Linking patterns and processes of species diversification in the cone flies Strobilomyia (Diptera: Anthomyiidae).

Abstract

Phytophagous insects provide useful models for the study of ecological speciation. Much attention has been paid to host shifts, whereas situations where closely related lineages of insects use the same plant during different time periods have been relatively neglected in previous studies of insect diversification. Flies of the genus Strobilomyia are major pests of conifers in Eurasia and North America. They are specialized feeders in cones and seeds of Abies (fir), Larix (larch) ,and Picea (spruce). This close association is accompanied by a large number of sympatric Strobilomyia species coexisting within each tree genus. We constructed a molecular phylogeny with a 1320 base-pair fragment of mitochondrial DNA that demonstrated contrasting patterns of speciation in larch cone flies, as opposed to spruce and fir cone flies; this despite their comparable geographic distributions and similar resource quality of the host. Species diversity is the highest on larch, and speciation is primarily driven by within-host phenological shifts, followed by allopatric speciation during geographical expansion. By contrast, fewer species exploit spruce and fir, and within-host phenological shifts did not occur. This study illustrates within-host adaptive radiation through phenological shifts, a neglected mode of sympatric speciation.

Full text

Page 1

Molecular Phylogenetics and Evolution 41 (2006) 606–621
www.elsevier.com/locate/ympev
1055-7903/$ - see front matter © 2006 Elsevier Inc. All rights reserved.
doi:10.1016/j.ympev.2006.06.005
Linking patterns and processes of species diversiWcation
in the cone Xies Strobilomyia (Diptera: Anthomyiidae)
Jean-Marie Sacheta,¤, Alain Roquesb, Laurence Desprésa
a Laboratoire d’Ecologie Alpine, UMR 5553, Université Joseph Fourier, BP 53, 38041 Grenoble Cedex 09, France
b Unité de Zoologie Forestière, INRA Centre d’Orléans, BP 20619, 45166 Olivet Cedex, France
Received 13 February 2006; accepted 7 June 2006
Available online 9 June 2006
Abstract
Phytophagous insects provide useful models for the study of ecological speciation. Much attention has been paid to host shifts,
whereas situations where closely related lineages of insects use the same plant during diVerent time periods have been relatively neglected
in previous studies of insect diversiWcation. Flies of the genus Strobilomyia are major pests of conifers in Eurasia and North America.
They are specialized feeders in cones and seeds of Abies (Wr), Larix (larch) ,and Picea (spruce). This close association is accompanied by a
large number of sympatric Strobilomyia species coexisting within each tree genus. We constructed a molecular phylogeny with a 1320
base-pair fragment of mitochondrial DNA that demonstrated contrasting patterns of speciation in larch cone Xies, as opposed to spruce
and Wr cone Xies; this despite their comparable geographic distributions and similar resource quality of the host. Species diversity is the
highest on larch, and speciation is primarily driven by within-host phenological shifts, followed by allopatric speciation during geograph-
ical expansion. By contrast, fewer species exploit spruce and Wr, and within-host phenological shifts did not occur. This study illustrates
within-host adaptive radiation through phenological shifts, a neglected mode of sympatric speciation.
© 2006 Elsevier Inc. All rights reserved.
Keywords: Adaptive radiation; Ecological speciation; Host shift; Mitochondrial DNA; Molecular phylogeny; Oviposition behavior; Phenological shift
1. Introduction
The study of species concepts and speciation has under-
gone many developments since Darwin (Pigliucci, 2003). In
the 1930s, research shifted away from the study of natural
selection as the driving force of speciation, and highlighted
the fundamental role of geographical reproductive isolation
in population diVerentiation (Bush, 1975; Coyne and Price,
2000; Mayr, 1963). Sympatric speciation was long consid-
ered impossible because the conditions required were
thought unlikely to occur in nature (Bush, 1994). In the past
few years, however, ecological shifts have been shown to
play a major role in the process of speciation. They are at
the heart of the mechanism of ecological speciation (Orr
and Smith, 1998; Rundle and Nosil, 2005; Schluter, 2001;
Wiens, 2004); according to which, populations living in
diVerent environments or using diVerent resources face con-
trasting selection pressures on traits that can ultimately
lead to reproductive isolation and the formation of new
species. Because of their tremendous biodiversity, phytoph-
agous insects have been particularly well studied with
respect to sympatric or ecological speciation (Berlocher and
Feder, 2002; Roderick and Gillespie, 1998). Much attention
has focused on host shifts and the evolution of host races
into new species (Abrahamson et al., 2003; Drès and Mal-
let, 2002; Emelianov et al., 2004; Filchak et al., 2000;
Thomas et al., 2003; Via, 1999), but other interesting situa-
tions arise when insect clades specialize in diVerent anatom-
ical parts of the same hosts, or use the same resource during
diVerent time periods (Groman and Pellmyr, 2000).
For example, the tephritid Xies of the genus Blepharone-
ura are highly host speciWc, and some species also specialize
*Corresponding author. Fax: +33 0 4 76 51 42 79.
E-mail address: Jean-Marie.Sachet@ujf-grenoble.fr (J.-M. Sachet).

Page 2

J.-M. Sachet et al. / Molecular Phylogenetics and Evolution 41 (2006) 606–621
607
on diVerent organs within a single host-plant species: male
Xowers, female Xowers, or seeds (Condon and Steck, 1997).
Field data and molecular phylogenetics show that organ
shifts have occurred many times during the evolution of the
Andricus gallwasps (Cook et al., 2002; Stone et al., 2002).
Allochronic speciation has been demonstrated in the peri-
odical cicadas of the genus Magicicada (Cooley et al., 2003;
Marshall and Cooley, 2000; Simon et al., 2000). Divergence
in the timing of development and reproduction at a sea-
sonal scale has also been documented. The North American
aphid Pemphigus obesinymphae forms galls on the same
host as Pemphigus populi-transversus, but has a truncated
life cycle without egg diapause (Sokal et al., 1991). It is sus-
pected that this divergence arose sympatrically without
host shift (Abbot and Withgott, 2004). Three cases of spe-
cies divergence by change in ovipositor length and
phenological shift have been demonstrated in the Apocryp-
tophagus parasitic Wg wasps (Weiblen and Bush, 2002). A
similar mechanism has been proposed to explain the origin
of three closely related sympatric species of Megarhyssa
parasitoid wasps (Gibbons, 1979; Ramadevan and Deakin,
1990). Three adaptive syndromes have been described on
the Wg-breeding Xies Lissocephala, corresponding to corre-
lated morphological, behavioral, and ecological traits,
including diVerences in the oviposition time and site on the
Wg (Harry et al., 1998). Finally, phylogenetic analyses of the
Chiastocheta Xies, pollinators, and seed predators of the
globeXower Trollius (Ranunculaceae), have shown that
host shifts cannot explain all speciation events (Després
et al., 2002). Several sister-species can be found on one host-
plant species, diVering only in their timing of oviposition
(Després and Jaeger, 1999; Pellmyr, 1989). At the genetic
level, laboratory experiments on the daily locomotive activ-
ity of Drosophila show that mutations on clock genes can
lead to allochronic isolation (Tauber et al., 2003).
Adaptive radiation, the evolution of ecological and phe-
notypic diversity within a rapidly multiplying lineage (Gav-
rilets and Vose, 2005), can be observed in several of the
preceding examples (Condon and Steck, 1997; Després and
Jaeger, 1999; Gibbons, 1979; Harry et al., 1998; Weiblen
and Bush, 2002). It produces a concentration of speciation
events soon after the evolution of key innovation or coloni-
zation of unoccupied habitats. Practically this results in
speciation bursts at the base of molecular phylogenies,
which can be used to reveal the historical factors underly-
ing these adaptive radiations (Gavrilets and Vose, 2005;
Lovette and Bermingham, 1999).
The objective of this study is to test the relative roles of
three mechanisms of speciation in one particular insect
genus: speciation by geographical isolation (vicariance),
speciation by host shift, and speciation by phenological
shift within a single host-plant. The cone Xies of the genus
Strobilomyia are excellent candidates for studying such
mechanisms because of their wide geographic distribution
and their extensive adaptive radiation (Michelsen, 1988;
Turgeon et al., 1994). Molecular phylogenies are a widely
used tool to understand the patterns and processes of speci-
ation (Barraclough and Nee, 2001), and here we present
results inferred from the sequence-analysis of a 1320 base-
pair (bp) fragment of mitochondrial DNA. We compared
this phylogeny with the taxonomy, distribution range and
ecology (host-plant, phenology, and oviposition behavior)
of the species to test the relative roles of geographical isola-
tion, host shifts, and phenological shifts in driving specia-
tion (Barraclough and Vogler, 2000; Berlocher and Feder,
2002; Losos and Glor, 2003). If geographical isolation is the
primary condition for speciation, then most sister-species
should be distributed across non-overlapping ranges; if
host shift is the main mechanism of speciation, then most
sister-species should occur on diVerent host-plants. Finally,
if phenological shift within a host-plant is the main mecha-
nism of speciation, sister-species should occur on the same
host-plant but with divergence in the seasonal timing of
reproduction and development.
2. Biology and taxonomy of Strobilomyia cone Xies
Cone Xies of the genus Strobilomyia Michelsen (Diptera:
Anthomyiidae) are among the most serious cone-and-seed
pests of conifers (Turgeon et al., 1994). They are taxonomi-
cally close to the genus Chiastocheta (Pellmyr, 1989), but
they are strict parasites and do not contribute to the polli-
nation of their hosts. Their larval instars develop exclu-
sively in cones of various genera of Pinaceae, namely Wr,
larch, and spruce (Michelsen, 1988). Until the 1980s, precise
knowledge of the speciWc distribution and life cycle of
Strobilomyia species was limited by taxonomic uncertain-
ties; for instance, most cone-Xy damage on larch across
Eurasia was erroneously attributed to a single species, Las-
iomma (Strobilomyia) laricicola (Roques et al., 2003). The
genus Strobilomyia was created in 1988 following a revision
of cone-Xy taxonomy by Michelsen, based on morphologi-
cal (especially genitalic) features. So far, 20 Strobilomyia
species are recognized (Michelsen, 1988; Roques et al.,
1996), of which 12 have been recorded on larch (Larix
Miller), 5 on Wr (Abies Miller), and 3 on spruce (Picea Ditr).
Several species groups have been deWned by Michelsen
(1988) using morphology (Table 1). The geographic distri-
bution of the genus is large, including boreal forests and
alpine regions of the Palearctic and the Neartic.
The biology of the diVerent species is similar. In spring,
the eggs are laid on developing cones, and the larvae tunnel
into the cone, feeding on both cone tissues and seeds. Full-
grown third-instar larvae drop to the ground where they
pupate and over-winter. Larval development is closely syn-
chronized with the phenology of their host’s cones (Brocker-
hoV and Kenis, 1997; Roques et al., 1984, 2003). One
noteworthy feature of the genus Strobilomyia is the number
of sympatric species that infest the same host, especially
those species exploiting larch. Three species that colonize
larch have been identiWed in the French Alps, and six in
Northern China (Table 1). A single cone can contain up to
four larvae, often belonging to diVerent species (Roques
et al., 1984, 2003). Several ecological and behavioral traits

Page 3

608
J.-M. Sachet et al. / Molecular Phylogenetics and Evolution 41 (2006) 606–621
contribute to resource partitioning; such as larval feeding
pathways inside the cone, oviposition behavior, and seasonal
timing of adult emergence and larval development. Although
the periods of activity largely overlap, three groups of species
(early, late, and intermediate; Table 1) have been described
on larch based on these phenological patterns (McClure
et al., 1996; Roques et al., 1984, 2003). Each of the three spe-
cies exploiting larch in the Alps is member of one of these
groups, and there is a time lag of about 2 weeks between the
oviposition periods of two successive species (Roques et al.,
1984). The same pattern can be found in northeastern China
(Roques et al., 2003; Yao et al., 1991), where 2 weeks separate
the oviposition periods of the early species (S. laricicola) and
the Wrst intermediate species (S. melaniola). The other four
species lay their eggs during the two following weeks, strictly
following the sequence, S. svenssoni, then S. baicalensis, S.
infrequens and S. luteoforceps. The eggs can be easily identi-
Wed by their position on the cone, which is related to the ovi-
positor length of the females. For example, in the Alps the
early species (S. laricicola) lays eggs at the base of the cone,
and its ovipositor is short. The intermediate and late species
(S. melania and S. infrequens) lay eggs inside the cone,
between scales, and they have longer ovipositors (Michelsen,
1988; Roques et al., 1983).
3. Methods
3.1. Sampling
All described species of Strobilomyia except 5 (S. abietis,
S. luteoforceps, S. macalpinei, S. svenssoni, and S. suwai)
were sampled. Except for S. abietis, there are few records of
these rare species (Anderton and Jenkins, 2001; Belova
et al., 1998; Roques et al., 2003). They do not form a partic-
ular group of Strobilomyia, as they exploit diVerent hosts
and are present in diVerent parts of the world (Table 2),
though two of the three larch-exploiting species of the late
ecotype are missing. Every eVort was made to collect at
least two individuals per species and locality. Adult Xies
were identiWed by genital examination (Michelsen, 1988;
Roques et al., 2003) and preserved in absolute alcohol
before DNA extraction, except samples from Krasnoyarsk
(Siberia), Japan and some of those from Canada, that were
kept dried and pinned in collection boxes.
3.2. Sequencing
Genomic DNA was extracted from the head of the Xy
using the DNeasy Tissue Kit (Qiagen). The head and tho-
rax were used in the extraction for those Xies that were not
conserved in alcohol. We used the Polymerase Chain Reac-
tion (PCR) to amplify a region of the mitochondrial
genome containing most of the cytochrome oxydase I
(COI) and II (COII) genes. The two fragments were sepa-
rately ampliWed using forward primers COI-2171 (5?-TTG
ATTTTTTGGTCAYCCNGAAGT-3?)
3023 (5?-GATTAGTGCAATGGATTTAGCTC-3?), and
reverse primers tRNAleu-3048 (5?-TGGAGCTTAAA
TCCATTGCAC-3?) and COII-3683 (5?-CCRCAAATTT
CTGAACATTGACC-3?), as deWned by Després and Jae-
ger (1999). No ampliWcation of the whole COI fragment
was possible with the dried specimens, so we used internal
and tRNAleu-
Table 1
ClassiWcation, host-plants, and distribution of the Strobilomyia spp
The ecological groups are those described on larch by Roques et al. (1984), McClure et al. (1996), and Roques et al. (2003). The species groups are those
described by Michelsen (1988) using morphological features, and only some species were attributed to a species group.
Ecological groupSpecies groupSpecies Host-plant Geographical distribution
Early species
laricicolaS. laricicola
S. laricis
Larix
Larix
Palearctic (Europe, Siberia, China, Kamchatka, Japan)
Nearctic
Intermediate species
S. baicalensis
S. svenssoni
Larix
Larix
Eastern Palearctic (Eastern Siberia, Manchuria, Kamchatka)
Manchuria, Mongolia
melania S. lijiangensis
S. melania
S. melaniola
S. sibirica
S. viaria
Larix
Larix
Larix
Larix
Larix
Yunnan
Europe
Manchuria
Western Siberia
Eastern Siberia, Nearctic
Late species
S. macalpinei
S. infrequens
S. luteoforceps
Larix
Larix
Larix
Western Canada
Palearctic (Europe, Siberia, China, Kamchatka)
Manchuria, Japan
abietis S. abietis
S. oriens
Abies
Abies
Nearctic
Japan
S. carbonaria
S. suwai
S. todocola
Abies
Abies
Abies
Europe
Japan
Japan
anthracinaS. anthracina
S. appalachensis
S. neanthracina
Picea
Picea
Picea
Palearctic
Nearctic
Eastern Nearctic

Page 4

J.-M. Sachet et al. / Molecular Phylogenetics and Evolution 41 (2006) 606–621
609
Table 2
Collection localities, host-plants, number of individuals sequenced (Ni), number of haplotypes (Nh), haplotype identiWcation numbers, and GenBank accession numbers
SpeciesCollection localityLatitude and longitudeRegion or countryHost-plantNiNh Haplotype id. Nos.GenBank Accession Nos.Collectors
S. baicalensis
Da Khinggan Mountains
Esso-Krapivnaïa
51°28?N, 124°29?E
55°53?N, 159°19?E
Manchuria
Kamchatka
Larix gmelini
Larix cajanderi (Larix gmelini)
1
2
1
2
14
4
6
DQ191578, DQ191526
DQ191571, DQ191519
DQ191573, DQ191521
A.R., J.-H.S.
A.R., P.K.
S. infrequens
Briançon44°53?N, 6°39?E France
Larix decidua
2235
38
15
16
30
9
DQ191589, DQ191537
DQ191591, DQ191539
DQ191579, DQ191527
DQ191580, DQ191528
DQ191586, DQ191534
DQ191575, DQ191523
A.R., J.-P.R.
Da Khinggan Mountains51°28?N, 124°29?EManchuria
Larix gmelini
73 A.R., J.-H.S.
Esso-Krapivnaïa55°53?N, 159°19?E Kamchatka
Larix cajanderi (Larix gmelini)11 A.R., P.K.
S. laricicola
Briançon 44°53?N, 6°39?E France
Larix decidua
22 10
34
20
13
20
21
88
97
DQ191576, DQ191524
DQ191588, DQ191536
DQ191581, DQ191529
DQ191577, DQ191525
DQ191581, DQ191529
DQ191582, DQ191530
DQ191604, DQ191552
DQ191610, DQ191558
A.R., J.-P.R.
Krasmoyarsk
Da Khinggan Mountains
58°05?N, 92°46?E
51°28?N, 124°29?E
Siberia
Manchuria
Larix sibirica
Larix gmelini
2
3
1
2
N.B.
A.R., J.-H.S.
Bibai 43°33?N, 141°86?EJapan
Larix kaempferi (Larix leptolepis)22 K.K.
S. laricis
Bruce Mines 49°22?N, 82°10?WCanada
Larix laricina
44 65
66
79
80
DQ191596, DQ191544
DQ191597, DQ191545
DQ191602, DQ191550
DQ191603, DQ191551
J.T.
S. lijiangensis
Mt Yuelongxueshan 27°01?N, 100°09?EYunnan
Larix potaninii
22 41
42
DQ191593, DQ191541
DQ191594, DQ191542
A.R., Y.-Z.P.
S. melania
Briançon44°53?N, 6°39?E France
Larix decidua
3332
37
39
39
DQ191587, DQ191535
DQ191590, DQ191538
DQ191592, DQ191540
DQ191592, DQ191540
A.R., J.-P.R.
Ry56°05?N, 9°45?EDanemark
Larix leptolepis
11A.R.
S. melaniola
Da Khinggan Mountains51°28?N, 124°29?E Manchuria
Larix gmelini
3324
26
27
DQ191583, DQ191531
DQ191584, DQ191532
DQ191585, DQ191533
A.R., J.-H.S.
S. sibirica
Krasmoyarsk58°05?N, 92°46?E Siberia
Larix sibirica
33 53 DQ191595, DQ191543
DQ191619, DQ191567
DQ191621, DQ191569
DQ191583, DQ191531
N.B.
113
120
24Da Khinggan Mountains51°28?N, 124°29?EManchuria
Larix gmelini
11 A.R., J.-H.S.
S. viaria
Esso-Krapivnaïa55°53?N, 159°19?EKamchatka
Larix cajanderi (Larix gmelini)431
5
7
DQ191570, DQ191518
DQ191572, DQ191520
DQ191574, DQ191522
DQ191599, DQ191547
DQ191600, DQ191548
DQ191601, DQ191549
A.R.
Bruce Mines 46°31?N, 83°79?W Canada
Larix laricina
33 75
77
78
J.T.
S. carbonaria
Briançon 44°53?N, 6°39?E France
Abies alba
2193 DQ191608, DQ191556A.R., J.-.P.R.
S. oriens
Ashibetsu43°52?N, 142°19?E Japan
Abies sacchalinensis
2289
98
DQ191605, DQ191553
DQ191611, DQ191559
K.K.
(continued on next page)

Page 5

610
J.-M. Sachet et al. / Molecular Phylogenetics and Evolution 41 (2006) 606–621
primers PE-301 (5?-GGAGTAGTATTAGCTAATTCT
TC-3?) and RPE-358 (5?-CTACATAATATGTGTCATG
AAG-3?), as deWned by Després et al. (2002). Thermal cycling
conditions for ampliWcation were: denaturation at 94°C for
30s, annealing at 48–50°C for 45s, and extension at 72°C for
90s. Double-stranded DNA products were puriWed using a
QIAquick PCR PuriWcation Kit or QIAquick Gel Extraction
Kit (Qiagen) after excision from a 1.6% agarose gel.
Each strand was sequenced using ampliWcation primers
and ABI PRISM BigDye Terminator Cycle Sequencing
Ready Reaction Kit (Applied Biosystems) following the
manufacturer’s instructions. Fluorescently labeled sequenc-
ing products were analyzed on an ABI PRISM 3100 Capil-
lary DNA Sequencer then aligned and corrected with
SeqScape version 2.0 (Applied Biosystems). The homolo-
gous sequence of Chiastocheta dentifera (Diptera, Antho-
myiidae; GenBank Accession No. AH010112) was used as
the outgroup. We tested other outgroups: Drosophila mela-
nogaster (Diptera: Drosophilidae; GenBank Accession No.
AJ400907) and Delia brassicae (Diptera: Anthomyiidae;
GenBank Accession Nos. AF325362 and AF325363).
Changing outgroup did not change the topology of the tree,
so we kept the phylogenetically closest outgroup. Sequences
were easily aligned (coding regions, no indels) and analyzed
with MEGA2 (Kumar et al., 2001).
3.3. Phylogenetic analysis
Phylogenetic analyses were performed using three
methods: for maximum parsimony (MP), we used PAUP*
version 4.0b10 (SwoVord, 2002), with a heuristic search
and tree-bisection-reconnection (TBR) branch swapping
with 10 random addition sequence replicates, applying
bootstrap analyses with 100 replicates. For maximum
likelihood (ML), we used the Akaike Information Crite-
rion as implemented in Modeltest version 3.5 (Posada and
Crandall, 1998) to select a general-time reversible model
of nucleotide substitution, accounting for among-site rate
variation using invariant sites and a gamma-distributed
rate correction across sites (GTR+I+G). Analysis was
performed using PHYML (Guindon and Gascuel, 2003)
with 1000 bootstrap replicates. This program was used
instead of PAUP* because it is more time-eYcient for ML
calculation. For Bayesian inference, we used MrBayes
version 3.0 (Huelsenbeck and Ronquist, 2001). We ran
four simultaneous Markov chains for 2,000,000 genera-
tions, starting from random initial trees, and sampled the
results of the analysis every 100 generations. The consen-
sus phylogeny and posterior probability of nodes were
calculated from data following the 1,000,000th genera-
tion, after conWrming that likelihood values of the four
Markov chains had stabilized. This process was repeated
four times independently to account for the possibility of
multiple optima. For each method, the consensus tree was
calculated using the 50% majority rule. We also tested
alternative topologies representing diVerent hypothesis of
Strobilomyia species relationships using the Shimodaira–
Table 2 (continued)
A.R., A. Roques; J.-H.S., J.-H. Sun; J.-P.R., J.-P. Raimbault; J.S., J. Sweeney; J.T., J. Turgeon; K.K., K. Kamijo; M.K., M. Kenis; N.B., N. Belova; P.K., P. Khomentovsky; Y.-Z.P., Y.-Z. Pan.
Species
Collection locality
Latitude and longitude
Region or country
Host-plant
Ni
Nh
Haplotype id. Nos.
GenBank Accession Nos.
Collectors
S. todocola
Ashibetsu
43°52?N, 142°19?E
Japan
Abies sacchalinensis
2
2
91
DQ191607, DQ191555
K.K.
100
DQ191612, DQ191560
S. anthracina
Briançon
44°53?N, 6°39?E
France
Picea abies
2
2
69
DQ191598, DQ191546
A.R., J.-P.R.
96
DQ191609, DQ191557
Ayent
44°51?N, 6°24?E
Switzerland
Picea abies
4
2
69
DQ191598, DQ191546
M.K.
103
DQ191614, DQ191562
Krasmoyarsk
58°05?N, 92°46?E
Siberia
Picea obovata
1
1
69
DQ191598, DQ191546
N.B.
Asahikawa
43°77?N, 142°36?E
Japan
Picea jezoensis
2
1
90
DQ191606, DQ191554
K.K.
S. appalachensis
Bettsburg
42°20?N, 75°55?W
Canada
Picea mariana
2
2
102
DQ191613, DQ191561
J.T.
107
DQ191616, DQ191564
105
DQ191615, DQ191563
S. neanthracina
Queensbury
46°00?N, 67°24?W
Canada
Picea glauca
4
4
111
DQ191617, DQ191565
J.S.
112
DQ191618, DQ191566
114
DQ191620, DQ191568

Page 6

J.-M. Sachet et al. / Molecular Phylogenetics and Evolution 41 (2006) 606–621
611
Hasegawa test (Shimodaira and Hasegawa, 1999) as
implemented in PAUP* version 4.0b10 (SwoVord, 2002).
The molecular clock hypothesis was tested using the
likelihood ratio test (LRT), by comparing the ML trees
obtained with or without assuming a molecular clock (Fel-
senstein, 1988). To test whether lineages evolve at similar
rates, relative-rate tests (RRT) were performed with the
program RRTree v1.1 (Robinson-Rechavi and Huchon,
2000). We considered only well-supported clades (sup-
ported by bootstrap values or posterior probabilities
>95%) and each one was compared with its sister groups
separately, using C. dentifera used as the outgroup in all
analyses. Rates of divergence were estimated in our phylog-
eny using Bayesian analysis as implemented in the program
BEAST v1.3 (Drummond and Rambaut, 2003; Drummond
et al., 2002). We used an uncorrelated exponential clock
model, which is considered to be more accurate than other
models of rate change (Ho et al., 2005). Rates were sampled
every 1000 generations from 10,000,000 Markov Chain
Monte Carlo steps, with a burn-in period of 1,000,000
cycles. We used a GTR+I+G model of molecular evolu-
tion and the ML tree as topology prior in the Bayesian
analysis. The analysis was repeated four times indepen-
dently, and convergence of the chains to the stationary dis-
tribution was conWrmed visually with the program Tracer
version 1.3 (Rambaut and Drummond, 2004). To calibrate
the relaxed clock model, we used a rate of evolution of 2%
pairwise divergence per million years (My), which is a
widely accepted rate for mitochondrial cytochrome oxy-
dase genes in insects (Brower, 1994; DeSalle et al., 1987).
Characters (continent of origin, host race, and phenol-
ogy in the case of the larch-attacking species) were mapped
using the Ancestral State Reconstruction Packages for the
program Mesquite v1.06 (Maddison and Maddison, 2005).
Character states were treated as unordered and most parsi-
moniously mapped on the consensus phylogeny.
4. Results
4.1. Sequence analysis
The total length of the sequenced fragments was 772
nucleotides for the COI region, and 548 nucleotides for the
COII region, resulting in a total of 1320 nucleotides. As
expected, there was no variation in length of the coding
sequences. We sequenced a total of 67 individuals, and
found 52 diVerent haplotypes. In total, 226 (17%) of the
sites were variable and 186 (14%) were parsimony informa-
tive. Among these sites, 178 (79%) variable sites were in the
third-codon position, 35 (15%) in the Wrst codon position,
and 13 (6%) in the second codon position. The COI and
COII regions showed similar levels of variability. As usu-
ally found in insect mitochondrial DNA (Brower and DeS-
alle, 1998), there was a strong bias in nucleotide
composition, with a majority of bases T (40.0%) and A
(33.2%). This was especially true for the third-codon posi-
tion where 47.1% of the nucleotides were T and 47.1% were
A. Base frequencies were homogeneous across individuals
(?2D15.64, dfD156, pD1.00). The corrected distances
between the Strobilomyia haplotypes ranged from 0.076%
to 7.98%, and the average transition/transversion ratio was
equal to 5.4. The translation of the nucleotide sequence
resulted in a 440 amino acid sequence; among them 32 (7%)
amino acid positions were variable and 17 (4%) were parsi-
mony informative. We observed no nonsense mutations.
4.2. Phylogenetic analysis
The trees obtained by majority rule consensus from the
three methods showed only minor diVerences, so we sum-
marized the results into one single tree (Fig. 1). The only
relevant diVerence was the position of the S. carbonaria
haplotype, which was found at the base of the tree with the
MP and Bayesian analyses, and as a sister-species of S. inf-
requens with the ML analyses, but none of these positions
were supported by bootstrapping or Bayesian posterior
probabilities. The MP analysis yielded 24 trees, each of 450
steps (CID0.673, RID0.886). The log likelihood score of
the best ML tree was ¡4399.98731.
Individuals belonging to the same species usually formed
monophyletic groups, but this was not always the case.
Three individuals described as S. sibirica formed a paraphy-
letic group, whereas the fourth was found among the S. mel-
aniola individuals, and had the same haplotype as one of
them (although it was not extracted or sequenced the same
day, excluding any contamination problems). Strobilomyia
viaria was divided in two geographic groups, one paraphy-
letic group with the Asian individuals, and one monophy-
letic group with the Canadian individuals. However, forcing
their monophyly did not produce a signiWcantly diVerent
tree from the best ML tree (Table 3). Strobilomyia anthra-
cina was also separated into two groups (clades G and I,
Fig. 1), the Wrst with Wve European individuals and the Sibe-
rian individual, and the second with one European individ-
ual and the two Japanese individuals. In this case, forcing
the monophyly of these two groups led to a tree that was
signiWcantly diVerent from the best ML tree (Table 3). The
11 other species were monophyletic, with bootstrap support
and Bayesian posterior probabilities of at least 99%.
If we consider the classiWcation of Michelsen (1988),
only one of the four described groups (the melania group)
was found to be monophyletic in our phylogeny (clade A):
it groups the species S. melania, S. melaniola, S. sibirica,
and S. viaria. Constraining the monophyly of the laricicola
group (formed by S. laricicola, clade B, and S. laricis, clade
D) gave a tree that was not signiWcantly diVerent from the
best ML tree. However, adding S. baicalensis (clade H) as a
sister group, as suggested by Michelsen (1988), was rejected
(Table 3). The anthracina group contains the three species
exploiting spruce, but the species are not clustered by host
genus in the consensus tree (Fig. 2A), and forcing the
monophyly of the species exploiting Wr or the monophyly
of the species exploiting spruce resulted in trees signiW-
cantly diVerent from the best ML tree (Table 3). By

Page 7

612
J.-M. Sachet et al. / Molecular Phylogenetics and Evolution 41 (2006) 606–621
Fig. 1. Cladogram of the Strobilomyia genus, obtained by computing the consensus of the 50% majority rule bootstrapped trees of the maximum parsi-
mony (MP), maximum likelihood (ML), and Bayesian analysis. Values on branches are bootstrap values from MP (100 replicates) and ML (1000 repli-
cates) analysis, and posterior probabilities from Bayesian inference. Bootstrap values or posterior probabilities less than 50% are not reported or indicated
by dashes. Each haplotype has an identiWcation number, and the number in parenthesis shows the number of individuals sharing this haplotype, if it is
higher than one.
26
24 (2)
27
7
5 (2)
1
113
120
53
78
77
75
37
32
39 (2)
42
41
20 (3)
13
21
97
88
34
10
15
9
16 (5)
30
38
35
93 (2)
96
69 (5)
91
100
6
4
14
103
90 (2)
Chiastocheta dentifera
114
111
105
112
107
102
79
65
66
80
98
89
S. melaniola
S. sibirica
(Manchuria)
S. viaria
(Kamchatka)
S. sibirica
(Siberia)
S. viaria
(Canada)
S. melania
(Europe)
S. lijiangensis
(Yunnan)
S. laricicola
(Manchuria and
Siberia)
S. laricicola
(Japan)
S. laricicola
(Europe)
S. infrequens
(Kamchatka and
Manchuria)
S. infrequens
(Europe)
S. carbonaria
(Europe)
S. laricis
(Canada)
S. oriens
(Japan)
S. neanthracina
(Canada)
S. appalachensis
(Canada)
S. anthracina
(Europe and
Siberia)
S. todocola
(Japan)
S. baicalensis
(Kamchatka and
Manchuria)
S. anthracina
(Europe and Japan)
59/67/0.96
56/62/0.93
-/59/0.71
56/58/0.78
-/-/0.54
100/100/1.00
82/71/-
A
64/61/0.92
100/100/1.00
100/99/1.00
100/99/0.98
100/100/1.00
100/100/1.00
54/80/1.00
70/86/0.55
90/86/0.87
91/97/0.98
B
98/99/1.00
100/100/1.00
60/67/0.99
83/97/0.83
C
100/100/1.00
85/96/0.98
66/77/0.95
93/96/1.00
100/100/1.00
D
100/100/1.00
E
62/66/0.83
100/100/1.00
F
93/96/1.00
100/100/1.00
100/100/1.00
G
100/100/1.00
100/99/1.00
-/-/0.60
64/90/0.86
100/100/1.00
H
88/89/0.99
I

Page 8

J.-M. Sachet et al. / Molecular Phylogenetics and Evolution 41 (2006) 606–621
613
contrast, the monophyly of the species exploiting larch
(which are more numerous) is not rejected. At least seven
host shifts are needed to explain the phylogeny (Fig. 2A).
Only a few species are grouped by region or even conti-
nent of origin (Fig. 2B). This is the case for the Asian spe-
cies S. melaniola, S. sibirica, and some S. viaria (in clade A),
also for S. appalachensis and S. neanthracina (clade F), two
Canadian species. Otherwise, species are not geographically
assembled in our phylogeny, and the most parsimonious
scenario implies 10 continent shifts (Fig. 2B).
Some clades with high bootstrap support or Bayesian
posterior probabilities match a classiWcation based on eco-
logical features of species exploiting larch (Fig. 2C), such as
the phenology of attack (Roques et al., 1983, 2003). The
early species (S. laricicola, clade B, and S. laricis, clade D)
are separated in the consensus tree, but as seen above, the
monophyly of these two species was not statistically
rejected. The intermediate species are also divided into two
groups (clades A and H), but forcing their monophyly did
not produce a statistically diVerent tree (Table 3). S. inf-
requens, the only late species present in our sample, is
monophyletic (clade C). The character state reconstruction
suggests Wve phenological shifts during the evolution of the
larch-exploiting species (Fig. 2C).
When incorporating the molecular clock hypothesis, we
obtained six trees with a log likelihood score of
¡4447.29896. This was signiWcantly diVerent from the best
ML tree (2?¡lnLD94.6233, dfD51, p<0.001), so we
rejected the molecular clock hypothesis. The RRT revealed
that two clades (A and C) evolved signiWcantly faster than
four other groups (E, F, H, and I; Table 4). This result con-
Wrms the LRT test, and shows that the rate of evolution is
heterogeneous in our phylogeny. The comparisons within
the clades deWned in Fig. 1 did not show any signiWcant
diVerences (data not shown).
We used four coalescent models to date interior nodes
with the relaxed clock model: constant population size,
exponential growth, logistic growth and expansion growth,
and we found that the estimation varied little between the
models. Means and 95% conWdence intervals of all estimate
dates for nodes of interest are indicated in Fig. 3. This
reveals an estimation for the Wrst Strobilomyia branching
event of about 3.65My. We also dated the Wrst nodes of
clades A, B, C, and F, which correspond to groups of allo-
patric species with similar ecologies, and of clade G, which
groups two diVerent host species.
5. Discussion
5.1. Molecular clock and divergence dates
The molecular phylogeny is overall well resolved, with
an important number of clades supported by high boot-
strap values or Bayesian posterior probabilities (Fig. 1). But
the resolution of the internal nodes of the tree is poor, with
short branch lengths and eight clades starting from the base
of the tree. This star-like pattern has already been described
in other studies on fast adaptive radiations (Baldwin and
Sanderson, 1998; Hughes and Vogler, 2004; Jordan et al.,
2003; Kambysellis et al., 1995; Lovette and Bermingham,
1999; Megens et al., 2004; Thomas and Hunt, 1991) and is
consistent with models of adaptive radiation which predict
a burst of speciation soon after colonization rather than a
more-or-less continuous process of speciation (Gavrilets
and Vose, 2005).
The pairwise genetic distances are high enough to diVer-
entiate the haplotypes and give good resolution, and low
enough to limit the occurrence of multiple substitutions
(DeFilippis and Moore, 2000), as demonstrated by the high
Table 3
Shimodaira–Hasegawa tests of alternate topologies based on hypothesis of Strobilomyia species relationships
¤SigniWcantly diVerent topology from the best ML tree.
Hypothesis
¡lnL of the best tree
4399.98731
4436.80003
4418.81578
4433.46363
4414.92133
4424.13524
p
Best tree overall
Monophyly of S. anthracina
Monophyly of S. viaria
Monophyly of S. baicalensis, S. laricicola, and S. laricis
Monophyly of the early larch species (S. laricicola and S. laricis)
Monophyly of the intermediate larch species (S. baicalensis, S. lijiangensis, S. melania, S. melaniola, S.
sibirica, and S. viaria)
Monophyly of the larch species (S. baicalensis, S. infrequens, S. laricicola, S. laricis, S. lijiangensis, S.
melania, S. melaniola, S. sibirica, and S. viaria)
Monophyly of the spruce species (S. anthracina, S. appalachensis, and S. neathracina)
Monophyly of the Wr species (S. carbonaria, S. oriens, and S. todocola)
0.034¤
0.321
0.049¤
0.457
0.188
4417.62311 0.341
4445.13379
4454.18549
0.007¤
0.002¤
Table 4
p-values of the relative-rate tests between Strobilomyia groups as described
in Fig. 1
S.c. means S. carbonaria.
¤Lineages evolving at signiWcantly diVerent rates.
ABC
S.c.
DEFGH
B
C
S.c.
D
E
F
G
H
I
0.053
0.546
0.067
0.102
0.005¤
0.025¤
0.161
0.044¤
0.002¤
0.233
0.888
0.796
0.398
0.693
0.764
0.828
0.251
0.186
0.291
0.037¤
0.083
0.364
0.184
0.017¤
0.215
0.481
0.830
0.649
0.958
0.347
0.175
0.483
0.953
0.624
0.120
0.612
0.268
0.521
0.778
0.454
0.853
0.330
0.578
0.092 0.268

Page 9

614
J.-M. Sachet et al. / Molecular Phylogenetics and Evolution 41 (2006) 606–621
CI value. Thus our molecular marker seems to be adequate
for this level of resolution, and has been extensively and
eVectively used for genus-scale phylogenies in insects (Cate-
rino et al., 2000; Mallarino et al., 2005). Homoplasy is low,
and the short distances at the base of the phylogeny reXect
fast divergence of the Wrst lineages rather than poor resolu-
tion due to saturation of the marker.
The molecular clock hypothesis is rejected, which is illus-
trated by the diVerences in branch lengths and rates of evo-
lution (Fig. 1). This has been documented in other rapid
radiations, especially in island endemic radiations (Brom-
ham and WoolWt, 2004; Jordan and McDonald, 1998),
where bottlenecks or relaxation of negative selection can
produce strong variation in the rates of molecular evolu-
tion among lineages. As no fossil evidence is known for
Strobilomyia or any related Anthomyiidae (Michelsen,
1988), and as the molecular clock hypothesis is rejected, we
used a relaxed clock model to estimate the Wrst major radia-
tion of the Strobilomyia genus: it occurred 3.65My ago
(with a 95% conWdence interval of 1.65–6.75My), during
the Pliocene (Fig. 3).
5.2. Phylogeny and taxonomy
We Wnd some strong diVerences from the accepted tax-
onomy and deWnitions of Strobilomyia species. The most
surprising discrepancy is the division of S. anthracina into
two well-diVerentiated groups (Fig. 1, Table 3): the Wrst
contains one Siberian and Wve European individuals, and
the second two Japanese and one European individual,
with a corrected genetic distance of 1.1% between them.
This genetic diVerentiation is not related to geography or
host-plant species (Table 2). Until now, a division within
this species has never been suspected (BrockerhoV and
Kenis, 1997; Michelsen, 1988) and no variation has been
noted in the genitalic structure. Nevertheless, it is probable
that we face two diVerent species, and further studies on
their anatomy or ecology could identify diVerences in sup-
port of the genetic evidence. S. viaria is also divided into
two groups, but these are geographical groups: Asian
against North American individuals. Accordingly, slight
anatomical diVerences were noted between Chinese and
American individuals of S. viaria (Michelsen, 1988), sug-
gesting that they could be treated as distinct species. On the
other hand, the three groups that constitute the S. melani-
ola, S. sibirica, and Asian S. viaria individuals are para- or
polyphyletic, with low genetic distances (0.1–0.6%) between
the haplotypes and low bootstrap support (Fig. 1). Assem-
bling these three groups forms a well-supported monophy-
letic clade. Michelsen (1988) assembled the species S.
sibirica and S. viaria in the melania group, and did not con-
sider S. melaniola as a diVerent species from S. viaria, con-
trary to Fan who Wrst described this species (Roques et al.,
2003): the present molecular results support Michelsen’s
Fig. 2. Schematics of the phylogenetic tree from Fig. 1, showing diVerent character states of the individuals sequenced. The most parsimonious ancestral
character states were reconstructed with Mesquite v1.06 (Maddison and Maddison, 2005). (A) Host-plant used: black, larch-exploiting species; dark gray,
Wr-exploiting species; light gray, spruce-exploiting species. (B) Geographical origins: black, North American individuals; dark gray, Asian individuals;
light gray, European individuals. (C) Phenological pattern (larch-exploiting species only): black, late species; dark gray, intermediate species; light gray,
early species; white, undeWned state.
larch
fir
spruce
North America
Asia
Europe
late
intermediate
early
ABC

Page 10

J.-M. Sachet et al. / Molecular Phylogenetics and Evolution 41 (2006) 606–621
615
hypothesis. Furthermore, they suggest that S. sibirica
should not be treated as a diVerent species from S. melani-
ola and the Asian populations of S. viaria. Further investi-
gation should be performed to conWrm this, and to
determine if the American populations of S. viaria can be
treated as a distinct species from this assemblage.
Polyphyletic species occur among insects and in other
taxa (Funk and Omland, 2003), and can result from inaccu-
rate description and demarcation of the species. But distinct
species or lineages subjected to convergent evolution of
similar morphologies also have been described among
insect groups (Kim et al., 2000; Rees et al., 2001a). Intro-
gression and incomplete lineage sorting represent other
sources of polyphyly, especially with mitochondrial mark-
ers (Ballard, 2000; Funk and Omland, 2003), and can aVect
the resolution of the phylogenic tree. A nuclear marker
could conWrm, or on the contrary invalidate, our interpreta-
tions (Megens et al., 2004; Shaw, 2002; Sota and Vogler,
2001). Introgression could indeed be the source of the low
diVerentiation between S. melaniola, S. sibirica, and S. via-
ria.
Several species show a strong within-species geographi-
cal structuration. The S. infrequens individuals are well sep-
arated into a European clade and an Asian clade, with high
bootstrap values and Bayesian posterior probabilities sup-
porting the two clades separated by low genetic distances
Fig. 3. Maximum likelihood phylogram of the Strobilomyia genus. Taxa are in the same order as in Fig. 1. Relaxed clock estimates are indicated for Wve
nodes in My, with 95% conWdence interval in brackets.
S. melaniola
S. sibirica
(Manchuria)
S. viaria
(Kamchatka)
S. sibirica
(Siberia)
0.01
Chiastocheta dentifera
S. viaria
(Canada)
S. melania
(Europe)
S. lijiangensis
(Yunnan)
S. laricicola
(Manchuria and Siberia)
S. laricicola (Japan)
S. laricicola
(Europe)
S. infrequens
(Kamchatka and Manchuria)
S. infrequens (Europe)
S. carbonaria (Europe)
S. laricis
(Canada)
S. oriens
(Japan)
S. neanthracina
(Canada)
S. appalachensis (Canada)
S. anthracina (Europe and Siberia)
S. todocola (Japan)
1.06
A
(0.51 – 1.70)
0.60
(0.22 – 1.10)
S. baicalensis
(Kamchatka and Manchuria)
S. anthracina (Europe and Japan)
0.29
(0.06 – 0.63)
0.62
(0.21 – 1.21)
3.65
(1.45 – 6.75)
B
C
D
E
F
G
0.43
(0.13 – 0.83)
H
I

Page 11

616
J.-M. Sachet et al. / Molecular Phylogenetics and Evolution 41 (2006) 606–621
(0.3–0.6%). The same pattern is found for S. laricicola,
which can be divided into three well-supported clades:
Europe, China and Siberia, and Japan. The genetic dis-
tances between the two Asian clades are low (0.3–0.5%), but
higher if compared with the European group (1.8–2.0%).
This last distance is equivalent to the genetic distances sep-
arating S. anthracina and S. appalachensis (1.3–2.0%), and
S. melania and S. sibirica (1.7–1.9%), which are considered
vicariant species in this genus (Michelsen, 1988; Roques
et al., 2003). It is then probable that in the case of the Asian
and European populations of S. laricicola, we face two
cryptic vicariant species, even if no anatomical diVerences
have been described (Michelsen, 1988). Other cases of cryp-
tic vicariant species with no morphological diVerences have
been described in insects and other taxa (Fukatsu et al.,
2001; Funk and Omland, 2003; Riddle et al., 2000). The
genetic distances between the European and Asian popula-
tions of S. infrequens, or within the Chinese-Siberian and
Japanese populations of S. laricicola, are too low to put
forward similar conclusions in these cases, but it is possible
that we face diverging allopatric populations that could
ultimately give birth to new species.
The discrepancies between the recognized taxonomy and
classiWcation of Strobilomyia and the molecular phylogeny
illustrate the diYculties and the bias in using genitalic traits
or molecular data alone to describe species and infer phylo-
genetic relationships in arthropods (Hosken and Stockley,
2004; Jocqué, 2002). Genitalic characters are often the only
variable traits available for taxonomists, but their associa-
tion with molecular methods can provide a more accurate
deWnition of species, and a better knowledge of the phylog-
eny and the relationships between the diVerent species (Will
et al., 2005).
5.3. Allopatric speciation
There is no relation between the molecular phylogeny
and the geographic distribution of the Strobilomyia species.
Three well-supported clades (the melania group, S. larici-
cola, and S. infrequens) group species or populations from
diVerent continents (Figs. 1 and 2B), so we can exclude the
hypothesis of independent radiations in Europe, Asia, and
North America. Considering that the genus displays its
highest species diversity in northeastern China, it was previ-
ously supposed to be the center of origin of the genus
(Michelsen, 1988). Our phylogeny supports this hypothesis,
as Asia is considered the continent of origin of the genus in
the most parsimonious evolutive scenario (Fig. 2C).
But changes in the distribution of species can strongly
limit the geographical interpretation of phylogenies (Bar-
raclough and Vogler, 2000; Losos and Glor, 2003), and
such events occurred steadily during the evolution of this
genus. With respect to our hypotheses, the more recent
cause of speciation in Strobilomyia seems to be geographi-
cal isolation. Instances of geographical isolation occurred
independently at least 10 times in the evolution of the genus
(Fig. 2C), and at diVerent periods for the involved groups:
the Wrst event of divergence in clade A is dated at 1.06,
0.62My for clade F, 0.60My for clade B, and 0.29My for
clade C (Fig. 3). The 95% conWdence intervals of these esti-
mates overlap, but this pattern does not support a scenario
in which the geographical expansion of the diVerent groups
was synchronous. These dates all correspond to the Pleisto-
cene, and periods of glaciation and deglaciation are likely
to have had strong eVects on the distribution ranges of the
Strobilomyia species, as they had on their hosts (Brubaker
et al., 2005). Thus extinction and recolonization events may
have occurred, but cannot be recovered in a phylogeny.
5.4. Host shifts
The Strobilomyia species are not grouped with respect to
their host-plant (Fig. 2A), and our results show that if mono-
phyly of the larch-exploiting species is a reliable hypothesis, it
is rejected in the case of Wr- and spruce-exploiting species
(Table 3). All host shifts seem to take place at the origin of
the genus (Fig. 2A), except one: S. todocola (a Japanese spe-
cies exploiting Wr) and one group of S. anthracina (exploiting
spruce) form a well-supported clade, but the Japanese haplo-
types of S. anthracina constitute another clade (Fig. 1). It is
diYcult to infer the geographical context (allopatry or symp-
atry) or the direction of the host shift in this case, given that
both the original host and the geographical origin of this
group are unknown.
The lack of resolution at the internal nodes and the poly-
phyly of at least two of the three groups of species attacking
the same host genus prevent us from drawing conclusions
about the ancestral host of the Strobilomyia genus (Barrac-
lough and Nee, 2001). We cannot infer it from the sister
groups of Strobilomyia, because none of the most closely
related genera, Acklandia, Chirosia, and Egle, exploit coni-
fers (Michelsen, 1988). Whatever this ancestral host, host
shifts have been numerous in the evolution of the genus (at
least Wve). Furthermore, divergence of the three host-plant
genera (Abies, Larix, and Picea) was dated from the begin-
ning of the Cretaceous (Wang et al., 2000), and divergence
of the Larix species from the Eocene (Semerikov et al.,
2003), estimates that are both far older than the radiation
of Strobilomyia (during the Pliocene). Thus the hypothesis
of cospeciation between Strobilomyia and their host-plants
can be excluded. Moreover, though the Strobilomyia species
are highly speciWc to a particular host-plant genus, they are
not speciWc to a particular species within this genus. Many
species have a geographic range including several host-
plant species, and have been recorded on most of them
(Michelsen, 1988; Roques et al., 2003). Our sampling took
this into account (Table 2), but in the results the individuals
were not separated by host species, or this eVect was con-
founded with the eVect of geography. This lack of speciWc-
ity at the species level has also been observed in seed
orchards where European larch-exploiting species can suc-
cessfully infest any imported species of larch, with the pos-
sible exception of the North American Larix laricina
(Roques, unpublished observations).

Page 12

J.-M. Sachet et al. / Molecular Phylogenetics and Evolution 41 (2006) 606–621
617
5.5. Allochronic isolation
If we consider only the larch-exploiting species, a strong
correlation is observed between the phylogeny and the eco-
logical groups based on the phenology of attack (Fig. 3C).
Two ecological groups are split into two in the phylogeny,
but their monophyly is not statistically rejected (Table 3).
Thus it appears that the divergences in the periods of attack
of sympatric species of Strobilomyia did not evolve inde-
pendently in diVerent regions, and are strongly implicated
in the fast adaptive radiation at the origin of the genus.
However, the sampling design may have biased our analy-
sis, because two late species out of three could not be
obtained. Such ecological groups have been described only
on larch, and seem to be the major cause of high species
diversity of Strobilomyia on this host-plant (12 described
species out of the 20 composing the genus; Michelsen, 1988;
Roques et al., 1996). On the other hand, phenological
groups have not been described for the species specialized
on spruce or Wr, and there is only one described species of
Strobilomyia exploiting spruce or Wr in one given locality,
except for the three Wr-exploiting species in Japan (S.
oriens, S. suwai, and S. todocola; Michelsen, 1988).
5.6. Patterns of species richness
DiVerences in species richness and modes of speciation
which depend on host-plant genus are noteworthy. Several
hypotheses based on resource traits have been proposed to
explain the patterns of species richness in phytophagous
insects, and more generally in animals exploiting patchy
resources: the resource distribution hypothesis (Kelly and
Southwood, 1999), the resource size hypothesis (Päivinen
et al., 2003), the resource abundance hypothesis (Brändle
and Brandl, 2001; Marques et al., 2000), the resource con-
centration (or resource distribution over the landscape)
hypothesis (Bukovinszky et al., 2005), and the enemy-free
hypothesis (JeVries and Lawton, 1984).
The resource distribution hypothesis predicts that wide-
spread plant species are able to support a richer local fauna
of herbivores; but Wr, larch, and spruce all have extremely
large distribution ranges, while that of larch appears to be a
bit smaller (VidakoviT, 1991). However, larch species have a
larger altitudinal distribution, at least in Europe (Vidak-
oviT, 1991), and this could in itself be a cause of the higher
richness of Strobilomyia species on larch. This large altitu-
dinal range could also promote phenological shifts and
have an indirect impact on the number of Strobilomyia spe-
cies: local Xy populations living at diVerent altitudes emerge
at diVerent times, in synchrony with cone development
(Roques et al., 1984).
According to the resource size and resource abundance
hypotheses, larger hosts that oVer more resources are able
to support more species of herbivore and are easier to Wnd
than smaller hosts that oVer limited resources. But this is
unlikely in our case because the tree species attacked in the
genera Abies and Picea usually oVer cones that are 2–5
times larger than these of the Larix species (VidakoviT,
1991), except L. griYthiana and L. potaninii in the Himala-
yas, but they sustain less species of Strobilomyia than any
other location. Moreover, the Asian Larix gmelini, which
has the smallest cones among the larches of Eurasia (1.5–
2.5cm; VidakoviT, 1991), is also the tree sustaining the
highest number of Strobilomyia species (Roques et al.,
2003). Cone abundance may however diVer both in space
and time among Abies, Picea, and Larix. Indeed, masting
phenomenon is common in both genera, with larger crops
generally occurring at 2–4 years intervals in Abies spp., 3–5
(England) to 12–13 (Finland) in Picea abies, and 3–10 in
Larix decidua (Shopmeyer, 1974). In support of the
resource abundance hypothesis, the total number of cones
per tree during years of large cone crops is usually higher in
larch (several thousands on mature trees) than in spruce
and Wr (several hundreds to one thousand maximum).
The resource concentration hypothesis predicts that
plant species that occur in high-density patches are able to
support high species richness of phytophages. But the three
Strobilomyia host-plant genera are dominant species of
subalpine and boreal forests, where their densities do not
strongly diVer (VidakoviT, 1991). Finally, according to the
enemy-free hypothesis, resources sustaining a lower num-
ber of competitors or predators provide more opportunity
for colonization and radiation (JeVries and Lawton, 1984;
Murphy, 2004). There is a huge diversity of insects and
associated parasitoids exploiting cones of Pinaceae, belong-
ing to seven orders, and showing diVerent degrees of spe-
cialization (Turgeon et al., 1994). At least in Europe, there
are no obvious diVerences in the number or the diversity of
phytophagous competitors between the three plant genera
exploited by Strobilomyia. In that area, a total of 13 species
exploit spruce cones and 9 species those of Wr, including one
Strobilomyia species for each, whereas 13 species exploit
larch cones including three Strobilomyia species (Roques,
1989). The scarcity of available data does not allow such a
comparison for parasites and predators. More precise stud-
ies should be performed to test all these hypotheses, and
could reveal interesting host-plant traits that could be
involved in the diVerent speciation patterns observed in
Strobilomyia groups exploiting diVerent host-plant genera.
5.7. Evolutionary scenario
The origin of the genus Strobilomyia can be dated to the
Pliocene. The Wrst step of the evolution of Strobilomyia cor-
responds to an adaptive radiation due to ecological factors,
when all the ecotypes appeared and all the host shifts took
place, except for the last one (divergence between S. todo-
cola and one group of S. anthracina). Such a pattern, with a
concentration of speciation events at the base of a phylog-
eny, is a signature of adaptive radiation (Gavrilets and
Vose, 2005; Lovette and Bermingham, 1999), and in our
case phenological and host shifts are the causes of this radi-
ation. It is possible that these speciation events occurred in
allopatry, with the divergent lineages adopting diVerent

Page 13

618
J.-M. Sachet et al. / Molecular Phylogenetics and Evolution 41 (2006) 606–621
adaptations during geographical isolation and reinforcing
these adaptations through secondary contacts, but this usu-
ally requires a long time to proceed (Berlocher and Feder,
2002). Phenological shifts represent a particular case,
because they can be direct consequences of geographical
isolation; for example, when there is a strong diVerence in
latitude or in altitude between isolated populations (Jiggins
and Bridle, 2004). But the present study involves three
groups of species with divergent phenologies (larch-exploit-
ing species), with fully overlapping distribution ranges. The
hypothesis that at least part of this adaptive radiation
occurred in sympatry seems more probable. There is a
growing literature showing that ecological shifts often
occur in sympatry (Abrahamson et al., 2003; Després and
Jaeger, 1999; Drès and Mallet, 2002; Filchak et al., 2000;
Schluter, 2001; Via, 1999), and this is supported by a large
number of theoretical models (Dieckmann and Doebeli,
1999; Ferdy et al., 2002; Gavrilets, 2003; Kirkpatrick and
Ravigné, 2002; Kondrashov and Kondrashov, 1999; Turelli
et al., 2001).
Vicariance was the major mechanism of speciation in the
Strobilomyia genus after this Wrst ecological radiation. The
genus probably extended its range from its center of origin
to all the places where host-plants were available. As the
divergence rates are not homogeneous across vicariant spe-
cies or groups, it is probable that the Pliocene glaciations
induced extinctions of some groups and secondary coloni-
zations, or on the contrary accelerated divergence by
increasing isolation (Bermingham and Martin, 1998;
Knowles, 2000).
5.8. Speciation patterns
The speciation pattern observed in Strobilomyia, involv-
ing Wrst adaptive radiation (through ecological shifts), then
geographical expansion, is extremely concordant with that
described for the globeXower Xies of the genus Chiastoch-
eta. They are members of the same family (Anthomyiidae)
and have a similar geographic range in Eurasia, but are
absent in North America. Their biology is also comparable,
because they are seed parasites of their host (genus Trol-
lius), and exhibit a similar phenological divergence between
three groups of sympatric species (early, intermediate, and
late), linked with diVerent oviposition behaviors (Després
and Jaeger, 1999). These three ecological groups Wrst
diverged in sympatry, probably in Europe, before the genus
expanded to all Eurasia. Several vicariant species appeared
during this expansion phase (Després et al., 2002). Esti-
mated ages of the adaptive radiations of the two genera are
not extremely diVerent (3–4My for Strobilomyia and 2My
for Chiastocheta).
Such fast adaptive radiations are well known in other
systems, especially in island endemic arthropods (Bromham
and WoolWt, 2004; Gillespie and Roderick, 2002; Kamby-
sellis et al., 1995; Rees et al., 2001b), birds (Sato et al., 1999),
lizards (Losos et al., 1998), plants (Baldwin and Sanderson,
1998), and in the African cichlid Wshes (KornWeld and
Smith, 2000). In these cases, the appearance of new species
or ecological groups follows colonization of a new island or
a new lake. The presence of vacant ecological niches in new
islands or lakes facilitates the evolution of a single specialist
species into a large array of diVerent ecotypes (Gillespie
and Roderick, 2002; Rees et al., 2001a), and a comparable
situation can occur when an insect colonizes a new host.
However, in some studies, the diVerent ecotypes were
shown to have independently colonized other islands of an
archipelago as they appeared, leading to secondary allopat-
ric speciation (Cryan et al., 2001; Gillespie, 2002; Jordan
et al., 2003; Roderick and Gillespie, 1998). These patterns
are similar to those observed in Strobilomyia and Chiast-
ocheta (Després et al., 2002), with a Wrst step of adaptive
radiation (likely in sympatry), and a second step of geo-
graphic expansion and vicariance of each of the previously
formed ecological groups.
The genus Strobilomyia represents an extreme situation
of coexistence of congeneric species, given that up to six
species can exploit the same resource in sympatry. This
resource (the cones of conifers) is available only for a short
period of time and is discontinuous in space. These condi-
tions promote competition between individuals, which may
favor allochronic divergence (Després and Cherif, 2004;
Ferdy et al., 2002). Interestingly, in the other cases of allo-
chronic speciation that involve insects specialized on repro-
ductive organs (Després et al., 2002; Groman and Pellmyr,
2000; Harry et al., 1998; Ramadevan and Deakin, 1990;
Weiblen and Bush, 2002), temporal divergences are com-
bined with the same ecological divergences observed in
Strobilomyia (Michelsen, 1988; Roques et al., 1983, 2003),
i.e., an increase in ovipositor length in late species. As the
cone or the fruit grows and its structure changes, seeds
become more diYcult to reach and availability of oviposi-
tion sites is modiWed. In response, associated insect species
have longer ovipositors and show diVerent oviposition
modes (Després and Cherif, 2004; Pellmyr and Leebens-
Mack, 2000; Roques et al., 1983). Reproductive organs also
produce diVerent volatile compounds according to their
phenological stage, and insect specialization for oviposition
on these diVerent stages implies perception and identiWca-
tion of the emitted volatiles (Irwin and Dorsett, 2002).
In conclusion, our study demonstrates the importance of
ecological speciation in fast adaptive radiation. It involves
both host shifts and phenological divergence, but the inten-
sity of this last mechanism seems to be strongly dependent
on the host, as it is found only in the larch-exploiting spe-
cies. Further studies on the Strobilomyia–larch interaction
as compared to the Strobilomyia–Wr and the Strobilomyia–
spruce interactions could provide a better understanding of
the mechanism underlying allochronic speciation.
Acknowledgments
We thank N. Belova, K. Kamijo, M. Kenis, P. Khomen-
tovsky, Y.-Z. Pan, J.-P. Raimbault, J.-H. Sun, J. Sweeney,
and J.-J. Turgeon for providing specimens; L. Gielly, C.

Page 14

J.-M. Sachet et al. / Molecular Phylogenetics and Evolution 41 (2006) 606–621
619
Miquel, and D. Rioux for technical assistance; I. Chinta-
uan-Marquier and C. Conord for help with the analyses;
and A. Bonin, S. Boyer, S. Jordan, C. Kerdelhué, and M.
Robson for helpful comments on the manuscript. This
study was partly supported by the Institut Français de la
Biodiversité.
References
Abbot, P., Withgott, J.H., 2004. Phylogenetic and molecular evidence for
allochronic speciation in gall-forming aphids (Pemphigus). Evolution
58, 539–553.
Abrahamson, W.G., Blair, C.P., Eubanks, M.D., Morehead, S.A., 2003.
Sequential radiation of unrelated organisms: the gall Xy Eurosta soli-
daginis and the tumbling Xower beetle Mordellistena convicta. J. Evol.
Biol. 16, 781–789.
Anderton, L.K., Jenkins, M.J., 2001. Cone entomofauna of whitebark pine
and alpine larch (Pinaceae): potential impact of Leptoglossus occiden-
talis (Hemiptera: Coreidae) and a new record of Strobilomyia macalpi-
nei (Diptera: Anthomyiidae). Can. Entomol. 133, 399–406.
Baldwin, B.G., Sanderson, M.J., 1998. Age and rate of diversiWcation of the
Hawaiian silversword alliance (Compositae). Proc. Natl. Acad. Sci.
USA 95, 9402–9406.
Ballard, J.W., 2000. When one is not enough: introgression of mitochon-
drial DNA in Drosophila. Mol. Biol. Evol. 17, 1126–1130.
Barraclough, T.G., Vogler, A.P., 2000. Detecting the geographical pattern
of speciation from species-level phylogenies. Am. Nat. 155, 419–434.
Barraclough, T.G., Nee, S., 2001. Phylogenetics and speciation. Trends
Ecol. Evol. 16, 391–399.
Belova, N.V., Baranchikov, Y.N., Roques, A., 1998. New observation of
cone Xies attacking cones of Picea obovata and Larix sibirica in central
Siberia. J. For. Res. (Harbin) 9, 256–260.
Berlocher, S.H., Feder, J.L., 2002. Sympatric speciation in phytophagous
insects: moving beyond controversy? Annu. Rev. Entomol. 47, 773–815.
Bermingham, E., Martin, A.P., 1998. Comparative mtDNA phylogeography
of neotropical freshwater Wshes: testing shared history to infer the evolu-
tionary landscape of lower Central America. Mol. Ecol. 7, 499–517.
Brändle, M., Brandl, R., 2001. Species richness of insects and mites on
trees: expanding Southwood. J. Anim. Ecol. 70, 491–504.
BrockerhoV, E.G., Kenis, M., 1997. Oviposition, life cycle, and parasitoids
of the spruce cone maggot, Strobilomyia anthracina (Diptera: Antho-
myiidae), in the Alps. Bull. Entomol. Res. 87, 555–562.
Bromham, L., WoolWt, M., 2004. Explosive radiations and the reliability of
molecular clocks: island endemic radiations as a test case. Syst. Biol. 53,
758–766.
Brower, A.V.Z., 1994. Rapid morphological radiation and convergence among
races of the butterXy Heliconius erato inferred from patterns of mitochon-
drial DNA evolution. Proc. Natl. Acad. Sci. USA 91, 6491–6495.
Brower, A.V.Z., DeSalle, R., 1998. Patterns of mitochondrial versus
nuclear DNA sequence divergence among nymphalid butterXies: the
utility of wingless as a source of characters for phylogenetic inference.
Insect Mol. Biol. 7, 73–82.
Brubaker, L.B., Anderson, P.M., Edwards, M.E., Lozhkin, A.V., 2005.
Beringia as a glacial refugium for boreal trees and shrubs: new perspec-
tives from mapped pollen data. J. Biogeogr. 32, 833–848.
Bukovinszky, T., Potting, R.P.J., Clough, Y., van Lenteren, J.C., Vet,
L.E.M., 2005. The role of pre- and post-alighting detection mechanisms
in the responses to patch size by specialist herbivores. Oikos 109, 435–
446.
Bush, G.L., 1975. Modes of animal speciation. Annu. Rev. Ecol. Syst. 6,
339–364.
Bush, G.L., 1994. Sympatric speciation in animals: new wine in old bottles.
Trends Ecol. Evol. 9, 285–288.
Caterino, M.S., Cho, S., Sperling, F.A.H., 2000. The current state of insect
molecular systematics: a thriving Tower of Babel. Annu. Rev. Entomol.
45, 55–81.
Condon, M.A., Steck, G.J., 1997. Evolution of host use in fruit Xies of the
genus Blepharoneura (Diptera: Tephritidae): cryptic species on sexually
dimorphic host plants. Biol. J. Linn. Soc. 60, 443–466.
Cook, J.M., Rokas, A., Pagel, M., Stone, G.N., 2002. Evolutionary shifts
between host oak sections and host-plant organs in Andricus gallwasps.
Evolution 56, 1821–1830.
Cooley, J.R., Simon, C., Marshall, D.C., 2003. Temporal separation and
speciation in periodical cicadas. Bioscience 53, 151–157.
Coyne, J.A., Price, T.D., 2000. Little evidence for sympatric speciation in
island birds. Evolution 54, 2166–2171.
Cryan, J.R., Liebherr, J.K., Fetzner Jr., J.W., Whiting, M.F., 2001. Evalua-
tion of relationships within the endemic Hawaiian Platynini (Coleop-
tera: Carabidae) based on molecular and morphological evidence. Mol.
Phylogenet. Evol. 21, 72–85.
DeFilippis, V.R., Moore, W.S., 2000. Resolution of phylogenetic relation-
ships among recently evolved species as a function of amount of DNA
sequence: an empirical study based on woodpeckers (Aves: Picidae).
Mol. Phylogenet. Evol. 16, 143–160.
DeSalle, R., Freedman, T., Prager, E.M., Wilson, A.C., 1987. Tempo and
mode of sequence evolution in mitochondrial DNA of Hawaiian Dro-
sophila. J. Mol. Evol. 26, 157–164.
Després, L., Jaeger, N., 1999. Evolution of oviposition strategies and speci-
ation in the globeXower Xies Chiastocheta spp. (Anthomyiidae). J. Evol.
Biol. 12, 822–831.
Després, L., Pettex, E., Plaisance, V., Pompanon, F., 2002. Speciation in the
globeXower Xy Chiastocheta spp. (Diptera: Anthomyiidae) in relation
to host plant species, biogeography, and morphology. Mol. Phylogenet.
Evol. 22, 258–268.
Després, L., Cherif, M., 2004. The role of competition in adaptive radia-
tion: a Weld study on sequentially ovipositing host-speciWc seed preda-
tors. J. Anim. Ecol. 73, 109–116.
Dieckmann, U., Doebeli, M., 1999. On the origin of species by sympatric
speciation. Nature 400, 354–357.
Drès, M., Mallet, J., 2002. Host races in plant-feeding insects and their
importance in sympatric speciation. Philos. Trans. R. Soc. Lond. B
Biol. Sci. 357, 471–492.
Drummond, A.J., Nicholls, G.K., Rodrigo, A.G., Solomon, W., 2002. Esti-
mating mutation parameters, population history and genealogy simul-
taneously from temporally spaced sequence data. Genetics 161, 1307–
1320.
Drummond, A.J., Rambaut, A., 2003. BEAST v1.0, Available from:
<http://evolve.zoo.ox.ac.uk/beast/>. University of Oxford, Oxford.
Emelianov, I., Marec, F., Mallet, J., 2004. Genomic evidence for divergence
with gene Xow in host races of the larch budmoth. Proc. R. Soc. Lond.
B Biol. Sci. 271, 97–105.
Felsenstein, J., 1988. Phylogenies from molecular sequences: inference and
reliability. Annu. Rev. Genet. 22, 521–565.
Ferdy, J.-B., Després, L., Godelle, B., 2002. Evolution of mutualism
between globeXowers and their pollinating Xies. J. Theor. Biol. 217,
219–234.
Filchak, K.E., Roethele, J.B., Feder, J.L., 2000. Natural selection and sym-
patric divergence in the apple maggot Ragoletis pomonella. Nature 407,
739–742.
Fukatsu, T., Shibao, H., Nikoh, N., Aoki, S., 2001. Genetically distinct
populations in an Asian soldier-producing aphid, Pseudoregma bam-
bucicola (Homoptera: Aphididae), identiWed by DNA Wngerprinting
and molecular phylogenetic analysis. Mol. Phylogenet. Evol. 18, 423–
433.
Funk, D.J., Omland, K.E., 2003. Species-level paraphyly and polyphyly:
frequency, causes, and consequences, with insights from animal mito-
chondrial DNA. Annu. Rev. Ecol. Evol. Syst. 34, 397–423.
Gavrilets, S., 2003. Models of speciation: what have we learned in 40
years? Evolution 57, 2197–2215.
Gavrilets, S., Vose, A., 2005. Dynamic patterns of adaptive radiation. Proc.
Natl. Acad. Sci. USA 102, 18040–18045.
Gibbons, J.R.H., 1979. A model for sympatric speciation in Megarhyssa
(Hymenoptera: Ichneumonidae): competitive speciation. Am. Nat. 114,
719–741.

Page 15

620
J.-M. Sachet et al. / Molecular Phylogenetics and Evolution 41 (2006) 606–621
Gillespie, R.G., 2002. Biogeography of spiders on remote oceanic islands
of the PaciWc: archipelagoes as stepping stones? J. Biogeogr. 29, 655–
662.
Gillespie, R.G., Roderick, G.K., 2002. Arthropods on islands: colonization,
speciation and conservation. Annu. Rev. Entomol. 47.
Groman, J.D., Pellmyr, O., 2000. Rapid evolution and specialization fol-
lowing host colonization in a yucca moth. J. Evol. Biol. 13, 223–236.
Guindon, S., Gascuel, O., 2003. A simple, fast, and accurate algorithm to
estimate large phylogenies by maximum likelihood. Syst. Biol. 52, 696–
704.
Harry, M., Solignac, M., Lachaise, D., 1998. Molecular evidence for paral-
lel evolution of adaptive syndromes in Wg-breeding Lissocephala (Dros-
ophilidae). Mol. Phylogenet. Evol. 9, 542–551.
Ho, S.Y.W., Phillips, M.J., Drummond, A.J., Cooper, A., 2005. Accuracy of
rate estimation using relaxed-clock models with a critical focus on the
early metazoan radiation. Mol. Biol. Evol. 22, 1355–1363.
Hosken, D.J., Stockley, P., 2004. Sexual selection and genital evolution.
Trends Ecol. Evol. 19, 87–93.
Huelsenbeck, J.P., Ronquist, F., 2001. MRBAYES: Bayesian inference of
phylogenetic trees. Bioinformatics 17, 754–755.
Hughes, J., Vogler, A.P., 2004. The phylogeny of acorn weevils (genus Cur-
culio) from mitochondrial and nuclear DNA sequences: the problem of
incomplete data. Mol. Phylogenet. Evol. 32, 601–615.
Irwin, R.E., Dorsett, B., 2002. Volatile production by buds and corollas of
two sympatric, confamilial plants, Ipomopsis aggregata and Polemo-
nium foliosissimum. J. Chem. Ecol. 28, 565–578.
JeVries, M.J., Lawton, J.H., 1984. Enemy-free space and the structure of
ecological communities. Biol. J. Linn. Soc. 23, 269–286.
Jiggins, C.D., Bridle, J.R., 2004. Speciation in the apple maggot Xy: a blend
of vintages? Trends Ecol. Evol. 19, 111–114.
Jocqué, R., 2002. Genitalic polymorphism—a challenge for taxonomy. J.
Arachnol. 30, 298–306.
Jordan, I.K., McDonald, J.F., 1998. Evolution of the copia retrotransposon
in the Drosophila melanogaster species subgroup. Mol. Biol. Evol. 15,
1160–1171.
Jordan, S., Simon, C., Polhemus, D., 2003. Molecular systematics and
adaptive radiation of Hawaii’s endemic damselXy genus Megalagrion
(Odonata: Coenagrionidae). Syst. Biol. 52, 89–109.
Kambysellis, M.P., Ho, K.F., Craddock, E.M., Piano, F., Parisi, M., Cohen,
J., 1995. Pattern of ecological shifts in the diversiWcation of Hawaiian
Drosophila inferred from a molecular phylogeny. Curr. Biol. 5, 1129–
1139.
Kelly, C.K., Southwood, T.R.E., 1999. Species richness and resource avail-
ability: a phylogenetic analysis of insects associates with trees. Proc.
Natl. Acad. Sci. USA 96, 8013–8016.
Kim, C.G., Zhou, H.Z., Imura, Y., Tominaga, O., Su, Z.H., Osawa, S., 2000.
Pattern of morphological diversiWcation in the Leptocarabus ground
beetles (Coleoptera: Carabidae) as deduced from mitochondrial ND5
gene and nuclear 28S rDNA sequences. Mol. Biol. Evol. 17, 137–145.
Kirkpatrick, M., Ravigné, V., 2002. Speciation by natural and sexual selec-
tion: models and experiments. Am. Nat. 159, S22–S35.
Knowles, L.L., 2000. Tests of Pleistocene speciation in montane grasshop-
pers (genus Melanoplus) from the sky islands of western North Amer-
ica. Evolution 54, 1337–1348.
Kondrashov, A.S., Kondrashov, F.A., 1999. Interactions among quantita-
tive traits in the course of sympatric speciation. Nature 400, 351–354.
KornWeld, I., Smith, P.F., 2000. African cichlid Wshes: model systems for
evolutionary biology. Annu. Rev. Ecol. Syst. 31.
Kumar, S., Tamura, K., Jakobsen, I.B., Nei, M., 2001. MEGA2: molecular
evolutionary genetics analysis software. Bioinformatics 17, 1244–1245.
Losos, J.B., Jackman, T.R., Larson, A., de Queiroz, K., Rodriguez-Schet-
tino, L., 1998. Contingency and determinism in replicated adaptive
radiations of island lizards. Science 279, 2115–2118.
Losos, J.B., Glor, R.E., 2003. Phylogenetic comparative methods and the
geography of speciation. Trends Ecol. Evol. 18, 220–227.
Lovette, I.J., Bermingham, E., 1999. Explosive speciation in the New
World Dendroica warblers. Proc. R. Soc. Lond. B Biol. Sci. 266, 1629–
1636.
Maddison, W.P., Maddison, D.R., 2005. Mesquite: a modular system for
evolutionary analysis. Version 1.06 <http://mesquiteproject.org>.
Mallarino, R., Bermingham, E., Willmott, K.R., Whinnett, A., Jiggins,
C.D., 2005. Molecular systematics of the butterXy genus Ithomia (Lepi-
doptera: Ithomiinae): a composite phylogenetic hypothesis based on
seven genes. Mol. Phylogenet. Evol. 34, 625–644.
Marques, E.S.A., Price, P.W., Cobb, N.S., 2000. Resource abundance and
insect herbivore diversity on woody fabaceous desert plants. Environ.
Entomol. 29, 696–703.
Marshall, D.C., Cooley, J.R., 2000. Reproductive character displacement
and speciation in periodical cicadas, with description of a new species,
13-year Magicicada neotredecim. Evolution 54, 1313–1325.
Mayr, E., 1963. Animal Species and Evolution. Belknap Press, Cambridge,
MA.
McClure, M., Quiring, D.T., Turgeon, J.J., 1996. Oviposition, temporal dis-
tribution, and potential impact of Strobilomyia laricis Michelsen and
S. viaria (Huckett) (Diptera: Anthomyiidae) on eastern larch, Larix
laricina (Du Roi) K. Koch. Can. Entomol. 128, 67–78.
Megens, H.J., van Moorsel, C.H., Piel, W.H., Pierce, N.E., de Jong, R.,
2004. Tempo of speciation in a butterXy genus from the Southeast
Asian tropics, inferred from mitochondrial and nuclear DNA sequence
data. Mol. Phylogenet. Evol. 31, 1181–1196.
Michelsen, V., 1988. A world revision of Strobilomyia gen.n.: the antho-
myiid seed pests of conifers (Diptera: Anthomyiidae). Syst. Entomol.
13, 271–314.
Murphy, S.M., 2004. Enemy-free space maintains swallowtail butterXy
host shift. Proc. Natl. Acad. Sci. USA 101, 18048–18052.
Orr, M.R., Smith, T.B., 1998. Ecology and speciation. Trends Ecol. Evol.
13, 502–506.
Päivinen, J., Ahlroth, P., Kaitala, V., Kotiaho, J.S., Suhonen, J., Virola, T.,
2003. Species richness and regional distribution of myrmecophilous
beetles. Oecologia 134, 587–595.
Pellmyr, O., 1989. The cost of mutualism: interactions between Trollius
europaeus and its pollinating parasites. Oecologia 78, 53–59.
Pellmyr, O., Leebens-Mack, J., 2000. Reversal of mutualism as a mecha-
nism for adaptive radiation in yucca moths. Am. Nat. 156, S62–S76.
Pigliucci, M., 2003. Species as family resemblance concepts: the (dis-)solu-
tion of the species problem? BioEssays 25, 596–602.
Posada, D., Crandall, K.A., 1998. MODELTEST: testing the model of
DNA substitution. Bioinformatics 14, 817–818.
Ramadevan, S., Deakin, M.A.B., 1990. The Gibbons speciation mecha-
nism. J. Theor. Biol. 145, 447–456.
Rambaut, A., Drummond, A.J., 2004. Tracer v1.3, Available from: <http://
evolve.zoo.ox.ac.uk/software.html>. University of Oxford, Oxford.
Rees, D.J., Emerson, B.C., Oromi, P., Hewitt, G.M., 2001a. Mitochondrial
DNA, ecology and morphology: interpreting the phylogeography of
the Nesotes (Coleoptera: Tenebrionidae) of Gran Canaria (Canary
Islands). Mol. Ecol. 10, 427–434.
Rees, D.J., Emerson, B.C., Oromi, P., Hewitt, G.M., 2001b. The diversiWca-
tion of the genus Nesotes (Coleoptera: Tenebrionidae) in the Canary
Islands: evidence from mtDNA. Mol. Phylogenet. Evol. 21, 321–326.
Riddle, B.R., Hafner, D.J., Alexander, L.F., 2000. Phylogeography and sys-
tematics of the Peromyscus eremicus species group and the historical
biogeography of North American warm regional deserts. Mol. Phylo-
genet. Evol. 17, 145–160.
Robinson-Rechavi, M., Huchon, D., 2000. RRTree: relative-rate tests
between groups of sequences on a phylogenetic tree. Bioinformatics 16,
296–297.
Roderick, G.K., Gillespie, R.G., 1998. Speciation and phylogeography of
Hawaiian terrestrial arthropods. Mol. Ecol. 7, 519–531.
Roques, A., Martinez, M., Delplanque, A., 1983. Les Diptères Anthomyii-
dae du genre Lasiomma Stein. ravageurs des cônes et graines de Mélèze
d’Europe (Larix decidua Mill.) en France. I. Taxonomie et répartition.
Z. Angew. Entomol. 95, 429–439.
Roques, A., Raimbault, J.-P., Delplanque, A., 1984. Les Diptères Antho-
myiidae du genre Lasiomma Stein. ravageurs des cônes et graines de
Mélèze d’Europe (Larix decidua Mill.) en France. II. Cycles biologi-
ques et dégâts. Z. Angew. Entomol. 98, 350–367.