EVOLUTION AND PHYLOGENY OF THE DIPTERA: A … · syst. biol. 46(4):674-698, 1997 evolution and...

Syst. Biol. 46(4):674-698, 1997

EVOLUTION AND PHYLOGENY OF THE DIPTERA: A MOLECULARPHYLOGENETIC ANALYSIS USING 28S RDNA SEQUENCES

MARKUS FRIEDRICH1 AND DIETHARD TAUTZ2

California Institute of Technology, Division of Biology 156-29, Pasadena, California 91125, USA;E-mail: [email protected]

2Zoologisches Institut der Universitat Miinchen, Luisenstrasse 14, 80333 Miinchen, Germany;E-mail: [email protected]

Abstract.—Portions of the large ribosomal subunit RNA gene (28S rDNA) encompassing the Dland the D7 region were obtained from 16 dipteran species and families to reconstruct earlyphylogenetic events in the order Diptera. For outgroup comparison, the corresponding sequenceswere used from representative taxa of the Siphonaptera, Mecoptera, and Lepidoptera. A subsetof 488 unambiguously alignable sites was analyzed with respect to important sequence evolutionparameters. We found (1) sequence variability is significantly higher in double-stranded sites thanin single-stranded sites, (2) transitions are close to saturation in most pairwise sequence compar-isons, (3) significant substitution rate heterogeneity exists across sites, and (4) significant substi-tution rate heterogeneity exists among lineages. Tree reconstruction was carried out with theneighbor joining, maximum parsimony, and maximum likelihood methods. Four major subgroupsare consistently and robustly supported: the Brachycera, the Culicomorpha, the Tipulomorphasensu stricto, and the hitherto controversial Bibionomorpha sensu lato, which includes the familiesSciaridae, Mycetophilidae, Cecidomyiidae, Bibionidae, Scatopsidae, and Anisopodidae. The phy-logenetic relationships within or among these subclades and the positions of the families Psy-chodidae and Trichoceridae were not robustly resolved. These results support the view that themouthparts of extant dipteran larvae evolved from a derived ground state characterized by sub-divided and obliquely moving mandibles. Furthermore, sequence divergence and the paleonto-logical record consistently indicate that a period of rapid cladogenesis gave rise to the majordipteran subgroups. [Character state polarity; Diptera; Drosophila; evolution; maximum likelihoodratio test; molecular phylogeny; ribosomal DNA; systematics.]

Because of their abundance and the dressed with molecular methods. Thischaracteristic modification of the hind scarcity of molecular studies addressingwings to halteres, representatives of the the higher systematics of the Diptera isholometabolan insect order Diptera are fa- contrasted by a rich body of morphologicalmiliar to all of us. The presumably best studies. The first accounts of dipteran phy-known dipteran is the fruitfly Drosophila logeny appeared more than 150 years agomelanogaster, which has become one of the (Lameere, 1906). A large number of studiesmost intensively studied model organisms followed, which derived phylogenetic evi-in modern biology (Ashburner, 1989; Law- dence from cytology, paleontology, or mor-rence, 1992). Evolution of the genus Dm- phology (White, 1949; Rohdendorf, 1964;sophila has been the subject of several mor- Hackman and Vaisanen, 1985). Most influ-phological and molecular phylogenetic ential were the numerous contributions bystudies (DeSalle and Grimaldi, 1991; Pe- Hennig (1948, 1968, 1981), which profitedlandakis and Solignac, 1993). In recent from the application of the principles ofyears, various dipteran clades other than phylogenetic systematics (Hennig, 1965).Drosophila have been studied using molec- Hennig's final conclusions, which wereular phylogenetic techniques (Vossbrinck largely drawn from the analysis of adultand Friedman, 1989; Xiong and Kocher, characters, predominated dipteran system-1991; Raich et al, 1993; Miller et al., 1996; atics for a long time. The first major at-Pawlowski et al., 1996; Tang et al., 1996; tempt to revise dipteran phylogeny wasMcPheron and Han, 1997; Smith and Bush, put forward by Wood and Borkent (1989),1997). Nonetheless, the basal relationships in which they considered both adult andwithin the Diptera have not been ad- larval characters. Their study provided the

674

1997 FRIEDRICH AND TAUTZ—DIPTERAN PHYLOGENY 675

background for a series of publications onmore specific aspects of dipteran phylog-eny (Courtney, 1990; Griffiths, 1990, 1994;Oosterbroek and Theowald, 1991; Krze-minski, 1992b; Sinclair, 1992; Sinclair et al.,1994). Most recently, Oosterbroek andCourtney (1995) undertook a new majoreffort towards dipteran phylogeny by re-vising and extending the morphologicalcharacters used by Wood and Borkent(1989), Courtney (1991), Krzeminski(1992b), and Sinclair (1992) and by apply-ing computer analysis in tree reconstruc-tion.

Nonetheless, many important issues ofhigher dipteran systematics have remainedcontroversial. There also is an ongoing in-terest in comparative and evolutionary as-pects of this group (Shaw and Meinertzha-gen, 1986; Schmidt-Ott et al., 1994; Curtiset al., 1995; Sander, 1996; Melzer et al.,1997). Because of this situation, we inves-tigated the potential of partial sequencesfrom the nuclear large ribosomal RNA(rRNA) gene (28S rDNA) to resolve theearly phylogenetic events in dipteran evo-lution.

Advances and Conflicts in HigherSystematics of the Diptera

The basic concepts of dipteran system-atics date back to the early 19th century.At that time, the Diptera were already sub-divided into Nematocera, flies with longand evenly segmented antennae, and Bra-chycera, flies with short antennae built ofmodified flagellomeres (Latreille, 1802;Macquart, 1834). The >85,000 extant bra-chycerous species, subdivided into 110families, stand out against the remaining35,000 species, thought to represent 27families (Schumann, 1992) traditionallyunited as Nematocera. This taxon is gen-erally assumed to be paraphyletic with re-spect to the Brachycera (Hennig, 1968,1981;Wood and Borkent, 1989; Oosterbroek andCourtney, 1995). Nonetheless, argumentshave been made that support a monophy-letic origin of the Nematocera (Ulrich,1991). The Brachycera is a well-supportedmonophyletic group, which is subdividedinto Cyclorrhapha and Orthorrhapha

(Woodley, 1989). The orthorrhaphous fliesare thought to represent a paraphyletic as-sembly with respect to the Cyclorrhapha,which form a conspicuous monophyleticsubgroup supported by numerous syna-pomorphies (McAlpine, 1989).

The relationships between the majordipteran subgroups are a matter of ongo-ing controversy. Part of the problem is theuncertainty about the sister group of theDiptera. Candidate groups are the Mecop-tera (scorpion flies) (Hennig, 1969; Micko-leit, 1969), the Siphonaptera (fleas) (Boud-reaux, 1979; Wood and Borkent, 1989), asubclade including the Mecoptera and theSiphonaptera (Kristensen, 1975), or themecopteran family Nannochoristidae (Till-yard, 1929; Imms, 1944). Unfortunately,molecular studies based on rDNA se-quences, which support either the Lepi-doptera (Pashley et al., 1993; Friedrich andTautz, 1997) or the Strepsiptera (Chalwatz-is et al., 1996; Whiting et al., 1997), havefailed to convincingly clarify this issue.These studies suffer from the presence ofsignificant substitution rate differencesand base compositional shifts in strepsip-teran and dipteran rRNA genes, which se-verely reduce the success of molecularphylogenetic tree estimation (Carmeanand Crespi, 1995; Friedrich and Tautz,1997; Huelsenbeck, 1997). A further reasonfor controversy over higher dipteran sys-tematics is the general lack of phylogenet-ically informative character states, al-though a large number of character statecomplexes such as general morphology,chaetotaxy, cytology, ecology, and paleon-tology have been explored (see Hackmanand Vaisanen, 1982; Wood and Borkent,1989; Oosterbroek and Courtney, 1995).

To outline the major controversies indipteran phylogeny, we focus on the con-cepts of Hennig (1969, 1973, 1981), Woodand Borkent (1989), and Oosterbroek andCourtney (1995). Furthermore, the range ofgroups discussed is restricted to those in-cluded in the present molecular analysis(Table 1). Setting aside terminological dif-ferences, all of these studies converge onfive major dipteran subgroups (Fig. 1):the Brachycera, the Tipulomorpha or

676 SYSTEMATIC BIOLOGY VOL. 46

TABLE 1. Dipteran and outgroup species includedthis study are given in bold.

Spedes

Tipula paludosa MeigenLimonia nebulosa MeigenBradysia coprophila LintnerBolitophila cinerea MayerDilophus febrilis L.Clinodiplosis cilicrus (Kieffer)Chironomus tentans F.Simulium euryadminiculum DaviesAedes albopictus Mit.Culex pipiens L.Psychoda cinerea BanksTrichocera regelationis L.Anapausis inermis RuthSylvicola fenestralis Scop.Drosophila tnelanogaster MeigenTabanus sudeticus Zell.Manduca sexta L.Archaeopsylla erinacei BoucheePanorpa communis L.

Family

TipulidaeLimoniidaeSciaridaeMycetophilidaeBibionidaeCecidomyiidaeChironomidaeSimuliidaeCulicidaeCulicidaePsychodidaeTrichoceridaeScatopsidaeAnisopodidaeDrosophilidaeTabanidae

in this study. New

Subgroup

TipulomorphaTipulomorphaBibionomorphaBibionomorphaBibionomorphaBibionomorphaCulicomorphaCulicomorphaCulicomorphaCulicomorphaPsychodomorpha

BrachyceraBrachycera

accession numbers associated with

Order

DipteraDipteraDipteraDipteraDipteraDipteraDipteraDipteraDipteraDipteraDipteraDipteraDipteraDipteraDipteraDipteraLepidopteraSiphonapteraMecoptera

EMBL accession no.

X93387, X93405X93369, X93378X93391, X93402X93365, X93373X93366, X93375X93363, X93372X93383, X93412X93368, X93377L22060X93384, X93403X93386, X93404X93370, X93379X93364, X93374X93367, X93376M21017X93362, X93371X93382, X93408X93381, X93407X93380, X93406

crane-fly-like flies, the Psychodomorpha ormoth-midge-like flies, the Bibionomorphaor fungus-gnat-like flies, and the Culico-morpha or mosquito-like flies. The com-position of some of these clades, however,differs considerably among dipterists. Thebasic incongruency concerns the Psycho-domorpha and the Bibionomorpha. TheBibionomorpha sensu stricto, which tradi-tionally includes the families Bibionidae(March-flies), Sciaridae (dark fungusgnats), Cecidomyiidae (gall midges), andMycetophilidae (fungus gnats), is general-ly accepted (Fig. 1). Hennig (1981) also in-cluded the families Anisopodidae (woodgnats) and Scatopsidae within the Bibion-omorpha, which one may then refer to asBibionomorpha sensu lato (Fig. la). Woodand Borkent, however, included the Ani-sopodidae and the Scatopsidae togetherwith the Trichoceridae (winter gnats) andthe Psychodidae (moth-midges) in the Psy-chodomorpha, referred to hereinafter asPsychodomorpha sensu W&B (Fig. lb). Al-though the Psychodomorpha in the systemof Hennig (1981) includes the Psychodidaeand a number of other families not includ-ed in this study (Fig. la), Oosterbroek andCourtney extended the Psychodomorphasensu W&B by including the families Ti-

pulidae and Limoniidae and the Brachyera(Fig. lc). Traditionally, however, the fami-lies Tipulidae, Limoniidae, and Trichoceri-dae are considered basal families of theDiptera. According to Hennig (1981), forexample, these three families represent themost basal subgroup, here termed Tipulo-morpha sensu lato (Fig. la). Because themonophyly of the Tipulomorpha sensulato was rejected by Wood and Borkent infavor of including the Trichoceridae in thePsychodomorpha sensu W&B (Fig. lb), thegeneral consensus reduces to consider theTipulidae and Limoniidae sister groups(Fig. 1). The corresponding clade is termedTipulomorpha sensu stricto. A further ma-jor subgroup, for which all authors agreeon component families, is the Culicomor-pha (Fig. 1). Regarding the culicomor-phous families Culicidae (mosquitos), Chi-ronomidae (water midges), and Simuliidae(black flies), most authors assume the lattertwo to be more closely related.

Given the fundamental differences in thecomposition the Bibionomorpha and thePsychodomorpha, a comparison concern-ing the relationships among the major dip-teran subdivisions is hardly possible.Nonetheless, Hennig (1981) and Wood andBorkent (1989) favored the Tipulomorpha


(a) TipulidaoLimoniidaeTrichoceridae

Tipulomorpha sensu stricto

Tipulomorpha sensu latoPsychodidae Psychodomorpha

- r £ChironomidaeSimuliidaeCulicidae

Culicomorpha

ScatopsidaeAnisopodidae Bibionomorpha sensu lato

(c)

TipulidaeLimoniidae

Tipulomorpha sensu stricto

BibionidaeMycetophilidaeCecidomyiidaeSciaridae

Bibionomorpha sensu stricto

ChironomidaeSimuliidaeCulicidae

Culicomorpha

PsychodidaeTrichoceridaeScatopsidaeAnisopodidae

Psychodomorpha W&B

DrosophilidaeTabanidae

Brachycera

ChironomidaeSimuliidaeCulicidae

Culicomorpha

BibionidaeMycetophilidaeCecidomyiidaeSciaridae

Bibionomorpha sensu stricto

ScatopsidaePsychodidaeTrichoceridas

LimoniidaaAnisopodidae

Psychodomorpha O&C

Tipulomorpha sensu lato

FIGURE 1. Phylogeny of the major dipteran subgroups and families covered in this study. Controversialsubgroups are printed bold. Major subgroups are indicated by shaded boxes. Subgroups nested within higherclades are indicated by darker boxes, (a) Hennig, 1973, 1981. (b) Wood and Borkent, 1989. (c) Oosterbroek andCourtney, 1995.


sensu lato or the Tipulomorpha sensustricto, respectively, as the most basal ma-jor dipteran subgroup, and Oosterbroekand Courtney (1995) found evidence forthe Culicomorpha as most basal.

The phylogenetic status of the Brachy-cera, which is of most general interest, issimilarly controversial. Hennig (1981) sug-gested a sister-group relationship betweenBibionomorpha sensu lato and the Brachy-cera (Fig. la). Wood and Borkent (1989) fa-vored the Psychodomorpha sensu W&B asthe sister group of the Brachycera (Fig. lb).Oosterbroek and Courtney (1995) placedthe Brachycera as sister group of the familyAnisopodidae, a grouping that had beenproposed previously by Woodley (1989).

MATERIALS AND METHODS

Samples

Each major subgroup was representedby at least one species (Table 1). Represen-tatives of families with uncontroversialsystematic position in terms of subgroupallocation are Tipula paludosa (Tipulidae)and Limonia nebulosa (Limoniidae) from theTipulomorpha sensu stricto, Bradysia copro-phila (Sciaridae), Bolitophila cinerea (Myce-tophilidae), Dilophus febrilis (Bibionidae),and Clinodiplosis cilicrus (Cecidomyiidae)from the Bibionomorpha sensu stricto, Chi-ronomus tentans (Chironomidae), Simuliumeuryadminiculum (Simuliidae), Aedes albopic-tus (Culicidae), and Culex pipiens (Culici-dae) from the Culicomorpha, and Psychodacinerea (Psychodidae) from the Psychodo-morpha. The systematic positions of thefamilies represented by the species Tricho-cera regelationis (Trichoceridae), Anapausisinermis (Scatopsidae), and Sylvicola fenes-tralis (Anisopodidae) are currently contro-versial (Fig. 1). To cover the evolutionarydivergence of the Brachycera, the pub-lished sequences of Drosophila melanogaster(Drosophilidae, Cyclorrhapha) (Tautz etal, 1988) were retrieved from the EMBLdatabase, and a representative (Tabanus su-deticus) of one of the basal families, the Ta-banidae (Woodley, 1989), was chosen forsequence analysis. Representatives of theinsect orders Lepidoptera (Manduca sexta),

Mecoptera (Panorpa communis), and Si-phonaptera (Archaeopsylla erinacei) were in-cluded for outgroup comparison.

MethodsExtraction of genomic DNA from dried

or alcohol-preserved specimens was car-ried out according to Gustincich et al.(1991) with the following modifications.DNA was not precipitated but was spin-dialyzed in Ultra free tubes (Millipore),washed three times with 300 |xl H2Q andthen recovered in 25-100 JULI Tris-EDTA(TE) buffer. This procedure increased theDNA yield and reduced inhibition in sub-sequent PCR reactions when dried or al-cohol-preserved insect specimens frommuseum collections were used. Concentra-tion and quality of the extracted genomicDNA was analyzed on 0.8% agarose gels.Dried insect specimens generally yieldedlow-molecular-weight DNA fragments of<600 bp. Alcohol-preserved insect speci-mens usually yielded high-molecular-weight DNA. Aliquots of the DNA extrac-tions were subjected to PCR reactions toamplify DNA fragments corresponding tothe following regions of the Drosophila me-lanogaster 28S rRNA sequence (Tautz et al.,1988): Dl fragment, 3338-3650 (primers: 5'-CCC(C / G)CGTAA(T / QTTAAGCATAT-3',5'-ACTCTCTATTCA(A / G)AGTTCTTT(G /C)-3', annealing temperature 60°C); D7fragment, 5000-5464 (primer: 5'-CTGAAGTGGAGAAGGGT-3', 5'-GACTTCCCTTACCTACAT-3', annealing temperature60°C). PCR reactions were set up with ap-proximately 10-100 ng DNA, 0.2 mMdNTPs, 2 \sM. of each primer, and 2 mg/mlbovine serum albumin in a buffer of 67mM Tris/HCl (pH 8.8), 2 mM MgCl2 (Paa-bo, 1990). Cycling was performed in aThermo Cycler (Perkin Elmer) with 5 minat 94°C for initial denaturation, followed bymanual addition of Taq DNA polymerase(Cetus) on the block, then 20 cycles with 1min denaturation at 93°C, 1 min annealingat the temperature specified for each prim-er pair above, and 1 min elongation at72°C. This reaction sequence was followedby 20 analogous cycles with a 2-min 72°Celongation step. Amplified products were


examined on 1.5% agarose gels. DNA frag-ments were excised from the agarose gel,eluted using glassmilk (Quiaex Kit, Qui-agen), and redissolved in 25 |xl TE buffer.About 100 ng of template DNA was usedper direct sequencing reaction, which wascarried out according to Casanova et al.(1990). DNA fragments were sequenced inboth directions with the amplificationprimers and the following internal primersfor the D7 fragment: 5'-AGGGTTTCGTGTGAACAG-3', 5'-TTCCAAACC(A/C)TATCTC-3', 5'-CGATTTTCAAGGTCC-3'. Asjudged by gel electrophoresis, the D7a ex-pansion segments (Hancock et al., 1988) ofTabanus sudeticus and Panorpa communis areabout 300 bp longer than those in all otherspecies investigated. These expansionregions were not sequenced completely be-cause no phylogenetically valuable infor-mation for the problem addressed herewas expected.

Initial multiple alignments were pro-duced by hand and, alternatively, usingthe CLUSTAL V program applying defaultsettings (Higgins and Sharp, 1988). The re-sulting alignments were compared and di-vergent regions were rechecked, takingconservation of secondary structure mod-els previously published for Drosophila me-lanogaster into account (Hancock et al.,1988; Rousset et al., 1991). Nucleotides162-188 in the Drosophila melanogaster 28SDl sequence provided here (Appendix 1)were accidentally omitted from the corre-spondent sequence and the secondarystructure published by Hancock et al.(1988) and (Tautz et al., 1988) but were ver-ified by direct sequencing (Friedrich, 1995).

Pairwise sequence divergence and pair-wise transition (Ts) and transversion (Tv)sequence divergence were determined us-ing subroutines of PAUP 3.0 (Swofford,1993). The numbers for variable and phy-logenetically informative sites were deter-mined using MacClade (Maddison andMaddison, 1992). Stationarity of nucleotidecomposition was tested using the STATIOprogram (Rzhetsky and Nei, 1995). Maxi-mum likelihood (ML) estimation of treelikelihoods and substitution parameterswere carried out using PAML (Yang,

1996b). ML tree search was carried out us-ing PHYLIP (Felsenstein, 1995) applyingthe Global Rearrangement option. Heuris-tic maximum parsimony (MP) tree searchwas carried out in PAUP applying the TBRoption for branch swapping. Estimation ofevolutionary distances and neighbor join-ing (NJ) analysis (Saitou and Nei, 1987)were carried out in PHYLIP. Branch prob-ability (BP) values were determined bynonparametric bootstrapping (Felsenstein,1985) based on 1,000 replicate data sets.

RESULTS

Multiple Alignment

Correct alignment of homologous posi-tions between sequences is the first prereq-uisite for reliable molecular phylogeneticreconstruction. In the present study, the28S Dl and D7 sequence regions had to bealigned among 16 distantly related dipter-an species and 3 outgroup species repre-senting lepidopterans, siphonapterans,and mecopterans. 28S rDNA expansionsegment regions characteristically accu-mulate insertion and deletion events overtime. These length-variable regions, whichwere dispersed among several stronglyconserved blocks in the present set of se-quences, were difficult to align andshowed little congruence between hand-and computer-generated alignment (Ap-pendix 1). Reexamination with respect toevolutionarily conserved secondary struc-ture elements as indicated by compensa-tory substitutions in stem-forming regionshelped to refine and confirm the finalalignment.

The secondary structure elements sup-ported in our set of sequences were com-pared with the recently published second-ary structure models of Drosophilamelanogaster and Aedes albopictus (Schnareet al., 1996). Other than some minor dif-ferences, the stem assumed to be formedby the sites 68-72 and 104-108 in ouralignment of the 28S D7 fragment is dif-ferent from these published models (Ap-pendix 1). The same applies for the stemassumed to be formed by the sites 157-162and 167-172, and a part of the stem


formed by sites 146-156 and 261-273 inour 28S D7 fragment alignment (Appendix1). Nonetheless, 488 sites, which includedno gaps in any of the dipteran taxa, werealigned with confidence. This subset wasconsidered for molecular phylogeneticanalysis.

Sequence Divergence

Of the 488 sites included in the analysis,211 (43%) are variable and 159 sites (33%)are phylogenetically informative for thedipteran taxa. Within the Diptera, se-quence divergence ranges from 2.9% to24.3% of pairwise sequence divergence(Table 2). Relatively high sequence diver-gences are observed when representativesof the Culicomorpha (Chironomus, Simu-lium, Culex, Aedes) are compared with oth-er dipteran species (17.4-24.3%). Other-wise, pairwise sequence divergence doesnot exceed 15.2% within the Diptera. Thesequence divergence among the culico-morphous families is also high (16.7-21.1%); as expected, sequence divergencebetween the closely related mosquito spe-cies Aedes albopictus and Culex pipiens is thelowest observed in the data set (2.9%). Therange of sequence divergence observed forthe culicomorphous representatives withinthe Diptera even overlaps with that ob-served between the dipteran species andthe outgroup taxa (22.7-31.0%), indicatingelevated substitution rates in the culico-morphous lineages.

If the extracted sequence sites are sub-divided into those derived from double-stranded (DS) and single-stranded (SS)regions, 281 sites (58%) are located in pu-tative DS regions and the remaining 207(42%) are in SS regions. About two thirdsof the variable sites (142) are found in DSregions. Sequence variability in DS regionsis significantly higher than that in SSregions (x2, P < 0.005) if the occurrence ofcompensatory substitutions in DS regionsis neclected. Assuming that all substitu-tions in DS regions are correlated with acomplementary substitution, the numberof independent substitution events in theDS regions would reduce to 71 versus 70complementary pairs of sites where no

OH

g

r\|CN(NCN|CN|CN|CNCNCN|(N{NCNfN|cOCNCN

j i ) m m c o n n f i ( O C 5 ( N O \ d t s o \CNIC\|CNCN(NC\|CN|{NCN|CN|CNrs|fNCO(N(N

o d a o a \ o 6 o r ot—I t—I CN r—I t—< f s | CN T-H

d N O O N O O O O N W O< N C N T — I t — I t — I r - I T — I T — I T — I < — I t — I

c N O O O o t c o c n t N OCN <N CN >—iCNCNfNCNCNCNCMCNI

K CN co p oq CNJ *qcs co *d vd in od K

T-H LO ON CNCN CO CO ^

O\ CNjCS LO

£ £ l-i~ B-8-S

Q E2 CQ U CQ Q


substitution is observed. Under this mostconservative assumption, DS regions arestill significantly more variable than SSregions (P < 0.05).

Nucleotide Composition

Most models of sequence evolution as-sume stationarity of nucleotide composi-tion over time (Gillespie, 1986). The impactof base compositional shifts between taxaon the reliability of phylogenetic recon-struction has been discussed (Hasegawaand Hashimoto, 1993; Lockhart et al.,1994). For the present data set, nucleotidecomposition was determined for DS andSS regions separately (Appendix 2). Con-sistent with the results of a recent study(Friedrich and Tautz, 1997), a considerabledifference between the dipteran taxa andthe outgroup taxa was observed in the DSregions. A mean (SD) AT content of 49.4%(2.4%) in the Diptera contrasts with 34.9%(1.3%) in the outgroup taxa. The analogouscomparison in the SS regions also shows atendency towards elevated AT content inthe Diptera with 64.4% (1.6%) versus59.9% (1.3%) in the outgroup taxa.

We tested stationarity of base composi-tion for DS and SS regions using the testdeveloped by Rzhetsky and Nei (1995),which takes into account the phylogeneticrelatedness of the sequences being com-pared. As expected, when the outgrouptaxa are included, stationarity of base com-position is rejected for DS sites (P < 0.001)but not for SS sites. However, stationarityis not rejected for both SS sites and DSsites when the three outgroup taxa are ex-cluded from the test. Because the station-arity assumption can be upheld for the in-group taxa, it is unlikely that nucleotidecomposition will have a strong influenceon the reconstruction of the higher in-group taxa.

Degree of Substitutional Saturation

Given the considerable degree of se-quence divergence between many of thedistantly related dipteran taxa, it appearednecessary to analyze the loss of phyloge-netic information due to multiple hits. Thedegree of substitutional saturation was

characterized by comparing observed pair-wise Ts and Tv divergences of all possiblespecies combinations to those expected atsubstitutional equilibrium, Tssat and Tvsat

using the approximating formulas: Tssat =2/(TTT<7TC + HATTG) and Tvsat = 2/7TCTirGA (Ha-segawa et al., 1985). In this context,/is theprobability that a given site is variable andTT is the stationary nucleotide compositionper base; IT was set equal to the mean per-centage of each nucleotide across all taxa,and/was roughly approximated by settingit equal to the partition of variable sites ob-served in a given set of sites. Because DSand SS sites differed significantly in se-quence variability, saturation analysis wascarried out for each subset separately.

Pairwise Ts divergences were plottedagainst Tv divergences in relation to Tssat/Tvsat (Fig. 2). The comparison indicates afaster increase of Ts over Tv in DS sites butnot in SS sites, which is consistent with thepresence of a strong DS site-specific Tsbias previously found in a study of 28SrRNA evolution in closely related Drosoph-ila species (Rousset et al., 1991). For se-quence saturation, it becomes obvious thatin the DS sites Ts have reached the rangeof substitutional equilibrium for the ma-jority of pairwise sequence comparisonswhereas Tv have remained clearly distantfrom Tvsat (Fig. 2a). In the SS sites, how-ever, Ts and Tv are affected to same degreeby multiple hits (Fig. 2b). In approximatelyone third of the pairwise sequence com-parisons, both classes of substitutions havereached the range of expected substitu-tional equilibrium.

To see which aspects of tree topologymight be most affected by the considerableamount of substitutional saturation ob-served, we subdivided the data accordingto taxon combinations (Fig. 2). The strong-est impact was on sequence divergencesbetween ingroup and outgroup taxa. Fur-thermore, those pairwise sequence com-parisons within the Diptera that involvedat least one culicomorphous taxon yieldeda distinct cluster of data points that arecloser to saturation than are any of thedata derived from the rest of the possiblepairwise sequence combinations among


(a) (b)

Ts

0.25 --

0.2 --

0.15

0.1

0.05 -

C O Ts

o

0.25 --

0.2 --

0.15

0.1

0.05 -

0

0.1 _ 0.2

Tv0.3 0.1 _ 0.2

Tv0.3

FIGURE 2. Observed average transition (Ts) and transversion (Tv) divergence per nucleotide. Light circles =pairwise comparisons between dipteran taxa except Culicomorpha; medium circles = pairwise comparisonsbetween dipteran taxa including culicomorphous taxa; heavy circles = pairwise comparisons between dipteranand outgroup taxa. Heavy diamond = Tssat/Tvsat. (a) DS sites./is assumed to be 0.51, and stationary nucleotidecomposition is TTA = 0.21, irc = 0.26, TTT = 0.25, TTG = 0.28. (b) SS sites./is assumed to be 0.33, and stationarynucleotide composition is TTA = 0.37, TTC = 0.14, TTT = 0.26, TTG = 0.23.

the Diptera. The only exception in this re-spect is the data point obtained from theclosely related culicomorphous taxa Aedesand Culex, which falls into the cluster ofdata derived from nonculicomorphousDiptera. Thus overall, the 28S rDNA se-quences seem to have retained phyloge-netic information in the majority of in-group taxa included. Nonetheless,ambiguous results may be anticipated con-cerning the position of the Culicomorphawithin Diptera and concerning the root ofthe tree based on the extremely divergedoutgroup sequences. The rooting problemreduces the chances of accurately recon-structing the most basal splits among theDiptera.

Rate Heterogeneity across SitesSubstitution rates across sites are un-

equally distributed in most sequences andmay best be fitted to a gamma distribution(Uzzell and Corbin, 1971). The shape pa-rameter a of the gamma distribution,which is negatively correlated with the ex-tent of rate heterogeneity, is commonlyused as a measure for rate heterogeneity

across sites. In real sequences, typical val-ues for a range from 0.1 to 1, and the im-portance of accommodating rate hetero-geneity across sites in tree estimation hasbeen well established in recent years(Yang, 1996a).

To test the present data for the signifi-cance of rate heterogeneity across sites, weemployed the likelihood ratio test (seeHuelsenbeck and Rannala, 1997). As thenull hypothesis (Ho), rate homogeneityacross sites was assumed; the alternativehypothesis (Hj) was rate heterogeneityacross sites. In this case, two times the dif-ference in logarithmic likelihood (2A logL)is approximately \2 distributed with onedegree of freedom. Rate heterogeneity wasaccommodated under the F84 model of se-quence evolution (Felsenstein, 1995) by ap-plying the discrete gamma model (Yang,1994), which approximates the gamma dis-tribution by allowing a number of differentrate categories with equal probability of oc-currence along the sequence (F84-dF). Weallowed four rate categories, which givessufficiently good approximation of the


TABLE 3.sequences.

Tree

T-MP19

T-MP19

T-MP16

T-MP12

Likelihood

Ho

Model

F84F84-drF84-drF84-df

ratio test

Clock

noyesyesyes

results for

logL0

-3681.6-3557.4-2934.1-2102.1

rate heterogeneity across

H,

Model

F84-drF84-drF84-drF84-df

Clock

nononono

sites and among lineages

logL,

-3454.5-3454.5-2915.1-2088.3

2A logL

454.2205.937.927.4

for dipteran

P

<0.001<0.001<0.001<0.005

gamma distribution (Yang, 1994). Becauserate heterogeneity is overestimated if an un-realistic topology such as a star tree is as-sumed (Sullivan et al., 1996), we used thesingle most-parsimonious tree derivedfrom the data (T-MP19) (not shown) for MLestimation of rate heterogeneity across sites.Accordingly, if the F84-dF sequence evolu-tion model is assumed, substitution ratesare gamma distributed across the 28SrDNA sequences with an a of 0.37 (rates:0.013, 0.16, 0.7, 3.13). As expected, the like-lihood ratio test shows that accounting forrate heterogeneity across sites (F84-dF) im-proves the likelihood of the T-MP19 topologysignificantly (Table 3). Thus, there is signif-icant rate heterogeneity across sites in thepresent set of 28S rDNA sequences thatmust be accounted for in tree estimation.

Rate Heterogeneity among Lineages

Strong differences in substitution ratesamong lineages can cause tree estimationmethods to become misleading (Felsen-stein, 1978; Huelsenbeck and Hillis, 1993).The rDNA of dipterans evolves significant-ly more quickly than does that of most oth-er insects (Carmean et al., 1992; Friedrichand Tautz, 1997). There was thus reason toassume significant taxon-specific substitu-tion rate differences between ingroup andoutgroup taxa in the present data. In addi-tion, the sequence variability analysis indi-cated accelerated substitution rates in theculicomorphous lineages (Table 2). Wetherefore also tested homogeneity of sub-stitution rates among lineages using likeli-hood ratio test statistics (Felsenstein, 1981).

The null hypothesis of rate homogeneityamong lineages, i.e., a molecular clock,was compared with the alternative hy-

pothesis, which allows for rate heteroge-neity among lineages. Thereby 2A logL isX2 distributed with number of species in-cluded minus two degrees of freedom. Ifall taxa are included (T-MP19), 2A logL isvery large and the molecular clock is re-jected with high significance (Table 3). Ifthe three outgroup taxa are excluded fromthe test (T-MP16), 2A logL drops consider-ably but the molecular clock is still rejectedwith very high probability. Even if the fourquickly evolving culicomorphous ingroupspecies are excluded from the test (T-MP12),the molecular clock is rejected, althoughwith reduced probability. These results in-dicate that the outgroup taxa and the culi-comorphous taxa contribute most of thesubstitution rate heterogeneity among lin-eages, reinforcing the caveats derived fromthe staturation analysis regarding the in-ference of the position of the Culicomor-pha and the accuracy of tree rooting.

The ML ratio test rejects rate homoge-neity among the dipteran lineages whenthe taxa with most obviously acceleratedsubstitution rates are excluded from thetest. Previous relative rate tests of 28SrDNA sequences, which were carried outaccording to Wu and Li (1985), did not re-ject the molecular clock when the substi-tution rates of nonculicomorphous dipter-an taxa where compared (Friedrich andTautz, 1997). This discordance might be ex-plained by the fact that the x2 distributionis an approximate null distribution of thelikelihood ratio test statistics. It thereforeseems advisable to determine the null dis-tribution of the likelihood ratio test statis-tics by Monte Carlo simulation (Goldman,1993) to confirm the present result.


Tree Estimation

Tree estimation was guided by the ob-jective of accounting for sequence evolu-tion parameters that have a possible bear-ing on the success of tree estimationmethods, such as rate heterogeneity acrosssites and unequal Ts and Tv rates. We em-ployed the MP method to construct an ini-tial topology for the estimation of the rel-evant substitution parameters. AlthoughMP is misleading in cases of extreme rateheterogeneity among lineages (Felsenstein,1978), this method has the advantage of re-quiring no explicit assumptions and beingefficient in recovering long true branches(Huelsenbeck and Hillis, 1993). To accountfor the correlated sequence evolution ofcomplementary sites in DS regions, MPtree reconstruction was carried out weight-ing DS regions 0.8 over SS regions as pro-posed by Dixon and Hillis (1993). Underthese assumptions, MP finds a singlemost-parsimonious tree (T-MP19) with alength of 5,686 (consistency index = 0.587;retention index = 0.621). Several clades inthis tree are highly supported, as indicatedby bootstrap analysis (Fig. 3).

T-MP19 was then used for ML estimationof sequence evolution parameters assum-ing the F84-dF model of sequence evolu-tion, which allows for different Ts and Tvrates, unequal stationary base composition,and rate heterogeneity across sites (Yang,1994; Felsenstein and Churchill, 1996). Un-der these assumptions, a Ts/Tv rate ratioparameter K of 1.44 and an a parameter of0.37 were estimated. Employing these pa-rameters in the ML tree search resulted ina best ML F84-dr tree (logL = -3442.5)that included all of the ingroup clades thatwere highly supported with MP (Fig. 3).To get an estimation of branch probabili-ties independent from MP, bootstrap anal-ysis was carried out applying the NJ meth-od. Evolutionary distances (K80-T) wereestimated according to Jin and Nei (1990),combining the two-parameter sequenceevolution model (Kimura, 1980) with acontinuous gamma distribution model toaccount for rate heterogeneity across sites.

The same substitution parameter settingswere used as in the ML tree search.

Both MP and NJ K80-F find strong sup-port for the monophyly of the Diptera ver-sus the outgroup taxa chosen (BP = 100)(Fig. 3). To some extent, this result may bedue to the episodically accelerated rate ofrDNA evolution in the stem group in theDiptera (Friedrich and Tautz, 1997), whichleads to an exceptionally high number ofmolecular synapomorphies as is clearly re-flected in the extremely long branch thatjoins the outgroup taxa with the Diptera inthe ML F84-dF phylogram.

Within the Diptera, three major clades ofhigher systematic level are very stronglysupported (BP > 95). One of these includesDrosophila melanogaster and Tabanus sudeti-cus, which represent the subgroup Brachy-cera. A second clade includes six taxa:Bradysia coprophila, Dilophus febrilis, Bolito-phila cinerea, Clinodiplosis cilicrus, Anapausisinermis, and Sylvicola fenestralis. The firstfour of these represent the families Sciari-dae, Bibionidae, Bolitophilidae, and Ceci-domyiidae, respectively, and thus the Bib-ionomorpha sensu stricto. The remainingtwo species represent the Scatopsidae andAnisopodidae, respectively. The wholeclade therefore corresponds to the contro-versial Bibionomorpha sensu lato. The thirdsignificantly supported major subclade in-cludes Chironomus teutons, Simulium euryad-miniculum, Culex pipiens, and Aedes albopic-tus, thus representing the Culicomorpha.A fourth major clade that is strongly sup-ported (BP > 80) is the Tipulomorpha sen-su stricto, represented by Tipula paludosaand Limonia nebulosa.

All branches interconnecting these fourmajor clades and the two remaining taxa,Psychoda cinerea and Trichocera regelationis,are poorly supported by the data (BP < 60)and to a large extent not consistently re-solved by the different methods. In the MLF84-dr tree for example, the Tipulomorphasensu stricto splits off most basally fromthe rest of Diptera, whereas in the NJ K80-F tree and in T-MP19 the Culicomorpha ismost basal (not shown). Obviously, thephylogenetic information available allowsinference of the composition of major dip-


p Panorpa

Manduca

Archaeopsylla

88/85

100/100

-Limonia

-TipulaTipulomorpha sensu stricto

i— Trichocera

Chlronomus

•Simulium

Culex

Aedes

Culicomorpha

—Difophus

'Sylvicola

Crinodlplosis95/98 Bibionomorpha sensu lato

Anapausis

•Bolitophila

Bradysia

—Psychoda

98/100

-Drosophila

-TabanusBrachycera

FIGURE 3. ML F84-dF dipteran phylogram. Bold branches are strongly supported by MP and NJ K80-F.Numbers at bold branches are BP for MP and NJ K80-F, respectively. Branches without numbers are supportedby BP < 80 and may not be consistently reconstructed by NJ K80T or MP. Significantly supported majorsubgroups are indicated by shaded boxes. Line at the bottom represents 0.1 substitutions per base.


teran infraorders but not the branchingevents among them. Such ambiguity in theestimation of the basal relationships wasexpected based on the results of the se-quence saturation analysis.

Within the Bibionomorpha sensu latoand the Culicomorpha, molecular phylo-genetic resolution is also rather low. Noneof the methods applied significantly sup-ported further subgrouping in the Bibion-omorpha sensu lato (BP < 50). In the caseof the Culicomorpha, the monophyly ofthe two mosquito species, Culex pipiens andAnopheles albopictus, is significantly andconsistently supported but their relation-ship to the Chironomidae {Chironomus ten-tans) and the Simuliidae (Simulium euryad-miniculum) is not consistently resolved.The ML F84-dF and MP favor a closer re-lationship between Culicidae and Simuli-idae, but NJ K80-F favors a relationship be-tween the Chironomidae and theSimuliidae (BP = 73) (not shown).

Robustness of the Bibionomorpha sensu lato

Because the tree estimation results ap-peared to have a strong bearing on thecontroversy about the composition of theinfraorder Bibionomorpha, we carried outadditional analyses of the robustness ofthis result. Given the difficulty in aligningsequence regions and the high degree ofsequence saturation observed in manypairwise sequence comparisons, the im-pact of these factors was investigated withsimple procedures.

To test for alignment effects, phyloge-netic reconstruction was carried out basedon the CLUSTAL V alignment and apply-ing NJ K80 after removal of all sites withgaps. This procedure produced the sameresults concerning major subgroups (notshown). Thus, the molecular inference ofthe major subgroups is robust againstchanges in the more divergent parts of thealignment.

Weighted MP was applied to investigatethe influence of homoplasy. Because Tvwere far from saturation for most pairwisesequence comparisons in both DS and SSsites (Fig. 2), this class of substitutions wasexpected to be less affected by homoplasy

than Ts. A search for MP trees based onTv only yielded two equally most-parsi-momious trees (Fig. 4a), which include thethree major subgroups that are significant-ly supported by unweighted MP or NJK80-F (Fig. 3). For comparison, MP recon-struction was also carried out based on Tsonly, which again yielded two most-par-simonious trees (Fig. 4b). These trees, how-ever, include only the Brachycera as oneout of the three otherwise significantlysupported higher clades. The effect of Tssaturation was expected to be most dra-matic for the Culicomorpha given thehigher saturation level in this subclade(Fig. 2). The lack of support by Ts-only MPfor the Culicomorpha, which is otherwisestrongly supported, is thus consistent withthe finding that Ts are largely saturated inthis group. Finding the Bibionomorphasensu lato to be recovered by Tv-only MPbut not by Ts-only MP, as with the Culi-comorpha, indicates that the strong sup-port in the data predominantly derivesfrom Tv and is thus based on phylogeneticinformation rather than on convergence atthe molecular level.

DISCUSSION

Tree Evaluation with Reference toMorphological Hypotheses of Higher

Dipteran Systematics

In addressing the earliest splits among themajor dipteran groups, no robust resolutioncould be achieved with any of the three treeestimation methods employed. Thus, themolecular data do not allow reliable infer-ences at this level, which proved similarlyhard to resolve using morphological char-acters. Nonetheless, there are three majordipteran infraorders that are significantlysupported by all of the tree reconstructionmethods applied. Given that consistencyamong tree reconstruction methods corre-lates positively with the accuracy of esti-mated trees (Kim, 1993) and that tree esti-mation bias problems are unlikely, thesemajor subgroups may be considered as re-liably inferred (Fig. 5). In each case, theircomposition is in accordance with groupsthat had been postulated previously on the


(a)I— Bradysia

• I DilophusI— BolitophilaP Crinodiplosis

"""•"L SyMcolaAnapausis

fDrosophilaTabanusTipulaLimonia

L Trichocera—— Psychoda

— AedesPL Cufex

__j l—. ChironomusI — Simulium

|- PanorpaL Archaeopsylla

— ^ Manduca

(b)DrosophilaTabanusSimuliumAedesCulexDilophusAnapausisBradysiaCrinodiplosisBolitophila

_ SylvicolaL- Chironomus_. 77pu/aL Limonia- . Trichocera

"L Psychoda_. PanorpaL Archaeopsylla

— ManducaFIGURE 4. Weighted MP dipteran trees. Higher taxa that are significantly supported when all character state

changes are considered are indicated by shaded boxes, (a) Strict consensus of two shortest MP trees based onTv only and weighting DS sites 0.8 over SS sites, (b) Strict consensus of two shortest MP trees based on Tsonly and weighting DS sites 0.8 over SS sites.

basis of morphological characters. Two ofthese clades, the Brachycera and the Culi-comorpha, are groups on which most au-thors agree because they are clearly sup-ported by a large number of morphologicalcharacters (Hennig, 1981; Wood and Bor-kent, 1989; Oosterbroek and Courtney, 1995).For the monophyly of the Bibionomorphasensu lato, however, no clearly derived mor-phological character states could be put for-ward in spite of overall similarity of theadult phenotype. Hennig (1981) favored theBibionomorpha sensu lato based on reduc-tional similarities in the wing venation. Morerecently, many authors have followed Woodand Borkent (1989), who rejected the mono-phyly of the Bibionomorpha sensu lato(Courtney, 1990; Griffiths, 1990; Sinclair,1992; Oosterbroek and Courtney, 1995).However, our results clearly support themonophyly of the controversial Bibionomor-pha sensu lato and indicate that the Aniso-podidae or Scatopsidae are less closely re-

lated to the Psychodidae than to theBibionomorpha sensu stricto, as implied inthe Psychodomorpha sensu W&B or in thePsychodomorpha sensu Oosterbroek andCourtney.

A fourth clearly and consistently support-ed higher infraorder is the Tipulomorphasensu stricto, which is consistent with thegeneral consensus on this group (Hennig,1981; Wood and Borkent, 1989; Oosterbroekand Theowald, 1991). Regarding the familyTrichoceridae, however, which according tomany authors is the sister group of the Ti-pulomorpha sensu stricto (Hennig, 1981;Griffiths, 1990; Oosterbroek and Theowald,1991; Oosterbroek and Courtney, 1995), ourresults provide no further clue. Nonetheless,the rejection of the Psychodomorpha sensuW&B rules out the only morphology-basedhypothesis that had questioned the Tipulo-morpha sensu lato.

Thus, despite limitations with respect tophylogenetic resolution and taxon choice,


-Gyclorrhapha

— "Orthorrhapha"

-Sciaridae

-Cecidomyiidae

-BoHtophiiidae

•Bibionidae

-Anisopodidae

-Scatopsidae

- Psychodidae

- Limoniidae

-Tipulidae

Brachycera

Bibionomorpha sensu lato

Tipulomorpha sensu strlcto

—Trichoceridae

-Chironomidae

-Culicidae Culicomorpha

-Simuliidae

FIGURE 5. Consensus tree of major dipteran subgroups reliably reconstructed in this study. Occurrence ofthe obliquely moving and subdivided larval mandible types in a lineage is indicated by boxed symbols. Highersubgroups are indicated by shaded boxes.

the present results have important impli-cations for morphological studies of dip-teran systematics. A further conclusion re-lates to the controversial discussion of thesister group of the Brachycera (Griffiths,1994). The family Anisopodidae was pro-posed as sister group of the Brachycera(Woodley, 1989; Oosterbroek and Court-ney, 1995), which would render the Bibion-omorpha sensu lato paraphyletic with re-spect to the Brachycera (Fig. lc). The highsupport for the monophyly of the Bibion-omorpha sensu lato, however, excludes thispossibility. Although the position of theBrachycera differed among the moleculartrees, in no tree was monophyly of the Ne-matocera supported. Thus, the Brachy-cera/Bibionomorpha sensu lato sister-group hypothesis remains, which for dif-

ferent reasons had previously been pro-posed by Hennig (1981) and by Collessand McAlpine (1970).

In conclusion, the molecular phylogeneticapproach was powerful enough to provide anew framework for future morphologicalstudies of dipteran phylogeny. The presentstudy is complemented by a recent analysisof the infraorder Culicomorpha based on adifferent set of 28S rDNA sequences (Paw-lowski et al., 1996). Although the basal re-lationships of the Culicomorpha have re-mained ambiguous as in our study,significant support for a sister-group rela-tionship between the Simuliidae and theproblematic family Thaumaleidae wasfound, which is noteworthy given that someauthors proposed to include the Thauma-leidae in infraorders different from the Cul-


icomorpha (Hackman and Vaisanen, 1982;Colless and McAlpine, 1991).

Evolution of the Dipteran Larval Mandible:A Paradigm Case for the Argument of

Commonality

Of central importance for the contem-porary discussion of higher dipteran sys-tematics is the interpretation of the evolu-tionary pathways of the larval mouthpartsin the Diptera. Many groups in the systemof Wood and Borkent (1989) rely on argu-ments derived from this character statecomplex. Because some of the characterssuch as labral shape and mandibularswinging orientation may well be correlat-ed (Teskey, 1981), the discussion may befocused on the long-standing dissent overthe evolution of the larval mandibles. Theunderlying problem of determining char-acter state polarity has been more or lessexplicitly discussed by several authors. Be-cause neither new paleontological nor newontogenetic data are at hand, one still de-pends on indirect evidence, which is com-monly derived from outgroup comparison.The larvae of other holometabolous insectorders usually possess solid and horizon-tally moving mandibles, as is also the casefor the nymphal stages of hemimetabolousinsect orders (Snodgrass, 1935). This typeof larval mandible is regarded as agroundplan feature of the holometabolousinsects. Similar types of mandibles can befound in representatives of several dipter-an families such as the Tipulidae, Limoni-idae, Bibionidae, Mycetophilidae, andSciaridae. Based on outgroup comparison,many comparative morphologists (Goeth-gebuer, 1925; Snodgrass, 1935; Cook, 1949;Gouin, 1959) have considered this charac-ter state ancestral for the Diptera (Fig. 6a).Some authors, however, emphasized theoccurrence of strikingly similar obliquelyor vertically moving and subdivided man-dibles in representatives of distantly relat-ed families such as the Trichoceridae, An-isopodidae, and Psychodidae (Edwards,1926; Anthon, 1943; Schremmer, 1951).They interpreted the phylogenetically widedistribution of this character state withinthe Diptera as due to evolutionary conser-

vation of an ancestral ground state (Fig.6b). Hennig (1948, 1981) realized the con-tradiction between adult and larval char-acter states in higher dipteran systematicsand explicitely adopted the view of thosewho favored obliquely and subdividedmandibles as symplesiomorphic. Formally,the argument underlying this hypothesisof character state polarity corresponds tothe principle of commonality or ingroupdistribution (de Jong, 1980). In general, thecommonality criterion is considered lessstringent than outgroup comparison(Kitching, 1992) because it depends onsome reliable knowledge of the ingroupphylogeny.

In the present case, this knowledge isprovided by the molecular phylogeneticanalysis. The distribution of larval mandi-ble character states can be mapped on acladogram that includes the dipteran sub-groups that were reliably inferred (Fig. 5).The combined occurrence of the two char-acter states, subdivision and oblique or ver-tical movement of the larval mandible, canbe noted in representatives of at least fourmajor dipteran lineages, the Brachycera, theBibionomorpha sensu lato, the Psychodo-morpha, and the Trichoceridae. The contro-versial homology of mandibular structuresin the Cyclorrhapha excludes this groupfrom the inference (Sinclair, 1992; Griffiths,1994). Moreover, larvae of the limoniid gen-era Pilaria and Ulomorpha possess sudividedand obliquely moving mandibles that areconsidered to be structurally very differentfrom the subdivided mandibles found inthe other dipteran inf raorders (Oosterbroekand Theowald, 1991; Oosterbroek andCourtney, 1995). Oblique mandible move-ment, however, has been proposed to be an-cestral for the Tipulomorpha sensu stricto(Oosterbroek and Theowald, 1991; Ooster-broek and Courtney, 1995). If restricted touncontroversial evidence, the wide distri-bution of subdivided and obliquely movingmandibles among the Diptera is confirmedand may be interpreted to indicate plesio-morphy according to the criterion of com-monality.

Thus, during the evolution of the dipter-an stem group the solid and horizontally


(a)

(b)

IFIGURE 6. Hypotheses of the evolution of the dipteran larval mandibles. Thick arrows indicate transfor-

mation pathways in the dipteran stem group; thin arrows indicate transformation pathways in the dipterancrown group, (a) Gradual evolution of solid, horizontally moving mandibles (I) of the dipteran stem group intosolid, obliquely or vertically moving mandibles (II) and subsequently into subdivided, obliquely moving man-dibles (III), (b) Transformation of solid, horizontally moving mandibles into subdivided, obliquely movingmandibles during evolution of the dipteran stem group, followed by diversification of this ground plan intovarious types of mandibles observed in modern dipterans, such as solid, vertically moving mandibles (IV).

moving larval mandible type seen in lar-vae of most other holometabolous insectorders was most probably transformedinto the derived type of subdivided andobliquely moving mandibles (Fig. 6b). Thesolid and horizontally moving biting man-dibles in representatives of the familiesBibionidae or Sciaridae must then be con-sidered examples of striking convergence.

Early Dipteran Diversification according toMolecular and Fossil Record

The lack of molecular phylogenetic res-olution among the major dipteran sub-groups is striking when contrasted with

the strong support for the Diptera as suchand for each major subgroup in this study.The considerable amount of sequence sat-uration in the data may obscure the basalrelationships. The branching pattern in themolecular phylogram however providesevidence for yet an alternative explanationfor the lack of resolution (Fig. 3). Thebranches that join the major groups arevery short, suggesting that the diversifi-cation of the major subgroups represents aperiod of rapid cladogenesis in the evolu-tion of the Diptera. A similar pictureemerged from a recent analysis of longer28S rDNA sequences from a sample of spe-


245

<EQ. T

riass

ic

g

Jura

ssic

ft

Cre

tace

ous

Cen

ozoi

c

TipulidaeLimoniidaeTrichoceridaePsychodidaeChironomidaeSimuliidaeCulicidaeBibionidaeMycetophilidaeCecidorriyiidaeSciaridaeScatopsidaeAnisopodidae"Orthorrhapha"Cyclorrhapha

FIGURE 7. Paleontological record of dipteran fam-ilies or subgroups included in this study. MYA = mil-lion years ago. Higher subgroups are indicated byshaded boxes.

ties included here (Friedrich and Tautz,1997). This pattern of divergence can becompared with the paleontological data onearliest occurrences of dipteran families inthe fossil record (Fig. 7).

The origin of the Diptera most probablydates back into the Upper Permian, 250million years ago (MYA), although the fos-sil evidence for flies from that period isvirtually nonexistent and difficult to inter-pret (Tillyard, 1929; Hennig, 1981; Will-mann, 1989; Wootton and Ennos, 1989;Krzeminski, 1992a). Nonetheless, the find-ing of representatives of various dipteransubgroups in the Late Triassic (Krzemin-ski, 1992b; Fraser et al., 1996) quite con-vincingly documents the diversification ofsome major dipteran sublineages by themiddle Mesozoic. Upper Triassic species ofthe crown group Diptera are known fromlocations in Australia (Evans, 1971) andNorth America (Olsen et al., 1978; Krze-minski, 1992b; Fraser et al., 1996). TheNorth American site and a Triassic site inRussia contain a fly related to the modernnematocerous family Anisopodidae(Krzeminski, 1992a; Fraser et al., 1996).

The North American site also contains anumber of additional fly taxa. Amongthose, Krzeminski (1992a) identified a sec-ond anisopodid-like family (Procrampton-omyiidae) and representatives of the mod-ern subgroups Limoniidae and Brachycera.Fraser et al. (1996) recently extended thelist of North American Triassic dipteransby adding descriptions of psychodid fos-sils. Thus, representatives of four majormodern dipteran subdivisions appear al-most simultanuously in the paleontologi-cal record 220 MYA. Although the possi-bility of sampling artifact must beconsidered, the consistency between thepaleontological and molecular data con-cerning a rapid diversification of the majorlineages is notable. The amount of phylo-genetic information that can accumulatefor a clade is correlated with the time spanof the existence of the respective stem lin-eage. The small amount of phylogenetic in-formation documenting the earliest splitsin the Diptera at both the molecular andthe morphological level is thus also consis-tent with a rapid diversification of the ma-jor lineages.

ACKNOWLEDGMENTS

We thank Friedrich Reiss and Wolfgang Schacht forsupplying specimens, for help with species determi-nation, and for their interest, Hans Meyer for deter-mining Clinodiplosis cilicrus, Monika Retzlaff for aDNA sample of Psychoda cinerea, Ralf Sommer forDNA samples of Bradysia coprophila, Simulium euryad-miniculum, and Chironomus teutons, Svante Paabo forsome of the primers, Alexander Riedel for help withPCR and sequencing, Hans Ulrich, Pjotr Oosterbroek,and Herman de Jong for help with systematic litera-ture, Ewa Krzeminska for detecting a mistaken spe-cies name, and Kimberlee Wollter for improving thelanguage of the final manuscript. We gratefully ac-knowledge critical reading of an early version of thismanuscript by Bernhard Hausdorf and Roland Mel-zer, suggestions on the final version by Pjotr Ooster-broek and Greg Courtney, and detailed comments byChris Simon, David Grimaldi, Timothy Collins, andan anonymous reviewer. This work was supported bya Ph.D. studentship awarded to M.F. by the Universityof Munich.

REFERENCES

ANTHON, H. 1943. Der Kopfbau der Larven einigernematoceren Dipterenfamilien: Rhyphidae, Tricho-ceridae, Psychodidae und Ptychopteridae. SpoliaZool. Mus. Haun. 3:1-61.


ASHBURNER, M. 1989. Drosophila. A laboratory hand-book. Cold Spring Harbor Laboratory Press, ColdSpring Harbor, New York.

BOUDREAUX, H. B. 1979. Arthropod phylogeny, withspecial reference to insects. Wiley & Sons, NewYork.

CARMEAN, D., AND J. B. CRESPI. 1995. Do long branch-es attract flies? Nature 373:666.

CARMEAN, D., L. S. KIMSEY, AND M. L. BERBEE. 1992.18S rDNA sequences and the holometabolous in-sects. Mol. Phylogenet. Evol. 2:270-278.

CASANOVA, J. L., C. PANNETIER, C. JAULIN, AND P.KOURILSKY. 1990. Optimal conditions for directlysequencing double-stranded PCR products with Se-quenase. Nucleic Acids Res. 18:4028.

CHALWATZIS, N., J. HAUF, Y. VAN DE PEER, R. KINZEL-BACH, AND R. K. ZIMMERMANN. 1996. 18S riboso-mal-RNA genes of insects—Primary structure of thegenes and molecular phylogeny of the Holometabo-la. Ann. Entomol. Soc. Am. 89:788-803.

COLLESS, D. H., AND D. K. MCALPINE. 1970. Diptera.Pages 656-740 in The insects of Australia, 1st edi-tion, Volume 2 (D. F. Waterhouse, ed.). CSIRO, Mel-bourne Univ. Press, Melbourne.

COLLESS, D. H., AND D. K. MCALPINE. 1991. Diptera.Pages 717-786 in The insects of Australia: A text-book for students and research workers, 2nd edi-tion, Volume 2 (I. D. Naumann, P. B. Carne, J. F. Law-rence, E. S. Nielsen, J. P. Spradberry, R. W. Taylor,M. J. Whitten, and M. J. Littlejohn, eds.). CSIRO,Melbourne Univ. Press, Melbourne.

COOK, E. F. 1949. The evolution of the head in thelarvae of the Diptera. Microentomology 14:1-57.

COURTNEY, G. W. 1990. Cuticular morphology of lar-val mountain midges (Diptera: Deuterophlebiidae):Implications for the phylogenetic relationships ofthe Diptera. Can. J. Zool. 68:556-578.

COURTNEY, G. W. 1991. Phylogenetic analysis of theBlephariceromorpha, with special reference tomountain midges (Diptera: Deuterophlebiidae).Syst. Entomol. 16:137-172.

CURTIS, D., J. APFELD, AND R. LEHMANN. 1995. Nanosis an evolutionarily conserved organizer of anterior-posterior polarity. Development 121:1899-910.

DE JONG, R. 1980. Some tools for evolutionary andphylogenetic studies. Z. Zool. Syst. Evolutions-forsch. 18:1-23.

DESALLE, R., AND D. A. GRIMALDI. 1991. Morpholog-ical and molecular systematics of the Drosophilidae.Annu. Rev. Ecol. Syst. 22:447-475.

DIXON, M. T., AND D. M. HILLIS. 1993. RibosomalRNA secondary structure: Compensatory mutationsand implications for phylogenetic analysis. Mol.Biol. Evol. 10:256-267.

EDWARDS, F. W. 1926. The phylogeny of nematocerousDiptera: A critical review of some recent sugges-tions. Verh. Int. Kongr. Entomol. 3:111-130.

EVANS, J. W. 1971. Some upper Triassic Hemipterafrom Mount Crosby, Queensland. Mem. Queensl.Mus. 6:145-152.

FELSENSTEIN, J. 1978. Cases in which parsimony orcompatibility methods will be positively mislead-ing. Syst. Zool. 19:401^110.

FELSENSTEIN, J. 1981. Evolutionary trees from DNAsequences: A maximum likelihood approach. J. Mol.Evol. 17:368-376.

FELSENSTEIN, J. 1985. Confidence limits on phyloge-nies: An approach using the bootstrap. Evolution39:783-791.

FELSENSTEIN, J. 1995. PHYLIP: Phylogenetic inferenceprograms, version 3.572. Univ. Washington, Seattle.

FELSENSTEIN, J., AND G. A. CHURCHILL. 1996. A Hid-den Markov Model approach to variation amongsites in rate of evolution. Mol. Biol. Evol. 13:93-104.

FRASER, N. C, D. A. GRIMALDI, P. E. OLSEN, AND B.AXSMITH. 1996. A Triassic Lagerstatte from easternNorth America. Nature 380:615-619.

FRIEDRICH, M. 1995. Phylogenetischer Informations-gehalt und evolutionare Eigenschaften von kernko-dierten ribosomalen Genen innerhalb der Euarthro-poda. Ph.D. Dissertation, Ludwig-MaximiliansUniv., Munich.

FRIEDRICH, M., AND D. TAUTZ. 1997. An episodicchange of rDNA nucleotide substitution rate has oc-curred at the time of the emergence of the insectorder Diptera. Mol. Biol. Evol. 14:644-653.

GILLESPIE, J. H. 1986. Rates of molecular evolution.Annu. Rev. Ecol. Syst. 17:637-665.

GOETHGEBUER, M. 1925. Contribution a l'etude des"premandibules" chez les larves des Dipteres ne-matoceres. Encycl. Entomol. B 2:142-152.

GOLDMAN, N. 1993. Simple diagnostic statistical testsof models for DNA substitution. J. Mol. Evol. 37:650-661.

GOUIN, F. J. 1959. Morphology of the larval head ofthe Chironomidae. Smithson. Misc. Coll. 157:175-201.

GRIFFITHS, G. C. D. 1990. Review of Manual ofNearcticDiptera, Volume 3, edited by J. F. McAlpine and D.M. Wood. Quaest. Entomol. 26:117-130.

GRIFFITHS, G. C. D. 1994. Relationships among themajor subgroups of the Brachycera (Diptera): A crit-ical review. Can. Entomol. 126:861-880.

GUSTINCICH, S., G. MANFIOLETTI, G. DEL SAR, AND C.SCHNEIDER. 1991. A fast method for high-qualitygenomic DNA extraction from whole human blood.BioTechniques 11:298-301.

HACKMAN, W., AND R. VAISANEN. 1982. Differentclassification systems in the Diptera. Ann. Zool.Fenn. 19:209-219.'

HACKMAN, W, AND R. VAISANEN. 1985. The evolutionand phylogenetic significance of the costal chaeto-taxy in the Diptera. Ann. Zool. Fenn. 22:169-203.

HANCOCK, J. M., D. TAUTZ, AND G. DOVER. 1988. Evo-lution of secondary structures and compensatorymutations of the ribosomal RNAs of Drosophila me-lanogaster. Mol. Biol. Evol. 5:393-414.

HASEGAWA, M., AND T. HASHIMOTO. 1993. RibosomalRNA trees misleading? Nature 361:23.

HASEGAWA, M., H. KISHINO, AND T. A. YANO. 1985.Dating of the human ape splitting by a molecularclock of mitochondrial DNA. J. Mol. Evol. 22:160-174.

HENNIG, W 1948. Die Larvenformen der Dipteren.Eine Ubersicht der bisher bekannten Jugendstadien


der zweifliigeligen Insekten, Volume 1. AkademieVerlag, Berlin.

HENNIG, W. 1965. Phylogenetic systematics. IllinoisUniv. Press, Urbana.

HENNIG, W. 1968. Kritische Bemerkungen uber denBau der Fliigelwurzel bei den Dipteren und die Fra-ge nach der Monophylie der Nematocera. Stuttg.Beitr. Naturk. 193:1-23.

HENNIG, W. 1969. Die Stammesgeschichte der Insek-ten. Senckenberg-Buch, Volume 49. Kramer, Frank-furt am Main, Germany.

HENNIG, W. 1973. Diptera (Zweifliigler). Pages 1-337in Handbuch der Zoologie. Eine Naturgeschidhtedes Tierreichs, Volume 4 (J. G. Helmcke, D. Starck,and H. Wermuth, eds.). De Gruyter, Berlin.

HENNIG, W. 1981. Insect phylogeny. Wiley & Sons,New York.

HIGGINS, D., AND P. M. SHARP. 1988. CLUSTAL: Apackage for performing multiple sequence align-ments on a microcomputer. Gene 73:237-244.

HUELSENBECK, J. P. 1997. Is the Felsenstein zone a flytrap? Syst. Biol. 46:69-74.

HUELSENBECK, J. P., AND D. M. HILLIS. 1993. Successof phylogenetic methods in the four-taxon case.Syst. Biol. 42:247-264.

HUELSENBECK, J. P., AND B. RANNALA. 1997. Phylo-genetic methods come of age: Testing hypotheses inan evolutionary context. Science 276:227-232.

IMMS, A. D. 1944. On the constitution of the maxillaeand labium in Mecoptera and Diptera. J. Microscop.Sri. N.S. 85:73-96.

JIN, L., AND M. NEI. 1990. Limitations of the evolu-tionary parsimony method of phylogenetic analysis.Mol. Biol. Evol. 7:607-633.

KIM, J. 1993. Improving the accuracy of phylogeneticestimation by combining different methods. Syst.Biol. 42:331-340.

KIMURA, M. 1980. A simple method for estimatingevolutionary rates of base substitutions throughcomparative studies of nucleotide sequences. J. Mol.Evol. 16:111-120.

KITCHING, I. J. 1992. The determination of characterpolarity. Pages 22-43 in Cladistics. A practicalcourse in systematics (P. L. Forey, C. J. Humphries,I. J. Kitching, R. W. Scotland, D. J. Siebert, and D.M. Williams, eds.). Clarendon Press, Oxford, En-gland.

KRISTENSEN, N. P. 1975. The phylogeny of hexapod"orders". A critical review of recent accounts. Z.Zool. Syst. Evolutionsforsch. 13:1^14.

KRZEMINSKI, W. 1992a. The oldest Polyneura (Dip-tera) and their importance to the phylogeny of thegroup. Acta Zool. Cracov. 35:45-52.

KRZEMINSKI, W. 1992b. Triassic and Lower Jurassicstage of Diptera evolution. Mitt. Schweiz. Entomol.Ges. 65:39-59.

LAMEERE, A. 1906. Notes sur la classification des Dip-teres. Mem. Soc. Entomol. Belg. 12:105-140.

LATREILLE, P. A. 1802. Histoire naturelle, generale etparticuliere, des Crustaces et des Insectes. Pages 1-468 in Histoire naturelle par Buffon, Volume 3 (C.S. Sonnini, ed.). Buffon, Paris.

LAWRENCE, P. A. 1992. The making of a fly. The ge-netics of animal design. Blackwell, London.

LOCKHART, P. J., M. A. STEEL, M. D. HENDY, AND D.PENNY. 1994. Recovering evolutionary trees undera more realistic model of sequence evolution. Mol.Biol. Evol. 11:605-612.

MACQUART, J. 1834. Histoire naturelle des insects.Pages 1-578 in Collection des suites a Buffon, Vol-ume 1 (N. E. Roret, ed.). Buffon, Paris.

MADDISON, W P., AND D. R. MADDISON. 1992.MacClade, version 3.0. Sinauer, Sunder land, Mas-sachusetts.

MCALPINE, J. F. 1989. Phylogeny and classification ofthe Muscomorpha. Pages 1397-1518 in Manual ofNearctic Diptera, Volume 3 (J. F. McAlpine and D.M. Wood, eds.). Agriculture Canada, Ottawa.

MCPHERON, B. A., AND H. Y. HAN. 1997. Phylogeneticanalysis of North American Rhagoletis (Diptera: Te-phritidae) and related genera using mitochondrialDNA sequence data. Mol. Phylogenet. Evol. 7:1-16.

MELZER, R. R., T. ZIMMERMANN, AND U. SMOLA. 1997.Modification of dispersal patterns of branched pho-toreceptor axons and the evolution of neural super-position. Cell. Mol. Life Sri. 53:242-247.

MICKOLEIT, G. 1969. Vergleichend-anatomische Unter-suchungen an der pterothorakalen Pleurotergal-muskulatur der Neuropteria und Mecopteria (Insec-ta, Holometabola). Z. Morphol. Tiere 64:151-178.

MILLER, B. R., M. B. CRABTREE, AND H. M. SAVAGE.1996. Phylogeny of fourteen Culex mosquito spe-cies, including the Culex pipiens complex, inferredfrom the internal transcribed spacers of ribosomalDNA. Insect Mol. Biol. 5:93-107.

OLSEN, P. E., C. L. REMINGTON, B. CORNET, AND K. S.THOMSON. 1978. Cyclic change in Late Triassic la-custrine communities. Science 201:729-733.

OOSTERBROEK, P. F. L. S., AND G. COURTNEY. 1995.Phylogeny of nematocerous families of Diptera (In-secta). Zool. J. Linn. Soc. 115:267-311.

OOSTERBROEK, P. F. L. S., AND B. THEOWALD. 1991.Phylogeny of the Tipuloidea based on characters ofthe larvae and pupae (Diptera, Nematocera) with anindex to the literature except Tipulidae. Tijdschr. En-tomol. 134:211-267.

PAABO, S. 1990. Amplifying ancient DNA. Pages 159-166 in PCR protocols. A guide to methods and ap-plications (A. M. Innis, D. H. Gelfand, J. J. Sninsky,and T. J. White, eds.). Academic Press, San Diego,California.

PASHLEY, D. P., B. A. MCPHERON, AND E. A. ZIMMER.1993. Systematics of holometabolous insect ordersbased on 18S ribosomal RNA. Mol. Phylogenet.Evol. 2:132-142.

PAWLOWSKI, J., R. SZADZIEWSKI, D. KMIECIAK, J. FAHR-NI, AND G. BITTAR. 1996. Phylogeny of the infraor-der Culicomorpha (Diptera: Nematocera) based on28 RNA sequences. Syst. Entomol. 21:167-178.

PELANDAKIS, M., AND M. SOLIGNAC. 1993. Molecularphylogeny of Drosophila based on ribosomal RNAsequences. J. Mol. Evol. 37:525-543.

RAICH, T. J., J. L. ARCHER, M. A. ROBERTSON, W. J. TA-BACHNICK, AND B. J. BEATY. 1993. Polymerase chainreaction approaches to Culicoides (Diptera: Cera-


topogonidae) identification. J. Med. Entomol. 30:228-232.

ROHDENDORF, B. B. 1964. The historical developmentof the Diptera. Univ. Alberta Press, Edmonton.

ROUSSET, E, M. PELANDAKIS, AND M. SOLIGNAC. 1991.Evolution of compensatory substitutions through G-U intermediate state in Drosophila rRNA. Proc. Natl.Acad. Sci. USA 88:10032-10036.

RZHETSKY, A., AND M. NEI. 1995. Tests of applicabil-ity of several substitution models for DNA sequencedata. Mol. Biol. Evol. 12:131-151.

SAITOU, N., AND M. NEI. 1987. The neighbor-joiningmethod: A method for reconstructing phylogenetictrees. Mol. Biol. Evol. 4:406-425.

SANDER, K. 1996. Variants of embryonic patterningmechanisms in insects—Hymenoptera and Diptera.Semin. Cell Dev. Biol. 7:573-582.

SCHMIDT-OTT, U., G. M. TECHNAU, AND K. SANDER.1994. Expression of engrailed in embryos of a beetleand five dipteran species with special reference tothe terminal regions. Arch. Dev. Biol. 203:298-303.

SCHNARE, M. N., S. H. DAMBERGER, M. W. GRAY, ANDR. R. GUTELL. 1996. Comprehensive comparison ofstructural characteristics in eukaryotic cytoplasmat-ic large subunit (23S-like) ribosomal RNA. J. Mol.Biol. 256:701-719.

SCHREMMER, E 1951. Die Mundteile der Brachycer-enlarven und der Kopfbau der Larven von Stratio-mys chamaekon L. Osterr. Zool. Z. 3:326-397.

SCHUMANN, H. 1992. Systematische Gliederung derOrdnung Diptera mit besonderer Beriicksichtigungder in Deutschland vorkommenden Familien. Dtsch.Entomol. Z. N.F. 39:103-116.

SHAW, S. R., AND I. A. MEINERTZHAGEN. 1986. Evo-lutionary progression at synaptic connections madeby identified homologous neurones. Proc. Natl.Acad. Sci. USA 83:7961-7965.

SINCLAIR, B. J. 1992. A phylogenetic interpretation ofthe Brachycera (Diptera) based on the larval man-dible and associated mouthpart structures. Syst. En-tomol. 17:233-252.

SINCLAIR, B. J., J. M. CUMMING, AND D. M. WOOD.1994. Homology and phylogenetic implications ofmale genitalia in Diptera—Lower Brachycera. En-tomol. Scan. 24:407-432.

SMITH, J. J., AND G. L. BUSH. 1997. Phylogeny of thegenus Rhagoletis (Diptera: Tephritidae) inferredfrom DNA sequences of mitochondrial cytochromeoxidase II. Mol. Phylogenet. Evol. 7:33^43.

SNODGRASS, R. E. 1935. Principles of insect morphol-ogy. McGraw-Hill, New York.

SULLIVAN, J., K. E. HOLSINGER, AND C. SIMON. 1996.The effect of topology on estimates of among-siterate variation. J. Mol. Evol. 42:308-312.

SWOFFORD, D. L. 1993. PAUP: Phylogenetic analysisusing parsimony, version 3.0r. Illinois Natural His-tory Survey, Champaign.

TANG, J., K. PRUESS, E. W. CUPP, AND T. R. UNNASCH.1996. Molecular phylogeny and typing of blackflies(Diptera: Simuliidae) that serve as vectors of humanor bovine onchocerciasis. Med. Vet. Entomol. 10:228-234.

TAUTZ, D, J. M. HANCOCK, D. A. WEBB, C. TAUTZ, ANDG. A. DOVER. 1988. Complete sequences of therRNA genes of Drosophila melanogaster. Mol. Biol.Evol. 5:366-376.

TESKEY, H. J. 1981. Morphology and terminology—Larvae. Pages 65-88 in Manual of Nearctic Diptera,Volume 1 (J. E McAlpine, B. V. Peterson, G. E. Shew-ell, H. J. Teskey, J. R. Vockeroth, and D. M. Wood,eds.). Agriculture Canada, Ottawa.

TILLYARD, R. J. 1929. Permian Diptera from Warner'sBay, N.S.W. Nature 123:778-779.

ULRICH, H. 1991. The present state of comparativemorphology of thoracic musculature in Diptera. Int.Congr. Diptera 2:315-326.

UZZELL, T., AND K. W. CORBIN. 1971. Fitting discreteprobability distributions to evolutionary events. Sci-ence 172:1089-1096.

VOSSBRINCK, C. R., AND S. FRIEDMAN. 1989. A 28Sribosomal RNA phylogeny of certain cyclorrha-phous Diptera based upon a hypervariable region.Syst. Entomol. 14:417-431.

WHITE, M. J. D. 1949. Cytological evidence on thephylogeny and classification of the Diptera. Evolu-tion 3:252-260.

WHITING, M. E, J. C. CARPENTER, Q. D. WHEELER, ANDW. C. WHEELER. 1997. The Strepsiptera problem:Phylogeny of the holometabolous insect orders in-ferred from 18S and 28S ribosomal DNA sequencesand morphology. Syst. Biol. 46:1-68.

WlLLMANN, R. 1989. Rediscovered: Permotipula patri-cia, the oldest known fly. Naturwissenschaften 76:375-377.

WOOD, D. M., AND A. BORKENT. 1989. Phylogeny andclassification of the Nematocera. Pages 1333-1370 inManual of Nearctic Diptera, Volume 3 (J. F. McAlpineand D. M. Wood, eds.). Agriculture Canada, Ottawa.

WOODLEY, N. E. 1989. Phylogeny and classification ofthe "orthorrhaphous" Brachycera. Pages 371-1395in Manual of Nearctic Diptera, Volume 3 (J. F.McAlpine and D. M. Wood, eds.). Agricultural Can-ada, Ottawa.

WOOTTON, R. J., AND A. R. ENNOS. 1989. The impli-cations of function on the origin and homologies ofthe dipterous wing. Syst. Entomol. 14:507-520.

Wu, C.-L, AND W.-H. Li. 1985. Evidence for higherrates of nucleotide substitution in rodents than inman. Proc. Natl. Acad. Sci. USA 82:1741-1745.

XIONG, B., AND T. D. KOCHER. 1991. Comparison ofmitochondrial DNA sequences of seven morpho-species of black flies (Diptera: Simuliidae). Genome34:306-311.

YANG, Z. 1994. Maximum likelihood phylogenetic es-timation from DNA sequences with variable ratesover sites: Approximate methods. J. Mol. Evol. 39:306-314.

YANG, Z. 1996a. Among-site rate variation and its im-pact on phylogenetic analysis. Trends Evol. Ecol. 11:367-372.

YANG, Z. 1996b. PAML: Phylogenetic analysis bymaximum likelihood, version 1.2. Department of In-tegrative Biology, Univ. California, Berkeley.

Received 30 January 1995; accepted 19 July 1997Associate Editor: Chris Simon

1 9 9 7 FRIEDRICH AND TAUTZ—DIPTERAN PHYLOGENY 6 9 5

APPENDIX 1ALIGNMENT OF 28S Dl AND 28S D7 R D N A SEQUENCES

Taxon names are abbreviated to the first three letters of the genus name. Sites extracted for phylogeneticanalysis are marked with a star above the alignment. DS regions are indicated by lines above the alignment.Dots indicate positions identical to that in the heading sequence. Dash = gap; N = ambiguous site. Regions ofmajor disagreement between the final alignments and CLUSTAL V alignments are situated at positions 23-25,39-69, 75-120, 143-150, and 176-200 of the final 28S Dl alignment and at positions 39-104, 124-130, 177-179,188-190, 226-230, 253, 263, 301-303, and 312-315 of the final 28S D7 alignment. The alignments can be retrievedfrom the EBI server under accession numbers DS25432 (Dl) and DS25433 (D7).

85

Dro AGTAGCGGCGAGCGAAAAGAAAA-CAGTTCAGCACTAAGTCACTTTGTC TATAT--GGCAAATGTGAGATGCAGTGTAT-GGTab -A T . T TA- - . T . T . -TAA A. A -A.Bra CAGG.T-A. .CC CTCTCCG . . . GTCC .--. A .TGG . AGAGCri CAGG.T-T. .CC TCCAA.T .T. . .--.ATTGG.ABol CAGG. .-A. .CC C.CTCCA. . . GTCC.--. . .TGG.AG.GAna T.GG.T-A. .C C.TTC.G. .T--C. .A.C--.A.T.G.TA.G -..Dil C.GG.T-A CTC.AA .TA. . TT.G-AG TSyl C.GG. .-A. .CC TCTC.AA. .T--GT.G.-G. .TT.G.AGAPsy C.GG. .-A. .CC T. .CCA. .G.CT—CATA.AC- . . .C. . .TG.GTip N. . .T.GG. . -A. . .C T CT . . .ACTG--AT. . CTC . .T.G.ATG. .A..T..A....AA..Lim C.GG.T-A. .CC. .T. . .A CTC. .TC--TT.TA.C-GA. ATTG.TG . .A. . . .T. . . .CAAT.Tri C.GG.T-T. .CC C CTC .A.C .T--.TA.C-G. .TT.G.AG G A. .Aed T C.GG. .-G. .C GC. .GGG.GA. .CA CTGG T.T.TC. .CCC.G.T.C CT . .Cul T N GG. .-G. .C GC. .GGGTGGC.C ACTACCC— . .GTC . ACCC .G. T .C CT . .Chi T A C.GG.T-. . .CC. .T. . .GT. .GAT.A.A.G. .TA. .T-.T.T.T. .ATC. ..T.C AA..Sim GG. . -A. . .C GT. .AAT.GCAACT GT . . . — . .TTGT. .ATT C CA . .Pan C. .GG. . -G. .CC G.A. . C C . A . C . G — A T . T — C . .AT.G..G..A...T A..Arc N. . .C. .GG. .-G. .CC A. .C.GCG.C.G--G.G. ~ C . .ACCG. .G..A...T TAG. .Man C. .G. .TAT. .CC G.A. .C.GC. . .TG--. .A C.GC.GC.G TG. . . .TCG. .

Dl 86

Dro AGCGTCAATATTCTAGTATGAGAAATTAACG ATTTAAGTCCTTCTTAAATGAGGC-CA-TTTACCCATAGAGGGTGCCAGGCTab .ATA. .T. .TA ATT TA G. .C G. ..-..- AT. ...Bra .AA. . .T. .TA.. .TA.GCA.CG.TAAT.T G. .C A....G....G...-..T.AACri .AA. ..T. .TA. . .G..G.TGT.TTCA.TAC G. .C A....G....G...-..A.A..T...CBol GAA. . .T. .TA. . .TA CGGTAGC . . G..C A....G....G...-..T..AAna .AAA. .T. .TA. . .TA. . . ATAT .CAGC . TAT--G . .C A. ...G....G. ..-..- ADil .AA. ..T. .TACT.TA.. .A.AG.TGAT. . G..C A....G....G...-..T..A....GCSyl . .A. . .T.CTA. . ,CA. . . .TT.CGGC. .TAC--G. .C G....G...C..-Psy . .T. ..C. .TAA. .CA.TAC.AT.GA.G.T.TACG. .C A....G....G...-..TTip . .TA. .T. .TA. .CTAA. . TTTT TAT G . CC A. ...G....G. ..-..- CLim T.AA.TC. .TA. . .TA. . . .CAT.TGGT. . GC T.A. ...G....G. ..-..- CTri TAAA. .T TA. . . .TTCC.A. .C G..C A....G....G...-..T..AAed . .A. .TCG.TA. . .GCCG.A.CCGG.CGCT. G. .C. .. .T.AA. .AG. .. .T. . .-TTT.ACT.. . .G AT. ...Cul .AAT.TTG.CA. . . GCC . . .CCCGGCGC.GT G . CC . . . . T . AA . . AG. . . .T. . .-TTT.ACT AT. ...Chi . .G. . .C. .TA. . .T.C.CATAT. .ACT.AT-- AG...G....C...T..A.A CSim TT. .G.C .TA. . .CA. . .T.T. .T.AGTAT G. .C G --.A.A C AT....Pan . .GA.TC . T . ATCCTG. GATC .TG. AGCGC G A. . . . G. . . .G. . . - . .-AArc . .GA. .C.AT.ATCTCGG. .C.TCGCGGCGA G..C A. ...G....G. ..-..-Man . ..GT. .CG-C . TCTCGTCACCG . TACTC. T G.CC ...T.G....G..C.G...-.G-...T...GA

Dl 171 255

Dro CCGTATAACGTTAATGAT TACTA-GATGATGT TTCCAAAGAGTCGTGTTGCTTGATAGTGCAGCACTAAGTGGGTGTab G. .ACA.T. .T.AA .'. ......A. ..A. ...TT T.T....Bra GC . G C ATTATCAG . .TA.- . .A.GCT. T. . . .Cri G. .GCATTGTT.TG CTC.-. .A.GCT.Bol GCACGTTACCAG . .TA.-..A...T. ....G T....Ana GT.ACATTGT. .TA . .TAG- . .G. . .TC A. . . .T. . . .Dil GT . A C ATTATCAA . .TA.A. .A.GCTCSyl GTGAC. GTTGTCCGTTTGC TG . - . .A. .CT. C. .TT T. . . .Psy ....GG...TACC..CTGT A.T.T-.TG..CTA C..TTTip GTGAC AT . ATC. TG .TTAG- ..A. ..A. C.T.G T....Lim - . GC. ACTATTGTA . .T. .-. .GA. .T. .G.ATCTri . . . . GGT . AC . ATTGTTAT ..TA.-..A...T. .G..TT GAed G.G. .GCGCACCGA . GG . .-..A...C TG A A.Cul GG. . .GCGC .. .GG.G-..AA..T. TG A A.Chi TGTA.G. . .C.TG . GTA. - . TG . GCTATCAAC . ..T.G A T A.Sim TGCT A ..TG.-..G.T... .A. .T.G A C...A.Pan A-.ACGTTACTGAT CTTGG- . .G. . . C . C . .TTT G G TA r c GCGACCGTTGCGCGT AT . G . - . . G . . . C C . . . T C A G TMan GCGACGG .GG .CGGT G G . G . - . .G.GAC. C. . .TC G G C


Dl 256 284

Dro GTAAACTCCATCTAAAACTAAATATAACCTabBra GG G. . .Cri GGBol GGAna G G T...DilSyl GG C . . . .Psy GG T. . .Tip G G. . .Lim GGTri T. . .GG NNAed T G C...Cul T G C...Chi T GG T. . .Sim T G T . . .Pan GGArc GG G. . .Man GG T. . .

D7

Dro GAACGAAAGGGAATACGGTTCCAATTCCGTAACCTGTT-GAGTATCCGTTTGTTATTAAATA TGGGCCTCGTTab -A A A. A T.TAATBra A.G - A TA.T. .A.A . . . .T. .TA.TACCri .GG - A TA.T. .A.G ....T..T. .TACBol . .G - A TA.T. .A.A . . . .T. .TA.TACAna ..G T A GA....A. ... . TTCAT . ATATATGCDil A.G - A TA.T. .A.A .. . . .T TACSyl ..G - A TAAT. .A.TA . . . .T.C.T.AGTPsy .TG A. .- A A. .G . . . .TACT.GGCCTTGCTTip ..G -....G.A..C. .T.A.GAT . . . ATA . TAACATTTTAALim . .C - A. .C. .AAA.GAT . . .ATA.TAACATGCAATTri . .G - A AAT . . . . TTCTT . TAATTACTAed ..G C A G.G -.G....A CA .T. . .G. .CCATACCGGCCul ..G C A G.G -.G....A CA .T. . .GGTAATACCGGTAChi ..G N A..- A.... . . TTAT . T . AATTCGTTTSim ..G C A ...G - A CAT. ATGTG .CCGCC . ACATT . ATGTAGGTTATPan .GG C T G. . . .C.GCA.C.N.A AACA.G. C CTCGAAAGAGA-Arc NNNNN C T G... .C.GCA.C.G.A ACA.G. C CTCGCAAGAGA-Man ..G C T G... .C.GCA.C.G.A CAATAAT C.TT. .CTCGTTTAATAG

D7 86 170

Dro GCTCATCCTGGCAACAGGAACG ACCATAAAGAAGCCGTCGAGAGATATCGGAAGAGTTTTCTTTTab G CA...A .T A GBra A G T CTCri A A G T TBol A T A G T CTAna A T A. . .A G T A. . .CT.ADil A T A G T CTSyl A A G T G. .CTPsy TGCAAGTGTTCCGGTTTTA A.G.T GG . . . .G AGCG . .C. .ATip GATGANT AT T T G. .G. .G. . .C G. .T. . .C. .ALim GATGAAT AT T A G...C..T....G..T ATri NTAATTAAGGT A T A GAed AACGGTAGCGCCTT TG.A. . .A T. . . TCCTTTT. T . CG AA.A CCul ACACGGTAGCGCCTT TG.A. . .A T. . .TCCTTT. .T.CG AA.AChi CGAA TAA. . . .TC. .T. . .GA. . .GCT--CT. . . .G. . .C.T.AAT.G. . .G ASim CCATT ATGT. . .TC GA. . . G C C — .AT TA . . .G. .TG . . A. AAPan .T. .G. .GG. .T. . .CCA. .AA GA.C.GG. . .C G CCA.AArc .T. .G. .GG. .T. . .CCA. .AT GA.C.GG. . .C G CG. .GMan CGAGT ,T. .GA.GG. .T. . .CCA. .GT GG.C.G. . . .C. . . .C G.CCG


D7 171 255

Dro TCTGTTT-TATAGCCGTA-CTACCATGGAAGTCTTTCGCAGAGAGATATGGTAGATG-GGCTAGAAGAGCATGACATATACT-GTTab - TT...-T A A... -AA AG - . CBra ....C.A-A.C.AA....-.C TAT G ...... CT . G T. ......Cri G-A.C.AT. . . . -TA A TA. . .- .AT.G A T. ......Bol G-A.C.AT. . . . -TC A GA. . .- .AT.G T. ......Ana G-A.C.ATT. . .-TC TA GA. . .-AAT.G T. ......Dil G-A.C.AT. . . . -TC AT GA. . .- .AT.G- T. ......Syl GTA. C . A . T. . . -TA AAA TA. . . -A. T . G T . . T . - . .Psy - T....-T A A...-.A G.C....T.-.CTip -C. . . .A.A. .- A.AA- -T T. ......Lim - AA. C . -TA T TA. . A- . T T .......Tri - T..C. -TC C A GA..T-.T G...T....-.CAed -. .C. .T.AC.CTAG AT T . GACA. . . .G.T GT. .TA.A.T.CCul -. .C. . . .AC.CTGG AT C.GACAA. . .G.T GT. .TA.A.T.CChi -..C.AT....CG AT T.ACA.AT.G.T TA.A.T. .Sim -. .C. .TT. . .CGC AT GT.ATA. ...G.T G...TC.G.T.CPan A.-G.GCATTCG.-G.T. .C TC . . . .A. . . .G G. . .TTGG--AATGC C-G. .GT.G.G-.CArc A.-G.GC.TTCG.-G.T. .C TC..C.A....G G.. .TTGG—AA.GC C-G. .GT.G.G-.CMan . . . .CC.-G.GC.TTCG.-G.T TC. .A.A.A. .G TCGG--AA.GC C-G. . .T.G.G-.C

D7 256 340

Dro TGTGTCG-ATATTTTCTCCTCGGACCTTGAAAATTTATGGTGGGG--ACACGCAAAC--TTCTCAACAGGCCGTACCAATATCCGTab - G A.. .—TT — ....TBra A-G T C —C -- A GCri A- T C A. . .--C. .N T-- A GBol A-G T C A. . .--C --A A GAna A-G. ...A T C A...—C — A A GDil A-G. ...A T C --C --A...N A G G.Syl A-G. .C T C --C --A A GPsy CG. .CG.CT C. . . .CCA. . . — C . . . .T. . . .A- T. .ATip -G....C..C.T C AA.. .--TT T A GLim - C..C....A C A.. .—CT T A GTri - C A...--C T ....N....A GAed - C...T..T CG.A.ACT. ..--G ..C CTT G-ACul - C...T..T CG.A.ACT. ..--G ..C CTT G-AChi -G...AC....A.T C . . . . AAA... - - AT . . A- - T . . C . T CGASim T.-T. .CAA TA AT — T A G-APan G T-GG. .C. . .C CC.G.AGA. . .—C. . . .TGG.GGTG. .G.GC.G.TT CArc G C-G. . .C. . .C CC.G.AGA. . .--C. . . .-GG.GGTG. .G.GC.G.TT CMan G C-GG. .AC. . . ,TG C.G.TGA. . .ATGT. . .TGG.AATG. .G.GC.G.TT C N.

D7 353

698 SYSTEMATIC BIOLOGY VOL. 4 6

APPENDIX 2. Percentage base composition of dipteran and outgroup 28S Dl and D7 rDNA sequences insingle-stranded (SS) and double-stranded (DS) sites.

Taxon

DrosophilaTabanusBradysiaCrinodiplosisBolitophilaAnapausisDilophusSylvicolaPsychodaTipulaLimoniaTrichoceraAedesCulexChironomusSimuliumPanorpaArchaeopsyllaManduca

A

39.339.839.839.839.339.838.839.338.337.836.737.839.339.339.839.334.734.732.7

T

25.528.124.024.024.524.524.525.026.526.026.026.524.525.026.525.526.525.525.5

SS

G

21.921.423.022.423.022.423.023.023.022.421.923.523.523.021.424.025.025.027.0

C

13.310.713.313.813.313.313.812.812.213.815.312.212.812.812.211.213.814.814.8

A

22.624.421.523.021.523.022.620.720.022.223.020.719.319.622.222.215.615.614.4

T

27.429.626.727.427.428.527.427.825.227.828.926.327.026.728.528.520.719.319.3

DS

G

26.324.126.725.626.725.926.327.027.825.925.627.830.029.627.026.734.434.835.9

C

23.721.925.224.124.422.623.724.427.024.122.625.223.724.122.222.629.330.430.4

EVOLUTION AND PHYLOGENY OF THE DIPTERA: A … · syst. biol. 46(4):674-698, 1997 evolution and...

Documents

Transcript of EVOLUTION AND PHYLOGENY OF THE DIPTERA: A … · syst. biol. 46(4):674-698, 1997 evolution and...