Molecular Phylogenetics and Evolution 38 (2006) 575–582www.elsevier.com/locate/ympev

Phylogenetic relationships among species groups of the ant genus Myrmecia

Eisuke Hasegawa a,b,c,¤, Ross H. Crozier b,c

a Department of Ecology and Systematics, Graduate School of Agriculture, Hokkaido University, Kita-ku, Sapporo 060-8589, Japanb School of Tropical Biology, James Cook University, Townsville, Qld 4811, Australia

c School of Biochemistry, Chemistry and Genetics, La Trobe University, Bundoora, Vic. 3083, Australia

Received 30 March 2005; revised 25 August 2005; accepted 20 September 2005

Abstract

Phylogenetic relationships among the nine species groups of the predominately Australian ant genus Myrmecia were inferred using 38Myrmecia species and an outgroup using DNA sequences from two nuclear genes (622 nt from 28S rRNA and 1907 nt from the long-wave opsin gene). Nothomyrmecia macrops was selected as the most appropriate outgroup based on recent reliable studies showing mono-phyly of Myrmecia with Nothomyrmecia. The four species groups with an occipital carina (those of gulosa, nigrocincta, urens, and picta)were found to form a paraphyletic and basal assemblage out of which the Wve species groups lacking an occipital carina (those of aber-rans, mandibularis, tepperi, cephalotes, and pilosula) arise as a strongly supported monophyletic assemblage. Monophyly was supportedfor four groups (those of gulosa, nigrocincta, picta, and mandibularis) but the situation is unclear for four others (those of urens, aberrans,tepperi, and pilosula). The aberrans group appears to be basal within the group lacking an occipital carina; a previous suggestion that it isthe sister group to the rest of the genus is thus not supported. 2005 Elsevier Inc. All rights reserved.

Keywords: Myrmecia ants; Nothomyrmecia ants; Molecular phylogeny; Nuclear DNA sequences; 28S rRNA; Long-wave opsin

1. Introduction

The ant subfamily Myrmeciinae comprises two extantgenera, Nothomyrmecia and Myrmecia (Bolton, 2003).Nothomyrmecia has a very restricted distribution in Austra-lia (Watts et al., 1998) and Myrmecia is widespread in Aus-tralia with just one of the 89 described species presentelsewhere, in New Caledonia (Clark, 1951; Ogata, 1991;Ogata and Taylor, 1991). Most species in both genera showa simple pattern of social life that is assumed to be theancestral one in the ants (Freeland, 1957; Haskins and Has-kins, 1950; Hölldobler and Taylor, 1983; Hölldobler andWilson, 1990; Sanetra and Crozier, 2000, 2001, 2002a,b;Wheeler, 1933). Molecular data conWrm that the two genera

* Corresponding author: Fax: +81 11 757 5595.E-mail address: ehase@res.agr.hokudai.ac.jp (E. Hasegawa).

1055-7903/$ - see front matter 2005 Elsevier Inc. All rights reserved.doi:10.1016/j.ympev.2005.09.021

are indeed closely related (Ohnishi et al., 2003; Ward andBrady, 2003).

Ogata (1991) recognized nine species groups within Myr-mecia, namely those of gulosa (42 spp.), nigrocincta (3 spp.),picta (2 spp.), urens (7 spp.), aberrans (5 spp.), mandibularis (7spp.), tepperi (5 spp.), cephalotes (3 spp.), and pilosula (15spp.). These groups can be further placed into two complexeson morphological grounds. The former four groups (Com-plex A) have a carina across the rear margin of the occiput,while the latter Wve groups (Complex B) do not. Ogata alsoprovided phylogenetic hypotheses relating the groups andcomplexes, but his arrangements were unstable and aVectedby the use of outgroup. When pseudomyrmecinae ants wereused as outgroups, Complex B appeared as the basal assem-blage and the aberrans group was the sister of the rest, butwhen Nothomyrmecia macrops was used Complex Aappeared basal and Complex B emerged from it. Ogataregarded the aberrans group as most plausibly reXecting the

576 E. Hasegawa, R.H. Crozier / Molecular Phylogenetics and Evolution 38 (2006) 575–582

ancestral state of this genus, because of various characteris-tics such as very small colony size (3–10 workers) and distinc-tive morphology and habitat preference (Ogata, 1991;Wheeler, 1933). Three independent molecular studies haveshown, however, that Nothomyrmecia is the sister group toMyrmecia with high conWdence (Ohnishi et al., 2003; Wardand Brady, 2003; Ward and Downie, 2005). Ward and Brady(2003) also strongly supported the Myrmecia species theyused as being monophyletic. Given the great importance ofMyrmecia in understanding the evolution of sociality in theants (see below), the phylogeny of its species groups andcomplexes, especially in relation to its close relative, Noth-omyrmecia, would be most worthwhile and this paper is thebeginning of such an endeavor.

Fossil Wnds show that the Myrmeciinae were once cos-mopolitan (see Ward and Brady, 2003), but the genus Myr-mecia is essentially restricted to Australia, suggesting that itmay have arisen and diversiWed there. While most Myrme-cia species have simple societies like those of monogynouspolisitine wasps, several species show derived social traitssuch as ergatoid queens (Clark, 1951), polymorphic work-ers (Gray, 1974), polygyny (Craig and Crozier, 1979),gamergates (Dietemann et al., 2004), and social parasitism(Brown and Douglas, 1959). All these derived states haveevolved within the monophyletic genus Myrmecia. The evo-lution of modiWcations is an important issue in social insectbiology (Bourke and Franks, 1995; Crozier and Pamilo,1996). The genus also presents features important to otherevolutionary concerns, such as the broadest range of chro-mosome numbers in Hymenoptera (Crosland and Crozier,1986; Hirai et al., 1996; Imai et al., 1994), extensive hybrid-ization (Crozier et al., 1995; Imai et al., 1994) and Müllerianmimicry complexes (Ogata, 1991). Therefore a robust phy-logeny of Myrmecia should contribute to the understand-ing not only of ant sociality but also of many otherbiological processes and it is our intent to provide such aresource.

In the present phylogenetic study, our main aims arethreefold: (1) to infer phylogenetic relationship among thenine species group, (2) to test the monophly of the speciesgroups, and (3) to test the hypothesis that the aberransgroup is the sister group to the rest of the genus. Weobtained DNA sequences from two nuclear genes, namely apart of the 28S rRNA gene (ca. 0.6 knt), and a large part ofthe long-wave opsin gene (1.96 knt, including introns). Eachgene, especially the latter, has regions varying widely insubstitution rates. Such variation enables phylogeneticinference over a broad taxonomic range, because rapidly orslowly changing sites should resolve shallow or deepbranches, respectively.

2. Materials and methods

2.1. Sample collection, DNA preparation, and sequencing

Thirty-Wve species of Myrmecia were collected (Table 1)at various location in Australia during 1999–2000. All the

nine species groups were included, and at least two speciesfrom every group included, other than that of cephalotes.All specimens were preserved into pure acetone (see Fuka-tsu, 1999) until DNA extraction.

Total DNA was extracted from the right middle leg ofeach specimen using a DNA extraction kit (DNeasy,Qiagen) according to the manufacturer’s instructions. Sev-eral pairs of primers (Table 2) were used to amplify the tar-get regions. We used a programmable thermocycler (PTC-100, MJ Research Inc.) for PCR ampliWcation. For the 28SrRNA gene ampliWcation involved 35 cycles of 1 min at94 °C, 1 min at 55 °C, and 3 min at 65 °C. The conditions forthe opsin gene was 32 cycles of 1 min at 94 °C, 1 min at42 °C, and 3 min at 65 °C. Each 100 �l reaction volume con-tained 10 �l of DNA sample solution (ca. 200 ng), 10�l of10£ Ex Taq buVer (Mg2+ plus), 10 �l of a solution contain-ing each dNTP (2.5 mM), 20 pmol of each primer of a pair,1.0 U of Taq DNA polymerase (TaKaRa Ex TaqTM;TaKaRa) and MilliQ water. AmpliWcation products werepuriWed using the QIamp PCR PuriWcation Kit(QIAGEN), and were suspended in 22 �l of TE buVer(0.1 mM EDTA, 10 mM Tris–HCl (pH 8.0)).

Automated sequencers (ABI PRISM 377, Applied Biosys-tems; CEQ2000XL, Beckman Coulter) were used tosequence the ampliWed fragments. We used direct sequencingkits (Dye Terminator Cycle Sequencing FS Ready ReactionKit, Applied Biosystems; DTCS Quick Start Kit, BeckmanCoulter) according to the manufacturer’s protocols. Theoptimal concentrations of template DNAs were checked bycomparing the brightness of ampliWed fragments with a sizemarker band (�/HindIII, TOYOBO) in 1.0% agarose gel elec-trophoresis with ethidium bromide staining.

2.2. Preparation of sequence data sets

Two alignments were made from Myrmecia and the out-group Nothomyrmecia macrops: the 28S rRNA gene(denoted the 28S Set), and the long-wave opsin gene (theOP Set). We selected N. macrops as the outgroup for thisstudy because of the accumulating strong evidence thatNothomyrmecia and Myrmecia form a monophyletic group(e.g., Ohnishi et al., 2003; Ward and Brady, 2003; Ward andDownie, 2005). Selection of a second outgroup was deemedproblematic, because no other ant subfamily is unequivo-cally close to these genera and we feared that selection of adistant second outgroup would cause problems, given thatour focus is the internal phylogeny of Myrmecia. ThePseudomyrmecinae may prove to be the closest subfamilyto Myrmecia + Nothomyrmecia, but are still distant fromthese genera (Ward and Downie, 2005). Thus, we do not usea second outgroup in this study.

We inferred the coding (exon) and non-coding (intron)regions of the opsin sequences from comparison withcDNA sequences from two ant species (Popp et al., 1996),using the fact that intron sequences do not occur in cDNAs.Alignment was made using CLUSTALX (Thompson et al.,1997) and checked by eye.

E. Hasegawa, R.H. Crozier / Molecular Phylogenetics and Evolution 38 (2006) 575–582 577

2.3. Character evaluation

Evolutionary rates were evaluated for recognizable par-titions. The OP Set was divided into four partitions, namelythe intron and the three codon positions, but we did notsubdivide the 28S set. For all partitions, we checked thedegree of saturation (multiple substitutions per site) forboth transitions and transversions. Such regions are wellknown to vary in evolutionary rate and in the pattern ofsubstitutions (Li, 1997), and saturation erodes phylogeneticsignal and may lead to erroneous Wndings. For each parti-tion, we plotted each substitution type against the Kimura2-parameter distance, K2D (Kimura, 1980), between eachpair of sequences. Saturation is indicated if there is adecline in substitution rate with time (using K2D as a sur-rogate for time), and thus information is obtained aboutthe decay of phylogenetic signal within the various regions(Misof et al., 2001). Heavily saturated partitions are oftenexcluded from analysis (e. g., Engstrom et al., 2004).

We used MODELTEST (Posada and Crandall, 1998) inconjunction with PAUP* (ver. 4.0b10; SwoVord, 2002) toinfer the most appropriate substitution model for each dataset using the Akaike Information Criterion (AIC).

2.4. Phylogenetic analysis

PAUP* 4.0b10 was used to infer phylogenies under threemethods, maximum likelihood (ML, see Section 3 for sub-stitution models used), neighbor-joining (NJ; Saitou andNei, 1987), using ML distances calculated using the samemodel as for the ML analyses, and maximum parsimony(MP). Branch-and-bound and heuristic searches were usedunder MP and ML, respectively. The robustness of eachclade was evaluated according to the bootstrap valuesresulting from 1000 replicates (Felsenstein, 1985) in MPand NJ, but because of computer time limitations we used100 bootstrap replicates for ML. For both MP and ML weused heuristic searches during in bootstrap analyses. Other

Table 1Species, used collection codes, and GenBank accession numbers

Subfamily Species Species group Sample No. Accession No.

Genus Species 28S Opsin

Myrmeciinae Nothomyrmecia macrops N512 AB208451 AB207135Myrmecia auriventris gulosa ok00l AB208449 —

simillima gulosa eh002 AB208459 —tarsata gulosa eh006 AB208460 AB207121brevinoda gulosa eh007 AB208461 AB207125pulchra gulosa eh021 — AB207123nigriscapa gulosa eh042 AB208462 —nigriscapa gulosa eh142 AB208465 AB207127ruWnodis gulosa eh049 — AB207124forceps gulosa eh053 AB208463 —forceps gulosa eh156 — AB207122esuriens gulosa eh060 AB208464 AB207126fulgida gulosa eh150 AB208467 —eungellensis gulosa eh173 — AB207117midas gulosa eh192 — AB207128comata gulosa eh208 AB208466 —fabricii gulosa eh212 — AB207119rowlandi gulosa eh222 — AB207120nigrocincta nigrocincta eh166 — AB207134nigrocincta nigrocincta eh214 AB208472 —petiolata nigrocincta eh215 AB208474 AB207133urens urens eh008 AB208470 AB207131loweryi urens eh158 AB208471 AB207132picta picta eh096 AB208468 AB207129fucosa picta eh073 AB208469 AB207130formosa aberrans eh163 AB208482 AB207116froggatti aberrans eh216 — AB207118fulviculis mandibularis ok002 AB208450 —fulviculis mandibularis eh209 AB208475 AB207115mandibularis mandibularis eh035 AB208474 AB207114sp. mandibularis eh211 AB208746 —callima cephalotes eh071 AB208481 AB207106testaceipes tepperi eh105 AB208479 AB207111swalei tepperi eh137 AB208480 AB207108croslandi pilosula HI9914 AB208478 AB207108chasei pilosula eh030 — AB207107michaelseni pilosula eh031 — AB207112harderi pilosula eh075 — AB207110chrysogaster pilosula eh206 AB208477 AB207113

578 E. Hasegawa, R.H. Crozier / Molecular Phylogenetics and Evolution 38 (2006) 575–582

than for the leading section of the OP set (see below) wetreated gaps as Wfth character state under MP.

3. Results and discussion

3.1. Gene structure and character evaluation

After alignment, we obtained 622 nt for the 28S, and1907 nt for the OP set. The 28S set has a short variableregion with indels. Comparisons with the known antopsin cDNAs show that the OP set includes the startcodon but not the stop codon. The region after the fourthexon contains many indels, indicating that the Myrmeciaopsin gene comprises at least four exons and Wve introns(Fig. 1). Although there may be further exons in the termi-nal region after the fourth exon and we treated it as theWfth intron. All new sequences are deposited in the DDBJ(for GenBank accessions numbers, see Table 1), and theresults of the alignment will be available from TREE-BASE.

We found no discernible saturation in any of the OPpartition (see Appendix), with the intron and third codonpositions showing the highest evolutionary rate, andtherefore weighted all changes equally as is usually donein such case (e.g., Yoshizawa and Johnson, 2003). Because

Table 2Primer sequences for the 28S rRNA and opsin genes

Degenerate nucleotides are indicated by the appropriate IUB code.

Gene Name Sequence (5�–3�)

For 28S rRNA28sf AAGGTAGCCAAATGCTCATC28sr AGTAGGGTAAAACTAACCTM28Sf TTAGTGACGCGCATGAATGGM28Sr GTCTAAACCCAGCTCTCGTT

For opsinOP03 ACAATAACAATTCAGCGAMOP01 AAGGTRAGARACATAGCTTCMOP21 TGTGGCACGCACTCCTTGGLwf AATTGCTATTACGAGACGTGGGTOPLwf CCCTGTTCGGATGTGGCTCCATOP11m GCCCCATTGTTYGGATGGAAMOP31 TGTTATTTCATGCCGYTGTTMOP12 ATAYTACCATGCCGTTACCGOP00m GTACCGCAGGCRGTCATGTTGCCMOP22 TGCTCGCGCATRTTCTTCTCRhr ATACGGAGTCCATGCCAWGAACCAOPLW2r CAAAGACAGAACCCCAGATAGTMOP32 CAKAATAATTTGATTAACTTOP16m AGAGCASCTCGRTACTTAGG

the 28S set presents very few variable sites (13), we did notsubject it to saturation analysis.

3.2. Separate analyses

Because the number of variable sites in the 28S set issmall (13, of which 4 are indels) we analyzed this set underMP alone, using indels as a Wfth character state. Few cladeswere resolved under this analysis (Fig. 2). The Complex Aspecies groups (with the occipital carina) did not form asingle clade but rather were paraphyletic, suggesting thatthis condition is ancestral. The picta groups form a com-pact monophyletic clade with high bootstrap support(95%), but monophyly of the urens group was not sup-ported. The Complex B species groups (lacking the occipi-tal carina) make a strong monophyletic clade (92%bootstrap support) there is no variable characters to resolvethese groups further. This result does not support the aber-rans group being the sister to the rest of the genus.

We were not able to amplify the leading 633 nt of theopsin gene for N. macrops, but many (110) parsimony-informative sites were found for the remaining species. Toenable use of the phylogenetic signal for resolving the inter-nal phylogeny of Myrmecia, we treated the leading 633 nt asmissing data for N. macrops in MP analysis. The MODEL-TEST analysis selected the TrN + I + G model and we usedthis in ML and NJ analyses. Fig. 3 gives the trees inferredusing the OP set; resolution is much higher than for the 28Sset. Each species group in the Complex A form a monophy-letic clade with high bootstrap support (100% for picta, 64–88% for urens, 87–94% for gulosa, and 100% for nigrocinctagroup). Relationships within the gulosa group, the largest inthe genus, were however relatively poorly resolved (47–70%), other than for a well-supported clade of M.pulchra + M. ruWnodis + M. brevinoda (100%). Monophylyof Complex B is strongly supported (72–94%), but within itresolution is poor and inconsistent with Ogata’s morpho-logical classiWcation into species groups, and only the man-dibularis group shows high bootstrap support (100%). Theaberrans group is paraphyletic and basal in the MP and NJresults (Figs. 3A and B), but is clearly part of Complex Band is not the sister group to the rest of the genus as awhole. The aberrans group is monophyletic under ML(Fig. 3C), but bootstrap support is not so high (68%).

Complex A is paraphyletic for the 28S set (for which weused only MP analysis) and for ML and MP analyses forthe OP set, but it is monophyletic with high support (93%)for the OP set under NJ anlysis. Loss of information in con-

Position of primers

0 500 1000 1500 2000

OP03MOP01

MOP21 MOP31OP11mLwf

MOP12 OP00m MOP22 Rhr OPLw2r MOP32 OP16m

463bp 333bp 100bp 243bp 82bp 264bp 110bp 168bp 140bp ca.

Non coding: Intron IStartExonI Int. II Exon II Int. III Exon III Int. IV Exon IV Int. V

End

Exon V

OPLwf

Fig. 1. Structure and primer position of the long-wave opsin gene of Myrmecia.

E. Hasegawa, R.H. Crozier / Molecular Phylogenetics and Evolution 38 (2006) 575–582 579

verting character information to a distance in the NJ analy-sis may be a cause of this diVerence.

3.3. Combined analysis

We concatenated the 28S and Op sets, but could onlyinclude 17 species in this analysis because we could notobtain one set or the other for several taxa (see Table 1).MODELTEST indicated the HYK + I + G model as theappropriate model and we used this for the ML analysisand for calculating the distances used for the NJ analysis.The results (Figs. 4A–C) resembled those of the separateanalyses. Again, MP and ML show paraphyly for ComplexA and NJ strongly supports monophyly (90%). The pictaand urens groups are returned as monophyletic by all threemethods with high bootstrap support (100%) and this cladeis found to be the sister to Complex B by both the MP (82%bootstrap support) and ML (63%) analyses. The MP andML results diVer in which group, nigrocincta or gulosa, isthe sister group to the rest of the genus. Although the NJmethod is fast and widely used as an initial step in analysesinvolving other methods, it is an approximate method andrelatively unreliable (Felsenstein, 2004), hence we incline tothe ML and MP results showing paraphyly of Complex Amore than to the NJ results showing monophyly of bothcomplexes. Further sequence is required to settle this ques-tion decisively and to resolve relationship within groups.

3.4. Evolutionary history of Myrmecia

The results have several implications. Firstly, they indi-cate monophyly of Complex B, the set of species lacking anoccipital carina.

Second, on balance we reach the provisional conclusionthat Complex A is paraphyletic, with Complex B arisingwithin it. The MP and ML results agree in also placingthe picta and urens groups as forming a sister clade to the

Complex B groups. Some morphological considerationssupport this arrangement. The allometries of head width vs.scape and of gaster width vs. postpetiole width broadlyoverlap for the nigrocincta and gulosa groups but thesediVer from those of the Complex B species (Ogata, 1991),while for these characteristics the picta and urens groupsresemble more Complex B species than the other ComplexA groups (see Figs. 29 and 49 of Ogata, 1991). Under thisscenario, the appearance of a picta + urens + Complex Bclade was associated with a change in allometry. ComplexB is united in lacking the occipital carina possessed by theother species of the genus; it is more likely that a morpho-logical feature would be lost than gained (Brown, 1965) aswould be required if Complex B were the basal one.

Third, contrary to Ogata’s (1991) morphological Wnd-ings, our results do not support the aberrans group beingsister to the rest of Myrmecia. The species of the aberransgroup present a number of morphological features diVerentfrom all other Myrmecia species such as a short robustbody, short scapes, and broad short mandibles, all of thewhich appeared to suggest ancestral feature (Ogata, 1991).Further, these ants have very small colony sizes and diVerin habitat preferences to their congeners (Wheeler, 1933).Ogata’s (1991) conclusion that the aberrans group is the sis-ter group to other Myrmecia depends critically on using apseudomyrmecine as an outgroup—when Nothomyrmeciais used instead, as we have shown is more appropriate, thenit is not. Hence, Ogata’s overall morphological Wndings areconsistent with the aberrans group falling well within thegenus and not being a sister to the other groups. The occur-rence of ergatoid queens in aberrans group species(Wheeler, 1932) shows that these are certainly not ancestralin every characteristic but derived for some.

Finally, despite its length the OP set yielded only poorresolution within the gulosa group and within Complex B.We investigated the situation further using opsin sequencesfor only the gulosa group or only Complex B species, but

Fig. 2. Inferred phylogeny within Myrmeciinae by using the 28S rDNA. Numbers on nodes represent bootstrap probabilities with 1000 resamplings.

Complex B

Complex A

eh071 callimaHI9914 croslandieh206 chrysogastereh035 mandibulariseh209 fulviculiseh211 sp8eh137 swaleieh105 testaceipeseh163 formosa

eh002 simillimaeh006 tarsataeh007 brevinodaeh042 nigriscapaeh053 forcepseh060 esurienseh142 nigriscapaeh208 comata

eh150 fulgidaeh214 nigrocinctaeh215 petiolata

eh096 pictaeh073 fucosa

eh008 urenseh158 loweryi

N512 N. macrops

0.5 changes

92

49

54

9547

69

cephalotes

Species group

pilosula

mandibularis

tepperi

aberrans

gulosa

nigrocincta

picta

urens

580 E. Hasegawa, R. H. Crozier I Molecular Phylogenetics and Evolution 38 (2006) 575-582

Species group

a. MP tree eh071 c a l l m a ~ h a l o t e s el1030 chawi

'~1,177 wain Hi9914 croiWi tepperi eh(f75 harden Complex B

IÑ i n n - eh096'mta - .

eh222 rowlandi ehWi6 tmala 8\11 56 forceps

100 49

eh049 rufinodis ehMf7 hrevznnda eh060 rsurtens

Complex A

b. NJ tree

Complex A

Species group

c. ML tree

Complex B

Complex A

Fig. 3. Inferred phylogeny within Myrmeciinae by using the opsin gene. Numbers on nodes represent bootstrap probabilities with 1000 resamplings (num- ber of resamplings was 100 in ML).

E. Hasegawa, R. H. Crozier 1 Molecular Phylogenetics and Evolution 38 (2006) 575-582

Species group em71 caUha - cephalotes

p

inpbcmohetsiEah

oor resolution. Thus, although the subfamily is thought to e old, with a putative fossil in the Aptian of the Creta- eous (Brand50 et al., 1989; Schultz, 2000), Myrmecia itself ay be relatively young or have had internal replacement

f groups leading to these being young. There is evidence of ybridization between some Myrmecia species (Crozier t al., 1995; Imai et al., 1994), which might also compromise he resolving power of gene sequences. Additional equences using mitochondria1 genes, or from nuclear genes ncreasingly used in insect molecular phylogeny such as F-la (Danforth et al., 2004), wingless (Brower and DeS- lle, 1998), and 18s rRNA (Ward and Brady, 2003), may elp resolve the remaining uncertainties.

epsaSR

A

fj

We thank R.W. Taylor, H.T. Imai, M. Kubota, M. San- tra, C. Peeters, IS. Masuko, F. Ito, and K. Ogata for roviding samples and useful information for sampling ites. We also thank K. Yoshizawa for discussion for data nalysis. This work was supported by grants from JMEC- ST to E.H. and the Australian Research Council to .H.C.

ppendix A. Supplementary data

Supplementary data associated with this article can be ound, in the online version, at doi: 10.10 161 .ympev.2005.09.02 1.

a. Strict consensus tree of the 4 MP trees HI9914 croslandi ehl05 testaceipes em35 mandiliuiaris eh2W fuhiculis mandibularis

82 eh206 chrysogaster

6s ehl63formosa - aberrans 100 em9 6picta

7 fucosa- picta 74

l o o e m wens 66 ehl58 1oweiyi+ urens

em07 brevinoda em60 esuriens

97

Complex A eM06 tarsata gul0sa ehl42 nigriscapa eh215 petiolata - nigrocincta N512 N. macrops

b. NJ tree em71 callha - cephalotes

eh206 chrysogaster ehl63 formosa aberrans

eh209fulviculis - mandibularis eM9 6picta

I em73 f u c o s a J Ã ‘ Ã ‘ Ã ‘ picta 1 eM08 urens

ehl58 loweryi urens 1

c. ML tree

pilosula eh209 fulviculis eh206 chryswaster ehl63 formosa aberrans -

em08 urens ehl58 loweryi urens

eh215 petiolata nigrocincta 7 - eMO7 brevinoda I

eM60 esuriens ehO0 6 tarsata

Complex B

Complex A

0.01 substitutions/sit~

Fig. 4. Inferred phylogeny within Myrmeciinae by using the combined data (opsin + 28s). Numbers on nodes represent bootstrap values with 1000 resam- lings (number of resamplings was 100 in ML).

mprovement in bootstrap value was not achieved (analyses ot shown). There are various possible reasons for this

Acknowledgments

582 E. Hasegawa, R.H. Crozier / Molecular Phylogenetics and Evolution 38 (2006) 575–582

References

Bolton, B., 2003. Synopsis and classiWcation of Formicidae. Mem. Am.Entomol. Inst. 71, 1–370.

Bourke, A.F.G., Franks, N.R., 1995. Social Evolution in Ants. PrincetonUniversity Press, Princeton, New Jersey.

Brandão, C.R.F., Martins-Neto, R.G., Vulcano, M.A., 1989. The earliestknown fossil ant (Wrst southern hemisphere Mesozoic record) (Hyme-noptera: Formicidae: Myrmeciinae). Psyche 96, 195–208.

Brower, A.V.Z., DeSalle, R., 1998. Patterns of mitochondrial versusnuclear DNA sequence divergence among nymphalid butterXies: theutility of wingless as a source of characters for phylogenetic inference.Insect Mol. Biol. 7, 73–82.

Brown Jr., W.L., 1965. Numerical taxonomy, convergence and evolution-ary reduction. Syst. Zool. 14, 101–109.

Brown Jr., W.L., Douglas, A., 1959. Myrmecia inquilina new species: theWrst parasite among the lower ants. Ins. Soc. 6, 13–19.

Clark, J., 1951. The Formicidae of Australia. Vol. 1. Subfamily Myrmecii-nae. CSIRO, Melbourne.

Craig, R., Crozier, R.H., 1979. Relatedness in the polygynous ant Myrme-cia pilosula. Evolution 33, 335–341.

Crosland, M.W.J., Crozier, R.H., 1986. Myrmecia pilosula: an ant with onepair of chromosomes. Science 231, 1278.

Crozier, R.H., Dobric, N., Imai, H.T., Graur, D., Cornuet, J.-M., Taylor,R.W., 1995. Mitochondrial-DNA sequence evidence on phylogeny ofAustralian Jack-jumper ants of the Myrmecia pilopsula complex. Mol.Phylogenet. Evol. 4, 20–30.

Crozier, R.H., Pamilo, P., 1996. Evolution of Social Insect Colonies. SexAllocation and Kin-selection. Oxford University Press, Oxford, UK.

Danforth, B.N., Brady, S.G., Sipes, S.D., Pearson, A., 2004. Single-copynuclear genes recover Cretaceous-age divergences in bees. Syst. Biol.53, 309–326.

Dietemann, V., Peeters, C., Hölldobler, B., 2004. Gamergates in the Aus-tralian ant subfamily Myrmeciinae. Naturewissenschaften 91, 432–435.

Engstrom, T.N., ShaVer, H.B., McCord, W.P., 2004. Multiple data sets,high homoplasy and the phylogeny of softshell turtles (Testudines:Tri-onychidae). Syst. Biol. 53, 693–710.

Felsenstein, J., 1985. ConWdence limits on phylogenies: an approach usingthe bootstrap. Evolution 39, 783–791.

Felsenstein, J., 2004. Inferring Phylogenies. Sinauer, Sunderland, MA.Freeland, J., 1957. Biological and social patterns in the Australian bulldog

ants of the genus Myrmecia. Aust. J. Zool. 6, 1–21.Fukatsu, T., 1999. Acetone preservation: a practical technique for molecu-

lar analysis. Mol. Ecol. 8, 1935–1945.Gray, B., 1974. A morphometric study of worker variation in three Myr-

mecia species (Hymenoptera: Formicidae). Ins. Soc. 20, 323–331.Haskins, C.P., Haskins, E.F., 1950. Notes on the biology and social behav-

ior of the archaic ponerine ants of the genera Myrmecia and Prom-yrmecia. Ann. Entomol. Soc. Am. 43, 461–491.

Hirai, H., Yamamoto, M., Taylor, R.W., Imai, H.T., 1996. Genomic disper-sion of 28S rDNA during karyotypic evolution in the ant genus Myr-mecia Formicidae. Chromosoma 105, 190–196.

Hölldobler, B., Taylor, R.W., 1983. A behavioral study of the primitive antNothomyrmecia macrops Clark. Ins. Soc. 30, 384–401.

Hölldobler, B., Wilson, E.O., 1990. The ants. Harvard Univ. Press, Cam-bridge, MA.

Imai, H.T., Taylor, R.W., Crozier, R.H., 1994. Experimental bases of theminimum interaction hypothesis: 1. Chromosome evolution in ants ofthe Myrmecia pilosula species complex (Hymenoptera: Formicidae:Myrmeciinae). Japan J. Genet. 69, 137–182.

Kimura, M., 1980. A simple method for estimating evolutionary rate ofbase substitution through comparative studies of nucleotide sequences.J. Mol. Evol. 16, 111–120.

Li, W.-H., 1997. Molecular Evolution. Sinauer Associates, Sunderland,MA.

Misof, B., Rickert, A.M., Buckley, T.R., Fleck, G., Sauner, K.P., 2001. Phy-logenetic signal and its decay in mitochondrial SSU and LSU rRNAgene fragments of Anisoptera. Mol. Biol. Evol. 18, 27–37.

Ogata, K., 1991. Ants of the genus Myrmecia Fabricius: a review of thespecies groups and their phylogenetic relationship (Hymenoptera:Formicidae: Myrmeciinae). Syst. Entmol. 16, 353–381.

Ogata, K., Taylor, R.W., 1991. Ants of the genus Myrmecia Fabricius: apreliminary review and key to the named species (Hymenoptera:Formicidae: Myrmeciinae). J. Nat. Hist. 25, 1623–1673.

Ohnishi, H., Imai, H.T., Yamamoto, M., 2003. Molecular phylogeneticanalysis of ant subfamily relationship inferred from rDNA sequences.Genes Genet. Syst. 78, 419–425.

Popp, M.P., Grisshammer, R., Hargrave, P.A., Smith, W.C., 1996. Antopsins: sequences from the Saharan silver ant and the carpenter ant.Invertebr. Neurosci. 1, 323–329.

Posada, D., Crandall, K.A., 1998. MODELTEST: testing the model ofDNA substitution. Bioinformatics 14, 817–818.

Saitou, N., Nei, M., 1987. The neighbor-joining method: a new method forreconstructing phylogenies. Mol. Biol. Evol. 4, 406–425.

Sanetra, M., Crozier, R.H., 2000. Characterization of microsatellite loci in theprimitive ant Nothomyrmecia macrops Clark. Mol. Ecol. 9, 2169–2170.

Sanetra, M., Crozier, R.H., 2001. Polyandry and colony genetic structurein the primitive ant Nothomyrmecia macrops. J. Evol. Biol. 14, 368–378.

Sanetra, M., Crozier, R.H., 2002a. Daughters inherit colonies from moth-ers in the ‘living fossil’ ant Nothomyrmecia macrops. Naturewissens-chaften 89, 71–74.

Sanetra, M., Crozier, R.H., 2002b. Population genetic structure in theprimitive ant Nothomyrmecia macrops using two diVerent types ofmarkers. In: XIV Intern. Congr. IUSSI. The Golden Jubilee Proceed-ings. Hokkaido University Coop, Hokkaido University, Sapporo,Japan, p. 129.

Schultz, T.R., 2000. In search of ant ancestors. Proc. Natl. Acad. Sci. USA97, 14028–14029.

SwoVord, D.L., 2002. PAUP*. Phylogenetic analysis using parsimony (*and other methods). Sinauer, Sunderland, MA.

Thompson, J.D., Gibson, T.J., Plewniak, F., Jeanmougin, F., Hoggins, D.J.,1997. The CLUSTAL-X windows interface: Xexible strategies for mul-tiple sequence alignment aided by quality analysis tools. Nucleic AcidsRes. 25, 4876–4882.

Ward, P.S., Brady, S.G., 2003. Phylogeny and biogeography of the ant sub-family Myrmeciinae (Hymenoptera: Formicidae). Invertebr. Syst. 17,361–386.

Ward, P.S., Downie, D.A., 2005. The ant subfamily Pseudomyrmecinae(Hymenoptera: Formicidae): phylogeny and evolution of big-eyedarboreal ants. Syst. Entomol. 30, 310–335.

Watts, C.H.S., MacArthur, A.J., Foster, R., 1998. Notes on the distributionof the dinosaur ant Nothomyrmecia macrops Clark (Hymenoptera:Formicidae) in South Australia. Aust. Entomol. 25, 29–31.

Wheeler, W.M., 1932. How the primitive ants of Australia start their colo-nies. Science 76, 532–533.

Wheeler, W.M., 1933. Colony-founding among ants with an account of someprimitive Australian species. Harvard Univ. Press, Cambridge, MA.

Yoshizawa, K., Johnson, K.P., 2003. Phylogenetic position of Phthiraptera(Insecta: Paraneoptera) and evaluated rate of evolution in mitochon-drial 12S and 16S rDNA. Mol. Phylogenet. Evol. 29, 102–114.