Group Title: BMC Evolutionary Biology
Title: Timing major conflict between mitochondrial and nuclear genes in species relationships of Polygonia butterflies (Nymphalidae: Nymphalini)
Full Citation
Permanent Link:
 Material Information
Title: Timing major conflict between mitochondrial and nuclear genes in species relationships of Polygonia butterflies (Nymphalidae: Nymphalini)
Physical Description: Book
Language: English
Creator: Wahlberg, Niklas
Weingartner, Elisabet
Warren, Andrew
Publisher: BMC Evolutionary Biology
Publication Date: 2009
Abstract: BACKGROUND:Major conflict between mitochondrial and nuclear genes in estimating species relationships is an increasingly common finding in animals. Usually this is attributed to incomplete lineage sorting, but recently the possibility has been raised that hybridization is important in generating such phylogenetic patterns. Just how widespread ancient and/or recent hybridization is in animals and how it affects estimates of species relationships is still not well-known.RESULTS:We investigate the species relationships and their evolutionary history over time in the genus Polygonia using DNA sequences from two mitochondrial gene regions (COI and ND1, total 1931 bp) and four nuclear gene regions (EF-1a, wingless, GAPDH and RpS5, total 2948 bp). We found clear, strongly supported conflict between mitochondrial and nuclear DNA sequences in estimating species relationships in the genus Polygonia. Nodes at which there was no conflict tended to have diverged at the same time when analyzed separately, while nodes at which conflict was present diverged at different times. We find that two species create most of the conflict, and attribute the conflict found in Polygonia satyrus to ancient hybridization and conflict found in Polygonia oreas to recent or ongoing hybridization. In both examples, the nuclear gene regions tended to give the phylogenetic relationships of the species supported by morphology and biology.CONCLUSION:Studies inferring species-level relationships using molecular data should never be based on a single locus. Here we show that the phylogenetic hypothesis generated using mitochondrial DNA gives a very different interpretation of the evolutionary history of Polygonia species compared to that generated from nuclear DNA. We show that possible cases of hybridization in Polygonia are not limited to sister species, but may be inferred further back in time. Furthermore, we provide more evidence that Haldane's effect might not be as strong a process in preventing hybridization in butterflies as has been previously thought.
General Note: Start page 92
General Note: M3: 10.1186/1471-2148-9-92
 Record Information
Bibliographic ID: UF00099945
Volume ID: VID00001
Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: Open Access:
Resource Identifier: issn - 1471-2148


This item has the following downloads:

PDF ( 1 MBs ) ( PDF )

Full Text

BMC Evolutionary Biology BioMed

Research article

Timing major conflict between mitochondrial and nuclear genes in
species relationships of Polygonia butterflies (Nymphalidae:
Niklas Wahlberg* 1, Elisabet Weingartner2, Andrew D Warren3,4 and
Soren Nylin2

Address: 'Laboratory of Genetics, Department of Biology, University of Turku, 20014 Turku, Finland, 2Department of Zoology, Stockholm
University, 106 91 Stockholm, Sweden, 3McGuire Center for Lepidoptera and Biodiversity, Florida Museum of Natural History, University of
Florida, SW 34th Street and Hull Road, PO Box 112710, Gainesville, FL 32611-2710, USA and 4Museo de Zoologia, Departamento de Biologia
Evolutiva, Facultad de Ciencias, Universidad Nacional Aut6noma de Mexico, Apdo. Postal 70-399, Mexico, DF 04510 Mexico
Email: Niklas Wahlberg*; Elisabet Weingartner;
Andrew D Warren; S6ren Nylin
* Corresponding author

Published: 7 May 2009
BMC Evolutionary Biology 2009, 9:92 doi: 10.1 186/1471-2148-9-92

Received: 20 August 2008
Accepted: 7 May 2009

This article is available from:
2009 Wahlberg et al; licensee BioMed Central Ltd.
This is an Open Access article distributed under the terms of the Creative Commons Attribution License (,
which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Background: Major conflict between mitochondrial and nuclear genes in estimating species
relationships is an increasingly common finding in animals. Usually this is attributed to incomplete
lineage sorting, but recently the possibility has been raised that hybridization is important in
generating such phylogenetic patterns. Just how widespread ancient and/or recent hybridization is
in animals and how it affects estimates of species relationships is still not well-known.
Results: We investigate the species relationships and their evolutionary history over time in the
genus Polygonia using DNA sequences from two mitochondrial gene regions (COI and ND I, total
1931 bp) and four nuclear gene regions (EF- I ~, wingless, GAPDH and RpS5, total 2948 bp). We
found clear, strongly supported conflict between mitochondrial and nuclear DNA sequences in
estimating species relationships in the genus Polygonia. Nodes at which there was no conflict tended
to have diverged at the same time when analyzed separately, while nodes at which conflict was
present diverged at different times. We find that two species create most of the conflict, and
attribute the conflict found in Polygonia satyrus to ancient hybridization and conflict found in
Polygonia oreas to recent or ongoing hybridization. In both examples, the nuclear gene regions
tended to give the phylogenetic relationships of the species supported by morphology and biology.
Conclusion: Studies inferring species-level relationships using molecular data should never be
based on a single locus. Here we show that the phylogenetic hypothesis generated using
mitochondrial DNA gives a very different interpretation of the evolutionary history of Polygonia
species compared to that generated from nuclear DNA. We show that possible cases of
hybridization in Polygonia are not limited to sister species, but may be inferred further back in time.
Furthermore, we provide more evidence that Haldane's effect might not be as strong a process in
preventing hybridization in butterflies as has been previously thought.

Page 1 of 16
(page number not for citation purposes)



Phylogenetics at the species-level is becoming increasingly
important in the study of processes underlying speciation
[1,2]. Most species-level phylogenies have until recently
been based on only mitochondrial DNA (mtDNA) due to
the ease of PCR amplification and its perceived suitability,
e.g. due to maternal inheritance (shorter time for coales-
cence than nuclear DNA (nDNA) because of smaller NJ,
lack of recombination and relatively high mutation rate.
However, species phylogenies are not necessarily the same
as gene phylogenies [3,4], as different genes might have
different histories. Genes involved with speciation, affect-
ing such traits as hybrid incompatibility, as well as sex
chromosomes should be more differentiated between
species and less likely to introgress than autosomes [5-8].
Different processes such as random sorting of homoplasy,
ancient polymorphism and hybridization with gene intro-
gression can obscure patterns of species relationships.
Information from different regions of genomes such as
mitochondrial DNA, nuclear DNA (from sex chromo-
somes as well as from autosomes) and microsatellites are
thus necessary in investigating the evolutionary history of
a group of closely related species.

As species-level molecular phylogenies based on both
mitochondrial and nuclear markers have become more
common, it has become clear that there is often well-sup-
ported conflict between the genomes for certain clades in
given phylogenies [9]. Recent work is pointing to major
conflict between mtDNA and nuclear DNA in species-
level phylogenetic analyses [6,9-15]. The conflict is often
attributed to ancient or recent hybridization [9,12-14] or
incomplete lineage sorting [15]. Hybridization, a well-
accepted process in plants, appears to be more common
also among closely related animals than previously
thought [16]. Kronforst [17] showed in Heliconius butter-
flies that gene flow between species can proceed for long
periods of time after divergence.

Although phylogenies give us a hypothesis of species rela-
tionships, they tell us little about the processes involved in
diversification on their own. More information is needed
to discover reasons for diversification, such as geographic
location of specimens used and times of divergence of lin-
eages. Contemporary sympatric species might have been
allopatric when the two lineages diverged and without
well-sampled species it is even harder to draw any conclu-
sions about movements of species and populations.
Knowledge about when divergences of lineages have hap-
pened in a given group of species may give insight to the
processes behind the conflicts in phylogenetic signal.
However, the temporal framework has rarely been studied
for such conflicts.

Here we study the relationships of species in the genus
Polygonia (Lepidoptera: Nymphalidae), which have been

used as model taxa in numerous studies on the evolution
of insect-host plant interactions [18-21], phenotypic plas-
ticity in life-history traits [22,23], effects of environment
on distribution [24] and insect physiology [25,26]. Polygo-
nia is a genus thought to include five Palaearctic species
(P. c-album, P. c-aureum, P. egea, P. gigantea and P. inter-
posita), and nine Nearctic species, seven in the United
States and Canada (P. comma, P. faunus, P. gracilis, P. inter-
rogationis, P. oreas, P. progne and P. satyrus [27,28]) and
two endemic to Mexico (P. g-argenteum and P. haroldii).
The taxonomic status of some of these species is disputed.
Polygonia interposita has been treated as a subspecies of P.
c-album [29] but was suggested to be a species-level taxon
by Churkin [30]. So far, this taxon has not been included
in any earlier molecular studies. We have tentatively
treated P. zephyrus as a species separate from P. gracilis [fol-
lowing [31]] although this status is unclear; P. zephyrus is
often considered conspecific with P. gracilis [32]. These
two taxa (P. zephyrus and P. gracilis) are morphologically
distinguishable at the extremes of their ranges but
between those "ends of a dine" a broad zone exists where
intermediate forms occur (ADW pers. obs.). This may be
an example of incipient speciation or secondary contact
between two species. Earlier, P. oreas was sometimes con-
sidered a subspecies of P. progne, but in a recent study
[33], P. oreas was found to be closely related to P. gracilis.

According to previous analyses, the ancestor of Polygonia
was distributed in the Palaearctic and there have been two
colonization events into the Nearctic region [33]. The
ancestral host plants were most likely "urticalean rosids"
(which includes the closely related plant families Urti-
caceae, Ulmaceae, Cannabaceae and Celtidaceae) [21,34].
Many Polygonia species are still restricted to plants from
this group but some species have included additional or
shifted completely to other plant families, such as Betu-
laceae, Ericaceae, Grossulariaceae and Salicaceae. In previ-
ous studies [21,34] phylogenetic trees have been used to
infer ancestral host plant ranges used by butterflies in the
subfamily Nymphalinae. The results imply that when host
plant range has expanded, an increase in the rate of net
diversification has followed. In order to understand in
more detail the dynamics of host plant use and diversifi-
cation in the Polygonia butterflies in particular, and insects
in general, it is necessary to generate a hypothesis of the
evolutionary history of the group.

Species of Polygonia have been included in several earlier
phylogenetic studies [33,35,36], but the relationships of
some species have not been stable and conflict between
datasets has been noted [36]. In this study, we present a
hypothesis of the evolution of this genus in which all cur-
rently accepted species and most subspecies have been
included. We then apply a temporal framework in order
to illuminate the causes of major conflicts between
genomic datasets.

Page 2 of 16
(page number not for citation purposes)

BMC Evolutionary Biology 2009, 9:92

Combined analysis with Partitioned Bremer Support
(PBS) of the 25-taxon dataset showed that there was
strong conflict between the mitochondrial partition and
the nuclear partition at almost all nodes within the genus
Polygonia (Figure 1). Analysis of the two genomic datasets
separately showed that the topology was different at these
conflicting nodes (Figure 2). Despite the conflict, there are
several well-supported clades at which there is no conflict
between datasets. The genus Polygonia without the species
Kaniska canace is strongly supported by all datasets. The
position ofK. canace as sister to the genus Nymphalis is not
well-supported in the cladistic analyses, but is supported
by all Bayesian analyses of combined and separate data.
The sister species relationships of P. egea and P. undina, as
well as P. comma and P. g-argenteum are well supported.
The clade containing P. c-album, P. interposita and P. fau-
nus, as well as the clade containing all the rest of the
Nearctic species (except P. faunus), are also well-sup-

The level of conflict between the different genomic data-
sets is particularly evident when comparing the separate
analyses, where species relationships are quite different
with relatively good support (Figure 2). These topologies
were not dependent on method used for analysis, and
thus the phylogenetic signal found within the mitochon-
drial and nuclear datasets appears to be strong. The excep-
tion is P. gigantea, which receives strong support as the
sister to P. undina+P. egea with the nuclear dataset, but lit-

- Aglais io
155 Aglais urticae
13-5 Aglais milberti
o Kaniska canace
I 2 Nymphalis I-album
S 6] -3 7 Nymphalis polychloros
77 1-4 Nymphalis antiopa
10 12 4 Nymphalis californica
36 1 Nymphalis xanthomelas
Polygonia c-aureum
.6 i- Polygonia gigantea
4- 4 _ Polygonia egea
S Polygonia undina
,- Plygonia faunus
~ Polygonia c-album
I -Polygonia interposita
Polygonia interrogations
12 1 Polygonia comma
5 -- i.j.,,; argenteum
-r -. I-- P->.,:,',;,,"' atyrus T T "
-, I.-., _- -.lygonia haroldii ,
|*I -- olygonia progne ,
[ lygonia oreas
1 Polygonia gracilis
a Polygonia zephyrus

Figure I
Combined analysis of all genes. Values above branches
are the Partitioned Bremer Support (PBS) values for the
combined mitochondrial gene partition and values below
branches are the PBS values for the combined nuclear gene
partition. Grey circles highlight nodes with strong conflict
within the Polygonia clade. Pictured butterflies are from top
to bottom Kaniska canace, Nymphalis polychloros, Polygonia c-
album, Polygonia satyrus and Polygonia zephyrus.

tie or no support as sister to P. c-aureum with the mito-
chondrial dataset (Figure 2). Interestingly, the estimated
times of divergence for clades which are common to the
two datasets are similar regardless of which dataset one
uses (with the caveat that the confidence intervals are very
wide). Thus the Polygonia clade is estimated to have started
diverging 18-19 million years ago (mya), the P. c-album to
Nearctic Polygonia clade between 13 and 16 mya, the
Nearctic Polygonia at 11-12 mya and the P. progne to P.
gracilis clade between 5 and 6 mya (Figure 2).

Of particular interest in the separate analysis of the mito-
chondrial and nuclear datasets is the position of P. inter-
posita as sister to P. c-album (indeed with almost identical
COI haplotypes) based on mtDNA, but as sister to P. c-
album+P. faunus based on nDNA. Also the position of P.
satyrus as very closely related to P. gracilis+P. oreas+P.
zephyrus based on mtDNA, but as sister to P. interroga-
tionis+P. g-argenteum+P. comma based on nDNA. Finally,
the position of P. oreas as part of the P. gracilis+P. zephyrus
clade based on mtDNA, but as sister to P. progne based on
nDNA (Figure 2).

Increasing the sample size for each gene region and ana-
lyzing them separately brings some light to these patterns.
The COI tends to have very little variation within species,
but substantial variation between species (Figure 3). The
exceptions are P. interposita, which is almost identical to P.
c-album; P. g-argenteum, which is very similar to P. comma;
and P. gracilis, P. zephyrus and P. oreas, which are all very
similar to each other even to the point of sharing haplo-
types between the three taxa. The position of P. satyrus is
consistent with the 25-taxon dataset, and shows some var-
iation within the species.

The nuclear gene regions show quite different topologies
when analyzed on their own, but several patterns are con-
sistent between them (Figure 4, Figure 5, Figure 6, Figure
7 and Figure 8). First of all, the haplotypes of GAPDH and
wgl are very similar in the taxa P. gracilis, P. zephyrus, P.
haroldii, P. oreas and P. progne (Figure 5, Figure 7 and Fig-
ure 8). The haplotypes of EF- 1 a, RpS5 and wgl in P. satyrus
are more related to P. comma, P. g-argenteum and P. inter-
rogationis than to the other Nearctic Polygonia (Figure 4,
Figure 6 and Figure 7). The nDNA haplotypes found in P.
interposita tend to not be especially close to P. c-album (Fig-
ure 4, Figure 5, Figure 6 and Figure 7). Interestingly, the
haplotypes of RpS5 found in P. oreas are very closely
related to those found in P. progne (Figure 6 and Figure
8d), while all other nDNA haplotypes are ambiguous
about this relationship. Finally, the subspecies P. e. undina
is differentiated from P. egea for all sequenced genes.

Regarding the haplotype networks, we have focused on
the unresolved clade of P. zephyrus, P. gracilis and P.

Page 3 of 16
(page number not for citation purposes)

BMC Evolutionary Biology 2009, 9:92

a) Mitochondrial genes Late Miocene Early Miocene Pleistocene
Oligocene Middle Miocene Pliocene

Aglais to
Aglais urticae
Aglais milberti
0.4 ---Kaniska canace
o. 40 Nymphalis I-album
Nymphalis californica
1 .76 Nymphalis xanthomelas
0 73 Nymphalis antiopa
0.71 Nymphalis polychloros
Polygonia gigantea
Polygonia c-aureum
Polygonia undina
S0.96 Polygonia egea
Polygonma faunus
77 Polygonia interposita
Polygonia c-album
o 99 Polygonia interrogationis
Ir-- Polygonta comma
1 1 Polygonia g-argenteum
............ .f Polygonia progne
Polygonta haroldii
1 [ Polygonia satyrus
S I Polygonia gracilis
Polygonia oreas
0.8 Polygonia zephyrus
b) nuclear genes

Agla's to
Si Aglais urticae
Aglais milberti
0.56 Kamska canace
Nymphalis I-album
S-I Ny mphaiis polychloros
--. 1 Nymphatis californica
______ Nymphalis antiopa
S096 Nymphalis xanthomelas
Polygonma c-aureum
Polygonia gigantea
___ Poly gonia egea
1 1 Polygonia undina
Polygonia interposita
1 Polygonia faunus
0.99, Polygonia c-album
Polygon'a satyrus
0.98 1 Polygonia interrogationis

S- Polygonia comma
0.99 1 Polygon'a progne
1 7 Polygonia oreas
Polygonia zephyrus
SPolygonia haroldii
0.99- Polygonia gracifis
32 30 28 26 24 22 20 18 16 14 12 10 8 6 4 2 0
Millions of years

Figure 2
Dated phylogenies of Polygonia. a) based on mtDNA b) based on nDNA. Values below the branches are posterior proba-
bilities for the nodes to the right of the numbers.

Page 4 of 16
(page number not for citation purposes)

BMC Evolutionary Biology 2009, 9:92

oreas (in COI) as well as the relationships between these
species and P. progne and P. haroldii (in RpS5, EF-la,
GAPDH and wgl) (Figure 8). None of the haplotype net-
works showed a star-like pattern, i.e., a "central" com-
monly shared haplotype from which other haplotypes
deviate by only a few mutational steps, indicative of a
rapid and recent diversification [37]. Most haplotypes
were only represented by one individual. A few haplo-
types were however shared, even between different spe-
cies. For instance, Polygonia oreas shared the same
haplotype with P. gracilis and P. zephyrus in the COI data-
set (Figure 8a). In the EF-la dataset one P. oreas haplo-
type is shared with P. gracilis (Figure 8b). Shared
haplotypes were found for P. haroldii, P. progne and P.
zephyrus in the GAPDH dataset as well as for P. progne, P.
zephyrus and P. gracilis (Figure 8c). In the RpS5 dataset P.
haroldii, P. zephyrus and P. gracilis shared the same hap-
lotype (Figure 8d). Three haplotypes were shared between
P. gracilis and P. zephyrus, one haplotype was shared
between P. progne and P. gracilis and one haplotype was
shared between P. oreas and P. zephyrus in the wgl dataset
(Figure 8e). In addition, in the wgl dataset one haplotype
is shared between P. comma, P. g-argenteum and all P.
satyrus. However, the P. g-argenteum individual had
many positions of missing data.

Major clades in the phylogeny
Unlinked genes are expected to have independent genea-
logical histories [3,38], and thus combining data may not
always be informative of species relationships [39]. In this
study we have found that gene regions from different
genomes (mtDNA and nDNA) give rather different esti-
mates of species relationships. The phylogenetic positions
of four taxa in particular need explanation: P. satyrus, P.
oreas, P. haroldii and P. interposita. Each of these is strongly
supported in different positions depending on which
dataset is analysed. Before we discuss these four anoma-
lous taxa, we will discuss the general findings for the other
species, as this is the most complete study of Polygonia
phylogeny to date, including several taxa that have never
been part of a phylogenetic systematic investigation.

Previous studies have shown conflicting results on the
position of the taxon "canace" [33,35,36], often placed in
the monotypic genus Kaniska, but suggested to be
included in Polygonia by Wahlberg & Nylin [36]. The
present study does not corroborate that finding, but it
should be noted that in contrast to the earlier study we did
not include morphological data here. However, it is clear
that the position of "canace" is not stable and it's sister
relationship either to Nymphalis or to Polygonia is weakly
supported. In such a case, we feel it is best to retain it in
the monotypic genus Kaniska, in order to highlight its

"oddity" and long history of independent evolution. This
is of course only valid if one accepts the validity of the
genera Polygonia and Nymphalis, which some consider to
be a single genus Nymphalis, along with our Aglais [e.g.,
[40,41]]. For reasons explained in Wahlberg & Nylin [36],
we feel that the genus Polygonia should be retained, and
thus we suggest that the taxon "canace" be retained in the
genus Kaniska, as is frequently done in the literature [e.g.,
[29], e.g., [42,43]].

There are several independent lineages within Polygonia.
The type species of the genus, Polygonia c-aureum, is the sis-
ter to the rest of Polygonia, as has been found in previous
studies [33,35,36]. Polygonia gigantea, included here for
the first time in a phylogenetic study, is an independent
lineage that is most likely sister to the P. egea+P. undina
clade, based on the well-supported result of the nDNA
dataset and the ambiguous result of the mtDNA dataset.
Polygonia undina has mainly been considered to be a sub-
species of P. egea, but our results show that it is genetically
very distinct and the common ancestor of the two
diverged as early as between 8-13 mya (Figure 2). This
makes the pair older than several other species pairs in
Polygonia, and we found no evidence of interbreeding (all
genes were clearly diverged for this pair of species). We
thus elevate P. undina to the species level Statt. nov.).

The clade containing P. c-album, P. interposita and P. faunus
is well-supported and quite clearly the sister to the Nearc-
tic clade. The interrelationships of these three species will
be discussed in more detail below. The Nearctic clade
including P. satyrus, P. interrogationis, P. comma, P. g-argen-
teum, P. progne, P. oreas, P. haroldii, P. gracilis and P. zephy-
rus, is also well-supported. Within this clade, P. g-
argenteum (here included for the first time in a phyloge-
netic analysis) is clearly the sister species to P. comma, and
apparently these two have diverged relatively recently (2-
4 mya). The position of P. interrogationis with regard to
these two species is different with the two genomic data-
sets. Mitochondrial DNA suggests that it is sister to the rest
of the Nearctic species, while nDNA suggests that it is sis-
ter to P. comma+P. g-argenteum. The latter sister relation-
ship is in fact suggested by morphological data as well
[35], giving more weight to this hypothesis of phylogeny.

Our data suggest that P. comma and P. g-argenteum have
diverged in the past 2-3 mya, during which time there has
been considerable morphological diversification between
them. Adults of g-argenteum are among the largest of Poly-
gonia (generally the same size as P. interrogationis), and
they lack the seasonal polyphenism (expression of dark
"summer" forms) seen in P. comma and P. interrogationis.
As a result, size excluded, adults of P. satyrus and P. g-
argenteum share a very similar superficial resemblance

Page 5 of 16
(page number not for citation purposes)

BMC Evolutionary Biology 2009, 9:92

BMC Evolutionary Biology 2009, 9:92

A iAglais iicae

Aglas utice 1.009 Kaniska canace EW19-11
0- Kaniska canace EW19-10
1.00 0.82 Kaniska canace NW164-1
11.00- Nymphalis I-album
S 097 Nymphalis californica
I 0 Nymphalis xanthomelas
Nymphalis antiopa
11.00 c uNymphalis polychloros
1.00 c-aureum EW13-10
c-aureum NW65-8
c-aureum EW13-12
0.1 1.00- gigantea NW166-5 1.00 egea NW 120-7
1.00 egea NW120-7
egea NW63-1
0.52 undina EW30-1 1.00 egea NW77-15
0.61 .1.00 n o- undina EW33-24
o[ undinaEW43-2
0.80 undinaEW43-1
1.00 r c-album NW70-3
.... / c-album EW13-1
0.83 interposita NW166-6
1.00 0.93t c-album EW26-32
c-album EW13-3
1 .00 faunus EW12-7
Sn Wfaunus EW19-1
pfaunus EW21-12
0.526 faunus NW74-12
1.00 faunus EW21-13
1.0 interrogations EW22-12
I ----'- .: 1 1"' JW 77-12
- l__-14
p comma EW21-10
1.00 g-argenteum NW165-2
Scomma NW65-6
1.00 comma EW21-11
0.57r- progne NW63-10
7 progne EW22-17
progne EW21-2
1.00 prone EW22-21
S." praogne EW22-16
progne EW21-4
progne EW21-3
0.63 1 progne EW21-16
progne EW22-20
progne EW22-15
r progne EW22-19
0.70 progne EW22-18

L haroldii NW165-1
satyrus EW11-11
satyrus EW16-7
satyrus EW23-13
satyrus EW11 -12
0.C 1 satyrus EW23-14
100 satyrus EW16-6
-satyrus EW15-11
93 -I-satyrus EW15-16
0.93 0 72_ 72 satyrus EW15-12
Ssatyrus EW15-13
satyrus NW65-9
Ssatyrus EW15-9
satyrus EW15-14
0.88 satyrus EW15-10
satyrus EW11-10
._Lsatyrus NW74-8
0.98 1.00- satyrus NW74-9

gracilis/zephyrus/oreas clade
0. 91

Figure 3
Topology of haplotypes from Bayesian analysis based on the mtDNA COI. Details of gracilis/zephyrus/oreas clade are
shown in the haplotype network in Figure 8a. Values below the branches are posterior probabilities for the nodes to the right
of the numbers.

(especially in the dorsal view), while adults of P. comma
(especially dark forms) and P. g-argenteum appear quite
different at first glance.

As an aside, it is interesting to note that apparently similar
patterns of differences between mitochondrial and
nuclear DNA are found in the genus Nymphalis (Figure 2).
This warrants a separate study to see if similar forces have
acted on the sister group of Polygonia.

Ancient mitochondrial introgression in P. satyrus
Mitochondrial DNA suggests that P. satyrus is closely
related to P. gracilis, P. zephyrus and P. oreas, whereas
nDNA suggests very strongly that P. satyrus is sister to P.
interrogationis, P. comma and P. g-argenteum (Figure 2).
Morphological and ecological features, however, suggest
that P. satyrus is more related to the latter clade. In addi-
tion to great overall phenotypic similarity between the
adults and immatures of P. satyrus and P. comma, larvae of

Page 6 of 16
(page number not for citation purposes)

Aglais io
1.00 Aglais urticae
Aglais milberti
1.00F Kaniska canace EW19-11
Kaniska canace NW164-1b
Kaniska canace NW164-1a
531 Nymphalis I-album
1.00 Nymphalis po ychloros
1.00 Nymphalis californica
1.00 Nymphalis antiopa
0.79 Nymphalis xanthomelas
1.00- c-aureum EW13-12
c-aureum NW65-8b
c-aureum NW65-8a
0.1 undina EW30-1
1.00 egea NW77-15
SI egea NW63-1
0.67 0.891 egea NW120-7b
Segea NW120-7a
comma NW65-6a
1.0 98 g-argenteum NW165-2a
g-argenteum NW165-2b
comma NW65-6b
C 174 .0 1.00 interrogationis NW77-12a
interrogationis NW77-12b
0.97 1.00 satyrus NW74-9
satyrus EW11-10
satyrus EW15-9a
0.67 satyrus EW15-12a
0.57 satyrus EW15-9b
0.99 satyrus EW15-12b
1.00- gigantea NW166-5a
gigantea NW166-5b
0o, -- c-album NW70-3a
c-album EW26-32
1.00 1.00 interposita NW166-6a
interposita NW166-6b
0.84 faunus NW74-12a
1.00 faunus NW74-12b
1.00 0.54 c-album EW13-1a
1.00 n c-album EW13-3
faunus EW19-1a
0.63 faunus EW21-13a
faunus EW12-7a
faunus EW12-7b
0.59 0.63 faunus EW21-13b
0.63 c-album EW13-1b
faunus EW19-1b

1.00 gracilis/zephyrus/oreas/progne/haroldii clade

Figure 4
Topology of haplotypes from Bayesian analysis based on the nDNA EF- Ia. Details of gracilis/zephyrus/oreas/prognel
haroldii clade are shown in the haplotype network in Figure 8b. Values below the branches are posterior probabilities for the
nodes to the right of the numbers.

those two taxa, as well as those of P. interrogationis and P.
g-argenteum [see [44]], feed on Ulmaceae, Moraceae and
Urticaceae as larvae, and late-instar larvae of P. satyrus and
P. comma make very similar larval nests out of altered host
plant leaves [45]. Polygonia satyrus is largely parapatric
with respect to P. comma, as the two fly in sympatry only
in a limited portion of northeastern North America, and
rarely in eastern Colorado, where P. comma is present only
as uncommon vagrant individuals from the east [46].

According to our estimates of times of divergence (Figure
2), P. satyrus diverged from the ancestral populations
between 7-8 mya based on nDNA, whereas the result
from mtDNA suggests that the divergence happened
much more recently, about 2 mya. Given that the nDNA
estimate of divergence time is older than that from

mtDNA, it is possible that the presence of an "alien"
mtDNA lineage in P. satyrus may be the result of ancient
introgression from the ancestor of P. gracilis+P. zephy-
rus+P. haroldii, which could have happened some 2-3
mya (prior to the onset of the Pleistocene glacial periods).
The current sympatric distribution of P. satyrus vs. P. graci-
lis+P. zephyrus (the geographic distribution of these taxa is
almost identical) highlights the potential for gene
exchange in the recent past and present. Given also that all
mtDNA haplotypes found to date in P. satyrus are very
similar, yet all nDNA haplotypes are more related to the
P. interrogationis clade, it is possible that repeated popula-
tion bottlenecks during the glacial cycles have wiped out
the original mtDNA lineages from P. satyrus, by chance
leaving the current introgressed lineage in extant popula-
tions. Lack of gene flow during the last 2 million years has

Page 7 of 16
(page number not for citation purposes)

BMC Evolutionary Biology 2009, 9:92

Aglais io
Aglais milberti
Aglais urticae
Aglais urticae Nmphalis I-album
1.oo Kaniska canace EW19-11
0.99- Kaniska canace NW164-1a
0.92 1--- -Kaniska canace NW164-1b
0.99 Nymphalis polychloros
9oF 91,- .Nymphalis californica
-'.68r- Nymphalis antiopa
1o mNymphalis xanthomelas
1 .oo1 I 1.oo1.00 c-aureum NW65-8
1.00 c-aureum EW13-12
0.54 -gigantea NW166-5
0o.94 undina EW30-1
0.1 1oo.- egea NW120-7a
S1 1.00 egea NW77-15
inerposia NW -b egea NW120-7b
i interposita NW166-6b
-0 oo c-album NW70-3b
0.64 1.00 faunus EW21-13b
c-album EW13-1b
interposita NW166-6a
c-album NW70-3a
faunus EW21-13a
0.94 1.00 faunus EW12-7
faunus EW19-1
c album EW13-1a
faunus 74-12b
1081 faunus 74-12a
8- zephyrus NW74-6b
1.00 interrogationis NW77-12a
interrogationis NW77-12b
1.0o comma NW65-6a
comma NW65-6b
1.00 .97 g-argenteum NW165-2a
g-argenteum NW165-2b
1.0- comma EW21-11a
comma EW21-11b
satyrus EW15-12b
satyrus EW15-9a
1.oo ).80 satyrus EW15-12a
- 1 satyrus EW15-9b
0.9"1 k o satyrus NW74-9
.89 satyrus EW11-10b
satyrus EW11-10a

-.74 gracilis/zephyrus/oreas/progne/haroldii clade

Figure 5
Topology of haplotypes from Bayesian analysis based on the nDNA GAPDH. Details of the gracilisizephyrus/oreas/
progne/haroldii clade are shown in the haplotype network in Figure 8c. Values below the branches are posterior probabilities for
the nodes to the right of the numbers.

now resulted in reciprocal monophyly to evolve in P.
satyrus and P. gracilis+P. zephyrus+P. haroldii. Such a specu-
lative scenario could be corroborated by more extensive
sampling of P. satyrus populations across North America,
which could make possible coalescense modeling to rule
out any possibility that the conflicting results can be
explained by ancient polymorphisms [39,47].

In butterflies females are the heterogametic sex and it is
accepted that "Haldane's rule" [48] is an important phe-
nomenon, ie. introgression of the maternally inherited
mtDNA will not enter the new gene pool due to low via-
bility or sterility of female F1 offspring [see [4911. Pres-

graves [50] showed that hybrid sterility and inviability are
common in Lepidoptera and evolve gradually. In those
studies of butterflies where both mtDNA and nDNA have
been screened, introgression in nDNA but not mtDNA
has been found between Papilio machaon and P. hospiton
(Papilionidae) [71 as well as between Heliconius cydno and
H. melpone (Nymphalidae) [13,14]. However, mtDNA
introgression has been found between the latter species
pair in another study [8], suggesting that the wide accept-
ance of Haldane's rule needs to be questioned. In the case
of Polygonia satyrus we have no knowledge of whether
hybrid female offspring are sterile or not, but even if
hybrids between contemporary P. satyrus and species from

Page 8 of 16
(page number not for citation purposes)

BMC Evolutionary Biology 2009, 9:92

BMC Evolutionary Biology 2009, 9:92

Aglais io
Aglais milberti
Aglais urticae



1.00- c-aureum NW65-8
c-aureum EW13-12
0.69 Nymphalis polychloros
0.99 Nymphalis I-album
0.92 nNymphalis californica
r--- Nymphalis antiopa

0.87 Nymphalis xanthomelas
1.00 Kaniska canace EW19-11
Kaniska canace NW164-1
faunus 74-12b
interposita NW166-6
0.90 0-95 c-album EW13-1a
0.90-- nI- c-album EW13-1b
faunus NW74-12a
faunus EW21-13
0.93 faunus EW12-7
faunus EW19-1
1.00 0.90 c-album NW70-3a
c-album NW70-3b
L interrogationis NW77-12
o1.00 g-argenteum NW165-2
0.98 0.96 satyrus NW74-9a
satyrus NW74-9b
0 981.00 comma NW65-6a
comma NW65-6b
1.00 egea NW120-7
1.00 egea NW77-15a
0.99 egea NW77-15b
0.99 undina EW43-1a
0.98 undina EW30-la
undina EW43-1b
undina EW30-1b
0.810 gigantea NW166-5
1.00 1.0- progne/oreas clade
0.99 gracilis/zephyrus/haroldii clade

Figure 6
Topology of haplotypes from Bayesian analysis based on the nDNA RpS5. Details of the progneloreas clade and the
gracilis/zephyrus/haroldii clade are shown in the haplotype network Figure 8d. Values below the branches are posterior probabil-
ities for the nodes to the right of the numbers.

the P. gracilis clade are inviable this may not have been the
case when (if) introgression occurred.

Recent mitochondrial introgression in P. oreas
The mtDNA haplotypes found in P. oreas are very similar
to those found in P. gracilis and P. zephyrus, and one hap-
lotype is shared between these species. In the nDNA data-
sets, haplotypes are shared between P. oreas, P. gracilis and
P. zephyrus for EF-la but not for the other genes, and in
the case of RpS5, haplotypes of P. oreas are clearly more
related to P. progne (Figure 4, Figure 5, Figure 6, Figure 7,
and Figure 8e). Polygonia oreas has been considered a sub-
species of P. progne by various authors [e.g., [51]], thus
once again, the nDNA dataset corroborates the morpho-
logical proposals of previous authors. Interestingly, both
the mtDNA and the nDNA datasets suggest that the clade
including the five taxa P. progne, P. oreas, P. haroldii, P. gra-

cilis and P. zephyrus began diverging about 5 mya at the
end of the Miocene. Based on nDNA, P. oreas and P. progne
began diverging about 3 mya, whereas the divergence of P.
oreas mtDNA is more recent. As with P. satyrus, no P. oreas
COI haplotypes were found to be more related to its prob-
able sister species P. progne, and it may be that bottlenecks
have wiped out the original mtDNA lineages, while cur-
rent introgression is introducing new genetic material into
P. oreas from the P. gracilis complex (most likely from
western P. zephyrus). On the other hand, we have sampled
only 5 individuals of P. oreas, and it may be that a denser
sampling would reveal mtDNA lineages closer to P.
progne. Polygonia oreas flies in sympatry and synchrony
with P. zephyrus throughout the vast majority of its range,
the latter usually being much more abundant locally and
regionally; thus there are ample opportunities for ongoing
introgression between the two taxa. It should be noted

Page 9 of 16
(page number not for citation purposes)

Aglais io
1.00 Aglais milberti


alis antiopa

I 1.00 gracilis/zephyrus/oreas/progne/haroldii clade

Figure 7
Topology of haplotypes from Bayesian analysis based on the nDNA wgl. Details of the gracilis/zephyrus/oreas/prognel
haroldii clade are shown in the haplotype network Figure 8e. Values below the branches are posterior probabilities for the
nodes to the right of the numbers.

that adults of some subspecies of P. oreas, especially nigro-
zephyrus, and some individuals of threatfuli, are so similar
to those of sympatric P. zephyrus that many experienced
lepidopterists cannot distinguish them (without life his-
tory information), and these two taxa were not described
until 1984 and 2001, respectively (adults of these taxa are
still hiding in museum series of P. zephyrus all over the

Recent speciation of P. haroldii and incipient speciation
of P. gracilis/zephyrus?
The taxa P. haroldii, P. gracilis and P. zephyrus appear to be
related to one another in a complicated way. Mitochon-

drial DNA suggests that P. haroldii is a distinct lineage sis-
ter to the P. gracilis/zephyrus lineage (that includes the
"alien" P. satyrus lineage) (Figure 3), yet the nDNA sug-
gests that P. haroldii is not distinct from P. zephyrus (note
that by chance, the P. zephyrus chosen for the 25-taxon
analyses is rather different from the other P. zephyrus) (Fig-
ure 4, Figure 5, Figure 6 and Figure 7). Here the classical
explanation for conflicts [52] of recent divergence with
not enough time for slowly evolving nuclear genes to have
segregated would appear to hold. Interestingly, P. haroldii
(endemic to mainland Mexico) and some P. zephyrus
(endemic to western United States and Canada, including
northern Baja California, Mexico) nDNA haplotypes

Page 10 of 16
(page number not for citation purposes)

BMC Evolutionary Biology 2009, 9:92

Aglais urticae
1.00 Kaniska canace NW164-1b
1.00 Kaniska canace EW19-11
0.75 Kaniska canace NW164-1a
0.99 Nymphalis I-album
099 --- Nymphalis polychloros
1.00 i Nymp
-- Nymphalis xanthomelas
SNymphalis californica
1.00 c-aureum EW13-12b
c-aureum EW13-10a
0.95 c-aureum NW65-8
c aureum EW13-12a
c aureum EW13-10b
zephyrus EW10-12a
gigantea NW166-5
3 interposita NW166-6a
interposita NW166-6b
0.98 egea NW77-15
egea NW120-7b
egea NW120-7a
undina EW30-1
1.00 undina EW43-2b
0.54 undina EW43-2a
-undina EW43-1
comma EW21-10
g-argenteum NW165-2
satyrus NW74-9
satyrus EW15-9
0.99 satyrus EW15-12
satyrus EW11-10
comma NW65-6
comma EW21-11b
comma EW21-11a
0.89 interrogationis EW22-14
1.00 interrogationis NW77-12a
interrogationis NW77-12b
0.96 gracilis EW22-9b
8 n c-album NW70-3b
--. -8 c-album NW70-3a
0.62 0.84 1 zephyrus EW10-12b
0.51 c-album EW13-1a
c album EW13-1b
Sc-album EW13-3a
0.99 c-album EW13-3b
S0.591 faunus EW21-13
0n .9 faunus EW12-7
S faunus NW74-12
faunus EW19-1





,,I% I

Polygonia zephyrus
* Polygonia oreas
* Polygonia progne
0 Polygonia gracilis
* Polygonia haroldii

Figure 8
Minimum spanning networks. a) COI b) EF-lIL c)
GAPDH d) RpS5 and e) wgl. Size of the circles are directly
proportional to the number of individuals with that haplo-
type. Small white circles indicate a missing haplotype. Each
branch is equivalent to one basepair change. Circles with
more than one pattern show the proportion of each species.
The colour coding are as follows; blue P. zephyrus, red P.
oreas, green P. progne, yellow P. gracilis and orange P.

appear to be more related to each other, perhaps suggest-
ing recent gene exchange between these western taxa dur-
ing the Pleistocene glacial periods. Currently, the two taxa
appear to be allopatrically distributed. A close relation-
ship between P. haroldii and P. zephyrus was suggested by
Krogen [53], based on morphological similarities. Beutel-
spacher [54] reported an unidentified species of Urti-
caceae as a larval foodplant for P. haroldii (presumably in
the Valley of Mexico), although this record seems
unlikely, since P. haroldii is usually found in immediate
association with Ribes species (Grossulariaceae) (ADW,
pers. obs.), the host plant genus utilized by P. zephyrus.

Our data suggest that morphological differentiation may
occur rapidly in Polygonia, once speciation has occurred.
Despite the essentially identical nDNA haplotypes

between P. zephyrus and P. haroldii, the latter has diversi-
fied morphologically to the point where it cannot be con-
fused with any other member of the genus. This was
perhaps achieved through evolution of a mimetic rela-
tionship with the presumably distasteful model Dione
moneta (Nymphalidae: Heliconiinae: Heliconiini); in
flight, adults of D. moneta and P. haroldii appear nearly
indistinguishable, since the metallic ventral spots of D.
moneta frequently are not visible (ADW, pers. obs.). Cur-
rently, these two species are very often found flying in
sympatry and synchrony throughout the Mexican distri-
bution of P. haroldii, although the presumed model, D.
moneta, is usually much more widespread and common
than P. haroldii. No obvious geographic variation in mor-
phology has been noted in P. haroldii.

The taxon pair P. gracilis and P. zephyrus has been treated
as two hypothetically separate species in this study, but
the current consensus is that these are subspecies of the
same species [27,28]. Our results are ambiguous about
whether these two taxa are currently diverging or merging.
Morphologically, populations of far western P. zephyrus
are separable from far eastern populations of P. gracilis,
but there is a clear dine between the extremes, and popu-
lations found in Alberta, Canada, consist mostly of adults
that cannot be confidently assigned to one or the other
taxon [55]. On the one hand, our data does not distin-
guish between the two taxa (haplotypes are shared regard-
less of gene or genome inspected), but on the other hand,
haplotypes are also shared with P. progne, P. haroldii, and
P. oreas, which are distinct species-level taxa. Thus,
detailed elaboration of the taxonomic status of P. zephyrus
and P. gracilis will only be possible once a thorough study
can be conducted, considering dozens of populations
from throughout the range of the complex. The current
distribution of these two taxa, with an apparently broad
zone in western Canada where their identities become
blurred, suggests ongoing gene flow between them, and a
careful study of populations in Alberta seems warranted.

Polygonia interposita, species or subspecies of P. c-album?
The rarely collected taxon P. interposita is found in central
Asian mountains and has often been considered a subspe-
cies of P. c-album [e.g., [43]]. This taxon has frequently
been confused with P. undina, due to the somewhat simi-
lar ventral wing pattern and similar distribution. We were
only able to get one specimen of P. interposita that gave
good quality DNA. The mtDNA of this specimen was
almost identical to P. c-album (which shows very little var-
iation in COI across its entire range; Weingartner et al. in
prep.). The nDNA, however, was quite distinct from P. c-
album, and indeed in the 25-taxon analysis, P. interposita
emerged as sister to P. c-album+P. faunus, with a diver-
gence time estimated at about 5 mya. Could this possibly
be a similar case to P. satyrus in North America? Only

Page 11 of 16
(page number not for citation purposes)

BMC Evolutionary Biology 2009, 9:92

Clear P undina+P egea

P faunus
P. comma+P. g-argenteum
P. c-album+P. interposita

The "grey" F progne
zone P. oreas
zone P haroldii

P v'acilis+P zephyrus

wnh a p .ieum
long lime P gigariea
sln.:,- P wlerroqationis
Speci3aIIol P saiI,,3

Figure 9
Species through time, a summary of our results. Spe-
cies below the "grey zone" are clear independent lineages
with no known closely related sister species. Species in the
"grey zone" are at various stages in the speciation process.
Species above the "grey zone" are closely related sister spe-
cies that are separate genetic entities. They have thus by this
time also become clear independent lineages ready for a
future new bifurcation, should the right circumstances arise.
Figure modified from de Queiroz [58].

more samples of P. interposita would shed light on this
question, but based on the current specimen, it is possible
that the mitochondrial lineage of P. c-album has invaded
the genome of P. interposita in recent times, resulting in a
situation where the two genomes give conflicting signals
regarding phylogenetic relatedness.

Species as lineages through time
The concept of species as lineages is fast gaining support
from the scientific community [56-59]. The concept takes
into account that species are part of an evolutionary con-
tinuum from diverging populations to already diverged,
well-defined species. Many of the multitude of proposed
species concepts lie along this continuum, but are not
general enough to explain the diversity we see in nature.
Here we present results for a small group of well-known
butterflies with a relatively stable taxonomy. Despite
molecular data from 6 gene regions for a total of 4879 bp
(much more than the standard in species level phyloge-
netic studies at the moment), we were unable to resolve
the relationships of the 16 species unambiguously,

mainly due to conflicts between mitochondrial and
nuclear gene regions.

Considering the lineage concept, it is clear that P. c-
aureum, P. gigantea, P. egea, P. undina, P. satyrus and P.
interrogationis have differentiated so long ago that there is
no question about their taxonomic status as species (Fig-
ure 9). The species-level status of P. comma and P. g-argen-
teum, as well as P. c-album, P. interposita and P. faunus also
is not really a question based on our results, but they have
speciated relatively recently and in the case of P. inter-
posita, may still hybridize in nature with P. c-album. In Fig-
ure 9, P. faunus is placed out of the grey zone due to the
clear separation of it's populations in North America from
those of P. c-album and P. interposita in Eurasia. Further
down the continuum closer to the divergence events are P.
progne, P. oreas and P. haroldii (Figure 9), which have spe-
ciated so recently that occasional gene flow may still occur
between P. oreas and P. progne and/or P. zephyrus, but
which remain taxonomic entities separate from their clos-
est relatives. Just above the divergence line, entering into
the grey zone of one or two species is the taxon pair P. gra-
cilis and P. zephyrus (Figure 9). To really be able to say
whether the two are above or below the line would
require population genetic methods to see whether gene
flow between the two populations is sufficiently high to
consider them conspecific.

In conclusion, although we now have included consider-
able amounts of new genetic information in an attempt to
interpret the evolutionary history of Polygonia butterflies,
we are still not able to fully understand the processes of
speciation in this taxon. Especially within the Nearctic
clade, more population genetic data is needed. However,
our results graphically demonstrate, first, that species in
this group evolve over time, sometimes over a very long
time, and, second, that evidently even well-differentiated
species can hybridize to the extent that different parts of
their genome may suggest strongly conflicting patterns of

The results from the present study do not change the main
conclusion from the study of host plant range in Polygonia
butterflies [21]. In that paper we introduced the idea that
based on the phylogeny it was possible to show that but-
terfly clades including species that use host plants addi-
tional to, or other than, the "urticalean rosids", included
more butterfly species than the sister clade (only feeding
on "urticalean rosids"). Our present results are still in
agreement with the former result and there is no case in
which the reverse is valid (that species restricted to "urti-
calean rosids" constitute more butterfly species than the
sister group of species with a broader host plant range).
Thus, we believe that being able to expand the host plant

Page 12 of 16
(page number not for citation purposes)

BMC Evolutionary Biology 2009, 9:92

Table I: Summary of number of individuals per species sequenced for a given gene.

Number of individuals sequenced





Outgroup taxa
Aglais io
Aglais milberti
Aglais urticae
Nymphalis antiopa
Nymphalis californica
Nymphalis I-album
Nymphalis polychloros
Nymphalis xanthomelas

Ingroup taxa
Kaniska canace
Polygonia c-album
Polygonia interposita
Polygonia c-aureum
Polygonia comma
Polygonia egea
Polygonia undina
Polygonia faunus
Polygonia g-argenteum
Polygonia gigantea
Polygonia gracilis
Polygonia zephyrus
Polygonia haroldii
Polygonia interrogationis
Polygonia oreas
Polygonia progne
Polygonia satyrus

See Additional File I for details of individuals, including collection locality and GenBank accession numbers.

range will enhance speciation through colonizations and
local adaptations, according to the oscillation hypothesis

We have shown that the species-level relationships
inferred from DNA sequence data may be strongly influ-
enced by the markers that have been chosen. This then
begs the question of how this phenomenon affects studies
aimed at looking at higher levels of phylogenetic relation-
ships, such as genera or families. Fortunately, our previ-
ous studies at higher levels have used the same markers as
we have in this study [33,61-66], and results show that the
COI is generally concordant with the nuclear markers at
taxonomic levels above genera. This is probably due to the
same stochastic processes that lead to reciprocal mono-
phyly at the species level, i.e. given enough time, lineages
(species) will go extinct, leaving sister entities that we call
genera (and by default higher taxa) reciprocally mono-
phyletic. This is not to say that all currently described gen-
era are monophyletic entities, simply because the majority
of genera have not been rigorously tested for monophyly
using phylogenetic analyses.

We sampled 96 individuals of all Polygonia species, as well
as 8 outgroup species belonging to the genera Nymphalis
and Aglais (see Table 1 and Additional File 1). Most indi-
viduals were collected by colleagues (see Acknowledg-
ments) and sent dry to Stockholm. Total genomic DNA
was extracted from two legs using QIAgen's DNEasy
extraction kit, according to the manufacturer's instruc-
tions, with the exception that individuals more than 2
years old at extraction were eluted into 50 Al of elution
buffer, rather than the recommended 200 il. Voucher
specimens are stored at the Department of Zoology,
Stockholm University and Laboratory of Genetics, Univer-
sity of Turku, and can be viewed at http://nymphali

We amplified 6 loci using PCR directly from the genomic
extracts. The loci were cytochrome oxidase subunit I (COI)
and NADH subunit 1 (ND1) from the mitochondrial
genome, and elongation factor-1 a (EF- l), wingless (wgl),
glyceraldehyde-3-phosphate dehydrogenase (GAPDH) and
ribosomal protein S5 (RpS5) from different nuclear

Page 13 of 16
(page number not for citation purposes)

BMC Evolutionary Biology 2009, 9:92

genomes. Primers and PCR protocols were taken directly
from Wahlberg and Wheat [67], except for ND1, for
which we followed the protocol described in Nylin et al.
[35]. PCR products were cleaned using exonuclease I and
calf intestine alkaline phosphatase (Fementas) and
sequenced directly, using either the PCR primers or uni-
versal tails attached to the primers [for details, see [67]],
on a Beckman-Coulter CEQ8000 capillary sequencer
(Stockholm), or an ABI PRISMR 3130xl capillary
sequencer (Turku) using dye terminator sequencing kits
according to the recommendations of manufacturers.

All six genes were initially amplified for a selection of 25
taxa (8 outgroup species and 17 taxa of Polygonia). In
order to verify patterns of strong conflict between the
mitochondrial and nuclear genes [33,36], a further 77
individuals of Polygonia and four individuals of Kaniska
canace were amplified and sequenced for COI, 27 individ-
uals of Polygonia for EF-la, 30 individuals of Polygonia for
wgl, 23 individuals of Polygonia for GAPDH and 20 indi-
viduals of Polygonia for RpS5. Two individuals of Kaniska
canace were amplified and sequenced for all nuclear

Resulting chromatograms were examined by eye in
BioEdit [68] and any heterozygous positions (two equally
sized peaks observed at one position) were coded with
IUPAC ambiguity codes. All sequences are from protein-
coding genes and thus alignment was trivial. As noted in
previous publications [33,35,36], a one-codon deletion
was inferred in the wgl sequence of the three species of
Aglais. Heterozygous sequences were separated manually
into haplotypes. For sequences with only one hetero-
zygous position, this was trivial. For those with two or
more heterozygous positions, one haplotype was
assumed to be identical to a common haplotype found in
other individuals of the same species. This was possible in
all cases.

The previously noted strong conflict between two mito-
chondrial and two nuclear genes [33,36] was investigated
with a total evidence approach and Partitioned Bremer
Support (PBS) on the 25-taxon dataset. Results suggested
that mitochondrial and nuclear partitions continued to
conflict with the addition of new nuclear gene regions. We
thus analysed the combined mitochondrial genes and the
combined nuclear genes to obtain estimates of relation-
ships based on the mitochondrial genome and the nuclear
genome, respectively. The two genome sets were analysed
separately, but combined within each set (ie. COI+ND1
and EF-la+GAPDH+RpS5+wgl) and will be referred to as
the mitochondrial data and the nuclear data, respectively.
Parsimony analyses were conducted using a heuristic
search algorithm in the program TNT [69] on the equally
weighted data set. The data were subjected to 100 random

addition rounds of successive Sectorial, Ratchet, Drift and
Tree Fusing searches [70-72]. We evaluated the character
support for the clades in the resulting cladograms using
Bremer support [73,74] and Partitioned Bremer support
[75,76]. The scripting feature of TNT was used to calculate
these values [see [6411.

Bayesian inference of phylogeny and times of divergence
were estimated using the program BEAST vl.4.6 [77].
Both datasets were analysed under the GTR+ F model with
a relaxed clock, allowing branch lengths to vary according
to an uncorrelated Lognormal distribution [78]. The tree
prior was set to the Yule process, and the "tree-
Model.RootHeight" prior (i.e., the age at the root of the
tree) was set to 33 million years (with a standard devia-
tion of 5 million years), in accordance with results from
Wahlberg [79]. All other priors were left to the defaults in
BEAST. Parameters were estimated using 2 independent
runs of 1 million generations each (with a pre-run burn-
in of 10000 generations), with parameters sampled every
1000 generations. Convergence was checked in the Tracer
vl.4.6 program and summary trees were generated using
TreeAnnotator vl.4.6, both part of the BEAST package.

To confirm that Bayesian analyses converged on the same
topology, the data were also analyzed with MrBayes 3.1
[80]. The Bayesian analysis was performed on the com-
bined data set with parameter values estimated separately
for each gene region using the "unlink" command and the
rate prior (ratepr) set to "variable". The analysis was run
twice simultaneously for 2 million generations, with four
chains (one cold and three heated) and every 500th tree
sampled. The first 500 sampled generations discarded as
bum-in (based on a visual inspection of when log likeli-
hood values reached stationarity), leaving 3501 sampled
generations for the estimation of posterior probabilities.
Results of the two simultaneous runs were compared for
convergence using Tracer vI.4.6 [77].

The expanded single-gene datasets were analysed sepa-
rately after separating heterozygotes into haplotypes.
These datasets were analysed using both parsimony and
Bayesian methods in TNT and MrBayes 3.1, respectively.
Search parameters were as above, except the single data-
sets were not partitioned in any way.

In order to further investigate the resulting polytomies, we
constructed a haplotype network in TCS [81], which
shows how haplotypes are connected to each other. In this
program, the gene genealogies from DNA sequences are
estimated with statistical parsimony according to Temple-
ton et al. [82]. We focused on the Nearctic Polygonia spe-
cies (excluding P. faunus). Regions of missing basepairs
were removed and we performed analyses of all Nearctic
taxa as well as subsets of clades. The datasets are com-

Page 14 of 16
(page number not for citation purposes)

BMC Evolutionary Biology 2009, 9:92

BMC Evolutionary Biology 2009, 9:92

prised of 1430 bp for COI, 1240 bp for EF-lIo, 392 bp for
wgl, 691 bp for GAPDH and 617 bp for RpS5.

Authors' contributions
NW, SN and EW conceived the study, EW did most of the
labwork, NW and EW carried out the analyses and wrote
the manuscript. ADW helped interpret phylogenetic pat-
terns and wrote part of the discussion. All authors partook
in discussions during analysis and writing, read and
approved the final manuscript.

Additional material

Additional file 1
List of specimens sampled in this study. Voucher codes, locality where
the taxa were collected and GenBank accession numbers for genes
sequenced. Photos of vouchers can be ,. .i n ,, "n' lI, ..' ,, i/
Click here for file

Thanks to Jorge Llorente, Armando Luis and Isabel Vargas (Museo de
Zoologia, UNAM, Mexico City) for access to specimens under their care,
literature and for providing permits and assistance to ADW. We are also
grateful to Anton Chichvarkin, Norbert Kondla, Zdenek Fric, Michel Tar-
rier, Vladimir Lukhtanov, Andrew B. Martynenko, David Threatful, Jim
Beck, Bob Parsons, Runar Krogen, Katy Prudic and Mike Leski for providing
specimens used in this study. We thank Bertil Borg and two anonymous ref-
erees for useful comments on a previous version of this ms. This study was
funded by grants to SN and NW from the Swedish Research Council, and
from the Academy of Finland to NW (grant number 118369). Partial fund-
ing to ADW was provided by DGAPA-UNAM (Mexico City).

I. Barraclough TG, Vogler AP, Harvey PH: Revealing factors that
promote speciation. Philos Trans R Soc Lond B Biol Sci 1998,
2. Barraclough TG, Vogler AP: Detecting the geographical pattern
of speciation from species-level phylogenies. Am Nat 2000,
3. Tajima F: Evolutionary relationship of DNA-sequences in
finite populations. Genetics 1983, 105:437-460.
4. Nichols R: Gene trees and species trees are not the same.
Trends Ecol Evol 200 1, 16:358-364.
5. Ting C-T, Tsaur S-C, Wu C-I: The phylogeny of closely related
species as revealed by the genealogy of a speciation gene,
Odysseus. Proceedings of the National Academy of Sciences USA 2000,
6. Bachtrog D, Thornton K, Clark A, Andolfatto P: Extensive intro-
gression of mitochondrial DNA relative to nuclear genes in
the Drosophila yakuba species group. Evolution 2006,
7. Cianchi R, Ungaro A, Marini M, Bullini L: Differential patterns of
hybridization and introgression between the swallowtails
Papilio machaon and P. hospiton from Sardinia and Corsica
islands (Lepidopera, Papilionidae). Mol Ecol 2003,
8. Salazar C,Jiggins CD, TaylorJE, Kronforst MR, Linares M: Gene flow
and the genealogical history of Heliconius heurippa. BMC Evol
Biol 2008, 8:132.

9. Chan KMA, Levin SA: Leaky prezygotic isolation and porous
genomes: rapid introgression of maternally inherited DNA.
Evolution 2005, 59:720-729.
10. Sota T, Vogler AP: Incongruence of mitochondrial and nuclear
gene trees in the carabid beetles Ohomopterus. Syst Biol 2001,
I I. Sota T: Radiation and reticulation: extensive introgressive
hybridization in the carabid beetles Ohomopterus inferred
from mitochondrial gene genealogy. Popul Ecol 2002,
12. Linnen CR, Farrel BD: Mitonuclear discordance is caused by
rampant mitochondrial introgression in Neodiprion
(Hymenoptera: Diprionidae) sawflies. Evolution 2007,
13. Bull V, Beltrin M, Jiggins CD, McMillan WO, Bermingham E, Mallet J:
Polyphyly and gene flow between non-sibling Heliconius spe-
cies. BMC Biology 2006, 4:1 I1.
14. Kronforst MR, Young LG, Blume LM, Gilbert LE: Multilocus analy-
ses of admixture and introgression among hybridizing Hell-
conius butterflies. Evolution 2006, 60:1254-1268.
15. McCracken KG, Sorensen MD: Is homoplasy or lineage sorting
the source of incongruent mtDNA and nuclear gene trees in
the Stiff-tailed Ducks (Nomonyx-Oxyura)? Syst Biol 2005,
16. Mallet J: Hybridization as an invasion of the genome. Trends
Ecol Evol 2005, 20:229-237.
17. Kronforst MR: Gene flow persists millions of years after speci-
ation in Heliconius butterflies. BMC Evol Biol 2008, 8:98.
18. Nylin S: Host plant specialization and seasonality in a polypha-
gous butterfly, Polygonia c-album (Nymphalidae). Oikos 1988,
19. Janz N, Nylin S: The role of female search behaviour in deter-
mining host plant range in plant feeding insects: a test of the
information processing hypothesis. Proceedings of the Royal Soci-
ety of London Series B Biological Sciences 1997, 264:701-707.
20. Janz N: Sex-linked inheritance of host-plant specialization in a
polyphagous butterfly. Proceedings of the Royal Society of London
Series B Biological Sciences 1998, 265:1675-1678.
21. Weingartner E, Wahlberg N, Nylin S: Dynamics of host plant use
and species diversity: a phylogenetic investigation in Polygo-
nia butterflies (Nymphalidae). j Evol Biol 2006, 19:483-491.
22. Nylin S: Seasonal plasticity in life-history traits growth and
development in Polygonia c-album (Lepidoptera, Nymphali-
dae). BiolJ Linn Soc 1992, 47(3):301-323.
23. Wiklund C, Wickman PO, Nylin S: A sex difference in the pro-
pensity to enter direct/diapause development a result of
selection for protandry. Evolution 1992, 46(2):519-528.
24. Bryant SR, Thomas CD, Bale JS: Thermal ecology of gregarious
and solitary nettle-feeding nymphalid butterfly larvae. Oeco-
logia 2000, 122:1-10.
25. Hiroyoshi S, Mitsuhashi J: Sperm reflux and its role in multiple
mating in males of a butterfly Polygonia c-aureum Linnaeus
(Lepidoptera: Nymphalidae). j Insect Physiol 1999,
45(2):107-1 12.
26. Tanaka D, Sakurama T, Mitsumasu K, Yamanaka A, Endo K: Separa-
tion of bombyxin from a neuropeptide of Bombyx mori show-
ing summer-morph-producing hormone (SMPH) activity in
the Asian comma butterfly, Polygonia c-aureum L. J Insect Phys-
iol 1997, 43(2):197-201.
27. Opler PA, Warren AD: Butterflies of North America. 2. Scien-
tific Names List for Butterfly Species of North America,
North of Mexico. Fort Collins, Colorado, USA: Gillette Museum
Publications; 2002.
28. Pelham JP: A catalogue of the butterflies of the United States
and Canada, with a complete bibliography of the descriptive
and systematic literature. j Res Lepid 2008, 40:.
29. Gorbunov PY: The Butterflies of Russia. Moscow, Russia: Russian
Academy of Sciences; 200 1.
30. Churkin SV: Taxonomic notes on Polygonia Hubner, [1818]
(Lepidoptera, Nymphalidae) with description of a new sub-
species. Helios (Moscow) 2003, 4:132-147.
31. Guppy CS, Shepard JH: Butterflies of British Columbia. UBC
Press in collaboration with the Royal British Columbia Museum; 200 1.
32. Layberry RA, Hall PW, LafontaineJD: The Butterflies of Canada.
Toronto: University of Toronto Press; 1998.

Page 15 of 16
(page number not for citation purposes)

BMC Evolutionary Biology 2009, 9:92

33. Wahlberg N, Brower AVZ, Nylin S: Phylogenetic relationships
and historical biogeography of tribes and genera in the sub-
family Nymphalinae (Lepidoptera: Nymphalidae). Biol] Linn
Soc 2005, 86:227-25 I.
34. Nylin S,Janz N: Butterfly host plant range: an example of plas-
ticity as a promoter of speciation? Evol Ecol 2009, 23:137-146.
35. Nylin S, Nyblom K, Ronquist F, Janz N, Belicek Kallersjo M: Phyl-
ogeny of Polygonia, Nymphalis and related butterflies (Lepi-
doptera: Nymphalidae): a total-evidence analysis. ZoolJ Linn
Soc 2001, 132:441-468.
36. Wahlberg N, Nylin S: Morphology versus molecules: resolution
of the positions of Nymphalis, Polygonia and related genera
(Lepidoptera: Nymphalidae). Cladistics 2003, 19:213-223.
37. Slatkin M, Hudson RR: Pairwise comparisons of mitochondrial-
DNA sequences in stable and exponentially growing popula-
tions. Genetics 1991, 129:555-562.
38. Maddison WP: Gene trees in species trees. Syst Biol 1997,
39. Knowles LL, Carstens BC: Delimiting species without mono-
phyletic gene trees. Syst Biol 2007, 56:887-895.
40. Kullberg J, Albrecht A, Kaila L, Varis V: Checklist of Finnish Lepi-
doptera Suomen perhosten luettelo. Sahlbergia 2001,
41. Aarvik L, Berggren K, Hansen LO: Catalogus Lepidopterorum
Norvegiae. As, Oslo: Lepidopterologisk arbeidsgruppe, Zoologisk
museum, Universitet i Oslo, Norsk instituttfor skogsforskning; 2000.
42. Corbet AS, Pendlebury HM: The Butterflies of the Malay Penin-
sula. Kuala Lumpur, Malaysia 4th edition. 1992.
43. Tuzov VK: Nymphalidae part I. In Guide to the Butterflies of the Pal-
earctic Region Volume 5. Edited by: Bozano GC. Milano: Omnes Artes;
44. Maza RGdl, Maza Jdl: Notas sobre el ciclo de vida de Polygonia
gargenteum (Dbl y Hew) (Nymphalidae). Revista de la Sociedad
Mexicana de Lepidopterologia 1977, 3:35-41.
45. Scott JA: The Butterflies of North America. Stanford: Stanford
University Press; 1986.
46. Ferris CD, Brown F: Butterflies of the Rocky Mountain States.
Norman, Oklahoma, USA: University of Oklahoma Press; 1981.
47. Good JM, Hird S, Reid N, Demboski JR, Steppan SJ, Martin-Nims TR,
Sullivan J: Ancient hybridization and mitochondrial capture
between two species of chipmunks. Mol Ecol 2008,
48. Haldane JBS: Sex ratio and unisexual sterility in animal
hybrids. JGenet 1922, 12:101-109.
49. Sperling FAH: Butterfly species and molecular phylogenies. In
Butterflies: Evolution and Ecology Taking Flight Edited by: Boggs CL, Watt
WB, Ehrlich PR. Chicago: University of Chicago Press; 2003:431-458.
50. Presgraves DC: Patterns of postzygotic isolation in Lepidop-
tera. Evolution 2002, 56:1168-1183.
51. Scott JA: A review of Polygonia progne (oreas) and P. gracilis
(zephyrus) (Nymphalidae), including a new subspecies from
the southern Rocky Mountains. j Res Lepid 1984, 23:197-210.
52. AviseJC: Phylogeography: The History and Formation of Spe-
cies. Cambridge, MA: Harvard University Press; 2000.
53. Krogen R: Records of Polygonia haroldi (Dewitz, 1877) in Son-
ora, Mexico. Atalanta 2000, 3 1:67-70.
54. Beutelspacher CR: Mariposas dirunas del Valle de Mexico. Mex-
ico City: Ediciones Cientificas La Prensa Medica Mexicana; 1980.
55. Bird CD, Hilchie GJ, Kondla NG, Pike EM, Sperling FA: Alberta But-
terflies. Edmonton: Provincial Museum of Alberta; 1995.
56. O'Hara RJ: Systematic generalization, historical fate, and the
species problem. Syst Biol 1993, 42:231-246.
57. O'Hara RJ: Evolutionary history and the species problem. Am
Zool 1994, 34:12-22.
58. de Queiroz K: The general lineage concept of species, species
criteria, and the process of speciation. In Endless Forms: Species
and Speciation Edited by: Howard DJ, Berlocher SH. Oxford: Oxford
University Press; 1998:57-75.
59. de Queiroz K: Species concepts and species delimitation. Syst
Biol 2007, 56:879-886.
60. Janz N, Nylin S, Wahlberg N: Diversity begets diversity: host
expansions and the diversification of plant-feeding insects.
BMC Evol Biol 2006, 6:4.
61. Wahlberg N, Weingartner E, Nylin S: Towards a better under-
standing of the higher systematics of Nymphalidae (Lepidop-
tera: Papilionoidea). Mol Phylogenet Evol 2003, 28:473-484.

62. Wahlberg N, Braby MF, Brower AVZ, de Jong R, Lee M-M, Nylin S,
Pierce N, Sperling FA, Vila R, Warren AD, et al.: Synergistic effects
of combining morphological and molecular data in resolving
the phylogeny of butterflies and skippers. Proc R Soc Lond B Biol
Sci 2005, 272:1577-1586.
63. Wahlberg N, Freitas AVL: Colonization of and radiation in
South America by butterflies in the subtribe Phyciodina
(Lepidoptera: Nymphalidae). Mol Phylogenet Evol 2007,
64. Peha C, Wahlberg N, Weingartner E, Kodandaramaiah U, Nylin S,
Freitas AVL, Brower AVZ: Higher level phylogeny of Satyrinae
butterflies (Lepidoptera: Nymphalidae) based on DNA
sequence data. Mol Phylogenet Evol 2006, 40:29-49.
65. Peha C, Wahlberg N: Prehistorical climate change increased
diversification of a group of butterflies. Biology Letters 2008,
66. Warren AD, Ogawa JR, Brower AVZ: Phylogenetic relationships
of subfamilies and circumscriptions of tribes in the family
Hesperiidae (Lepidoptera: Hesperioidea). Cladistics 2008,
67. Wahlberg N, Wheat CW: Genomic outposts serve the phylog-
enomic pioneers: designing novel nuclear markers for
genomic DNA extractions of Lepidoptera. Syst Biol 2008,
68. Hall TA: BioEdit: a user-friendly biological sequence align-
ment editor and analysis program for Windows 95/98/NT.
Nucl Acids Symp Ser 1999, 4 1:95-98.
69. Goloboff PA, Farris JS, Nixon KC: T. N. T. (Tree Analysis using
New Technology). Published by the authors 1.0th edition. 2004
70. Goloboff PA: Analyzing large data sets in reasonable times:
solutions for composite optima. Cladistics 1999, 15:415-428.
71. Nixon KC: The parsimony ratchet, a new method for rapid
parsimony analysis. Cladistics 1999, 15:407-414.
72. Moilanen A: Searching for most parsimonious trees with sim-
ulated evolutionary optimization. Cladistics 1999, 15:39-50.
73. Bremer K: The limits of amino acid sequence data in
angiosperm phylogenetic reconstruction. Evolution 1988,
74. Bremer K: Branch support and tree stability. Cladistics 1994,
75. Baker RH, DeSalle R: Multiple sources of character information
and the phylogeny of Hawaiian drosophilids. Syst Biol 1997,
76. Gatesy O'Grady P, Baker RH: Corroboration among data sets
in simultaneous analysis: hidden support for phylogenetic
relationships among higher level artiodactyl taxa. Cladistics
1999, 15:271-313.
77. Drummond AJ, Rambaut A: BEAST: Bayesian evolutionary anal-
ysis by sampling trees. BMC Evol Biol 2007, 7:214.
78. Drummond AJ, Ho SYW, Phillips MJ, Rambaut A: Relaxed phyloge-
netics and dating with confidence. Plos Biology 2006,
79. Wahlberg N: That awkward age for butterflies: insights from
the age of the butterfly subfamily Nymphalinae. Syst Biol 2006,
80. Ronquist F, Huelsenbeck JP: MrBayes 3: Bayesian phylogenetic
inference under mixed models. Bioinformatics 2003,
81. Posada D, Crandall KA: TCS: a computer program to estimate
gene genealogies. Mol Ecol 2000, 9:1657-1660.
82. Templeton AR, Crandall KA, Sing CF: A cladistic analysis of phe-
notypic associations with haplotypes inferred from restric-
tion endonuclease mapping and DNA sequence data. Ill.
Cladogram estimation. Genetics 1992, 132:619-633.

Page 16 of 16
(page number not for citation purposes)

University of Florida Home Page
© 2004 - 2010 University of Florida George A. Smathers Libraries.
All rights reserved.

Acceptable Use, Copyright, and Disclaimer Statement
Last updated October 10, 2010 - - mvs