signal through the noise? phylogeny of the tachinidae

Systematic Entomology (2014), 39, 335–353

Signal through the noise? Phylogeny of the Tachinidae(Diptera) as inferred from morphological evidence

P I E R F I L I P P O C E R R E T T I 1,2, J A M E S E . O ’ H A R A3,D . M O N T G O M E R Y W O O D3, H I R O S H I S H I M A4, D I E G O J .I N C L A N5 and J O H N O . S T I R E M A N III6

1Department of Biology and Biotechnology ‘Charles Darwin’, University of Rome ‘Sapienza’, Rome, Italy, 2Centro NazionaleBiodiversita Forestale – Corpo Forestale dello Stato, Verona, Italy, 3Canadian National Collection of Insects, Agriculture andAgri-Food Canada, Ottawa, Canada, 4Kyushu University Museum, Kyushu University, Fukuoka, Japan, 5DAFNAE-Entomology,University of Padova, Padova, Italy and 6Department of Biological Sciences, Wright State University, Dayton, OH, U.S.A.

Abstract. The oestroid family Tachinidae represents one of the most diverselineages of insect parasitoids. Despite their broad distribution, diversity and importantrole as biological control agents, the phylogeny of this family remains poorlyknown. Here, we review the history of tachinid systematics and present the firstquantitative phylogenetic analysis of the family based on morphological data.Cladistic analyses were conducted using 135 morphological characters from 492species belonging to 180 tachinid genera, including the four currently recognizedsubfamilies (Dexiinae, Exoristinae, Phasiinae, Tachininae) and all major tribes. Weused characters of eggs, first-instar larvae and adults of both sexes. We examined theeffects of implied weighting by reanalysing the data with varying concavity factors.Our analysis generally supports the subfamily groupings Dexiinae + Phasiinae andTachininae + Exoristinae, with only the Exoristinae and the Phasiinae reconstructedas monophyletic assemblages under a wide range of weighting schemes. Underthese conditions, the Dexiinae, which were previously considered a well-establishedmonophyletic assemblage, are reconstructed as being paraphyletic with respect tothe Phasiinae. The Tachininae are reconstructed as a paraphyletic grade fromwhich the monophyletic Exoristinae arose. The Exoristinae are reconstructed as amonophyletic lineage, but phylogenetic relationships within the subfamily are largelyunresolved. We further explored the evolution of oviposition strategy and found thatthe oviparous groups are nested within ovolarviparous assemblages, suggesting thatovipary may have evolved several times independently from ovolarviparous ancestors.This counterintuitive pattern is a novel hypothesis suggested by the results of thisanalysis. Finally, two major patterns emerge when considering host associations acrossour phylogeny under equal weights: (i) although more than 60% of tachinids areparasitoids of Lepidoptera larvae, none of the basal clades is unambiguously associatedwith Lepidoptera as a primitive condition, suggesting that tachinids were slow tocolonize these hosts, but then radiated extensively on them; and (ii) there is generalagreement between host use and monophyly of the major lineages.

Correspondence: Pierfilippo Cerretti, Department of Biology andBiotechnology ‘Charles Darwin’, University of Rome ‘Sapienza’,Piazzale A. Moro 5, I-00185 Rome, Italy. E-mail: [email protected]

Introduction

Within the order Diptera, the oestroid family Tachinidae re-

presents the largest lineage of endoparasitoids. Tachinid larvae

develop within their hosts (mainly insects, but also chilopods

© 2014 The Royal Entomological Society 335

336 P. Cerretti et al.

and scorpions) by feeding first on haemolymph and non-vital tissues and organs, and then on vital parts, with fewexceptions killing their hosts before completing their larvaldevelopment (Herting, 1960; Wood, 1987b; Mellini, 1991;Tschorsnig & Richter, 1998; Dindo, 2011). This finely tunedway of life is undoubtedly the most important ecologicaltrait that distinguishes these flies from other oestroids,which are mostly sarcophagous, coprophagous, parasitic onvertebrates, or kleptoparasitic on Hymenoptera (with only afew being parasitoids, notably the Rhinophoridae and certainSarcophagidae) (Feener & Brown, 1997).

According to the recent phylogenetic reconstruction ofWiegmann et al. (2011), cyclorrhaphan Diptera had an earlyTertiary explosive radiation. Tachinids were probably radiat-ing rapidly by the mid-Tertiary, becoming widespread andabundant in virtually all terrestrial ecosystems (Wood, 1987b;Tschorsnig & Richter, 1998). There are nearly 8200 namedtachinid species (and thousands more undescribed), accountingfor about 6.5% of the total diversity of Diptera. Over 90% oftachinid species attack phytophagous insects, and about 60% ofthese are parasitoids of ditrysian Lepidoptera (Stireman et al.,2006; Cerretti, 2010), which comprise the most diverse groupof plant-feeding organisms.

Given their broad distribution, great diversity and ecologicalrole as enemies of primarily herbivorous insects, tachinidshave attracted considerable attention from basic and appliedresearchers, as well as taxonomists (Stireman et al., 2006;O’Hara, 2013b). However, the classification and identificationof tachinids have posed formidable challenges, and thesehave hindered the understanding of their biology, ecology andpotential usefulness in controlling pests in managed systems.The group is widely regarded as one of the more difficultfamilies of Diptera in terms of identification (Crosskey, 1976),and a universal classification scheme remains elusive despiterecent progress.

Here, as part of a larger initiative to understand thephylogeny and evolution of the family Tachinidae on a globalscale (Stireman et al., 2013), we present the first quantitativephylogenetic analysis of the family based on morphologicaldata. We briefly review the history of tachinid classificationand the hypothesized relationships that key workers haveadvanced, noting some of the important character statesthat have been used to define various tachinid clades. Wethen present an analysis of the phylogenetic relationshipsof the Tachinidae based on a broad range of taxa (180genera) and a diverse collection of morphological characters(135). Our aim is to assess the strength of the phylogeneticsignal of morphology in defining major tachinid clades andtheir relationships, despite the great degree of morphologicalhomoplasy in the family. We use our resulting phylogeneticreconstruction to evaluate previous hypothesized relationshipsand classifications, and to propose novel ones that canbe evaluated with subsequent morphological and molecularphylogenetic studies. Finally, we examine the implicationsof our phylogenetic results for understanding evolutionarypatterns of reproductive strategy and host use within thefamily.

History of the classification of Tachinidae

The purpose of this section is to briefly review thehistory of tachinid classification and to consider some ofthe morphological characters that have been used to definefundamental groups in recent classification schemes. For amore comprehensive and detailed account of the history oftachinid classification, see O’Hara (2013b).

Coquillett (1897), in his revision of the Tachinidae ofAmerica north of Mexico, recognized the Tachinidae almostas it is known today but without the subfamily Dexiinae. Atabout the same time, in Europe, Girschner (1893, 1896) waselaborating on a system of chaetotaxy developed by Osten-Sacken (1881, 1884) and used the presence of ‘hypopleural’(meral) setae to delimit the Tachinidae, a concept for the familyequalling that of the present-day superfamily Oestroidea. Thisclassification was largely followed in the influencial Katalogder palaarktischen Dipteren (Bezzi & Stein, 1907). Not untilthe discovery of the value of the ‘metanotum’ (subscutellum;character state 60:1 below) by Malloch (1923) did authorsbegin to use this character to delimit the Tachinidae alongmodern lines.

The progressive elements of Girschner (1893, 1896) werefurther refined by Villeneuve (1924, 1933), whose work was,in turn, further advanced by Mesnil (1939) in his Essai sur lesTachinaires .

Mesnil (1939) recognized six subfamilies of the Tachinidae:Dexiinae, Phasiinae, ‘Larvaevorinae’ (Tachininae), Ameniinae(currently in Calliphoridae), ‘Salmaciinae’ and ‘Phorocerinae’(the last two comprising present-day Exoristinae). He broughttogether in his Phorocerinae those tachinids possessing a hairedprosternum (41:0) (i.e. character 41, state 0 in the present study,see File S2) and a short and fine ‘prealar’ (first postsuturalsupra-alar) seta (49:1). The Phorocerinae were subdividedinto the ‘Phorocerini’ (Exoristini), Blondeliini and ‘Crocutini’(Siphonini). Each was further characterized in part by: thePhorocerini by vein M (as ‘4e’) having an angular bend (71:3)and shadow fold (72:1), and subapical scutellar setae divergent(character not coded in analysis below); the Blondeliini byvein M having a rounded bend and no shadow fold (71:1), andsubapical scutellar setae divergent; and Crocutini by vein Mhaving a rounded bend and no shadow fold (71:1, 2; 72:0), andsubapical scutellar setae convergent. The current concepts ofthe Exoristini and Blondeliini largely date from Mesnil (1939).The Exoristini are probably monophyletic (Stireman, 2002;Tachi & Shima, 2010), but the Blondeliini, according to Wood(1985), are probably not. Wood (1985: 7) noted that the smallprealar seta (49:1) ‘is probably a plesiomorphic character’.

Villeneuve’s contemporary in the New World wasTownsend, a prolific describer of tachinid genera and species.Townsend’s greatest achievement was his 12-volume Manualof Myiology in which the genera of Tachinidae and allieswere diagnosed and assigned to a multitude of tribes andfamilies. This work was also Townsend’s greatest downfall,for unrelated genera were frequently grouped together intotribes. The catalogue of the Tachinidae of America south ofthe United States by Guimaraes (1971) did not significantly

© 2014 The Royal Entomological Society, Systematic Entomology, 39, 335–353

Phylogeny of Tachinidae 337

improve the classification of the region’s Tachinidae. Thefauna of America north of Mexico has fared better in thesuccessive catalogues of Sabrosky & Arnaud (1965) andO’Hara & Wood (2004), the latter much influenced by theclassification of Herting (1984).

Mesnil dominated tachinid taxonomy in Europe for decadesafter his Essai, particularly through instalments to DieFliegen der Palaearktischen Region (Mesnil, 1944–1975).During this time, Herting published early works on thefemale postabdomen of calyptrate flies and the biology ofWest Palaearctic Tachinidae (Herting, 1957, 1960). Implicitin these works was an attempt to order the Tachinidaephylogenetically. The Salmaciinae and Phorocerinae of Mesnil(1939) [also Mesnil (1944–1956, 1956–1965) as Salmaciiniand Phorocerini] were restructured to form the near-modernExoristinae with tribes Exoristini, Blondeliini, Acemyini (as‘Acemyiini’), Siphonini, Ethillini, Winthemiini and Goniini.The Goniini were later split to become the modern Goniiniand Eryciini, with the former restricted to species producingmicrotype eggs (6:1) that are ingested by their hosts (9:1)(Mesnil, 1975a,b; Herting, 1984). Thus constituted, the Goniiniare generally considered monophyletic; the Eryciini areregarded as paraphyletic. The Siphonini were later moved tothe Tachininae (Herting, 1984).

The modern concept of the Dexiinae, with the main tribesof Dexiini, Voriini and Dufouriini, dates from Herting (1957,1960). Herting used the as-yet-unpublished discoveries ofVerbeke (1962, 1963) to define the subfamily accordingto unique features of the male terminalia: phallus witha hinged connection between basiphallus and distiphallus(Verbeke’s ‘type II’ connection) (115:1) and pregonite strap-like and connective (‘type C’ of Vebeke’s ‘posterior paramere’)(103:1).

In the first instalment on the ‘Larvaevorini oder Tachinini’portion of Die Fliegen der Palaearktischen Region , Mesnil(1966) revised the classification of the tachinids he had dealtwith in previous instalments and presented the classification hewould follow for the Tachinini. The higher classification wassummarized in a table (Mesnil, 1966: 882) wherein present-dayTachinidae were treated as a subfamily with six tribes: Voriini,Dexiini and Tachinini (on the left in the table) and Exoristini,Goniini and Phasiini (on the right in the table). Both Mesnil(1966) and Herting (1966) expressed the same views about thetwo main branches of Tachinidae:

(1) The three groups on the right in Mesnil’s table areoviparous, or originally so, producing planoconvex eggswith a thick curved dorsal surface and a thin and flatunderside (state 2:1).

(2) The three groups on the left in Mesnil’s table areovolarviparous, producing eggs with a thin membrane andwithout differentiation into an upper and a lower surface(2:0).

Mesnil (1966) recognized the Tachinini as the most diversegroup of tachinids and subdivided it into about 30 subtribes.Herting (1984) treated these subtribes as tribes of Tachininae,

but reduced them in number and recognized 14 in thePalaearctic Region.

The treatments of Old World non-Palaearctic Tachinidae byCrosskey (1973, 1976, 1980, 1984) and Cantrell & Crosskey(1989) were conservative in their taxonomic approach,treating the Voriini in the Tachininae and the Neaerini andSiphonini in the Exoristinae, recognizing the Dufouriinaeas a subfamily, and not dividing the Goniini and Eryciiniaccording to egg type.

The ‘gold standard’ of classifications remains that of Herting(1984), which was followed by Herting & Dely-Draskovits(1993). Although this classification specifically applies to thePalaearctic Region, its hierarchy of subfamilies and tribesprovides a good framework for a global classification (Table 1),at least at present (see the Discussion). Only portions of theclassification are stable from a phylogenetic perspective, anda greater understanding of tachinid relationships may indicatewhere changes should be made to better reflect the evolutionaryhistory of the Tachinidae.

Recently, Stireman (2002) and Tachi & Shima (2010)provided the first attempts at reconstructing phylogeneticrelationships within the Tachinidae using molecular data.Stireman (2002) used two genes, EF1α and 28S rDNA, whileTachi & Shima (2010) used four (16S, 18S, 28S rDNA andwhite). These studies focused on the subfamily Exoristinae,generally supporting its monophyly and that of its constituenttribes (with some exceptions). Notably, Stireman (2002) failedto reconstruct a monophyletic Goniini (the microtype egglayers, state 6:1), while Tachi and Shima did find supportfor this clade. Both studies found a similar progression ofclades, with Winthemiini basal, then Exoristini branching off,followed by the Blondeliini and, finally, the Eryciini andGoniini forming a clade. As suggested by this progression,Tachi & Shima (2010) found, using character reconstruction,that ovolarviparous forms (state 7:1) have arisen repeatedlyfrom oviparous ancestors (state 7:0).

Materials and methods

Monophyly of terminal taxa

All terminal taxa included in the analysis are assumed, apriori, to be valid genera following the definitions given in allmajor monographs (Crosskey, 1984; Tschorsnig, 1985; Wood,1987b; Tschorsnig & Herting, 1994; Tschorsnig & Richter,1998; Cerretti, 2010; Cerretti et al., 2012a). We also assumedthat each species chosen for character coding is congenericwith the type species of the genus in which it is placed. It ispossible that some of these genera are in fact not monophyletic,but assessing this is beyond the scope of this study.

Selection of taxa, coding of character states and terminology

Specimens belonging to 180 tachinid genera (total of 492species examined) were selected for character state coding



Table 1. Subfamilial and tribal level classification of the Tachinidae, excluding tribes given in Table 2.

Dexinae Phasiinae Tachininae Exoristinae

Dexiini Catharosiini Bigonichetini Myiophasiini AnacamptomyiiniDoleschallini Cylindromyiini Brachymerini Myiotrixini BlondeliiniDufouriinia Gymnosomatinib Ernestiinic Neaerini EryciiniEpigrimyiini Hermyini Germariini Nemoraeini EthilliniRutiliini Leucostomatini Germariochaetini Occisorini ExoristiniSophiini Parerigonini Glaurocarini Ormiini GoniiniTelothyriini Phasiini Graphogastrini Pelatachinini ThrixioniniUramyini Iceliini Polideini WinthemiiniVoriini Leskiini Protohystriciini

Macquartiini SiphoniniMegaprosopini TachininiMinthoini

a Including Freraeini.b Including Trichopodini.c Including Linnaemyini and Loewiini.Taxa with at least one representative included in the present study are shown in bold.

(File S1). These taxa represent the four currently recognizedtachinid subfamilies (Dexiinae, Exoristinae, Phasiinae, Tachin-inae) and all of the major widespread tribes (Tables 1, 2). Blowflies of the genera Calliphora Robineau-Desvoidy, LuciliaRobineau-Desvoidy, Pollenia Robineau-Desvoidy, the rhiniidRhyncomya Robineau-Desvoidy and the rhinophorids PhytoRobineau-Desvoidy and Rhinophora Robineau-Desvoidy (FileS1) were selected as outgroups, as previous analyses sug-gest a close affinity between at least some members of theRhinophoridae, Calliphoridae and Tachinidae (Wood, 1987a;Kutty et al., 2010; Marinho et al., 2012; Singh & Wells, 2013).Wherever possible the egg, first-instar larva and adults of bothsexes were examined for each species, or information on allstages was retrieved from reliable sources in the literature (FileS1). The rhinophorid and most of the tachinid specimens cho-sen for coding were identified to species by one of us (P.C.),and the calliphorids were identified by Knut Rognes (Oslo,Norway). These specimens are deposited in P.C.’s collectionat the Museum of Zoology, University of Rome ‘Sapienza’.Specimens of a few rarer species were examined and codedfrom other collections: CNC, Canadian National Collection ofInsects, Agriculture and Agri-Food Canada (Ottawa, Canada);SMNS, Staatliches Museum fur Naturkunde (Stuttgart, Ger-many); and TAU, Department of Zoology, Tel Aviv University(Tel Aviv, Israel) (File S1). The majority of species are fromthe Palaearctic Region, but species from other regions areincluded to provide a global representation of tachinid tribes(File S1).

We selected 135 morphological characters, distributedas follows: egg, nine; first-instar larva, four; and adult,122 (including 36 in the male postabdomen and 10 in thefemale postabdomen) (File S2). Morphological terminol-ogy essentially follows that proposed by McAlpine (1981)(adult external morphology and chaetotaxy), Tschorsnig(1985) and Tschorsnig & Richter (1998) (male and femaleterminalia) (see also File S2 and explanatory figurestherein).

Table 2. List of exemplar ‘enigmatic’ taxa and their subfamilyplacement by different authors.

Tribe/genusHerting1984

Tschorsnig1985

O’Hara & Wood2004

Imitomyiini P P DEutherini P P DAcemyini E E TMasiphyini – ?P EStrongygastrini P Pa DEuthelairini – Eb TPalpostomatini T T DLitophasia P D –

a Including Rondaniooestrini.b As Neominthoini.P, Phasiinae; D, Dexinae; T, Tachininae; E, Exoristinae.Taxa with at least one representative included in the present study areshown in bold.

Phylogenetic analysis

Cladistic analyses using parsimony were conducted witha data matrix of the states of 135 characters belonging tospecies of 180 tachinid genera, as well as three calliphorid,one rhiniid and two rhinophorid genera (Files S1–S3). Thematrix was produced in mesquite version 2.74 (Maddison &Maddison, 2010) and analysed in tnt version 1.1 (Goloboffet al., 2003). The ‘New Technology Search’ (NTS) optionswere used because of the large number of taxa. To find the mostparsimonious trees, all analyses were run with the followingparameters: general RAM of 1000 megabytes and memory setto hold 600 000 trees; multistate characters were treated asunordered and zero-length branches collapsed. The ‘drivensearch’ was performed by using 14 replications as the startingpoint for each hit [which is the maximum value as suggestedby the program for the most exhaustive search, i.e. when theinitial level is set as 100 from ‘set initial level’ dialog box(= maximum value) or 10 from command line (= maximum



value)]. Each replicate, initially autoconstrained (previous andWagner), was conducted with: constraint and random basedsectorial searches, no ratchet, tree-drifting (six iterations) andtree-fusing (10 rounds); finding minimum length trees 30 times,not consensusing trees during search, multiplying trees byfusing after hitting best score and saving one tree per replicate.tnt commands from command line were as follows: ‘hold600000; xmult = level 10 hits 30 rss css fuse 10 drift 6;’.

Analyses were conducted with equal weighting and also withimplied weights under the default function k /(es + k ), wherek = concavity constant and es = extra steps, with k varyingcontinuously between 1 and 10, then for k = 15, 20, 40 (tntcommand: hold 600000; Piwe = 1, 2, 3, . . . n; xmult = level 10hits 30;) in order to examine and compare tree length, total fitand variation in tree topology under various weighting schemes(File S5). Total fit was calculated for most parsimonious trees(MPT) obtained under equal weights for each k -value testedand compared with total fit of MPTs obtained under impliedweighting for corresponding k -values (File S5).

Bremer support values were calculated in tnt from 10 000trees up to 10 steps longer than the shortest trees obtained froma ‘traditional search’, using the ‘trees from RAM’ setting. Ajackknife resampling was carried out with a traditional searchproducing 1000 replicates each of 100 random taxa additionsubreplicates applying tree bisection-reconnection branchswapping and saving 10 trees per replication. Jackkniferemoval probability was set at the default value of 0.36,and resampling percentiles were calculated as frequencydifferences [group present/contradicted (GC)]. Characteroptimization (using unambiguous changes only) was carriedout in winclada version 1.00.08 (Nixon, 1999). Ancestralstate parsimony reconstruction of host associations was carriedout in mesquite version 2.74 (Maddison & Maddison, 2010).

Results

Analysis under equal weights

The analysis under equal weights and with all charac-ters unordered produced 87 most-parsimonious trees (treelength = 1109 steps, consistency index for matrix, CI = 0.181,retention index for matrix, RI = 0.731): the strict consensustree is shown in Fig. 1. We selected one of the three shortesttrees with the highest total fit under the whole range of thek -values tested (Fig. 2) (i.e. tree number 7 of the nexus file,File S6) to depict inferred character state changes (File S4), asdiscussed in the following.

Tachinid monophyly. The Tachinidae (clade A) are recon-structed as a monophyletic assemblage based on eight non-homoplasious apomorphies (1:1, egg dorsal plastron absent;3:2, egg shell uncoloured; 7:1, ovolarviparous; 10:1, first-instarmandibles reduced; 12:1, first-instar labrum strongly devel-oped; 60:1, subscutellum strongly convex; 133:1, female ovisacvery long and generally spiralled when full; 134:1, femaleovisac enveloped by a well-developed tracheal network) and

two homoplasious apomorphies (80:0; 120:1). Monophyly iswell supported by a Bremer support value of 10 and a resam-pling probability of 95. The calliphorid outgroups formed aparaphyletic grade, consistent with the findings of Rognes(1997), and the two rhinophorids + Rhyncomya (Rhiniidae)were reconstructed as sister group to Tachinidae based on threehomoplasious apomorphies (51:1; 52:0; 71:1).

Basal taxa. The Coleoptera-parasitizing clade B(Gnadochaeta Macquart + Palpostomatini) is reconstructed assister group of clade C (i.e. all the remaining tachinids). Mono-phyly of clade C is weakly supported by two homoplasiousapomorphies (30:0; 43:2). The macquartine MacroprosopaBrauer & Bergenstamm (clade D) is sister group ofclade E ((Ormiini + (Dexiinae + Phasiinae)) + (Tachini-nae + Exoristinae)) based on three homoplasious apomorphies(57:0; 79:1; 111:2).

Monophyly of clade B is well supported by two homopla-sious apomorphies (33:1, 2; 111:1) and one non-homoplasiousapomorphy (105:1, medial surface of male pregonite mem-branous). Taxa comprising this clade have been placed in theTachininae by several authors and here occupy the most basalposition within the Tachinidae.

Clade F (Ormiini + (Dexiinae + Phasiinae)), which is sup-ported by three homoplasious apomorphies (51:0; 65:1; 106:1),is divided into two subclades: clade G (the tribe Ormiini, hererepresented by Aulacephala Macquart and Therobia Brauer)and clade H (Dexiinae + Phasiinae). Monophyly of clade Gis strongly supported by five homoplasious (13:1; 85:0; 88:2;96:1; 113:1) and three non-homoplasious apomorphies (14:1,ocelli missing or posterior pair strongly reduced; 42:1, femalewith a very swollen and convex prosternal region; 110:1,basiphallus dorsobasally thickened (Tschorsnig, 1985)).

The Dexiinae. Clade H (Dexiinae + Phasiinae) is sup-ported by two homoplasious (109:1; 114:1) and two non-homoplasious (115:1, dorsal connection between basiphallusand distiphallus membranous; 116:1, basiphallus and distiphal-lus joined at a right angle) apomorphies. The dorsal mem-branous connection between basiphallus and distiphallus haslong been considered a clear synapomorphy of the subfamilyDexiinae (see historical review). In the present analysis, thischaracter state is interpreted as having undergone a reversal,being secondarily lost in most Phasiinae.

Within clade H, the monophyly of clade I (i.e. the currentsubfamily Dexiinae excluding the Dufouriini) is supported bythree homoplasious apomorphies (34:0; 53:0; 95:1). Althoughadditional research is needed to corroborate the phylogeneticposition of the genus Phyllomya Robineau-Desvoidy as sistertaxon of clade K (Stomina Robineau-Desvoidy + Dexiini), itis interesting to point out that the Dexiini (clade L) are nestedwithin the Voriini sensu Herting (1984) and that the genusStomina takes up a position as sister group of the Dexiinibased on six homoplasious apomorphies (8:1; 13:1; 30:1; 51:1;65:0; 85:0). The monophyly of clade L is well supportedby four synapomorphic character states: one homoplasious



95/4

62/395/10

/371

/292

/4

/2/2

/2

/3

/3

99/8

/6

/2

59/665/4

/3 91/459/2

/275/2

/255/2

/2

/2 89/3 79/2

61/3

/3

/2

/3

/2/2 /3

67/296/5

59/2

/3

96/5/2

/271/4

59/

/4

92/6

64/2

/2

/2

90/8

93/7

62/3

81/2

/2

/3

61/3

/2

70/2

98/652/

Pel

atac

hina

Rondania

Rossimyiops

Bill

aea

Met

acem

yia

Baum

haueria

Chr

ysos

omop

sis

Besseria

Goniocera

arohpillaC

Cia

la

Demoticus

Thelym

orpha

Pol

leni

a

Campy

loche

ta

ParatrypheraA

ulac

epha

la

Ceranthia

Eut

rixop

sis

Hypovo

ria

Hemyda

Brachymera

Eloceria

Esth

eria

Phy

to

Winthemia

aireteleP

Freraea

Pseudogonia

Exorista

Siphona

Atylomyia

Rhaphiochaeta

aina

znall

apS

Pales

Cle

onic

e

Brullaea

Synactia

Clairvillia

Ger

mar

ia

spot cel poeN

Sto

min

a

Microphthalm

a

Erynnia

Gonia

Trichopoda

Cer

acia

Mintho

Phorocera

Eumeella

Gymnosoma

Hyleorus

Cerom

ya

Tria

rthr

iaRed

tenb

ache

ria

Rhi

noph

ora

Mel

ison

eura

Lophosia

ErythroceraO

cytata

Eugymnopeza

Sch

iner

ia

Carcelia

Phebe

llia

Leskia

Parerigone

Drino

Dex

ia

Actia

Phy

llom

ya

Clytiomya

Bothria

Platymya

Xysta

Bithia

Trig

onos

pila

Isto

chet

a

Pros

ena

Hya

lurg

us

Ace

mya

Anthomyiopsis

Senometopia

Petagnia

Litophasia

Tryphera

Opesia

Ptesiomyia

Trixa

Ethilla

Lydella

Tac

hina

Frontina

Ectophasia

Smidtia

Clausicella

Phryx

e

Onychogonia

Aphria

Gaedia

Ther

obia

Elodia

Erioth

rix

Staurochaeta

Sturmia

Allophorocera

Oswaldia

Nealsomyia

Weberia

Lomachantha

Voria

Myxexoristops

Luci

lia

Phytomyptera

Zeux

ia

Phania

Pan

zeria

Pandelleia

Lydina

Bessa

Heraultia

Plesina

Nem

orae

a

Dolichocolon

Blepha

romyia

Blon

delia

Hebia

Athrycia

Chetoptilia

Pseudomintho

Gym

noph

ryxe

Graphogaster

Zenillia

Tlephusa

Cylindromyia

Phorinia

Phorocerosoma

Dufouria

ImitomyiaLypha

Prosethilla

Comps

ilura

Pachystylum

CatagoniaPeria

rchiclops

Catharosia

Pseudoperic

haeta

Nemorilla

Thel

aira

Thecocarcelia

Sarrom

yia

Linn

aem

ya

Entomophaga

Eumea

Rhy

ncom

ya

Dexiosom

a

area

eN

Lecanipa

ChetogenaHermya

Lige

ria

Euth

era

Hubneria

Nilea

Phryno

CyzenisMicrosoma

Erycia

Medina

Hyperaea

Masicera

Perisc

epsia

Loewia

Euexorista

Pic

coni

a

Gna

doch

aeta

Leiophora

Masistylum

Palesisa

Adm

ontia

Prosopea

Mac

ropr

osop

a

Zaira

Strongygaster

Pseudopachystylum

Chetina

Phasia

OutgroupsE

xo

r is

t in

ae

Tachininae

Ph

as

iina

e

Dexiinae

Fig. 1. Strict consensus cladogram obtained from 87 most parsimonious trees under equal weights. Branch support: left, maximum parsimony %jackknife (cut 50%); right, Bremer support values.

(71:3) and three non-homoplasious (97:1, ear-shaped medialside of surstylus; 101:1, boundary between hypandrium andhypandrial apodeme very distinct; 122:1, acrophallus with agranulate or lamellate structure).

Clade M is supported by three homoplasious apomorphies(47:3; 48:0; 79:3,4). The Voria-group (clade N) is wellsupported by four homoplasious apomorphies (43:4; 45:3;56:0; 73:0) and one non-homoplasious apomorphy (76:1,crossvein dm-cu exceptionally oblique).

The Phasiinae and Dufourini. Clade O consists of the Phasi-inae + Dufouriini. The monophyly of clade O is supported bytwo homoplasious apomorphies (73:3; 90:1).

The Dufouriini sensu Herting (1984) (Chetoptilia Rondani,Dufouria Robineau-Desvoidy, Rondania Robineau-Desvoidy,Pandelleia Villeneuve, Microsoma Macquart, EugymnopezaTownsend, Freraea Robineau-Desvoidy), which are par-asitoids of adult beetles, are paraphyletic with respect tothe Phasiinae sensu Herting (1984) (clade Q). Clade P



A

B

D

CE

F

G

H

I

O

J

M

KL

N

Q

P

R ST

U

VW

X

YZ

AA

ABAC

ADAE

αα

AF AG

AHAI

AL

AJ

AK

AM

AN AP

AQ

AOAS

AR

ATAU

AVAXAZ

AWAY

BB

BA

BC

BD

BEBF

BH

BG

BI

BJ

Hermya

Clausicella

ThecocarceliaBesseria

Neoplectops

Gymnophryxe

Pel

atac

hina

Gna

doch

aeta

Hyleorus

Drino

airen

ihc

S

Smidtia

Eumea

Weberia

Mac

ropr

osop

a

Petagnia

Eloceria

Triarthria

Prosethilla

Rhy

ncom

ya

Pandelleia

Dexiosom

a

Perisc

epsia

Clytiomya

Goniocera

Gymnosoma

AphriaDem

oticus

Winthemia

Rhi

noph

ora

Phy

llom

ya

Phytomyptera

Phorinia

Platymya

Isto

chet

a

Gaedia

Blephar

omyia

Hyperaea

airet

ele

P

Nem

orae

a

Masistylum

Pol

leni

a

Mintho

Aul

acep

hala

TrichopodaC

hrys

osom

opsi

s

Chetoptilia

Met

acem

yia

Pic

coni

a Microphthalm

a

Graphogaster

Zeux

ia

Nilea

Phania

Leskia

Ethilla

Dex

ia

Phasia

Ptesiomyia

Bithia

CatharosiaC

erac

ia

Tlep

husa

Phorocerosoma

Spallanzania

Pachystylum

Onychogonia

Synactia

Ectophasia

Cerom

ya

Bessa

Cylindromyia

Siphona

Hypovoria

Microsoma

Loewia

Baum

haueria

Lophosia

Brachymera

Eryc

ia

Brullaea

Anthomyiopsis

Frontina

Litophasia

Cia

la

Chetogena

Pan

zeria

Trig

onos

pila

Erynnia

Carcelia

Esth

eria

Bothria

Hya

lurg

us

Hubneria

Dufouria

Leiop

hora

Ther

obia

Oswald

ia

Lydella

Lige

ria

Tryphera

Erythrocera

Pseudogonia

Hemyda

Freraea

Leca

nipa

Xysta

ZairaEu

ther

a

Eugymnopeza

Eumeella

Dolichocolon

Lom

acha

ntha

Phebellia

Sturmia

Euexorista

Clairvillia

Mel

ison

eura

Medina

Lypha

Pales

Atylomyia

Catagonia

Allophorocera

Paratryphera

Hebia

Pseudopachystylum

RondaniaA

dmon

tia

Imitomyia

Elodia

Cle

onic

e

Entomophaga

Neaera

arohpillaC

Ocytata

Nealsom

yia

Linnaemya

Sarromyia

Blond

elia

Trixa

Myxexoristops

Athry

cia

Strongygaster

Opesia

Senometopia

Rossimyiops

Thel

aira

Rhaphiochaeta

Red

tenb

ache

ria

Erioth

rix

Prosopea

Plesina

Zenillia

Palesisa

Pros

ena

Campy

loch

eta

Actia

Cyzenis

Exoris

ta

Pseudoperichaeta

Germ

aria

Phryno

Pseudomintho

Masicera

Thelymorpha

Lydina

Phryxe

Sto

min

a

Ceranthia

Bill

aea

Chetina

ailicuL

Eut

rixop

sis

Parerigone

Nemorilla

Gonia

anihcaT

Heraultia

StaurochaetaPhoroce

ra

Phy

to

Voria

Ace

mya

Com

psilu

ra

Periarchiclops

Ex

ori

st i

na

e

Tachininae

Ph

as

iina

e

Dexiinae

Outgroups

Lepidoptera

Coleoptera adults

Coleoptera larvae

Dermaptera Diptera larvae

Hemiptera

Chilopoda - Lithobiomorpha

Orthoptera Caelifera

Orthoptera Ensifera

Embioptera

Isopoda

Myriapoda Insecta: HolometabolaInsecta: Heterometabola

Hymenoptera 'Symphyta'

(folivores/xylophagous)

Crustacea

Ambiguous

non-parasitoids

Fig. 2. One of the 87 most parsimonious trees obtained under equal weights with ancestral parsimony reconstruction of host associations mappedin colour. Clades discussed in the text are labelled with letters.



(Rondania-group), supported by one non-homoplasiousapomorphy (128:1, female tergite six long and tubular), takesup the position as sister group of clade Q, which is, in turn,supported by four homoplasious apomorphies (20:0; 65:0;69:0; 116:0).

Litophasia Girschner, Imitomyia Townsend and Strongy-gaster Macquart represent rather basal phasiine lineages. Ofthese taxa, hosts are unknown for the first two, while the lasthas a rather complex host spectrum that includes adult Formi-cidae (S. globula (Meigen); see Herting, 1960), Heteropteraand a plethora of adult Coleoptera (S. triangulifera (Loew);see Reeves & O’Hara, 2004).

Four apomorphic character states support monophyly ofclade R (Phasiinae – Litophasia): two homoplasious (16:1;114:0) and two non-homoplasious (100:1, medial plate ofhypandrium elongated; 102:1, pregonite posteriorly connectedto hypandrium). Clade S (clade R – Imitomyia) is supportedby three homoplasious (103:0; 106:0; 115:0) and two non-homoplasious (91:1, posterior margin of sternite 5 not notched;93:1, male sternite 6 symmetrical) apomorphies. The mono-phyly of clade T (i.e. the Phasiinae sensu O’Hara & Wood,2004), the members of which are all parasitoids of Heteroptera,is rather well supported by five homoplasious apomorphies(7:0; 117:1; 131:1; 133:0; 134:0). Within clade V, the tribePhasiini sensu Tschorsnig (1985), here represented by Ope-sia Robineau-Desvoidy, Phasia Latreille and Xysta Meigen,is not recovered as monophyletic. The Gymnosomatini (cladeW, thus including Trichopoda Berthold), however, are mono-phyletic and supported by two homoplasious apomorphies (2:1;96:1) and one non-homoplasious apomorphy (123:1, spermduct developed into three well-sclerotized ducts).

Monophyly of clade X is supported by two homoplasiousapomorphies (65:1; 96:1). Parerigone Brauer and HermyaRobineau-Desvoidy represent basal lineages to clade Y (Cylin-dromyiini + Leucostomatini) whose monophyly is supportedby one non-homoplasious apomorphy (125:1, phallus smoothand tubular). The tribe Cylindromyiini (clade AA) is supportedby two homoplasious (95:1; 131:0) and two non-homoplasious(130:1, female tergite 7 and sternite 7 fused; 132:1, femalepostgenital plate stylet-like) apomorphies. Monophyly of theLeucostomatini (clade Z) is supported by two homoplasious(16:0; 18:2) and one non-homoplasious (127:1) apomorphies.

The tachinine grade. Clade AB (all remaining Tachinidae,i.e. Tachininae + Exoristinae) is weakly supported by threehomoplasious apomorphies (27:1; 35:3; 58:0). Clade AC, sup-ported by one homoplasious apomorphy (73:3), reconstructsthe Coleoptera-parasitizing genus Anthomyiopsis Townsendas sister group of the polyphagous tribe Graphogastrini(Graphogaster Rondani, Heraultia Villeneuve, PhytomypteraRondani). Within clade AD, which is supported by threehomoplasious apomorphies (81:1; 119:0; 121:1), the currenttribe Minthoini (Rossimyiops Mesnil, Hyperaea Robineau-Desvoidy, Plesina Meigen, Pseudomintho Brauer & Bergen-stamm, Mintho Robineau-Desvoidy), although representingrather basal lineages, forms a grade due to lack of any synapo-morphic character states.

Monophyly of the large clade AF is supported by threehomoplasious apomorphies (48:0; 81:0; 85:0). This assem-blage is divided into two large subclades, AG and AV, rep-resenting the remaining Tachininae and entire Exoristinae,respectively.

The tachinine clade AG is supported by threehomoplasious apomorphies (32:2; 65:1; 114:1). The Leskiini(Clausicella Rondani, Leskia Robineau-Desvoidy, BithiaRobineau-Desvoidy, Aphria Robineau-Desvoidy, DemoticusMacquart) are retrieved as a paraphyletic grade of lineagesat the base of clade AH. Clade AH, supported by threehomoplasious apomorphies (52:1; 78:2; 114:0), is divided intotwo subclades (AI and AL).

Clade AI, composed of the tribes Brachymerini andSiphonini, is supported by three homoplasious apomorphies(33:2; 85:3; 112:0). Monophyly of the tribes Brachymerini(clade AJ) and Siphoniini (clade AK) is supported, respec-tively, by four homoplasious apomorphies (34:1,2; 43:5;44:1; 65:0) and by five homoplasious (21:2; 41:0; 73:3; 74:1;106:1) and two non-homoplasious (104:1, anterior portion ofpregonite membranous; 135:1, female with two spermathecae)apomorphies.

Clade AL represents the Tachinini in the broad sense ofTschorsnig (1985: 70), i.e. an expanded concept with respectto Herting (1984) and most other authors, plus PelatachinaMeade. This clade is supported by five homoplasious apomor-phies (95:1; 111:1; 113:1; 117:0; 119:1). The Tachinini s.l. aredivided into three main subclades: clades AM and AN sup-ported by two homoplasious apomorphies [(24:4; 66:1,2) and(58:0; 79:2,3), respectively]; and clade AO supported by fourhomoplasious apomorphies (43:5; 49:0; 65:0; 80:0).

In clade AM, the highly apomorphic and moth-parasitizinggenus Sarromyia Pokorny (cf. Cerretti, 2010) is quitesurprisingly retrieved as sister to the beetle-parasitizingtribe Megaprosopini (Dexiosoma Rondani, MicrophthalmaMacquart).

Within clade AN, the genus Loewia Egger clusters with thePolideini sensu O’Hara (2002) (clade AP). The monophylyof clade AP is supported by two homoplasious apomorphies(43:4; 69:2). This clade is, in turn, sister to clade AQ, whichis composed of genera currently included in Herting’s (1984)tribes Neaerini (Neaera Robineau-Desvoidy, NeoplectopsMalloch), Ernestiini in part (Eloceria Robineau-Desvoidy,Synactia Villeneuve) and Bigonichetini (= Triarthriini)(Triarthria Stephens). The first two of these tribes arereconstructed as polyphyletic.

Clade AO is divided into two subclades: clade AR, sup-ported by three homoplasious apomorphies (39:1; 52:0; 73:0)and clade AS, supported by one homoplasious apomorphy(15:1). In clade AR, Germaria Robineau-Desvoidy and Lin-naemya Robineau-Desvoidy (clade AT) cluster together, form-ing the sister group of clade AU (Tachinini s.s.). Monophylyof clade AT is supported by just one homoplasious apomorphy(117:1). Clade AU is supported by three homoplasious apomor-phies (24:4; 38:1; 44:1) and one non-homoplasious apomorphy(94:1, postabdomen capsule-like). In clade AS, the generaNemoraea Robineau-Desvoidy (Nemoraeini) and Pelatachina



(Pelatachinini) are nested within a group of Ernestiini sensuHerting (1984), in part.

A monophyletic Exoristinae. Monophyly of the Exoristi-nae + Eutherini (clade AV) is supported by four homoplasiousapomorphies (2:1; 3:0; 8:0; 45:3). Within this clade the tribeAcemyini (clade AW) is corroborated as monophyletic withstrong support of six apomorphic character states: four homo-plasious (58:0; 88:1; 103:2; 112:0) and two non-homoplasious(40:1, first flagellomere sharply pointed; 124:1, pronouncedshortening of phallus). The Acemyini are sister to clade AX,which represents all the remaining Exoristinae.

Clade AX (Blondeliini, Eryciini, Ethillini, Exoristini, Goni-ini, Winthemiini, Eutherini) is supported by two homoplasiousapomorphies (52:1; 121:0).

Phylogenetic reconstruction of the lineages of subfamilyExoristinae is ambiguous in our analysis. The Eutherini(clade BA), consisting of two known genera (both includedin the present analysis), have been placed in either thePhasiinae or the Dexiinae, as discussed earlier. Here, the tribeoccupies a nested position within the Exoristinae. Monophylyof the Eutherini is strongly supported by nine homoplasiousapomorphies (19:0; 51:0; 78:1; 103:1; 106:1; 111:1; 114:1;117:0; 119:1) and one non-homoplasious apomorphy (5:1, eggwith a posterodorsal ‘window’).

Monophyly of each of the tribes Winthemiini, Ethillini,Eryciini and Blondeliini traditionally ascribed to the Exoristi-nae is not supported in any of the 87 MPTs, suggesting thatthese taxonomic groups should be studied in more detail todelineate monophyletic groups and revise their classificationaccordingly. In the tree chosen for discussion (Fig. 2), theBlondeliini, Eryciini and Ethillini are each reconstructed aspolyphyletic. There is some support for monophyly of Exoris-tini (clade BD), Winthemiini (+ Phorocerosoma Townsend)(clade BH) and Goniini (clade BJ).

A sister-group relationship is supported betweenExoristini (clade BD) and clade BE (Medina Robineau-Desvoidy + (Winthemiini + Ethillini (in part))). The tribeExoristini is supported by three homoplasious apomorphies(71:3; 72:1; 96:1). Clade BF, supported by two homoplasiousapomorphies (50:1; 121:1), reconstructs clade BG (EthillaRobineau-Desvoidy + Prosethilla Herting) as sister group ofclade BH (Winthemiini and the ethilline Phorocerosoma).The Goniini (clade BJ) are reconstructed as monophyleticwith support from three non-homoplasious apomorphies(3:1, egg shell brown or dark brown; 6:1, egg microtype;9:1, egg hatching stimulated by proteolytic enzymes of hostmesenteron) and one homoplasious (8:1) apomorphy.

Analysis under implied weights

Using implied weights analyses resulted in different con-sensus topologies depending on the k -values of the weightingfunction (Fig. 3). As a consequence, it is not possible to com-ment in detail upon all the various relationships observed in theresulting trees (especially for distal nodes). Instead, a concise

description of the more interesting and significant variations isgiven in the following (Fig. 3).

k = 1 (Fig. 3a). Under these extreme weighting conditions,the genus Litophasia is reconstructed as sister group of allother tachinids due its lack of a well-developed subscutellum.The presence of a strongly convex subscutellum (60:1) repre-sents the only non-ambiguous, non-homoplasious apomorphypresent on the consensus cladogram that supports the remain-ing Tachinidae. In this analysis, Eutherini are reconstructed as arather basal lineage of the Exoristinae, while the Dexiinae takeup a position as sister to clade Phasiinae (sensu Herting, 1984;thus including Imitomyiini and Strongygastrini) + Tachininae(paraphyletic). The Ormiini take up a position of sister groupto the Phasiinae.

2 ≤ k ≤ 5 (Fig. 3b). Here, the Phasiinae (sensu O’Hara &Wood, 2004, thus excluding Imitomyia and Strongygaster)take up a sister-group position to the remaining Tachinidae(cf. Richter, 1991). The subfamily Exoristinae is brokenup into a paraphyletic grade of lineages from which themonophyletic Acemyini and clade Dexiinae + Tachininaearose. Under these k -values of the weighting function,oviparous groups [i.e. Phasiinae, Winthemiini, Ethillini (inpart), Exoristini and the blondeliine genus Medina] arereconstructed as a basal paraphyletic grade of taxa fromwhich a monophyletic ovolarviparous group arose. Morespecifically, under these k -values, the laying of fully embry-onated eggs (7:1), together with the very long ovisac (133:1)and the presence of a network of tracheae enveloping theovisac (134:1), seems to have characterized the commonancestor of all the ovolarviparous tachinids, thus rejectingboth the reconstruction we obtained under equal weights andprevious hypotheses that ovolarvipary evolved several timesindependently from oviparous ancestors (cf. Stireman et al.,2006; Tachi & Shima, 2010). The membranous, translucentegg shell (3:2) is the only non-ambiguous, non-homoplasiousapomorphy supporting clade Dexiinae + Tachininae. TheEutherini are reconstructed as nested within a paraphyleticExoristinae (3 ≤ k ≤ 5) or as sister group of clade Tachin-inae + Dexiinae (k = 2). The tribe Ormiini and the genusStrongygaster represent rather basal lineages of clade (Imito-myia + (Litophasia + (Dexiinae))), which is in turn supportedby two non-ambiguous, non-homoplasious apomorphies (87:4,abdominal tergite 5 more than 1.5 times as long as tergite 4;115:1, membranous dorsal connection between basiphallus anddistiphallus).

k = 6 (Fig. 3c). Under these k -values of the weightingfunction, the oviparous groups noted earlier are still recon-structed as a basal paraphyletic grade of taxa from whichovolarviparous groups arose, with the subfamily Phasiinae rep-resenting the most basal tachinid lineage. The most notabledifference from the scenario discussed earlier is that the tribeOrmiini takes up the position of sister group of Imitomyia +



Outgroups

Litophasia

EXORISTINAE + Acemyini

Eutherini

DEXIINAE(including Dufouriini)

PHASIINAE

Ormiini

Imitomyia

Strongygaster

7:1

PHASIINAE s.s.

EXORISTINAE + Acemyini + Eutherini

Ormiini

DEXIINAE(includingDufouriini)

TACHININAE(monophyleticwhen k = 2)

Strongygaster

Imitomyia

87:4

EXORISTINAEoviparous

ovolarviparous

Litophasia

133:1134:1

14:142:1

110:1

99:1124:1

1:12:1

10:112:160:1

3:2

115:1116:1

PHASIINAE

EXORISTINAE

TACHININAE

Ormiini

Imitomyia

Strongygaster

Litophasia

DEXIINAE(includingDufouriini)

3:2

116:1

s.s.

EXORISTINAE

oviparous

ovolarviparous

Acemyini

Eutherini

40:1

124:1

k = 1 k = 2 – 5 k = 6

k = 9–15k = 7–8

oviparous

oviparous

k ≥ 20

Eutherini

Ormiini

DEXIINAE(including Dufouriini)

Litophasia

Imitomyia

Strongygaster

PHASIINAE

TACHININAE

EXORISTINAE

Myiophasiini +Palpostomatini

Acemyini

Outgroups

Eutherini

Ormiini

DEXIINAE (in part)

Litophasia

Imitomyia

Strongygaster

PHASIINAE

TACHININAE

EXORISTINAE


Acemyini

105:1

Outgroups

91:1100:1

102:1

115:1

116:1

Eutherini

Ormiini

DEXIINAE

Litophasia

Imitomyia

Strongygaster

PHASIINAE

TACHININAE

EXORISTINAE


Acemyini

105:1

Outgroups

91:1100:1

102:1

115:1116:1

Macroprosopa

DufouriiniDufouriini

ovolarviparous

ovolarviparous

7:1133:1134 :1

TACHININAE

Outgroups Outgroups

1:13:2

10:112:160:180:1

105:1

93:1

100:1102:1

14:142:1

110:1

1:110:112:157:0

3:2

91:193:1

100:1102:1

116:11:1

10:112:160:180:0

116:1

60:1

(in part)

(a) (b) (c)

(d) (e) (f)

Fig. 3. Simplified representation of the strict consensus cladograms (with non-ambiguous, non-homoplasious apomorphies mapped) obtained underimplied weights (default function) and concavity factor k varying from 1 to 40: (a) k = 1; (b) 2 ≤ k ≤ 5; (c) k = 6; (d) 7 ≤ k ≤ 8; (e) k = 9, 10, 15;k ≥ 20.

Strongygaster and that this clade represents a basal lineage ofclade (Eutherini + (Litophasia + (Dexiinae))).

7 ≤ k ≤ 8 (Fig. 3d). The general topology resulting fromthese analyses is close to that obtained under equal weights,such that the basal tachinid lineage is represented by cladeGnadochaeta + Palpostomatini and the remaining Tachinidaeare divided into two large assemblages: Dexiinae + Phasiinae(clade F of Fig. 2) and Tachininae + Exoristinae (clade AB ofFig. 2, but excluding Eutherini). Remarkably, under these con-ditions, subfamilies Dexiinae (sensu Herting, 1984), Phasiinae

(including Imitomyia and Strongygaster) and Exoristinae arereconstructed as monophyletic, while the subfamily Tachininaeremains paraphyletic with respect to Exoristinae. The tribesOrmiini and Eutherini are recovered as sister groups of Phasi-inae and Dexiinae, respectively.

k ≥ 9 (Fig. 3e, f). Under these conditions the topology isnearly the same as that obtained under equal weights. CladeGnadochaeta + Palpostomatini is retrieved as sister groupof all the remaining Tachinidae (see earlier for k = 7 and8) and the Dexiinae and Tachininae are both reconstructed



as paraphyletic with respect to Phasiinae and Exoristinae,respectively. The most notable difference is that the tribeEutherini does not cluster within the Exoristinae, taking up thesister-group position to clade Dexiinae + Phasiinae (clade Hof Fig. 2). For k ≥ 7, which we examined, Ethilla , Prosethilla ,Paratryphera Brauer & Bergenstamm and Atylomya Brauer(Ethillini) form a monophyletic group.

Discussion

Reconstructing the phylogeny of this large family of para-sitoids, a group characterized by extraordinary morphologicaldiversity and impressive radiation in nearly all terrestrialecosystems, remains a difficult challenge. As mentioned in theintroductory sections, the current classification of Tachinidaeat the tribal and subfamilial levels is the result of observationson external morphology and male and female terminalia.The phylogenetic value of these traits has never before beenanalysed using a rigorous cladistic approach on a broad scaleat the world level. This analysis, using an extensive mor-phological data set for a large cohort of taxa, represents thefirst explicit hypothesis of tachinid phylogenetic relationships.This first analysis has yielded a number of intriguing results(discussed later) and also provides phylogenetic data that canbe tested, modified and supplemented in future studies ontachinid evolution.

Despite the widely accepted subdivision of the Tachinidaeinto four subfamilies (Table 1), clear-cut definitions of thesegroups remain elusive and specialists are aware that somesubfamilies are not monophyletic. As a consequence, certaingenera and tribes have been arbitrarily moved from one sub-family to another, based simply upon the opinions of special-ists (however informed they may be) (Table 2). For example,in comparing the regional catalogues of Herting (1984) andO’Hara & Wood (2004) – whose authors largely agree onbasic ideas about tachinid classification – the following dif-ferences are noted (see also historical review and Table 2):the Eutheriini and Imitomyini were treated in the Phasiinae bythe former and in the Dexiinae by the latter (as proposed byShima, 1989); the Acemyini were classified in the Exoristinaeby the former and moved to the Tachininae by the latter; theStrongygastrini were treated in the Phasiinae by the former andin the Tachininae by the latter.

In addition to inconsistencies in the placement of taxa atthe tribal and subfamilial levels among authors, the placementof a key species or small genus into the wrong (though widelyaccepted) tribe or subfamily is sometimes overlooked. Forinstance, the genus Litophasia , which is consistently placed inthe phasiine tribe Catharosiini, is characterized by a very dif-ferent, dexiine type of male terminalia (cf. Tschorsnig, 1985).

Equal vs implied weights

Increasing k -values of the default weighting function in tntresulted in a gradual loss of concavity of the function, and

hence a reduction in the strength of the weighting function,which at the limit converges on the unweighted linear function.Under implied weighting, a tree that could increase the fit incharacters with few homoplasies by reducing the number ofsteps could also result in an increased number of steps incharacters with many homoplasies, i.e. the total fit may beincreased at the cost of an increased tree length (Goloboff,1993; Goloboff et al., 2008). The total fit calculated forMPTs obtained under equal weights for each k -value tested(tnt commands: piwe = 1, 2, . . . 40; fit*;) resulted in treesunder implied weighting displaying higher fit than the shortestobtained under equal weights (File S5). Also, by increasingvalues of k , the difference in steps between trees under impliedweighting and the shortest under equal weights diminishes, asdo the differences in total fit (File S5).

With a k -value of 1 (Fig. 3a), hence with strongest penaltyto homoplasy, the analysis yielded trees 87 steps longer thanthe shortest trees under equal weights. This extreme degreeof concavity gives certain characters excessive influence onthe tree, resulting in clades strongly contradicted by numerouscharacters. For example, with k = 1 the ‘dexiine-phasiine’genus Litophasia takes up the position of sister group toall the remaining tachinids due its unique lack of a stronglyconvex subscutellum (60:1). This character state undoubtedlyrepresents a reversal for Litophasia , which shares a large suiteof states (i.e. external morphology, male and female terminalia)with the dufouriines and the phasiines, thus making the entirereconstruction quite unreliable.

Trees obtained under implied weights with k -values varyingfrom 2 to 6 (which are still much better in terms of total fitwith respect to those obtained under equal weights; see File S5)reconstructed oviparous Exoristinae as a paraphyletic grade oftaxa from which ovolarviparous groups such as the Tachini-nae and Dexiinae arose (Fig. 3b, c). This hypothesis conflictsstrongly with previous classification schemes and morpholog-ical hypotheses as well as with preliminary molecular data. Infact, the Exoristinae, which are composed of both oviparousand ovolarviparous lineages, have been consistently recoveredas monophyletic (see Tschorsnig, 1985; Richter, 1991; Stire-man, 2002; Tachi & Shima, 2010). Here, characters associatedwith egg type and incubation, which characterize large groups(some tribes and subfamilies), are drawing together otherwiseunrelated groups and dominating clade structure due to theconcave weighting scheme.

For k > 6, the deeper nodes begin to stabilize around ageneral pattern (Fig. 3d–f) with the same three basal clades[(Gnadochaeta + Palpostomatini), (Dexiinae + Phasiinae) and(Tachininae + Exoristinae)] as obtained under equal weights(Figs 1, 2). The notable differences with respect to theanalysis under equal weights are that: (i) the tribe Eutheriniis recovered as sister group of (Dexiinae + Phasiinae) (fork ≥ 9) (Fig. 3e, f) or, within this assemblage, as sister toclade (Dexiinae + Litophasia) (for k = 7 and 8) (Fig. 3d);(ii) for k -values varying from 7 to 15, the macquartine genusMacroprosopa takes up a position as sister to Anthomyiopsiswithin the Tachininae grade.



Analyses conducted giving the strongest penalty to homo-plasy (i.e. 1 ≤ k ≤ 6) appear to be providing excess confidencein very unlikely trees by too severely overweighting charac-ters that change little and underweighting those that are morelabile. In the following discussion we focus mainly on theresults obtained from the analyses under equal weights andunder implied weighting with k ≥ 7, referring occasionally tocertain topologies obtained under implied weighting and k ≤ 6that require specific comments.

Tachinid monophyly and sister group

The current analysis does not allow us to fully test themonophyly of Tachinidae or determine their sister group,given the inclusion of a relatively small number of outgrouptaxa. However, the monophyly of the Tachinidae is herecorroborated by ten synapomorphic character states, andsupported by a Bremer support value of 10 and a jackkniferesampling probability of 95 (Fig. 1; File S4). Surprisingly,some recent phylogenetic reconstructions of calyptrate Diptera,employing mitochondrial and nuclear genes but includingonly a few tachinid species, have not strongly supported themonophyly of Tachinidae (Kutty et al., 2010; Singh & Wells,2013).

Studies aimed at elucidating phylogenetic relationshipsamong all Diptera (McAlpine, 1989; Wiegmann et al., 2011;Caravas & Friedrich, 2013; Lambkin et al., 2013) or withinthe oestroid Calyptratae in particular (Pape, 1992; Rognes,1997; Kutty et al., 2010; Marinho et al., 2012; Singh &Wells, 2013) have yielded contrasting results in reconstructingrelationships among oestroid families. Thus, the sister groupof Tachinidae has been identified as the Rhinophoridae byWood (1987a) and McAlpine (1989) and the Sarcophagidaeby Pape (1992) and Rognes (1997) using morphology-based inference, while more recently, molecular data seemto converge on some member of the Calliphoridae s.l.[Kutty et al., 2010 (Polleniinae); Wiegmann et al., 2011;Marinho et al., 2012 (Mesembrinellinae); Singh & Wells,2013 (Polleniinae)] as possible sister to the Tachinidae. Inour analyses the Calliphoridae are consistently reconstructedas paraphyletic (see also Rognes, 1997) and, as alreadyhypothesized by Pape & Arnaud (2001), the rhiniid genusRhyncomya is sister group to the Rhinophoridae (Figs 1, 2;File S4) based on three homoplasious apomorphies (48:0;62:1; 81:1) and one non-homoplasious apomorphy (61:1, lowercalypter tongue-shaped). Clade Rhyncomya + Rhinophoridaeis retrieved as sister group to Tachinidae under the entirerange of weighting schemes, giving more credence to thehypothesis of Wood (1987a), although not entirely rejectingthat of Rognes (1997), who suggested the rhiinids as sister toclade Sarcophagidae + Tachinidae. Interestingly, since Rognes(1997), the sister-group relationship between Sarcophagidaeand Tachinidae has not been corroborated. Additional studiesusing both morphological and molecular data and broadertaxon sampling of both tachinids and oestroid outgroups will beneeded to definitively establish the sister group of Tachinidae

and understand the development of parasitoid lifestyles withinthe oestroid lineage.

Phasiines from dexiines?

Our analysis generally supports the subfamily groupingsDexiinae + Phasiinae and Tachininae + Exoristinae (Figs 1, 2)(as first proposed by Shima, 1989; see earlier), with only theExoristinae and the Phasiinae reconstructed as monophyleticlineages under a wide range of weighting schemes. Thiscontrasts with the views of Mesnil (1966) and Herting (1966) inwhich the Tachinidae were grouped into Tachininae + Dexiinaeand Exoristinae + Phasiinae based on their oviposition andegg types. Interestingly, the Dexiinae, which were previouslyconsidered a well-established monophyletic assemblage, arereconstructed as paraphyletic with respect to the Phasiinae,under both equal and implied weights (k ≥ 9) (Fig. 3e,f). Following the traditional definition of the Dexiinae(Tschorsnig, 1985; Wood, 1987b; Shima, 1989; Tschorsnig &Richter, 1998), all members of the subfamily share a strap-likepregonite (103:1) and have the basiphallus and distiphallusjoined at a right angle (116:1) through a membranousconnection (115:1). In particular, states 115:1 and 116:1,characterizing the dexiine-type phallus, are here reconstructedas autapomorphic support for the clade Dexiinae + Phasiinae.These states are interpreted here for the first time ashaving undergone a reversal, being secondarily lost inthe Phasiinae. The strongly autapomorphic phallus of mostPhasiinae (Tschorsnig, 1985), often characterized by the lackof a clear border between basiphallus and distiphallus, lendssupport to this hypothesis. Interestingly, our results supportthe affinities between Dufouriini and Phasiinae identified byVerbeke (1962, 1963) and Tschorsnig (1985), reconstructingthe former as a grade from which the latter arose (Figs 1, 2;File S4).

Although previously suggested by Herting (1984), ouranalysis provides the first rigorous support for the subfamilyaffiliation of Litophasia , Imitomyia and Strongygaster withinthe Phasiinae. But, as hypothesized by Tschorsnig (1985),Litophasia takes up a position of sister group of the Dexiinaeunder a wide range of weighting schemes (Fig. 3a–d), inparticular, when the Dexiinae are recovered as monophyletic.Interestingly, under both equal weighting and for the lowestk -values of the weighting function (Fig. 3a, b), the Eutherinido not cluster with either the Dexiinae or the Phasiinae. Thetribe has long been considered a member of the Phasiinae,in part because Euthera species (and later confirmed forRedtenbacheria insignis Egger) are parasitoids of true bugs(Acanthosomatidae, Pentatomidae) (cf. Herting, 1966; Shima,1999). All Phasiinae, with the exception of Strongygaster , areparasitoids of true bugs and all other Tachinidae are parasitoidsof other arthropods. But unlike other Phasiinae, EutheraLoew and Redtenbacheria Schiner are ovolarviparous (7:1)(Shima, 1999). Herting (1966) and Mesnil (1966) were bothconvinced that the egg of Euthera is of the planoconvex type(2:1) (File S2, Figures S1, S2) and placement in the Phasiinae



is warranted despite its ovolarviparous habit. Tschorsnig (1985)was uncertain about the affinities of the Eutherini, recognizingcertain similarities with the Dexiinae in the male terminalia,but ultimately retaining the tribe in the Phasiinae. Shima(1989) placed Euthera at the base of the Dexiinae, and Shima(1999, 2006), O’Hara & Wood (2004) and Cerretti (2010)continued the placement of the Eutherini in this subfamily.We have observed that the long and platiform pregonite in theEutherini recalls the pregonite of the Dexiinae, as does theposition and orientation of the epiphallus in this tribe (FileS2, Figure S37). In fact, the only marked difference in themale terminalia of the Eutherini with respect to the Dexiinaeis the sclerotized dorsal connection between basiphallus anddistiphallus, as noted by O’Hara & Wood (2004). The unusualcombination of a planoconvex egg shape (2:1) and possessionof a large and tracheated ovisac in which eggs are embryonated(7:1; 133:1; 134:1) appears to play a role in reconstructingthe Eutherini at the base of Exoristinae (Fig. 3a), or nestedwithin the Exoristinae (Figs 1, 2, 3b). However, for k -valuesranging from 6 to 40, the Eutherini are retrieved as sister toDexiinae or sister to Dexiinae + Phasiinae (Fig. 3c–f). Theformer relationship is only inferred when the Dexiinae arereconstructed as monophyletic (6 ≤ k ≤ 8) (Fig. 3c, d). Forhigher k -values (Fig. 3e, f), the Dexiinae fragment into aparaphyletic grade from which the Phasiinae arose and theEutherini become the sister group of Dexiinae + Phasiinae. Inall these reconstructions, characters of the male terminalia playa crucial role in defining phylogenetic relationships over thoseof the egg and the female genitalia, supporting the hypothesisof Shima (1989).

Adults of the small tribe Strongygastrini resemble membersof Phasia Latreille, and other characters have led mostauthors to place the genus in the Phasiinae (e.g. Herting,1960; Verbeke, 1962; Mesnil, 1966; Herting, 1984; Tschorsnig,1985), where it is placed in our analysis. However, unlike otherphasiines, species of Strongygaster are not parasitoids of truebugs (see earlier) and are ovolarviparous (7:1). Shima (1989)placed Strongygaster at the base of the Phasiinae. O’Hara &Wood (2004) and Cerretti (2010) doubted the phasiine affinityof Strongygaster and placed the genus in the Tachininae.

The diverse and complicated Tachininae

It is generally agreed that the Tachininae are the mostweakly supported tachinid subfamily (cf. Tschorsnig, 1985),being composed of a diversity of morphologically heteroge-neous taxa. This is corroborated by our analysis, in whichthe subfamily is fragmented into a number of para- andpolyphyletic lineages. The Coleoptera-parasitizing clade B(Gnadochaeta + Palpostomatini) and the highly derived Ormi-ini (clade G) are widely separated from Tachininae (Figs 2, 3),the subfamily where they are traditionally assigned.

Although not supported by resampling, clade B is recoveredas sister to all the remaining tachinids under a widerange of weighting schemes. Interestingly, this hypothesis issimilar to that advanced by Mesnil (1966: 881–885). Mesnil

(1966) discussed a multitude of morphological and ecologicalcharacters of both adults and first instars of all major tachinidtribes and subtribes, and provided a ‘non-rooted’ graphicalrepresentation of the relationships among taxa. Mesnil (1966:881) depicted Myiophasiini and Palpostomatini in a rather‘central’, although ambiguous, position from which all othermajor groups seem to arise. The long and slender distiphallicstructure shared by members of clade B recalls that of mostdexiines; but the shape of the pregonite and postgonite, aswell as the shape and position of the epiphallus, recall those ofmany Tachininae and Exoristinae. Several highly homoplasiouscharacter states of the adult chaetotaxy seem to be involved inplacing clade B as sister to all remaining tachinids (File S4).

Macroprosopa is particularly mobile across phylogeneticanalyses, taking up positions at the base of clade C, or, underimplied weighting, nested within the Tachininae. It is worthnoting that the distiphallic structure in Macroprosopa is verysimilar to that of Gnadochaeta and of the Palpostomatini,which is more consistent with the phylogeny obtained underequal weights and for k -values ≥ 20 (Fig. 3f).

The phylogenetic position of the Ormiini remains ambiguousregardless of the weighting scheme, taking up positions asthe sister to the Phasiinae (Fig. 3a, d), sister to Dexiinae(Fig. 3b, c) or sister to (Phasiinae + Dexiinae) (Figs 1–3e,f). This ambiguity may be due to the highly autapomorphicmorphology of this group and we hesitate to make any strongclaims as to where this taxon may belong.

All remaining Tachininae are reconstructed as a paraphyleticgrade from which the monophyletic Exoristinae arose. Nearlyall the currently recognized tachinine tribes are reconstructedas para- or polyphyletic, as already hypothesized by severalauthors (Tschorsnig, 1985). Only a few tribes (Graphogas-trini, Brachymeriini, Siphonini, Megaprosopini, Tachinini s.s.),nested within this large grade, received support as representingmonophyletic groups (File S4). The Siphonini are well sup-ported, confirming conclusions of Andersen (1983) and O’Hara(1989), based largely upon the possession of only two (ratherthan the usual three) spermathecae in the female reproductivesystem (135:1). The Polideini (sensu O’Hara, 2002) are notretrieved as monophyletic here due to the (weakly supported)inclusion of Loewia in this clade, while the other examinedmembers of the Loewiini (sensu O’Hara & Wood, 2004), Tri-arthria and Eloceria , are included in a clade with Neoplectopsand Neaera of the Neaerini. However, the two synapomorphiesof the male terminalia used to define the recently restruc-tured Polideini (‘sternite 5 V-shaped with microspines on innermargins, and distiphallus with short lateral arms and special-ized anterior sclerite’; O’Hara, 2002: 20) were not included inthe present analysis. Support for the Tachinini s.s. (PeleteriaRobineau-Desvoidy, Tachina Meigen) is relatively strong, sup-porting Mesnil (1966) and most subsequent authors (but notGuimaraes, 1971) in defining this clade as those Tachininaewith the posterior (here posterolateral) margin of the hind coxasetose (83:1). Some other tachinids also have a setose hindcoxa, but the position of the setae is not the same through-out the family: in tachinine genera, the setae arise from theposterolateral margin of the hind coxa (83:1), whereas in the



Exoristinae and the genus Sarromyia , they arise from the pos-terodorsal margin of the hind coxa (82:1) (see File S2). Thisdifference in the position of the setae suggests that the two setalpatterns are non-homologous and thus two characters were rec-ognized in our analysis.

A monophyletic Exoristinae

Although weakly supported, the clade Exoristi-nae + Acemyini (AV) is reconstructed as a monophyleticlineage under a wide range of weighting schemes (Figs 1, 3a,d–f). Notably, the Acemyini, which split from Exoristinaeunder k -values varying from 2 to 6 (Fig. 3b, c), nevercluster within the Tachininae but rather at their base. Recentquantitative phylogenetic analyses of the Exoristinae usingmolecular data have also found evidence of monophyly forthe subfamily. Stireman (2002) recovered the Exoristinae asmonophyletic in most analyses, with the possible exceptionof certain ‘rogue’ taxa, including Ceracia Rondani, MasiphyaBrauer & Bergenstamm, and Phyllophilopsis Townsend. Intheir more comprehensive analysis, Tachi & Shima (2010)were able to resolve a monophyletic Exoristinae as sister toTachininae, similar to the results found here. These resultssuggest that the seemingly labile character of a ‘haired’prosternum (41:0) used by taxonomists to help define groupsis remarkably phylogenetically conserved, although it ispresent in some Tachininae and has been lost from someBlondeliini and a few Goniini.

Phylogenetic relationships within the Exoristinae are largelyunresolved and only the Exoristini and Goniini are recon-structed as monophyletic in all analyses. Although the Exoristi-nae are the most diverse subfamily of Tachinidae in terms ofnumbers of described species, they are the most morpholog-ically homogeneous subfamily as well. Most of the variationthat does exist appears to be highly homoplasious, with similarcharacter states recurring over and over again. The lack of res-olution and rampant homoplasy in this subfamily may reflectits relatively recent and explosive evolutionary radiation.Although employing far fewer taxa than here, Tachi & Shima’s(2010) molecular phylogenetic analyses generally supportedmonophyly of each of the tribes of Exoristinae and indicatedsister-group relationships as follows: ((((Goniini + Eryciini) +Blondeliini) + Exoristini) + Ethillini). Stireman (2002) reachedsimilar, if less clear, conclusions, except he was unable torecover a monophyletic Goniini. Neither the Eryciini nor theBlondeliini can be defined on the basis of non-homoplasiousapomorphies, and our results suggest that they may not bemonophyletic groups. As noted by Wood (1985), the short firstpostsutural, supra-alar seta (49:1) that is so useful for separat-ing the Blondeliini and Exoristini from most other Exoristinaemight be the primitive state of the character.

Another issue emerging from this analysis is the probablepolyphyly of the Ethillini as currently defined (Cerretti et al.,2012b). Ethillini are traditionally defined as sharing threecharacter states: strongly convex lower calypter, katepimeronsetulose, and ovipositor short and unmodified. The first

state is shared with several winthemiines, the second is acharacteristic feature of the Winthemiini, and the last representsa plesiomophic condition with respect to the long telescopingovipositor of winthemiines. These features indicate that thecomposition of the Ethillini has been weakly founded, andits monophyly was questioned by Tschorsnig (1985). Cerrettiet al. (2012b) argued that three monophyletic groups can beidentified within Ethillini, and these groups are recognized asmonophyletic in the current analysis: a Phorocerosoma group(here represented only by Phorocerosoma); an Ethilla group(clade BG), characterized by the laying of unembryonatedeggs (7:0); and a Paratryphera group (clade BI), whichlays embryonated eggs (7:1). The laying of unembryonatedeggs provides homoplasious autapomorphic support to cladeBC, wherein the Exoristini are positioned as sister groupof (Medina + ((Winthemiini + Phorocerosoma) + Ethillagroup)). The degree of egg development at oviposition, andhence the reproductive strategy of these flies, plays a key rolein the non-monophyly of the Ethillini in our reconstruction,but the resultant pattern of relationships is still not convincing.In fact, the Paratryphera and Ethilla groups share possessionof an operculate egg shell (4:1), which is a very rare conditionin tachinids (Cerretti et al., 2012b). Under implied weightingthis character state is reconstructed as a unique autapomorphysupporting monophyly of the Paratryphera group + Ethillagroup, under a wide range of k -values (i.e. k = 1 and k ≥ 7).

Which came first, the larva or the egg?

Within the Tachinidae, ovipary (deposition of unembry-onated eggs) is generally interpreted as a plesiomorphic con-dition with respect to ovolarvipary (deposition of embryonatedeggs), and it is presumed that the latter evolved several timesindependently (Wood, 1987b; Tschorsnig & Richter, 1998;Stireman et al., 2006; Tachi & Shima, 2010). In our analysisunder equal and implied weights (k ≥ 7), all treated oviparousgroups [all the Phasiinae s.s. (clade T) and clade BC, i.e. theexoristine taxa Exoristini, Winthemiini, some Ethillini (cladeBG and Phorocerosoma) and the blondeliine genus Medina]are nested within ovolarviparous assemblages (Figs 1, 2), sug-gesting the possible need to re-evaluate the evolutionary historyof ovolarvipary vs ovipary. The counterintuitive interpretationthat ovolarvipary could have characterized an early tachinidancestor and that ovipary developed several times indepen-dently from an ovolarviparous ancestor is a novel hypothesisgenerated from this analysis.

The laying of fully embryonated eggs is the most commonreproductive strategy in the family and our parsimonyreconstruction shows that its occurrence may be more primitivewithin the Tachinidae than previously thought. Ovolarviparyhas several major advantages, the primary being that the eggencloses a fully formed first instar the moment it is laid. Onceevolved, ovolarvipary presumably opened new opportunitiesfor the Tachinidae, for instance, the ability to take advantageof an indirect oviposition strategy, which is always associatedwith ovolarvipary. This means that eggs laid in habitats



frequented by potential hosts hatch almost immediately afterdeposition and the larvae actively search for a host (e.g.Dexiini) or wait for a host that might eventually pass by(e.g. Tachinini). In Goniini, a special type of ovolarviparyhas evolved in which eggs are laid on foliage and must beswallowed by a phytophagous host, subsequently hatching inits midgut through the stimulation of proteolytic enzymes.

The evolution of indirect oviposition undoubtedly allowedTachinidae to broaden their host spectrum to species that wouldnot be encountered by a searching female, because they are, forexample, hidden, buried or nocturnal. In groups characterizedby a direct oviposition strategy, the laying of ready-to-hatcheggs is still advantageous because it minimizes the length oftime that eggs are exposed and vulnerable during the delicateembryogenesis process.

There are also potential costs associated with ovolarvipary,especially for the adult female parasitoid. Tachinidae thatlay fully embryonated eggs have a long and distensibleovisac that serves as an incubator and, as a consequence,the time between mating and oviposition is longer than inoviparous females. Indirect oviposition is an inherently riskystrategy because the majority of first instars will perish beforeencountering a host, but inevitably this greater risk of mortalityis balanced by higher fecundity (O’Hara, 1985; Mellini, 1991;Stireman et al., 2006). The laying of unembryonated eggscomes at a cost, because such eggs are exposed to anincreased mortality risk while the first instar develops, butthis risk might be countered by the benefit of the egg beingdeposited directly onto a potential host. Minimizing the timenecessary to produce and lay eggs also minimizes the risk ofmortality for females before they can reproduce. Furthermore,these taxa can afford to produce fewer, larger eggs becauseall that are laid can be certain of reaching the host. Wethink that various combinations of these factors could havecontributed towards making the switch to ovipary sometimesadvantageous.

Aside from Goniini, a few other Exoristinae, Dexiini andmost Tachininae, which lay eggs ready to hatch in habitatsfrequented by hosts, all other ovolarviparous tachinids lay eggsdirectly on or inject eggs directly into the body of the host, asin all oviparous tachinids. Given our results, it is conceivablethat reversals to oviparity occurred in ancestors characterizedby a direct oviposition strategy.

Interestingly, with k -values varying from 2 to 6, analysesunder implied weights reconstruct ovipary as a plesiomorphicstate with respect to ovolarviparity (Fig. 3b, c), more in linewith previous hypotheses (see earlier), but with some importantdifferences. Under these implied weights, ovolarviparousgroups are retrieved as a monophyletic lineage, rejecting thehypothesis that ovolarvipary may have evolved several timesindependently from oviparous ancestors. As discussed earlier,this result has been achieved at the cost of reconstructingExoristinae as a paraphyletic grade. Morphological andmolecular evidence (Tschorsnig, 1985; Stireman, 2002; Tachi& Shima, 2010) converges on a monophyletic Exoristinae,making this scenario quite unlikely.

Host associations

Two major patterns emerge when considering host associa-tions across our phylogenetic reconstruction: (i) there is generalagreement between host use and monophyly of major lin-eages (see the Results section); and (ii) the host preferences ofmore basal groups tend to be either hemimetabolous orders orColeoptera (Fig. 2). More than 60% of tachinids are parasitoidsof lepidopteran larvae (cf. Cerretti, 2010: 43), but none of thefour basal clades (Fig. 2; B, D, F, AB) or the sister groups of allmajor, more distal, assemblages, are unambiguously associatedwith Lepidoptera as the primitive condition.

Gnadochaeta and the Palpostomatini (clade B) are bothColeoptera parasitoids. Gnadochaeta species attack concealedendophytic weevil larvae by means of actively seekingfirst-instar planidia, as do the Dexiini. Palpostomatines arenocturnal, or semi-nocturnal, parasitoids of adult scarabaeoids.Clade F (Ormiini + (Dexiinae + Phasiinae)) is ambiguous,being composed of three assemblages (clades G, I and O)with diverse host spectra. Clade G (Ormiini) are nocturnalparasitoids of Orthoptera; female ormiines are able to locatesinging male hosts by means of an unusual tympanicorgan located between the fore coxae (42:1). Within cladeI (Dexiinae – Dufourini), the Voriini s.l. (clade M) arecomposed of Lepidoptera parasitoids, without exception, whileclade J has the sawfly parasitoid Phyllomya as sister groupof clade K (Stomina + Dexiini). The hosts of Stomina areunknown, while the Dexiini are, with a few exceptions,parasitoids of soil or wood-dwelling beetle larvae. One of theexceptions is the genus Trixa Meigen, which is unusual amongDexiini in parasitizing caterpillars (Belshaw, 1993; Tschorsnig& Herting, 1994). The nested position of Trixa in clade Lsuggests a relatively recent shift to larval Lepidoptera in thistaxon. However, ecologically, Trixa is similar to other dexiinesin that its Hepialidae hosts (Belshaw, 1993; Tschorsnig &Herting, 1994) are soil dwellers. Within clade O, the Dufouriiniform a paraphyletic grade of adult Coleoptera parasitoids fromwhich the Phasiinae arose (clade Q). The only dufourines thatdo not directly attack adult beetles are those belonging to thegenus Dufouria , which lay eggs on larvae but emerge fromthe adults (Kaufmann, 1933; Mellini, 1964). Unfortunately,there are no host records for two key basal groups, Litophasiaand Imitomyia , but the position of Strongygaster , as sisterto the Phasiinae s.s. (clade T), is particularly interesting.Strongygaster has a diverse host range that encompasses adultbeetles as well as adult ants (Tschorsnig & Herting, 1994;Reeves & O’Hara, 2004; Shima, 2006), whereas the Phasiinaes.s. are exclusively parasitoids of Heteroptera. A shift fromadult Coleoptera to Heteroptera may have characterized thedevelopmental strategy of the phasiine ancestor. The only othertachinids known as Heteroptera parasitoids are the Eutherini.Under equal weights, the Eutherini are nested within theExoristinae, thus suggesting that parasitism of Heteropteraevolved independently in the Eutherini and Phasiinae. Asdiscussed earlier, analyses under implied weighting andk ≥ 6 yield an entirely different and probably more likelyphylogenetic scenario, reconstructing the Eutherini as more



closely related to Dexiinae and Phasiinae (Fig. 3c–f), althoughnever as sister to the Phasiinae. Interestingly, however, allancestral state reconstructions of host associations acrossphylogenetic analyses for k ≥ 6 agree in interpreting parasitismof Heteroptera as independently evolved in these two tachinidlineages.

Non-lepidopteran hosts also characterize basal lineages ofclade AB (Tachininae + Exoristinae). Species in the genusAnthomyiopsis parasitize chrysomelid beetle larvae (cf. Hert-ing, 1960). Similar to Dufouria , Anthomyiopsis species cancomplete development in the adult host, although in this genusit seems to be facultative, being triggered by specific environ-mental conditions (Mellini, 1957, 1964). It is worth notingthat, among parasitoids of Holometabola, larval-adult para-sitoids have been recorded almost exclusively in coleopteranhosts (Mellini, 1964), suggesting that these groups may haveretained some degree of presumably ancestral plasticity.

Within clade AB (Tachininae + Exoristinae), parasitoidsof Lepidoptera larvae become dominant at clade α (whichis also the most species-rich clade worldwide) and thishost preference may have characterized the ancestor of thisclade. As a consequence, associations with sawflies [cladePseudopachystylum Mik, Hyalurgus Brauer & Bergenstamm,Staurochaeta Brauer & Bergenstamm, Blondelia Robineau-Desvoidy (in part), Catagonia Brauer & Bergenstamm, Phe-bellia Robineau-Desvoidy (in part), Myxexoristops Townsend],crickets and grasshoppers (clade AZ, Leiophora Robineau-Desvoidy, Phorocerosoma), beetles [Microphthalma , Cleon-ice Robineau-Desvoidy, Chetogena Rondani (in part), Med-ina , Istocheta Rondani, Picconia Robineau-Desvoidy, ZairaRobineau-Desvoidy, Erythocera Robineau-Desvoidy], earwigs(Triarthria , Ocytata Gistel) and chilopods (Eloceria , Loewia)(Fig. 2) probably represent secondary shifts from primitivelyLepidoptera-parasitizing ancestors.

Interestingly, our cladograms agree in reconstruct-ing clade α as arising from a grade of taxa for whichhosts are mostly unknown (several Graphogaster , Her-aultia , Hyperaea , Plesina) or from taxa characterized byassociations with saprophagous micro-moths (most Phyto-myptera , a few Graphogaster and some minthoines) or withEmbioptera (Rossimyiops , one Phytomyptera , some unde-scribed Graphogaster) (Cerretti et al., 2009; and unpublisheddata from the E.S. Ross collection at California Academy ofSciences, San Francisco, CA, USA). Except for Phytomyptera ,most of these taxa are very rarely collected and it is possiblethat they may parasitize rarely reared groups of arthropods.

The possibility that the major groups of ditrysian parasitoids(i.e. clades M and α) may have arisen independently fromancestors that parasitized Heterometabola or Coleoptera leadsto the hypothesis that tachinids could have had an early Tertiaryorigin and that their massive radiation took place later in theTertiary and Quaternary, after that of angiosperms (which dateback to the Jurassic–Cretaceous boundary; Sun et al., 2002)and parallel to that of major ditrysian Lepidoptera lineages(Grimaldi & Engel, 2005). However, Tachinidae are exception-ally rare in the fossil record and there are no unquestionablefossils from the Eocene or older (O’Hara, 2013a).

A recent molecular phylogenetic analysis of the Dipterasuggests that the Tachinidae may have arisen as recently as30 million years ago (Wiegmann et al., 2011). Under this sce-nario, the radiation of tachinids would have experienced a timelag with radiations that took place in ditrysian host groups.Such non-parallel clade diversification, where parasitoids radi-ated into an existing and independently diversified resource,has recently been documented for other host/parasite associa-tions (Gomez-Zurita et al., 2007) and would be necessary toaccount for host-use patterns in other tachinid lineages, mostnotably for heterometabolous insect parasitoids (e.g. Ormiini,Phasiinae, Acemyini).

Caterpillars would seem to be excellent hosts for parasitoidsin general. They are often abundant, exposed (exophytic) andunable to flee from attack. Given these traits, it is expected thatthey would be colonized by parasitoids at the first opportunity.But, if there was a lag in their colonization by tachinids, as issuggested by the current phylogenetic reconstruction and theirestimated age, what was the reason? Also, it is interestingto note that many other families of Diptera have adopted theparasitoid lifestyle (e.g. acrocerids, nemestrinids, pipunculids,conopids, rhinophorids, many phorids and bombyliids, somemuscids, sarcophagids and calliphorids), but none havecolonized lepidopteran hosts as extensively as have tachinids.There are several factors that probably played a role in theradiation of Tachinidae on lepidopteran hosts. The groundplanof the Tachinidae was likely characterized by the formationof a respiratory funnel within the host (Clausen, 1940), whichis the method by which most tachinids avoid encapsulationduring their larval endoparasitic stage. Even with this preadap-tation, the parasitization of larval Lepidoptera must have posedsignificant problems. Due to their long evolutionary historywith hymenopteran endoparasitoids, caterpillars may haveevolved particularly effective defenses (e.g. encapsulationresponses; Strand, 2008; Smilanich et al., 2009), which makeit difficult for non-coevolved parasitoid taxa to colonize them.These defenses may have been partly overcome by furtherspecialized refinements to respiratory funnel formation and/orother means to circumvent the hosts’ immune response (e.g.Ichiki & Shima, 2003). Additionally, many caterpillars areknown to use secondary chemicals from their host plants fortheir own defense (Bowers, 1993; Dyer, 1997) and it is pos-sible that tachinids have had to develop special physiologicaladaptations to avoid or tolerate these chemical defenses beforethey could effectively colonize and radiate upon plant-feedinglepidopteran larvae. Similar and independently evolved adapta-tions may have allowed other tachinid lineages to radiate ontoother groups of plant-feeding and chemically protected insectssuch as Heteroptera and certain Coleoptera (e.g. chrysomelids).

In addition to surviving the immune response and chemicaldefenses of lepidopterans, other adaptations have been neces-sary for tachinids to take full advantage of this diverse lineageof potential hosts. These include the complex synchroniza-tion of development between parasitoid and host, which asa consequence is one of the determinants of host range amongTachinidae. Often, hormonal changes within the host that trig-ger developmental changes are also used as cues by tachinid



larvae to alter their behaviour, location, feeding and growth(Mellini, 1975, 1991; Baronio & Sehnal, 1985).

Innovations and adaptations in the adult stage have probablyalso contributed to the successful radiation of Tachinidae onLepidoptera and insects in general. In particular, the evolutionof specialized mechanisms of host location (e.g. use ofspecific plant volatiles by Exoristinae and host pheromones byPhasiinae; Roland et al., 1995; Aldrich & Zhang, 2002; Ichikiet al., 2012) and the diversification of oviposition strategies(O’Hara, 1985; Stireman et al., 2006) may have contributed tothe rapid, host-related diversification of tachinids.

Supporting Information

Additional Supporting Information may be found in the onlineversion of this article under the DOI reference:10.1111/syen.12062

File S1. Exemplar taxa.

File S2. Character list.

File S3. Data matrix.

File S4. Tree with character states traced.

File S5. Equal vs implied weighting.

File S6. Data matrix in nexus format.

Acknowledgements

We would like to thank Amnon Friedberg (TAU) and Hans-Peter Tschorsnig (SMNS), who enabled us to study thevaluable material preserved in their institutions (the TAUtachinid collection was examined by P.C. within the project‘Tachinidae of Israel’, funded by the Israel Academy ofSciences and Humanities). Thanks to Knut Rognes (Oslo,Norway) and to an anonymous referee for useful commentsand valuable suggestions on an early draft of this manuscript.P.C. is grateful to H.-P. Tschorsnig, Franco Mason (Verona,Italy) and Daniel Whitmore (Natural History Museum, London,UK) for stimulating discussions on tachinid systematics andecology. D.J.I. was supported by a PhD Fellowship from theCARIPARO Foundation. This work was supported by U.S.NSF DEB-1146269 to principal investigators J.O.S. and J.E.O.and to major collaborator P.C.

References

Aldrich, J.R. & Zhang, A. (2002) Kairomone strains of Euclytiaflava (Townsend), a parasitoid of stink bugs. Journal of ChemicalEcology , 28, 1565–1582.

Andersen, S. (1983) Phylogeny and classification of Old World generaof Siphonini (Diptera: Tachinidae). Entomologica Scandinavica , 14,1–5.

Baronio, P. & Sehnal, F. (1985) Dependence of the parasitoid Goniacinerascens on the hormones of its lepidopterous hosts. Journal ofInsect Physiology , 26, 619–626.

Belshaw, R. (1993) Tachinid flies. Diptera: Tachinidae. Handbooks forthe Identification of British Insects , Vol. 10, Part 4a(i), 169 pp. RoyalEntomological Society, London.

Bezzi, M. & Stein, P. (1907) Cyclorrapha Aschiza. CyclorraphaSchizophora: Schizometopa. Katalog der palaarktischen Dipteren ,Band III (ed. by T. Becker, M. Bezzi, K. Kertesz and P. Stein), pp.1–747. Budapest.

Bowers, M.D. (1993) Aposematic caterpillars: life-styles of thewarningly colored and unpalatable. Caterpillars: Ecological andEvolutionary Constraints on Foraging (ed. by N.E. Stamp and T.M.Casey), pp. 331–371. Chapman & Hall, New York, New York.

Cantrell, B.K. & Crosskey, R.W. (1989) Family Tachinidae. Catalogof the Diptera of the Australasian and Oceanian Regions , BishopMuseum Special Publication 86 (ed. by N.L. Evenhuis), pp.733–784. Bishop Museum Press, Honolulu and EJ Brill, Leiden.

Caravas, J. & Friedrich, M. (2013) Shaking the Diptera tree oflife: performance analysis of nuclear and mitochondrial sequencedata partitions. Systematic Entomology , 38, 93–103. DOI: 10.1111/j.1365-3113.2012.00657.x.

Cerretti, P. (2010) I tachinidi della fauna italiana (Diptera Tachinidae),con chiave interattiva dei generi ovest-paleartici . Cierre Edizioni,Verona.

Cerretti, P., De Biase, A. & Freidberg, A. (2009) Systematic studyof the genus Rossimyiops Mesnil (Diptera, Tachinidae). Zootaxa ,1984, 31–56.

Cerretti, P., Tschorsnig, H.P., Lopresti, M. & Di Giovanni, F. (2012a)MOSCHweb – a matrix-based interactive key to the genera of thePalaearctic Tachinidae (Insecta, Diptera). ZooKeys , 205, 5–18.

Cerretti, P., Wood, D.M. & O’Hara, J.E. (2012b) Neoethilla , a newgenus for the first record of the Ethillini from the New World(Diptera, Tachinidae, Exoristinae). ZooKeys , 242, 25–41.

Clausen, C.P. (1940) Entomophagous Insects . Hafner, New York, NewYork.

Coquillett, D.W. (1897) Revision of the Tachinidae of America northof Mexico. A family of parasitic two-winged insects. United StatesDepartment of Agriculture, Division of Entomology Technical Series ,7, 1–156.

Crosskey, R.W. (1973) A conspectus of the Tachinidae (Diptera) ofAustralia, including keys to the supraspecific taxa and taxonomicand host catalogues. Bulletin of the British Museum (Natural History)Entomology Supplement , 21, 221 pp.

Crosskey, R.W. (1976) A taxonomic conspectus of the Tachinidae(Diptera) of the Oriental region. Bulletin of the British Museum(Natural History) Entomology Supplement , 26, 357 pp.

Crosskey, R.W. (1980) Family Tachinidae. Catalogue of the Dipteraof the Afrotropical Region (ed. by R.W. Crosskey), pp. 822–882.British Museum (Natural History), London.

Crosskey, R.W. (1984) Annotated keys to the genera of Tachinidae(Diptera) found in tropical and southern Africa. Annals of the NatalMuseum , 26, 189–337.

Dindo, M.L. (2011) Tachinid parasitoids: are they to be considered askoinobionts? BioControl , 56, 249–255.

Dyer, L.A. (1997) Effectiveness of caterpillar defenses against threespecies of invertebrate predators. Journal of Research on theLepidoptera , 34, 48–68.

Feener, D.H. Jr. & Brown, B.V. (1997) Diptera as parasitoids. AnnualReview of Entomology , 42, 73–97.

Girschner, E. (1893) Beitrag zur Systematik der Musciden. BerlinerEntomologische Zeitschrift , 38, 297–312.

Girschner, E. (1896) Ein neues Musciden-System auf grund der Thora-calbeborstung und der Segmentierung des Hinterleibes. JllustrierteWochenschrift fur Entomologie, 1, 12–16, 30–32, 61–64, 105–112.

Goloboff, P.A. (1993) Estimating character weights during tree search.Cladistics , 9, 83–91.



Goloboff, P., Farris, S. & Nixon, K. (2003) T.N.T.: Tree Analysis UsingNew Technology , Version 1.1 [WWW document]. URL http://www.zmuc.dk/public/phylogeny [accessed on 12 September 2013].

Goloboff, P.A., Carpenter, J.M., Arias, J.S. & Miranda Esquivel,D.R. (2008) Weighting against homoplasy improves phylogeneticanalysis of morphological data sets. Cladistics , 24, 1–16. DOI:10.1111/j.1096-0031.2008.00209.x.

Gomez-Zurita, J., Hunt, T., Kopliku, F. & Vogler, A.P. (2007) Recal-ibrated tree of leaf beetles (Chrysomelidae) indicates independentdiversification of angiosperms and their insect herbivores. PLoSONE , 2, e360. DOI: 10.1371/journal.pone.0000360.

Grimaldi, D. & Engel, M.S. (2005) Evolution of the Insects . CambridgeUniversity Press, New York.

Guimaraes, J.H. (1971) Family Tachinidae (Larvaevoridae). A Cata-logue of the Diptera of the Americas South of the United States ,Vol. 104., 333 pp. Museu de Zoologia, Universidade de Sao Paulo,Sao Paulo.

Herting, B. (1957) Das weibliche Postabdomen der calyptraten Fliegen(Diptera) und sein Merkmalswert fur die Systematik der Gruppe.Zeitschrift fur Morphologie und Okologie der Tiere, 45, 429–461.

Herting, B. (1960) Biologie der westpalaarktischen Raupenfliegen(Dipt., Tachinidae). Monographien zur Angewandten Entomologie,16, 1–188.

Herting, B. (1966) Beitrage zur Kenntnis der Europaischen Raupen-fliegen (Dipt. Tachinidae). IX. Stuttgarter Beitrage zur Naturkunde,146, 12 pp.

Herting, B. (1984) Catalogue of Palearctic Tachinidae (Diptera).Stuttgarter Beitrage zur Naturkunde, Serie A (Biologie), 369, 228pp.

Herting, B. & Dely-Draskovits, A. (1993) Family Tachinidae.Catalogue of Palaearctic Diptera , Volume 13: Anthomyiidae –Tachinidae (ed. by A. Soos and L. Papp), pp. 118–458. HungarianNatural History Museum, Budapest.

Ichiki, R. & Shima, H. (2003) Immature life of Compsilura concinnata(Meigen) (Diptera: Tachinidae). Annals of the Entomological Societyof America , 96, 161–167.

Ichiki, R.T., Ho, G.T.T., Wajnberg, E., Kainoh, Y., Tabata, J. & Naka-mura, S. (2012) Different uses of plant semiochemicals in host loca-tion strategies of the two tachinid parasitoids. Naturwissenschaften ,99, 687–694.

Kaufmann, O. (1933) Der glanzstreifige Schildkafer (Cassida nobilisL.) nebst einigen Bemerkungen uber den nebligen Schildkafer(Cassida nebulosa L.). Arbeiten aus der Biologischen Reichsanstaltfur Land- und Forstwirtschaft Dahlem , 20, 457–516.

Kutty, S.N., Pape, T., Wiegmann, B.M. & Meier, R. (2010) Molecularphylogeny of the Calyptratae (Diptera: Cyclorrhapha) with anemphasis on the superfamily Oestroidea and the position ofMystacinobiidae and McAlpine’s fly. Systematic Entomology , 35,614–635. DOI: 10.1111/j.1365-3113.2010.00536.x.

Lambkin, C.L., Sinclair, B.J., Pape, T. et al. (2013) The phylogeneticrelationships among infraorders and superfamilies of Diptera basedon morphological evidence. Systematic Entomology , 38, 164–179.DOI: 10.1111/j.1365-3113.2012.00652.x.

Maddison, W.P. & Maddison, D.R. (2010) Mesquite: A ModularSystem for Evolutionary Analysis , Version 2.73 [WWW document].URL http://mesquiteproject.org [accessed on 3 January 2012].

Malloch, J.R. (1923) A new character for differentiating the familiesof Muscoidea (Dipt.). Entomological News , 34, 57–58.

Marinho, M.A.T., Junqueira, A.C.M., Paulo, D.F., Esposito, M.C., Vil-let, M.H. & Azeredo-Espin, A.M.L. (2012) Molecular phylogeneticsof Oestroidea (Diptera: Calyptratae) with emphasis on Calliphoridae:insights into the inter-familial relationships and additional evidencefor paraphyly among blowflies. Molecular Phylogenetics and Evo-lution , 65, 840–854.

McAlpine, J.F. (1981) Morphology and terminology – adults. Man-ual of Nearctic Diptera , Vol. 1, Research Branch, AgricultureCanada, Monograph 27 (ed. by J.F. McAlpine, B.V. Peterson,G.E. Shewell, H.J. Teskey, J.R. Vockeroth and D.M. Wood),pp. 9–63. Canadian Government Publishing Centre, Hull.

McAlpine, J.F. (1989) Phylogeny and classification of the Muscomor-pha. Manual of Nearctic Diptera , Vol. 3, Research Branch, Agricul-ture Canada, Monograph 32 (ed. by J.F. McAlpine and D.M. Wood),pp. 1397–1518. Canadian Government Publishing Centre, Hull.

Mellini, E. (1957) Studi sui Ditteri Larvevoridi. IV. Ptilopsinanitens Zett. parassita di Plagiodera versicolor Laich. (ColeopteraChrysomelidae). Bollettino dell’Istituto di Entomologia della Uni-versita degli Studi di Bologna , 22, 135–176.

Mellini, E. (1964) Biologia dei ditteri parassiti di coleotteri adulti.Bollettino dell’Istituto di Entomologia della Universita degli Studidi Bologna , 27, 171–192.

Mellini, E. (1975) Studi sui Ditteri Larvevoride. XXV. Sul determin-ismo ormonale delle influenze esercitate dagli ospiti sui loro paras-siti. Bollettino dell’Istituto di Entomologia della Universita degliStudi di Bologna , 31, 165–203.

Mellini, E. (1991) Sinossi di biologia dei Ditteri Larvevoridi. Bollettinodell’Istituto di Entomologia della Universita degli Studi di Bologna ,45 (1990), 1–38.

Mesnil, L.P. (1939) Essai sur les tachinaires (Larvaevoridae). Mono-graphies Publiees par les Stations et Laboratoires de RecherchesAgronomiques , 7, 67 + v pp.

Mesnil, L.P. (1944–1956) 64g. Larvaevorinae (Tachininae). [I.Salmaciini.]. Die Fliegen der Palaearktischen Region , 10, 21–554+ pls. I–XVII.

Mesnil, L.P. (1944–1975) 64g. Larvaevorinae (Tachininae). DieFliegen der Palaearktischen Region , 10, 1–1435 + pls. I–XXVII.

Mesnil, L.P. (1956–1965) 64g. Larvaevorinae (Tachininae). [II.Phorocerini oder Exoristini.]. Die Fliegen der PalaearktischenRegion , 10, 555–879 + pls. XVIII–XXVII.

Mesnil, L.P. (1966) 64g. Larvaevorinae (Tachininae). Die Fliegen derPalaearktischen Region , 10 (Lieferung 263), 881–928.

Mesnil, L.P. (1975a) 64g. Larvaevorinae (Tachininae). Die Fliegen derPalaearktischen Region , 10 (Lieferung 309), 1305–1384.

Mesnil, L.P. (1975b) 64g. Larvaevorinae (Tachininae). Die Fliegen derPalaearktischen Region , 10 (Lieferung 312), 1385–1435.

Nixon, K.C. (1999) Winclada , Version 1.00.08 [WWW document].URL http://www.cladistics.com [accessed on 31 August 2012].

O’Hara, J.E. (1985) Oviposition strategies in the Tachinidae, a familyof beneficial parasitic flies. University of Alberta Agriculture andForestry Bulletin , 8, 31–34.

O’Hara, J.E. (1989) Systematics of the genus group taxa of theSiphonini (Diptera: Tachinidae). Quaestiones Entomologicae, 25,1–229.

O’Hara, J.E. (2002) Revision of the Polideini (Tachinidae) of Americanorth of Mexico. Studia Dipterologica. Supplement , 10, 170 pp.

O’Hara, J.E. (2013a) Where in the world are all the tachinid genera?The Tachinid Times , 26, 10–16.

O’Hara, J.E. (2013b) History of tachinid classification (Diptera,Tachinidae). ZooKeys , 316, 1–34.

O’Hara, J.E. & Wood, D.M. (2004) Catalogue of the Tachinidae(Diptera) of America north of Mexico. Memoirs on Entomology,International , 18, iv + 410 pp.

Osten-Sacken, C.R. (1881) An essay of comparative chaetotaxy, or thearrangement of characteristic bristles of Diptera. Mittheilungen derMunchener Entomologischen Vereins , 5, 121–138.

Osten-Sacken, C.R. (1884) An essay of comparative chaetotaxy, or thearrangement of characteristic bristles of Diptera. Transactions of theEntomological Society of London , 1884, 497–517.



Pape, T. (1992) Phylogeny of the Tachinidae family-group. Tijdschriftvoor Entomologie, 135, 43–86.

Pape, T. & Arnaud, P.H. Jr. (2001) Bezzimyia – a genus of nativeNew World Rhinophoridae (Insecta, Diptera). Zoologica Scripta , 30,257–297. DOI: 10.1046/j.1463-6409.2001.00064.x.

Reeves, W.K. & O’Hara, J.E. (2004) First report of Strongygastertriangulifera (Diptera: Tachinidae) as parasitoid of a natharidbeetle, Chauliognathus pennsylvanicus (Coleoptera: Cantharidae).Canadian Entomologist , 136, 661–662.

Richter, V.A. (1991) A new tribe, new and little known speciesof the tachinid flies (Diptera, Tachinidae) of the fauna of theUSSR. Entomologiceskoe Obozrenie, 70, 229–246 (in Russian)[English version published in 1992, Entomological Review , 70(8),133–150].

Rognes, K. (1997) The Calliphoridae (blowflies) (Diptera: Oestroidea)are not a monophyletic group. Cladistics , 13, 27–66.

Roland, J., Denford, K.E. & Jimenez, L. (1995) Borneol as an attractantfor Cyzenis albicans , a tachinid parasitoid of the winter moth,Operophtera brumata L. (Lepidoptera: Geometridae). CanadianEntomologist , 127, 413–421.

Sabrosky, C.W. & Arnaud, P.H. Jr. (1965) Family Tachinidae (Lar-vaevoridae). A Catalog of the Diptera of America North of Mex-ico, Agriculture Handbook 276 (ed. by A. Stone, C.W. Sabrosky,W.W. Wirth, R.H. Foote and J.R. Coulson), pp. 961–1108.United States Department of Agriculture, Agricultural Research Ser-vice, Washington, DC.

Shima, H. (1989) Parasitic way of life in tachinid flies. Insectarium ,26, 4–9, 46–51, 88–94, 120–126 (in Japanese; cladogram on p.125 reproduced with English caption in The Tachinid Times , 3, 9(1990)).

Shima, H. (1999) Host-parasite catalog of Japanese Tachinidae(Diptera). Makunagi/Acta Dipterologica, Supplement , 1, 1–108.

Shima, H. (2006) A host-parasite catalog of Tachinidae (Diptera) ofJapan. Makunagi/Acta Dipterologica, Supplement , 2, 1–171.

Singh, B. & Wells, J.D. (2013) Molecular systematics of the Cal-liphoridae (Diptera: Oestroidea): evidence from one mitochondrialand three nuclear genes. Journal of Medical Entomology , 50, 15–23.DOI: http://dx.doi.org/10.1603/ME11288.

Smilanich, A.M., Dyer, L.A. & Gentry, G.L. (2009) The insect immuneresponse and other putative defenses as effective predictors ofparasitism. Ecology , 90, 1434–1440.

Stireman, J.O. III (2002) Phylogenetic relationships of tachinid fliesin subfamily Exoristinae (Tachinidae: Diptera) based on 28S rDNAand elongation factor-1α. Systematic Entomology , 27, 409–435.

Stireman, J.O. III, O’Hara, J.E. & Wood, D.M. (2006) Tachinidae:evolution, behavior, and ecology. Annual Review of Entomology ,51, 525–555.

Stireman, J.O. III, O’Hara, J.E., Moulton, J.K., Cerretti, P. & Winkler,I.S. (2013) Progress towards a phylogeny of world Tachinidae. Year1. The Tachinid Times , 26, 4–9.

Strand, M.R. (2008) The insect cellular immune response. InsectScience, 15, 1–14.

Sun, G., Ji, Q., Dilcher, D.L., Zheng, S., Nixon, K.C. & Wang, X.(2002) Archaefructaceae, a new basal angiosperm family. Science,296, 899–904.

Tachi, T. & Shima, H. (2010) Molecular phylogeny of the subfamilyExoristinae (Diptera, Tachinidae), with discussions on the evolution-ary history of female oviposition strategy. Systematic Entomology ,35, 148–163.

Tschorsnig, H.P. (1985) Taxonomie forstlich wichtiger Parasiten:Untersuchungen zur Struktur des mannlichen Postabdomens derRaupenfliegen (Diptera, Tachinidae). Stuttgarter Beitrage zurNaturkunde. Serie A (Biologie), 383, 1–137.

Tschorsnig, H.P. & Herting, B. (1994) Die Raupenfliegen (Diptera:Tachinidae) Mitteleuropas: Bestimmungstabellen und Angaben zurVerbreitung und Okologie der einzelnen Arten. Stuttgarter Beitragezur Naturkunde, Serie A (Biologie), 506, 1–170.

Tschorsnig, H.P. & Richter, V.A. (1998) Family Tachinidae. Contribu-tion to a Manual of Palaearctic Diptera (With Special Reference toFlies of Economic Importance), Vol. 3: Higher Brachycera (ed. byL. Papp and B. Darvas), pp. 691–827. Science Herald, Budapest.

Verbeke, J. (1962) Contribution a l’etude des Tachinidae africains(Diptera). Exploration Hydrobiologique des Lacs Kivu, Edouardet Albert (1952–1954). Resultats Scientifiques , 3, 77–187+ 25 pls.

Verbeke, J. (1963) The structure of the male genitalia in Tachinidae(Diptera) and their taxonomic value. Stuttgarter Beitrage zurNaturkunde, 114, 1–5.

Villeneuve, J. (1924) Contribution a la classification des “Tachinidae”palearctiques. Annales des Sciences Naturelles. Zoologie, Series 10 ,7, 5–39.

Villeneuve, J. (1933) Contribution a la classification des Tachinariaepalearctiques. Travaux des Ve Congres International d’Entomologie.Paris, 18–24 juillet 1932 , Vol. 2, pp. 243–255.

Wiegmann, B.M., Trautwein, M.D., Winkler, I.S. et al. (2011) Episodicradiations in the fly tree of life. Proceedings of the National Academyof Sciences of the United States of America , 108, 5690–5695. DOI:10.1073/pnas.1012675108.

Wood, D.M. (1985) A taxonomic conspectus of the Blondeliiniof North and Central America and the West Indies (Diptera:Tachinidae). Memoirs of the Entomological Society of Canada , 132,130 pp.

Wood, D.M. (1987a) Rhinophoridae. Manual of Nearctic Diptera , Vol.2, Research Branch, Agriculture Canada, Monograph 28 (ed. by J.F.McAlpine, B.V. Peterson, G.E. Shewell, H.J. Teskey, J.R. Vockerothand D.M. Wood), pp. 1187–1191. Canadian Government PublishingCentre, Hull.

Wood, D.M. (1987b) Tachinidae. Manual of Nearctic Diptera , Vol. 2,Research Branch, Agriculture Canada, Monograph 28 (ed. by J.F.McAlpine, B.V. Peterson, G.E. Shewell, H.J. Teskey, J.R. Vockerothand D.M. Wood), pp. 1193–1269. Canadian Government PublishingCentre, Hull.

Accepted 26 November 2013


signal through the noise? phylogeny of the tachinidae

Documents