (0.009) 0.061 (0.082)
LYJ2 2.82 0.246 (0.272) 0.192 0.223 0.012 (0.017) 0.060 (0.074)
MAR 2.27 0.229 (0.228) 0.106 0.289* 0.003 (0.010) 0.100 (0.111)
SLE 3.00 0.264 (0.219) 0.177 0.162 0.086 (0.108) 0.086 (0.089)
SSU1 2.36 0.202 (0.241) 0.159 0.217 0.000 0.039 (0.065)
SSU2 2.27 0.237 (0.245) 0.193 0.186 0.000 0.066 (0.100)
SSU3 2.45 0.268 (0.248) 0.248 0.073 0.012 (0.016) 0.038 (0.061)
TAK 2.36 0.187 (0.199) 0.134 0.045 0.006 (0.013) 0.065 (0.084)
TDR 2.36 0.234 (0.233) 0.190 -0.006 0.003 (0.009) 0.069 (0.090)
TRE 2.18 0.167 (0.214) 0.144 -0.103 0.000 0.030 (0.066)
ZPD1 3.00 0.259 (0.197) 0.191 0.264* 0.013 (0.023) 0.074 (0.104)
ZPD2 2.36 0.218 (0.252) 0.180 0.179 0.018 (0.031) 0.042 (0.064)
Average number of alleles per locus; H
n.b.
expected unbiased heterozygosity, H
obs
observed heterozygosity, F
is
Weir and
Cockerham’s (1984) estimate of Wright’s (1951) fixation index; proportion of missing data; mean estimated proportion of null alleles
* Multilocus deviations from HWE (experiment-wise P \ 0.05 after Bonferroni correction)
Ecological plasticity, permanent and structured populations
123
likelihood smaller than 6.7 (Online Resource 4),
indicating that they could be assigned to one cluster
or to another with only a small change in the likelihood
(Kass and Raftery 1995). In contrast, the other
populations were strongly affected to their respective
cluster with absolute values greater than 11.1.
Discussion
Our study not only characterized the genetic diversity
of the most damaging mango gall midge, P. mangife-
rae, but also gave insights into its ecology as an invader
in a subtropical insular environment. This study was
based on data from population dynamics surveys and
from molecular analyses performed on individuals
sampled in 17 sites in Reunion Island, on two mango
organs and during two different seasons. Our results
confirmed that a single species, P. mangiferae, was
present all year round in different environments
(climatic and cultural), and that it fed on both
inflorescences and young leaves without any genetic
difference. Populations were, however, structured into
two clusters in sympatry and present in different
proportions at each site.
No morphological differentiation was observed
from a pool of individuals collected from both types
of galled organs, young leaves and inflorescences.
Furthermore, a phylogenetic analysis on mtDNA
showed only one dominant haplotype all year round
and throughout Reunion Island, regardless of the
mango cultivar and galled organ. Finally, the low
frequency of null alleles observed in the genetic
analysis of the populations confirmed the presence of a
single species (Chapuis and Estoup 2007; Malausa
et al. 2007).
The1-year survey of gall midges in mango orchards
confirmed that P. mangiferae can feed all year round
on both inflorescences and young leaves, as reported
by Prasad (1971) and Grover and Kashyap (1984).
This result was not expected in Reunion Island since
its presence was only reported during the flowering
season in previous studies (Vincenot and Normand
2009). In subtropical India, the native region of
P. mangiferae, Prasad (1971) found that the entire
populations entered into diapause at the end of the
flowering period. Our results are the first record of
permanently-active populations in a subtropical cli-
mate, thanks to the ability of this gall midge to feed on
different organs of its host plant and to the ability of
Table 3 Analyses of molecular variance (AMOVA) of the 27 populations of P. mangiferae sampled on Reunion Island
Group Populations df SS Percentage of
variation
Variance
components
Fixation Indices
Area (3) All (27) Among groups 2 2.71 0.26 0.0013 FCT 0.0026*
Among populations within
groups
24 17.46 0.83 0.0040 FSC 0.0084*
Within populations 1,669 793.42 98.90 0.4754 FST 0.0197*
Cultivar
(3)
All (27) Among groups 2 1.95 0.08 0.0004 FCT 0.0008
Among populations within
groups
24 18.21 0.94 0.0045 FSC 0.0094*
Within populations 1,669 793.42 98.98 0.4754 FST 0.0102*
Season (2) CPEA, LYC,
SSU
Among groups 1 0.82 0.11 0.0005 FCT 0.0011
Among populations within
groups
8 5.45 0.65 0.0032 FSC 0.0066*
Within populations 612 296 99.24 0.4837 FST 0.0076*
Organ (2) CPEA, LYC,
SSU
Among groups 1 0.67 -0.02 -0.0001 FCT -0.0002
Among populations within
groups
8 5.60 0.72 0.0035 FSC 0.0072*
Within populations 612 296 99.30 0.4837 FST 0.0070*
Genetic variation was broken down according to different factors (number of levels per factor in brackets)
The asterisks indicate the significance of fixation index
P. Amouroux et al.
123
Table 4 Genetic differentiation between sites where populations of P. mangiferae were sampled
Area East South West
Site SSU TAK ANS TRE MAR BP HOU ESH CIL SLE LAS LYC LYJ TDR CPEA ZPD
East
TAK 0.033
ANS 0.031* 0.046
TRE 0.031 0.028 0.006
MAR 0.050*** 0.044 0.025 0.060
South
BP 0.032*** 0.025 0.034*** 0.024 0.054**
HOU 0.016* 0.026 0.040*** 0.023 0.061** 0.018
ESH 0.014 0.045 0.064 0.052 0.083 0.016 0.020
CIL 0.024 0.025 0.052 0.031 0.070* 0.011 0.026 0.009
West
SLE 0.011 0.061 0.041*** 0.059 0.044** 0.040 0.030 0.021 0.037
LAS 0.023** 0.079* 0.069*** 0.078** 0.062*** 0.050 0.028 0.035 0.058 0.008
LYC 0.031*** 0.067*** 0.058*** 0.078*** 0.046*** 0.040* 0.049 0.028 0.045 0.002 0.020
LYJ 0.004 0.037 0.041*** 0.039*** 0.045*** 0.025 0.006 0.012 0.021 0.004 0.005 0.021
TDR 0.026* 0.048 0.053*** 0.054*** 0.058*** 0.010 0.027 0.012 0.012 0.015 0.024 0.010 0.012
CPEA 0.021*** 0.053** 0.046*** 0.054*** 0.051*** 0.024* 0.032 0.015 0.025 0.007 0.017 0.004 0.013 -0.003
ZPD 0.007 0.045 0.023*** 0.037*** 0.029*** 0.02*** 0.017 0.015 0.025 -0.001 0.011 0.011 0.002 0.009 0.008
DLP 0.023 0.041 0.039*** 0.054 0.032*** 0.014 0.030 0.015 0.028 0.009 0.020 0.002 0.012 -0.002 -0.002 0.005
Values of pairwise estimators of F
ST
(Weir 1996) with ENA correction, F
ST
{ENA}
, are presented with their level of significance (Fisher’s exact test for genotypic differentiation with
the Bonferroni correction)
Study-wise type 1 error level: * 0.01 \ P \0.05; ** 0.001 \ P \ 0.01; *** P \ 0.001
Ecological plasticity, permanent and structured populations
123
the mango tree to sporadically produce new inflores-
cences or leaves throughout the year. Consequently,
P. mangiferae exhibits a large ecological plasticity since
it feeds on different organs of a unique host plant and
lives in contrasted climatic and cultural environments.
Other gall midge species produce generations on a
continuous basis, especially in warm and wet regions of
the neotropics, but generally on herbaceous or bush host
plants that grow all year round (Gagne
´
1994). Another
strategy to maintain permanent populations is host shift,
as reported by Freeman and Geoghagen (1989) with
Asphondylia boerhaaviae Mohn. This gall midge pro-
duces about 17 generations per year on three plant
species belonging to the Nyctaginaceae family in
Jamaica.
Despite its permanent presence, populations of
P. mangiferae showed considerable variations in size,
from large numbers of individuals during mango
flowering in winter to barely detectable numbers of
individuals during mango vegetative growth in sum-
mer. The large variation in population size is a factor
that reduces genetic diversity (Nei et al. 1975).
However, genetic analyses did not support the pres-
ence of a bottleneck for the populations sampled at
different sites and months. The genetic diversity of
P. mangiferae was thus maintained throughout the
year. Such a lack of a genetic bottleneck despite a
drastic reduction of the effective population size
during adverse conditions has already been reported
for Anopheles arabiensis Patton (Diptera: Culicidae)
(Kent et al. 2007) and A. funestus Giles (Michel et al.
2006). These authors assumed that this was due to the
0.0 0.2 0.4 0.6 0.8 1.0
A
0.0 0.2 0.4 0.6 0.8 1.0
B
Fig. 3 Results of admixture analysis in STRUCTURE for K = 2
clusters (a) and for K = 3 clusters (b). Individuals are sorted by
percentage of assignation
SSU
TA K
ANS
TRE
MAR
BP
HOU
ESH
CIL
SLE
LAS
LY C
LY J
TDR
CPEA
ZPD
DLP
0.0
0.2
0.4
0.6
0.8
1.0
East South West
Fig. 4 Percentage of P. mangiferae individuals assigned by
STRUCTURE at more than 70 % to the cluster A (black) or to the
cluster B (gray) at each site. Hatched areas correspond to hybrid
individuals that were not assigned at more than 70 % to a
cluster. The horizontal dotted lines indicate the mean percentage
of the cluster A in each area
−0.5 −0.3 −0.1 0.1 0.3 0.5 0.7
−0.9
−0.7
−0.5
−0.3
−0.1
0.1
0.3
F2 (30.3%)
F1 (69.7%)
Fst=0.053
Cluster A
Cluster B
Badly assignated
mtDNA sequenced
Fig. 5 First factorial plane of the genetic correspondence
analysis (AFC 3D per population on
GENETIX). Individuals
belonging to clusters A, B and low assignated individuals, as
identified by
STRUCTURE, are represented by different symbols;
the individuals sequenced on mitochondrial DNA are indicated
by an open circle
P. Amouroux et al.
123
small number of generations produced during the
adverse season, but with a large effective population
size on a large spatial scale, plus migration. These
processes are probably not involved in the mainte-
nance of the genetic diversity of P. mangiferae due to
its short life-span and probably a low migration
capacity (Kolesik 2000). In our case, a first hypothesis
could be that a proportion of the P. mangiferae
populations enter into diapause at each generation.
Backcrosses would then occur during the summer
generations between gall midges that emerge after
diapause and individuals in the process of fulfilling
their normal short life cycle. A second hypothesis
could be that the resources offered by young mango
leaves support sufficient populations to maintain a
high level of genetic diversity in spite of low
individual density.
Pairwise estimators of F
st
, AMOVA and clustering
analyses conducted with three different Bayesian
analyses showed the presence of two clusters in the
populations of P. mangiferae at the Reunion Island
sites. These two clusters were separated in the first
factorial plane of a COA, an independent type of
analysis that makes no assumption regarding Hardy–
Weinberg and linkage equilibrium (Delatte et al.
2006). Individuals belonging to each cluster were
scattered at most of the sample sites, with well-
balanced frequencies of the two clusters in the western
area. Six hypotheses are proposed to explain the
presence of the two clusters in sympatry at each site,
and are discussed. First, clustering may be due to the
specialization of individuals on one mango organ. Our
results showed that populations feeding on different
mango organs were not differentiated by mitochon-
drial or nuclear DNA. Moreover, individuals feeding
on inflorescences and leaves were distributed in the
two clusters. In the second hypothesis, individuals
with continuous development may be different from
diapausing individuals. However, we verified, on 144
adults from six populations (four from CPEA, one
from LYC and one from LYJ), that the pairwise
estimator of F
st
was not significant (F
st
=-0.003;
P [ 0.05; data not shown) and that the allelic richness
was well balanced between adults emerging without
diapause and after diapause. In the third hypothesis,
the gene flow between the two clusters might be
limited by post-zygotic barriers such as secondary
endosymbionts (Zimmer 2001). The fourth hypothesis
invokes the geographic and climatic conditions of
Reunion Island, and the fifth hypothesis is related to
the availability of resources, which is a factor that
could promote population structuring, as was shown
for the Hessian fly (Mayetiola destructor) (Morton
et al. 2011). However, the higher frequency of cluster
A in the western area cannot be explained by climate,
geography or cultural practices alone because of the
close relationship between these three factors. The
western area is subjected to dry conditions and
corresponds to the most suitable area for mango
production. About 70 % of mango orchards are
planted in this area (Vincenot and Normand 2009).
In contrast, the eastern area is the most humid,
unsuitable for economically-viable mango production.
Last hypothesis, agriculture is known to facilitate
invasions by offering large resources, by creating
disturbed sites for colonization and by inducing
population structure (Sakai et al. 2001). In agro-
ecosystems, pests are exposed to the selection pressure
of agricultural practices, inducing traits for persistence
and noxiousness (Sakai et al. 2001). The selection
pressures on insects could be due to insecticide
treatments or environmental perturbations, and they
rapidly induce a genetic structure (Bre
´
vault et al.
2011; Endersby et al. 2006; Franck et al. 2007; White
and Czajkowska 2009). The high density of resources
and agricultural practices could have induced selec-
tion pressures in the mango production area that
allowed diversification after the occurrence of a
bottleneck at the moment of its introduction on the
island. The first mention of this pest was done in 1974
in the western coast where we observed the highest
genetic diversity. The abundant mango resources in
this area may support large population size that
favours the maintaining of its genetic diversity.
Despite a higher diversity in the mango production
area, the entire genetic diversity of this species in
Reunion Island remains low. We probably observed
the beginning of population structuring for this
recently introduced pest species.
As defined by Davis (2009), the successful estab-
lishment of invasive populations requires the accom-
plishment of four tasks by the immigrant individuals:
(1) finding an environment with abiotic conditions
they can tolerate; (2) having access to resources
necessary for their maintenance, growth and repro-
duction; (3) finding a mate; and (4) avoiding pre-
reproductive mortality. Our results showed that
P. mangiferae has the assets that allow it to fulfill
Ecological plasticity, permanent and structured populations
123
the two first tasks. First, the ecological plasticity of
mango blossom gall midge suggests that it could
encounter environments with favorable abiotic condi-
tions. Second, regardless of the phenological season of
mango in the invaded country, it could be capable of
accessing the resources necessary for its reproduction
since larvae are able to feed on mango inflorescences
as well as on young leaves. Moreover, introduced
organisms experience ‘founder effects’, including
genetic bottlenecks that result in significant reductions
in genetic diversity (Ahern et al. 2009). The fact that
P. mangiferae is able to feed on different organs
despite a low genetic diversity is proof of the phe-
notypic plasticity of this species. It contributes to the
ability of this gall midge to establish itself at new sites
and to be invasive (Agrawal 2001). Furthermore,
concerning persistence and spread, P. mangiferae has
other adaptive advantages, including ‘r-selected’ traits
(Davis 2009) such as a high rate of reproduction (each
female can lay up to 150 viable eggs (Prasad 1971))
and a short generation time. Our results showed that
P. mangiferae therefore has numerous characteristics
that define a good invader and could explain its
distribution in mango areas worldwide.
Conclusion
This study showed that despite its low genetic diversity,
the invasive pest P. mangiferae exhibited a large
ecological and phenotypic plasticity by adapting to
different ecological conditions (cultivated vs. wild trees,
dry vs. humid areas, winter vs. summer) and by feeding
on different organs of a unique host plant. It was also
shown that despite low population size in summer, there
was no loss of genetic diversity throughout the year.
Populations were found to be structured according to
their geographic localization, agricultural practices or
climatic conditions. The low genetic diversity suggested
a recent introduction on Reunion Island. The ecological
plasticity and the ability of this species to generate
permanent populations, with a part of diapausing
individuals that preserve the genetic diversity in case
of adverse conditions, are factors favoring its establish-
ment in new areas and contributing to its invasive
success worldwide.
Acknowledgments We are especially grateful to B. Facon
and V. Ravigne
´
for their helpful comments and discussions. We
are also grateful to E. Pierre of the Center for Biology and
Management of Populations (CBGP) for the morphological
identification of P. mangiferae and for his enlightening
discussion on Cecidomyiidae. We thank an anonymous
reviewer and the editor for their critical reading and
suggestions. We would also like to thank L. Maillary for his
helpful assistance during sampling and C. Simiand for his
technical assistance in the laboratory. This work was funded by
the European Agricultural Fund for Rural Development
(EAFRD, Bilan de Sante
´
de la PAC, No 11111D974000019)
and by CIRAD.
References
Agrawal AA (2001) Phenotypic plasticity in the interactions and
evolution of species. Science 294(5541):321–326
Ahern R, Hawthorne D, Raupp M (2009) Founder effects and
phenotypic variation in Adelges cooleyi, an insect pest
introduced to the eastern United States. Biol Invasions
11(4):959–971
Amouroux P, Normand F (2010) Survey of mango pests on
Re
´
union Island, with a focus on pests affecting flowering.
In: Paper presented at the 9th international mango sym-
posium, Sanya, 8–12 April
Amouroux P, Normand F, Nibouche S, Delatte H (2012) Iso-
lation and characterization of microsatellite markers from
Procontarinia mangiferae (Felt). Original manuscript. Mol
Ecol Resour 12(1):185–189
Belkhir K, Borsa P, Chikhi L, Raufaste N, Bonhomme F (1996)
GENETIX 4.05, logiciel sous WindowsTM pour la ge
´
ne
´
-
tique des populations. Laboratoire Ge
´
nome, populations,
interactions CNRS UMR 5000, Universite
´
de Montpellier
II, Montpellier
Blair CP, Abrahamson WG, Jackman JA, Tyrrell L (2005)
Cryptic speciation and host-race formation in a purportedly
generalist tumbling flower beetle. Evolution 59(2):304–
316
Boubou A, Migeon A, Roderick G, Navajas M (2011) Recent
emergence and worldwide spread of the red tomato spider
mite, Tetranychus evansi: genetic variation and multiple
cryptic invasions. Biol Invasion 13(1):81–92
Bretz F, Hothorn T, Westfall P (2011) Multiple comparisons
using R. CRC Press, Boca Raton
Bre
´
vault T, Carletto J, Tribot J, Vanlerberghe-Masutti F (2011)
Insecticide use and competition shape the genetic diversity
of the aphid Aphis gossypii in a cotton-growing landscape.
Bull Entomol Res 101(4):407–413
CAB I (2004) Crop Protection Compedium 2004. CAB Inter-
national, Wallingford
Carletto J, Lombaert E, Chavigny P, Bre
´
vault T, Lapchin L,
Vanlerberghe-Masutti F (2009) Ecological specialization
of the aphid Aphis gossypii Glover on cultivated host
plants. Mol Ecol 18(10):2198–2212
Castillo Del RF (1994) Factors influencing the genetic structure
of Phacelia dubia, a species with a seed bank and large
fluctuations in population size. Heredity 72(5):446–458
Chapuis M-P, Estoup A (2007) Microsatellite null alleles and
estimation of population differentiation. Mol Biol Evol
24(3):621–631
P. Amouroux et al.
123
Chen C, Durand E, Forbes F, Franc¸ois O (2007) Bayesian
clustering algorithms ascertaining spatial population
structure: a new computer program and a comparison
study. Mol Ecol Notes 7:747–756
Cook MA, Ozeroff SN, Fitzpatrick SM, Roitberg BD (2011)
Host-associated differentiation in reproductive behaviour
of cecidomyiid midges on cranberry and blueberry. Ento-
mol Exp Appl 141(1):8–14
Corander J, Siren J, Arjas E (2008) Bayesian spatial modeling of
genetic population structure. Comput Stat 23(1):111–129
Cornuet JM, Luikart G (1996) Description and power analysis of
two tests for detecting recent population bottlenecks from
allele frequency data. Genetics 144(4):2001–2014
Davis MA (2009) Invasion biology. Oxford University Press,
New York
Delatte H, Reynaud B, Granier M, Thornary L, Lett JM, Gold-
bach R, Peterschmitt M (2005) A new silverleaf-inducing
biotype Ms of Bemisia tabaci (Hemiptera: Aleyrodidae)
indigenous of the islands of the south-west Indian Ocean.
Bull Entomol Res 95:29–35
Delatte H, David P, Granier M, Lett JM, Goldbach R, Peters-
chmitt M, Reynaud B (2006) Microsatellites reveal
extensive geographical, ecological and genetic contacts
between invasive and indigenous whitefly biotypes in an
insular environment. Genet Res 87:109–124
Dlugosch KM, Parker IM (2008) Invading populations of an
ornamental shrub show rapid life history evolution despite
genetic bottlenecks. Ecol Lett 11(7):701–709
Dorchin N (2008) Cecidomyiidae (Diptera). In: Capinera JL
(ed) Encyclopedia of entomology, 2nd edn. Springer Sci-
ence, Dordrecht, pp 1576–1580
Durand E, Chen C, Franc¸ois O (2009) TESS version 2.3—
Reference Manual. 30
Endersby NM, McKechnie SW, Ridland PM, Weeks AR (2006)
Microsatellites reveal a lack of structure in Australian
populations of the diamondback moth, Plutella xylostella
(L.). Mol Ecol 15(1):107–118
Etienne J (1977) Notes sur les ce
´
cidomyies du manguier a
`
la
Re
´
union. vol fiche n°27. I.R.A.T., Ile de la Re
´
union, France
Etienne J, Roura A (1974) Note sur l’entomofaune des
manguiers a
`
la Re
´
union. Rapport I.R.A.T. Ile de la
Re
´
union, France
Evanno G, Regnaut S, Goudet J (2005) Detecting the number of
clusters of individuals using the software STRUCTURE: a
simulation study. Mol Ecol 14(8):2611–2620
Excoffier L, Lischer HEL (2010) Arlequin suite ver 3.5: a new
series of programs to perform population genetics analyses
under Linux and Windows. Mol Ecol Resour 10(3):564–
567
Facon B, Genton BJ, Shykoff J, Jarne P, Estoup A, David P
(2006) A general eco-evolutionary framework for under-
standing bioinvasions. Trends Ecol Evol 21(3):130–135
Felt EP (1918) A study of gall midges, VI. Family Itonididae.
Tribe Itonididinariae. In: 33rd report of the state ento-
mologist on injuious and other insects of the State of New
York 1917, vol 202. New York State Museum Bulletin,
pp 76–205
Franck P, Reyes M, Olivares J, Sauphanor B (2007) Genetic
architecture in codling moth populations: comparison
between microsatellite and insecticide resistance markers.
Mol Ecol 16(17):3554–3564
Freeman BE, Geoghagen A (1989) A population study in
Jamaica on the gall-midge Asphondylia boerhaaviae:a
contribution to spatial dynamics. J Anim Ecol 58(2):
367–382
Funk DJ, Futuyma DJ, Orti G, Meyer A (1995) Mitochondrial
DNA sequences and multiple data sets: a phylogenetic
study of phytophagous beetles (Chrysomelidae: Ophrael-
la). Mol Biol Evol 12(4):627–640
Gagne
´
RJ (1989) Family Cecidomyiidae. In: Evenhuis NL (ed)
Catalog of the Diptera of Australasia and Oceania. HI:
Bishop Museum & E. J. Brill., Honululu, pp 152–163
Gagne
´
RJ (1994) The gall midges of neotropical region. Cornell
University Press, New York
Gagne
´
RJ (2004) A catalog of the Cecidomyiidae (Diptera) of
the world, vol 25. Memoirs of the Entomological Society of
Washington Entomological Society of Washington,
Washington, pp 5–350
Gagne
´
RJ (2010) Update for a catalog of the Cecidomyiidae
(Diptera) of the world. Digital version 1.http://www.ars.
usda.gov/SP2UserFiles/Place/12754100/Gagne_2010_
World_Catalog_Cecidomyiidae.pdf. Accessed 2012-04-10
Gagne
´
RJ, Medina CD (2004) A new species of Procontarinia
(Diptera : Cecidomyiidae), an important new pest of
mango in the Philippines. Proc Entomol Soc Wash 106(1):
19–25
Grover P, Kashyap V (1984) On the study of bio-ecology of
certain gall-midges (Cecidomyiiade, Diptera)—a discus-
sion. Cecid Indica V(3):73–127
Hugon R (1979) The complex of gall-forming organisms and
cecidomyids of mango in the Island of Re
´
union. In: Mar-
seilles CoCaIo (ed) Congress on the control of insects in the
tropical environment, Marseille, 13–16 March 1979,
pp 413–422
Jakobsson M, Rosenberg NA (2007) CLUMPP: a cluster
matching and permutation program for dealing with label
switching and multimodality in analysis of population
structure. Bioinformatics 23(14):1801–1806
Jones RG, Gagne
´
RJ, Barr WF (1983) Biology and taxonomy of
the Rhopalomyia gall midges (Diptera: Cecidomyiidae) of
Artemisia tridentata Nuttall (Compositae) in Idaho. Con-
trib Am Entomol Inst 21:1–79
Joy JB, Crespi BJ (2007) Adaptive radiation of gall-inducing
insects within a single host-plant species. Evolution 61(4):
784–795
Kass RE, Raftery AE (1995) Bayes factors. J Am Stat Assoc
90(430):773–795
Kent RJ, Mharakurwa S, Norris DE (2007) Spatial and temporal
genetic structure of Anopheles arabiensis in Southern
Zambia over consecutive wet and drought years. Am J Trop
Med Hyg 77(2):316–323
Kolbe JJ, Glor RE, Rodriguez Schettino L, Lara AC, Larson A,
Losos JB (2004) Genetic variation increases during bio-
logical invasion by a Cuban lizard. Nature 431(7005):
177–181
Kolesik P (2000) Distribution of infestation by lentil gall midge
Contarinia lentis (Dipt., Cecidomyiidae) in lentil fields:
statistical model. J Appl Entomol 124(1):7–10
Launey S, Brunet G, Guyomard R, Davaine P (2010) Role of
introduction history and landscape in the range expansion
of brown trout (Salmo trutta L.) in the Kerguelen Islands.
J Hered 101(3):270–283
Ecological plasticity, permanent and structured populations
123
Le Bellec F, Renard V (1997) Le grand livre des fruits tropicaux.
Editions Orphie, Saint-Denis, Ile de la Re
´
union, France,
pp 189
Lee CE (2002) Evolutionary genetics of invasive species.
Trends Ecol Evol 17(8):386–391
Malausa T, Dalecky A, Ponsard S, Audiot P, Streiff R, Chaval
Y, Bourguet D (2007) Genetic structure and gene flow in
French populations of two Ostrinia taxa: host races or
sibling species? Mol Ecol 16(20):4210–4222
Michel A, Grushko O, Guelbeogo W, Sagnon NF, Costantini C,
Besansky N (2006) Effective population size of Anopheles
funestus chromosomal forms in Burkina Faso. Malar J
5(1):115
Morton PK, Foley CJ, Schemerhorn BJ (2011) Population
structure and spatial influence of agricultural variables on
hessian fly populations in the Southeastern United States.
Environ Entomol 40(5):1303–1316
Nei M (1987) Molecular evolutionary genetics. Columbia
University Press, New york
Nei M, Maruyama T, Chakraborty D (1975) The bottleneck
effect and genetic variability in populations. Evolution
29(1):1–10
Pen
˜
a JE, Mohyuddin AI, Wysoki M (1998) A review of the pest
management situation in mango agroecosystems. Phyto-
parasitica 26(2):129–148
Perdereau E, Dedeine F, Christide
`
s JP, Dupont S, Bagne
`
res AG
(2011) Competition between invasive and indigenous
species: an insular case study of subterranean termites. Biol
Invasions 13(6):1457–1470
Pezhman H, Askari M (2004) A study on the biology of mango
inflorescence midge in Hormozgan province. Appl Ento-
mol Phytopathol 72(1):19–29
Phillips BL, Brown GP, Webb JK, Shine R (2006) Invasion
and the evolution of speed in toads. Nature 439(7078):
803–803
Pierre E (2011) Les dipte
`
res ne
´
matoce
`
res : la pre
´
paration micro-
scopique. http://ephytia.inra.fr/caracterisation/caracterisation_
utilisateur/index_appli.php?portail=bioagresseurs&
produit=caracterisation&main=56&ssrub1=36&ssrub2=
37. Accessed 2012 1 13
Posada D, Crandall KA (1998) Modeltest: testing the model of
DNA substitution. Bioinformatics 14(9):817–818
Prasad SN (1971) The mango midge pests. Cecidological
Society of India, Allahabad
Prasad SN, Grover P (1974) Population fluctuation of Eros-
omyia indica Grover. Cecid Indica IX(1&2):1–38
Pritchard JK, Stephens M, Donnelly P (2000) Inference popu-
lation structure using multilocus genotype data. Genetics
155:954–959
Raman A, Burckhardt D, Harris KM (2009) Biology and
adaptive radiation in the gall-inducing Cecidomyiidae
(Insecta Diptera) and Calophyidae (Insecta Hemiptera) on
Mangifera indica (Anacardiaceae) in the Indian subconti-
nent. Trop Zool 22:27–56
Raunet M (1991) Le milieu physique et les sols de
l’ı
ˆ
le de la Re
´
union, conse
´
quences pour la mise en valeur
agricole. CIRAD, Saint-Denis, Ile de la Re
´
union, France,
pp 438
Ronquist F, Huelsenbeck JP (2003) MrBayes 3: Bayesian phy-
logenetic inference under mixed models. Bioinformatics
12(19):1572–1574
Rosenberg NA (2004) DISTRUCT: a program for the graphical
display of population structure. Mol Ecol Notes 4(1):137–
138
Rousset F (2008) GENEPOP ‘007: a complete re-implementa-
tion of the GENEPOP software for Windows and Linux.
Mol Ecol Resour 8(1):103–106
Sakai AK, Allendorf FW, Holt JS, Lodge DM, Molofsky J, With
KA, Baughman S, Cabin RJ, Cohen JE, Ellstrand NC,
McCauley DE, O’Neil P, Parker IM, Thompson JN, Weller
SG (2001) The population biology of invasive species. Ann
Rev Ecol System 32:305–332
Stireman JO, Janson EM, Carr TG, Devlin H, Abbot P (2008)
Evolutionary radiation of Asteromyia carbonifera
(Dip-
tera: Cecidomyiidae) gall morphotypes on the goldenrod
Solidago altissima (Asteraceae). Biol J Linn Soc 95(4):
840–858
Stuart JJ, Chen MS, Harris MO (2008) Hessian fly. In: Hunter
W, Kole C (eds) Genome mapping and genomics in
arthropods, vol 1. Genome mapping and genomics in ani-
mals. Springer, Berlin, pp 93–102
Suarez AV, Tsutsui ND (2008) The evolutionary consequences
of biological invasions. Mol Ecol 17(1):351–360
Sun XJ, Xiao JH, Cook JM, Feng G, Huang DW (2011) Com-
parisons of host mitochondrial, nuclear and endosymbiont
bacterial genes reveal cryptic fig wasp species and the
effects of Wolbachia on host mtDNA evolution and
diversity. Bmc Evol Biol 11:86–94
Swofford DL (1998) PAUP. Phylogenetic analysis using par-
simony, version 3.1.1. Computer program distributed by
the Illinois National History Survey Champaign
Tabashnik BE (1983) Host range evolution: the shift from native
legume hosts to alfalfa by the butterfly, Colias philodice
eriphyle. Evolution 37(1):150–162
Tamura K, Dudley J, Nei M, Kumar S (2007) MEGA4:
molecular evolutionary genetics analysis (MEGA) soft-
ware version 4.0. Mol Biol Evol 24:1596–1599
R Development Core Team (2011) R: a language and environ-
ment for statistical computing. R foundation for statistical
computing. ISBN 3-900051-07-0. Website http://www.R-
project.org/. Vienna
Uechi N, Yukawa J (2006) Life history patterns and host ranges
of the genus Asphondylia (Diptera: Cecidomyiidae). In:
Galling Arthropods and Their Associates. Springer, Tokyo,
pp 275–285
Van Oosterhout C, Hutchinson WF, Wills DPM, Shipley P
(2004) MICRO-CHECKER: software for identifying and
correcting genotyping errors in microsatellite data. Mol
Ecol Notes 4(3):535–538
Vincenot D, Normand F (2009) Guide de production inte
´
gre
´
ede
mangues a
`
la Re
´
union. CIRAD et Chambre d’Agriculture,
Saint-Pierre, Ile de la Re
´
union, France
Ward NL, Masters GJ (2007) Linking climate change and spe-
cies invasion: an illustration using insect herbivores. Glob
Change Biol 13(8):1605–1615
Weir BS (1996) Genetic data analysis II. Sinauer Associates,
Sunderland
Weir BS, Cockerham CC (1984) Estimating F-statistics for the
analysis of population structure. Evolution 38(6):1358–
1570
White PF, Czajkowska M (2009) Effects of methoprene and
permethrin on paedogenetic larvae of Heteropeza pygmaea
P. Amouroux et al.
123
and Mycophila speyeri (Diptera: Cecidomyiidae). J Econ
Entomol 93(4):1222–1226
Whitwell AC (1993) The pest/predator/parasitoid complex on
mango inflorescence in Dominica. Acta Hortic 341:421–
432
Wright S (1951) The genetical structure of populations. Ann
Eugen 15:323–354
Yukawa J (2000) Synchronization of gallers with host plant
phenology. Popul Ecol 42(2):105–113
Yukawa J, Uechi N, Horikiri M, Tuda M (2003) Description of
the soybean pod gall midge, Asphondylia yushimai sp n.
(Diptera: Cecidomyiidae), a major pest of soybean and
findings of host alternation. Bull Entomol Res 93(1):73–86
Zimmer C (2001) Wolbachia: a tale of sex and survival. Science
292(5519):1093–1095
Ecological plasticity, permanent and structured populations
123
