Hennig’s orphans revisited: Testing morphological hypotheses in the “Opomyzoidea” (Diptera: Schizophora)

Molecular Phylogenetics and Evolution 54 (2010) 746–762

Contents lists available at ScienceDirect

Molecular Phylogenetics and Evolution journal homepage: www.elsevier.com/locate/ympev

Hennig’s orphans revisited: Testing morphological hypotheses in the ‘‘Opomyzoidea” (Diptera: Schizophora) Isaac S. Winkler a,d,*, Alessandra Rung b,c,d, Sonja J. Scheffer c,d a

Department of Entomology, North Carolina State University, Raleigh, NC 27695, USA Plant Pest Diagnositics Branch, California Department of Food and Agriculture, Sacramento, CA 95832, USA c Systematic Entomology Laboratory, USDA, Agricultural Research Service, Beltsville, MD 20705, USA d Department of Entomology, University of Maryland, College Park, MD 20742, USA b

a r t i c l e

i n f o

Article history: Received 26 February 2009 Revised 15 December 2009 Accepted 21 December 2009 Available online 28 December 2009 Keywords: Acalyptrate Phylogeny Higher classification Rapid radiation

a b s t r a c t The acalyptrate fly superfamily Opomyzoidea, as currently recognized, is a poorly-known group of 14 families. The composition of this group and relationships among included families have been controversial. Furthermore, the delimitation of two opomyzoid families, Aulacigastridae and Periscelididae, has been unstable with respect to placement of the genera Stenomicra, Cyamops, and Planinasus. To test the monophyly of Opomyzoidea, previously proposed relationships between families, and the position of the three problematic genera, we sequenced over 3300 bp of nucleotide sequence data from the 28S ribosomal DNA and CAD (rudimentary) genes from 29 taxa representing all opomyzoid families, as well as 13 outgroup taxa. Relationships recovered differed between analyses, and only branches supporting wellestablished monophyletic families were recovered with high support, with a few exceptions. Opomyzoidea and its included subgroup, Asteioinea, were found to be non-monophyletic. Stenomicra, Cyamops, and Planinasus group consistently with Aulacigastridae, contrary to recent classifications. Xenasteiidae and Australimyzidae, two small, monogeneric families placed in separate superfamilies, were strongly supported as sister groups. Ó 2009 Elsevier Inc. All rights reserved.

1. Introduction 1.1. General overview Willi Hennig is rightly known as the founder of modern phylogenetic systematics, but his equally groundbreaking contributions to Diptera systematics are less widely appreciated (Meier, 2005). Hennig’s extensive insight on relationships of the ‘‘acalyptrate” flies is particularly relevant today, because these relationships remain largely unresolved. In several of his major works on Diptera, Hennig gave much attention to several obscure acalyptrate families with especially problematic relationships, which he first listed as ‘‘families with unclear affinities” (Hennig, 1958), but later provisionally grouped together as the ‘‘Anthomyzoidea” (Hennig, 1971). In the classification which is most widely used today (McAlpine, 1989), this grouping remains largely intact in the superfamily Opomyzoidea, though the evidence supporting it is still weak. ‘‘Acalyptratae” is a probably paraphyletic group within Schizophora of 27,000 described species in about 70 families, representing approximately 20% of total dipteran species (Evenhuis et al., * Corresponding author. Address: Deptartment of Entomology, North Carolina State University, Campus Box 7613, Raleigh, NC 27695, USA. Fax: +1 919 515 7746. E-mail address: [email protected] (I.S. Winkler). 1055-7903/$ - see front matter Ó 2009 Elsevier Inc. All rights reserved. doi:10.1016/j.ympev.2009.12.016

2008). The much-studied model organism Drosophila melanogaster represents a typical acalyptrate to many biologists, but the group also includes a few large and economically important families such as Tephritidae (fruit flies) and Agromyzidae (leaf-mining flies), as well as obscure and poorly-known families such as Strongylophthalmyiidae (Shatalkin, 1993), Gobryidae (McAlpine, 1997), and Australimyzidae (Brake and Mathis, 2007). Acalyptrate flies are very diverse and generally poorly collected; as a result, new taxa are continuously being added from all zoogeographic regions (e.g., Buck, 2006; Barraclough and McAlpine, 2006). Determining phylogenetic relationships among families in this assemblage is arguably the most difficult problem in the systematics of Diptera (McAlpine, 1989), and a consensus classification has not yet been achieved. According to the ‘‘Manual of Nearctic Diptera” (McAlpine, 1989), Acalyptratae is monophyletic and is divided into 10 superfamilies. Of these, Nerioidea, Conopoidea, Sciomyzoidea (sometimes including Lauxanioidea), and Ephydroidea are also recognized in competing classification systems (Hennig, 1958, 1971, 1973; Colless and McAlpine, 1991; Griffiths, 1972). The limits, composition and phylogenetic relationships of the remaining superfamilies have been historically more problematic. This work is directed towards testing the monophyly of the superfamily Opomyzoidea, one of the most problematic acalyprate superfamilies, and testing previous hypotheses of relationships be-

747

I.S. Winkler et al. / Molecular Phylogenetics and Evolution 54 (2010) 746–762

tween its component families. The superfamily, as currently recognized, contains 14 families (excluding Acartophthalmidae; see below) and about 120 extant genera (see Table 2 and references therein). The composition of Opomyzoidea, the limits of some of its included families, and its position with respect to the remaining acalyptratae have been strongly debated. In order to advance knowledge of this little-known and controversial superfamily, we present in this paper the first molecular phylogenetic analysis of opomyzoid families, using partial sequences of two nuclear genes, the 28S ribosomal DNA and the protein-coding gene CAD (rudimentary). We focus particularly on the composition of two included families, Periscelididae and Aulacigastridae. We then use our results to evaluate the phylogenetic hypothesis of J.F. McAlpine (1989) against alternative proposals. 1.2. Taxonomic history and alternative hypotheses As defined by J.F. McAlpine (1989), Opomyzoidea includes 13 families (Table 1, Fig. 1) that are grouped into four suprafamilies: Clusioinea (Clusiidae and Acartophthalmidae), Agromyzoinea (Agromyzidae, Fergusoninidae, and Odiniidae), Opomyzoinea (Opomyzidae and Anthomyzidae), and Asteioinea (Asteiidae, Aulacigastridae, Xenasteiidae, Teratomyzidae, Neurochaetidae, and Periscelididae). Two families named more recently have been associated with Opomyzoidea after McAlpine (1989): Neminidae (Freidberg, 1994; described as a subfamily of Aulacigastridae by McAlpine, 1983) and Marginidae (newly described and provisionally placed in Opomyzoidea by McAlpine, 1991). In addition, McAlpine (1989; following Hennig, 1971) suggested that Paraleucopis Malloch (currently in Chamaemyiidae) and the unplaced Neotropical genera Mallochianamyia Santos–Neto (= Gayomyia Malloch; see Wheeler, 2000) and Schizostomyia Malloch probably represent a separate lineage closely related to Asteiidae. Acartophthalmidae is now generally considered to belong in the superfamily Carnoidea (Brake, 2000; Buck, 2006; see also Griffiths, 1972).

Hennig (1971, 1973) grouped the majority of opomyzoid families under the name ‘‘Anthomyzoidea”. It included Acartophthalmidae, Anthomyzidae, Asteiidae, Aulacigastridae, Clusiidae, Opomyzidae, Periscelididae, Teratomyzidae, and Chyromyidae. Hennig’s classification placed two remaining opomyzoid families, sensu J.F. McAlpine (1989), into a separate superfamily, Agromyzoidea (Odiniidae and Agromyzidae), and left Fergusoninidae unplaced. Xenasteiidae and Neurochaetidae were described after Hennig’s publications. D.K. McAlpine’s most comprehensive classification system (Colless and McAlpine, 1991) defined Opomyzoidea as containing the families Acartophthalmidae, Agromyzidae, Clusiidae, Fergusoninidae, Odiniidae, Opomyzidae, Xenasteiidae and Carnidae (Table 1). The remaining families of Opomyzoidea sensu J.F. McAlpine (1989) were classified in the superfamily Asteioidea. This classification partially follows results of previous studies (e.g. McAlpine, 1978, 1983), but other proposed groupings were not explicitly justified. D.K. McAlpine (1991) also tentatively added the Afrotropical family Marginidae to Opomyzoidea. Later, D.K. McAlpine (1997) reaffirmed his earlier separation of Opomyzoidea and Asteioidea and questioned the putative synapomorphies leading to J.F. McAlpine’s broader concept of Opomyzoidea. Contrasting with Hennig (1971) and Colless and McAlpine (1991), Griffiths’ general classification system has several unique characteristics. Griffiths (1972) divided Schizophora into five superfamilies, Lonchaeoidea, Lauxanioidea, Drosophiloidea, Nothyboidea and Muscoidea. Muscoidea was further divided into nine ‘‘prefamilies,” one of which was Calyptratae. Anthomyzoinea constituted another of the muscoid prefamilies in Griffiths’ system, including the following families currently placed in Opomyzoidea (sensu McAlpine, 1989): Aulacigastridae, Anthomyzidae, Asteiidae, and Opomyzidae. Griffiths’ Anthomyzoinea also included Chyromyidae, Heleomyzidae, Trixoscelididae, Rhinotoridae, and Sphaeroceridae, all in the superfamily Sphaeroceroidea of McAlpine (1989) (the last four families were recently included in a single

Table 1 Composition of the ‘‘Opomyzoidea” according to different authors. The original name of the superfamily is given below the author’s name. The families considered to belong in the ‘‘Opomyzoidea,” according to each author, are marked with an ‘‘X”. An ‘‘” indicates that the family was not described at the time. Putative outgroups (according to McAlpine) are listed below the line. Literature references are as follows: Hennig (1971, 1973), Griffiths (1972), McAlpine (1989), and Colless and McAlpine (1991). Family

Acartophthalmidaea Agromyzidae Asteiidae Anthomyzidae Aulacigastridae Clusiidae Fergusoninidae Neurochaetidae Odiniidae Opomyzidae Periscelididae Teratomyzidae Xenasteiidae Carnidae Chyromyidae Heleomyzidaeg Sphaeroceridaeg a

Author (Group name) Hennig (Anthomyzoidea)

Griffiths (Anthomyzoinea)

J.F. McAlpine (Opomyzoidea)

D.K. McAlpine (Opomyzoidea)

X (Agromyzoidea) X X Xb X (Unplaced) d (Agromyzoidea) X X X  (Chloropoidea) X (unplaced) (unplaced)

(Tephritoinea) (Agromyzoinea) X X Xc (Agromyzoinea) (Unplaced)  (Tephritoinea) X (Nothyboidea) (Nothyboidea)  (Tephritoinea) X X X

X X X X Xb,e X X X X X X X X (Carnoidea) (Sphaeroceroidea) (Sphaeroceroidea) (Sphaeroceroidea)

X X (Asteioidea) (Asteioidea) (Asteioidea)e X X (Asteioidea) X X (Asteioidea)b (Asteioidea) X Xf (Heleomyzoidea) (Heleomyzoidea) (Heleomyzoidea)

Moved to Carnoidea by recent authors (Brake, 2000; Buck, 2006). Includes Cyamops, Stenomicra, and Planinasus. c Includes Cyamops and Planinasus. d Hennig (1971) suggested the Baltic amber fossil Anthoclusia may represent separate lineage in Anthomyzoidea. This species was placed in the new family Neurochaetidae by McAlpine (1978). e Including Nemininae. J.F. McAlpine (1989) mentions D.K. McAlpine’s inclusion of Ningulus and Nemo, but the putative synapomorphies he discusses for Aulacigastridae are not found in these genera. f Including Australimyza. g Subsumed within the expanded family Heteromyzidae by McAlpine (2007). b

748

I.S. Winkler et al. / Molecular Phylogenetics and Evolution 54 (2010) 746–762

Malloch, Parascutops Mathis and Papp, Periscelis Loew, and Scutops Coquillett), and Aulacigastridae sensu stricto (containing Aulacigaster Macquart and Curiosimusca Rung, Mathis, and Papp) are generally recognized as monophyletic units supported by a handful of putative synapomorphies (see Mathis and Rung, 2004; Rung et al., 2005). The debate has largely centered on the position of Stenomicra Coquillett, Cyamops Melander, and Planinasus Cresson, comprising the subfamily Stenomicrinae (Mathis and Papp, 1992; Mathis, in press). This group also now includes two newly described taxa closely related to Cyamops and Stenomicra, the fossil Procyamops succini Hoffeins and Rung from Dominican amber (Hoffeins and Rung, 2005) and the extant genus Stenocyamops Papp from Thailand (Papp et al., 2006). Cyamops had been previously placed in either Anthomyzidae or Periscelididae. Planinasus was originally described in Ephydridae and later transferred to Periscelididae. The genus Stenomicra has a particularly convoluted classification history, having been placed into six different families since its description in the early twentieth century or classified as the type genus of a separate family (Papp, 1984; Khoo and Sabrosky, 1989). Prior to 1971 it was often included in Anthomyzidae. Hennig (1958, 1969, 1971), after a detailed discussion of these genera, was the first to include Cyamops, Planinasus, and later Stenomicra in Aulacigastridae, a placement followed by Griffiths (1972) and J.F. McAlpine (1989), the former with the exclusion of Stenomicra (provisionally retained in Anthomyzidae). Later, D.K. McAlpine (1978, 1983) transferred all three genera to the Periscelididae on the basis of similar antennal structure, a placement that has been followed by recent authors (Grimaldi and Mathis, 1993; Baptista and Mathis, 1994; Mathis and Papp, 1998; Mathis, in press). 1.3. Biology and diversity of Opomyzoidea Fig. 1. Phylogeny and classification of the Opomyzoidea according to J.F. McAlpine (1989). McAlpine’s Aulacigastridae includes Cyamops, Stenomicra, and Planinasus, and presumably also Ningulus and Nemo (Nemininae of D.K. McAlpine, 1983), though the synapomorphies he discusses for Aulacigastridae are not present in the latter two genera.

family ‘‘Heteromyzidae” by McAlpine, 2007). Griffiths classified the families Agromyzidae and Clusiidae separately in the prefamily Agromyzoinea. A particularly innovative aspect of Griffiths’ work is the classification of Periscelididae and Teratomyzidae in the superfamily Nothyboidea, with the families Psilidae and Nothybidae (currently classified in the superfamily Diopsoidea; J.F. McAlpine, 1989; D.K. McAlphine, 1997), instead of in Anthomyzoinea. Griffiths’ classification has been criticized on the grounds that it emphasizes a single, highly variable character system (the male postabdomen; see e.g. McAlpine, 1978), used inadequate taxon sampling, and some relationships he proposed are at odds with those currently accepted. However, Griffiths’ work remains influential, and evidence for many competing hypotheses is scant. J.F. McAlpine (1989) summarized previous work on higher classification and phylogeny of cyclorrhaphan (‘‘higher”) flies, and presented a fully resolved family-level phylogenetic hypothesis for the Acalyptratae. His work was the first to propose Opomyzoidea as a monophyletic group including the taxa (except Chyromyidae) placed by Hennig (1971) in Anthomyzoidea and Agromyzoidea. We used J.F. McAlpine’s (1989) phylogeny in this study as the major source of hypotheses to be tested. Contrasting with the lack of robust resolution of interfamilial relationships, the composition of families within Opomyzoidea is now considered relatively stable, with the exception of two families: Aulacigastridae and Periscelididae. Genera have been transferred between these two families and sometimes transferred to seemingly distantly-related groups. Periscelididae sensu stricto (containing Diopsosoma Malloch, Marbenia Malloch, Neoscutops

Although a few opomyzoids, such as the Agromyzidae, are widespread and abundant, most opomyzoid families are generally rarely collected and poorly-known, being among the most obscure groups of Diptera. In fact, five of the fifteen families putatively belonging to Opomyzoidea were undescribed and mostly unknown prior to 1970. For example, the diminutive Xenasteiidae were independently described from Mediterranean and Pacific seashores less than three decades ago (Hardy, 1980; Papp, 1980) and are still largely unknown. Neminidae, dubbed ‘‘nobody flies” by their discoverer (Latin nemo = ‘‘nobody”; McAlpine, 1983, 1998), are now known from southern Africa, Australia, and New Guinea, but have been very rarely collected, and nothing is known about their biology. Although easily recognized by the laterally elongate head produced into eyestalks (Mathis and Rung, 2004), the bizarre periscelidid genus Diopsosoma, with only one included species, D. primum Malloch, is still known only from the type series collected in Peru in 1931 (Mathis and Rung, 2004). Some opomyzoid taxa are restricted to very specific microhabitats (such as sap fluxes of wounded trees or phytotelmata on specific plant species) and are thus infrequently encountered. Furthermore, many opomyzoid families exhibit limited diversity and distribution, especially in northern temperate regions (Table 2). Significant undescribed species diversity probably exists in many opomyzoid families. For example, about twenty species of the widespread genus Stenomicra (Periscelididae) have been described, including a few in north temperate regions, but over 100 undescribed tropical species are represented in museum collections. Because several opomyzoid families are either rare or simply poorly collected, knowledge about their behavior and biology are fairly incomplete (Table 2). Phytophagy is known to occur in five families: Agromyzidae (mostly leaf-miners on many different plants) (Spencer, 1990; Scheffer et al., 2007), Anthomyzidae and Opomyzidae (larvae feeding in the stems of grasses) (Vockeroth, 1987; Brunel, 1998; Rohácˇek, 1998a,b), Asteiidae (species of Asteia

749

I.S. Winkler et al. / Molecular Phylogenetics and Evolution 54 (2010) 746–762

Table 2 Biology, diversity, and distribution of families assigned to Opomyzoidea by McAlpine (1989) and subsequent authors. Species diversity estimates were obtained from a working version of the Biosystematic Database of World Diptera (Evenhuis et al., 2008), in addition to references listed below. Family

Diversity

Distribution

Known larval biologies

References

Acartophthalmidae Agromyzidae

1 gen., 4 spp. 28 gen., ,800 spp.

3 holarctic, 1 Australian Widespread

Papp and Ozerov (1998) Spencer (1990), Scheffer et al. (2007)

Anthomyzidae

19 gen., 94 spp.

Asteiidae

10 gen., 135 spp.

Widespread (predominantly holarctic) Widespread

Aulacigastridae sensu stricto Clusiidae

Fergusoninidae

2 gen., ca 58 spp. (most undescribed) 14 gen., 520 spp. (some currently in litt.) 1 gen., 31 + spp.

Rotting fungi and carrion in damp woodlands Phytophagous in many plants (leaf-miners, stem-borers, etc.) Mostly phytophagous or saprophagous in graminoids, larvae of 2 spp. in fungi Fungivorous (Leiomyza), under bark of dead trees (Phlebosotera), saprophagous or phytophagous in plants Sap fluxes of decidous trees, phytotelmata of bromeliads Rotting fungi in damp wood

Marginidae

1 gen., 2 spp.

Neminidae Neurochaetidae

3 gen., 14 spp. 4 gen., 20 spp.

Odiniidae

11 gen., 63 spp.

Opomyzidae

4 gen., 59 spp.

Periscelididae sensu stricto Stenomicrinae (Aulacigastridae or Periscelididae) Teratomyzidae

6 gen., 27 spp.

Xenasteiidae

Widespread (except Australia and Oceania) Widespread

Mostly Australia, also N. Guinea, N. Zealand, tropical Asia Southern Africa, Madagascar Australia, southern Africa Australia, southern Africa, tropical Asia Worldwide

4 genera, 52 spp. (+ many undescribed)

Mostly holarctic, also in Africa Widespread (Periscelis), others Neotropical Widespread, mostly tropics

7 gen., 35 + spp. (only 8 described) 1 gen., 13 spp.

Australia, N. Zealand, E. Asia, S. America Mediterranean, Oceania

Rohácˇek, (1998a,b, 2006)

Gallers on Myrtaceae in association with nematodes

Murphy (1990), Papp (1998b), Freidberg (2002); Freidberg, personal communication Papp (1998a), Rung et al. (2005); Rung and Mathis, unpublished Lonsdale and Marshall (2006, 2007, 2008); Lonsdale, personal communication Scheffer et al. (2004), Taylor et al. (2004, 2005)

Unknown

McAlpine (1991)

Unknown (adults on tree trunks) Watery medium inside Alocasia inflorescence; probably phytotelmata of other plants Saprophages and predators in galleries of wood-boring insects Phytophagous (in grass stems)

McAlpine (1983), Freidberg (1994) McAlpine (1978, 1988a,b, 1993)

Sap fluxes of decidous trees, wood borer tunnels Plant phytotelmata (Stenomicra)

Mathis and Papp (1998), Mathis (in press) Fish (1983), Baptista and Mathis (1994), Papp et al. (2006)

Associated with ferns; larvae feed externally on microflora? Unknown (vegetation near seashore)

McAlpine and de Keyzer (1994)

Meigen reared as primary phytophages from Chenopodiaceae, Asteraceae, and Acanthaceae; Murphy, 1990; Freidberg, personal communication) and Fergusoninidae (gallers on Myrtaceae) (Currie, 1937; Taylor et al., 2004, 2005). Fungal associations are documented for Acartophthalmidae (Papp and Ozerov, 1998), some Asteiidae (Leiomyza Macquart; Papp, 1998b) and two species of Anthomyzidae (Fungomyza Rohácˇek; Rohácˇek, 2006). Larvae and adults of temperate species of Periscelididae sensu stricto (Mathis and Papp, 1998) and Aulacigastridae (Papp, 1998a), as well as some adult Odiniidae (Gaimari and Mathis, in press) have been collected from sap fluxes of deciduous trees. Odiniid flies display mixed feeding strategies. Some species apparently can feed and fully develop solely on frass and fungi within galleries of wood-boring beetles, lepidopterans and flies, but they sometimes attack and predate upon its residents. Some species have also been reared from the egg masses of mealybugs associated with tree bark (Gaimari and Mathis, in press). Aquatic or semi-aquatic larvae are found within three families: Aulacigastridae, Periscelididae and Neurochaetidae. Undescribed New World species of Aulacigaster (Rung and Mathis, unpublished) have fully aquatic larvae, developing in the phytotelmata of bromeliads. The same may be true for several undescribed species of Stenomicra (Periscelididae) collected from these plants (see Fish, 1983). Two other genera of Periscelididae, Cyamops and Planinasus, have been collected near streams, bogs, and marshes, and larvae may be aquatic or semi-aquatic. Larvae of Neurochaeta inversa McAlpine (Neurochaetidae) develop within a watery liquid between developing fruitlets of Alocasia (Schott) (McAlpine, 1978). Other species are presumed to similarly develop in phytotelmata in other plants on which the adults have been collected, including Pandanus Parkinson, Ravenala Adans., Zingiber Boehm., Musa L., Strelitzia Aiton, and Xanthorrhoea Smith (McAlpine, 1988b; Britton, 2009).

Gaimari and Mathis (in press) Vockeroth (1987), Brunel (1998)

Hardy (1980), Papp (1980, 1998c)

2. Materials and methods 2.1. Taxon sampling In accordance with our stated goals, the set of exemplar species selected (Appendix A) comprises several elements. First, a putative ingroup sample was constructed that included 29 opomyzoid taxa distributed in all 15 putative component families. Outgroups were chosen initially on the basis of McAlpine’s (1989) hypothesis of a sister-group relationship between Opomyzoidea and Carnoidea. Five genera of the latter, representing the families Carnidae, Australimyzidae, Milichiidae, and Chloropidae were sampled. To this list of putative outgroups were added exemplars included in other acalyptrate superfamilies by McAlpine (1989), which have been postulated by other authors to be related to various ‘‘opomyzoid” families. These are as follows: (1) Representatives of Chyromyidae (two genera), placed in Sphaeroceroidea by McAlpine (1989), were included to test Hennig’s (1971) and Griffiths’ (1972) suggestion that the family instead lies close to Aulacigastridae. (2) A single representative of Sphaeroceridae was added to test Griffiths’ (1972) placement of this and related taxa (‘‘Heteromyzidae”; see McAlpine, 2007) in Anthomyzoinea. (3) Representatives of Psilidae (two genera) were included to test Griffiths’ (1972) hypothesis that the families Teratomyzidae and Periscelididae sensu stricto are more closely related to this family than to other opomyzoid families. (4) We also included one species of Tephritidae (superfamily Tephritoidea) to test Griffiths’ (1972) expanded concept of this superfamily, which included Odiniidae and Acartophthalmidae, as well as members of McAlpine’s ‘‘Carnoidea”. Finally, sequences of Musca domestica L. (Calyptratae: Muscoidea: Muscidae) and Rhingia nasica Say, belonging to the non-schizopho-

750

I.S. Winkler et al. / Molecular Phylogenetics and Evolution 54 (2010) 746–762

ran family Syrphidae, were obtained from Genbank and included to allow unambiguous rooting of the tree. 2.2. Sequence data collection Adults representing the various species included (see Appendix A) were collected and preserved in 70–100% ethanol for various amounts of time. In the lab, specimens were transferred to vials containing 100% ethanol and stored in a 80 °C freezer until DNA extraction. Two rare taxa (Margo clausa McAlpine and Nemula longarista Freidberg) were sequenced from recently collected pinned specimens and DNA extracted by soaking the entire specimen in a buffer/proteinase solution for 36–48 h, followed by removal of the specimen and adherence to standard protocol using the DNeasy Blood and Tissue Kit (QIAGEN, Valencia, CA). DNA from remaining specimens was extracted from individual fly specimens following standard protocols in the DNEasy Blood and Tissue Kit. Before DNA extraction, the postabdomen and wings were preserved in microvials containing glycerin. These vials have been kept as vouchers and are deposited at the National Museum of Natural History (Smithsonian Institution, USA) collection. Polymerase chain reaction (PCR) was carried out using TaKaRa ExTaq (TaKaRa, U.S.A.) following the manufacturer’s protocol with 1–2 ll of DNA extract in a total of 50 ll solution. A large (2700 bp) region of 28S ribosomal DNA was amplified in 2–3 overlapping pieces and sequenced for all taxa, and a small (680 bp) region of CAD (rudimentary) representing the initial 5’ portion was obtained for 29 of the 42 taxa. PCR primers used were standard primer pairs used for other Diptera. These were rc28C/

Table 3 Sequencing primers designed for this study. Position numbers indicate the annealing site of the 50 end to the Drosophila melanogaster ribosomal RNA sequence published by Tautz et al. (1988). Primer direction is indicated by an F (forward) or R (reverse) in the primer name. Primer

Sequence

Position

28P1F 28P2F 28P2R 28JR

AAC AAC TCA CCT GCC GAA GCA ACT AG GTG AAC AGT GGT TGA TCA CGA G CTC GTG ATC AAC CAC TGT TCA C GGT YGT TCC TGT TRC CAG GAT GAG

4747 5005 5027 5200

28H and rc28D/28Z for the 28S gene (Wiegmann et al., 2000), and 54F/405R for CAD (Moulton and Wiegmann, 2004). For 28S, a standard touchdown PCR protocol was used: 94 °C for 4 min followed by 30 cycles of 94 °C for 1 min, 48–56 °C for 1 min, and 72 °C for 1 min 30 s with a final extension at 72 °C for 4 min. The same program was used for CAD, with the addition of five extra cycles before the final extension to obtain a higher yield of product. PCR product was purified using the QIAquick PCR purification kits (QIAGEN). Sequencing reactions were carried out with PE Applied Biosystems Big Dye Terminator sequencing kits (Perkin-Elmer Applied Biosystems, Foster City, CA). Besides the standard 28S and CAD primers (Wiegmann et al., 2000; Moulton and Wiegmann, 2004), a few additional primers were used for sequencing portions of 28S in some taxa (Table 3). Sequence data were obtained by analyzing samples on ABI 377 and ABI 3130 automated sequencers. Sequences from both strands were assembled and edited as necessary using Sequencher (Gene Codes Corp., Ann Arbor, Michigan, U.S.A.). Multiple alignment of 28S was carried out with Muscle v.3.6 (Edgar, 2004) using default parameters. The program Gblocks (Castresana, 2000, 2002) was then used to identify regions of ambiguous alignment to exclude. Several different parameter combinations were tried in Gblocks, but the final data set used the parameters 27 (minimum number of sequences for a conserved position, out of 42 total), 39 (minimum number of sequences for a flanking position), 5 (maximum number of contiguous nonconserved positions), 3 (minimum length of a block), and half (allowed gap positions). The set of included characters was adjusted slightly by hand to account for incongruous handling of positions with gaps. Nucleotide sequences for CAD were aligned using Clustal X (Thompson et al., 1994). The portion of CAD used excluded a nearby intron, and very minimal length variation was observed. All sequences are deposited in Genbank under accession numbers GU299217–GU299280.

2.3. Data analysis Data sets were first examined for base composition bias in PAUP* 4.0b10 (Swofford, 2001). Uncorrected (P) distances and corrected (ML, GTR + I + G) distances were also calculated in PAUP* for all pairwise comparisons and plotted to examine levels of diver-

Fig. 2. Uncorrected pairwise (P) distances plotted against corrected (ML, GTR + I + G) distances for each of the three gene partions (28S, CAD positions 1 and 2, CAD position 3). Third positions of CAD show very high divergence levels and evidence of saturation at deep (interfamilial) comparisons.

751

I.S. Winkler et al. / Molecular Phylogenetics and Evolution 54 (2010) 746–762 Table 4 Base frequecies observed for the gene partitions examined in this study. 28S was analyzed both with the entire included sequence (i.e. excluding hypervariable regions), and with only those sites which showed variation (also excluding hypervariable regions). Base frequencies are exceptionally variable in third codon positions of CAD, which were excluded from further analyses. Taxon

Proportion G + C 28S (all excluding 28S (included CAD pos. CAD ypervariable) variable sites) 1&2 pos. 3

Rhingia Musca Mumetopia Pterogramma Anthomyza Xenasteia Australimyza Teratomyza Neurochaeta Leiomyza Asteia Chyliza Periscelis Loxocera Procecidochares Margo Acartophthalmus Opomyza Geomyza Aphaniosoma Gymnochiromyia Nemula Meoneura Fergusonina sp. F. purcelli Odinia Neoalticomerus Agromyza Cerodontha Liriomyza Clusiodes Clusia Paramyia Chaetochlorops Thaumatomyia Planinasus Stenomicra Cyamops sp. C. nebulosus Aulacigaster ecuadoriensis A. leucopeza A. neoleucopeza Mean

0.428 0.434 0.433 0.433 0.435 0.430 0.430 0.428 0.429 0.422 0.429 0.423 0.417 0.428 0.429 0.424 0.430 0.429 0.431 0.422 0.431 0.425 0.432 0.424 0.422 0.428 0.430 0.425 0.416 0.420 0.422 0.418 0.436 0.431 0.425 0.424 0.425 0.425 0.423 0.420 0.422 0.423 0.426

0.407 0.434 0.434 0.434 0.438 0.416 0.417 0.410 0.413 0.392 0.411 0.385 0.366 0.409 0.411 0.389 0.420 0.416 0.420 0.399 0.420 0.395 0.437 0.388 0.389 0.414 0.418 0.394 0.354 0.372 0.389 0.363 0.441 0.419 0.393 0.388 0.395 0.399 0.385 0.378 0.378 0.386 0.403

0.463 0.442 — 0.449 — — 0.486 0.461 0.438 0.464 0.462 0.443 0.472 0.448 0.426 — 0.464 0.465 0.460 — 0.490 — 0.455 0.451 0.457 0.434 — 0.461 — 0.438 0.470 0.477 0.452 0.424 0.428 — 0.360 — — 0.421 0.430 0.424 0.447

0.460 0.476 — 0.422 — — 0.467 0.342 0.235 0.586 0.472 0.293 0.398 0.381 0.333 — 0.482 0.592 0.565 — 0.553 — 0.414 0.450 0.317 0.264 — 0.450 — 0.296 0.566 0.435 0.265 0.140 0.246 — 0.094 — — 0.274 0.238 0.202 0.367

gence and possible signal saturation. Based on these results, third codon positions of CAD, which were highly divergent and extremely biased (Fig. 2, Table 4), were excluded from further analyses. The final data set included 2889 bp (2433 of 28S and 456 of CAD) of which 688 were parismony informative (367 of 28S and 321 of CAD). Analyses were then performed on the separate and combined data sets using maximum parsimony (MP) in PAUP* with TBR swapping on 100 random addition trees. The maximum likelihood (ML) criterion was implemented with a GTR + I + G model in Garli 0.951 (Zwickl, 2006). One hundred separate runs were performed with each data set, as this was found to be necessary due to differences in topology (with small differences in likelihood score) obtained in successive runs. Bootstrap analyses (500 reps.) were conducted using both MP and ML criteria. Next, the combined data set was subjected to Bayesian inference (BI) using Mr. Bayes version 3.1.2p (Ronquist and Huelsenbeck, 2003) with two independent runs of four chains each run on eight processors. For this analysis, the genes were modeled separately, each using the GTR + I + G model. The analysis was extended to 10 million gener-

Table 5 Results of topology tests in CONSEL (Shimodaira and Hasegawa, 2002). Results significant at the 0.05 level are marked by , and at the 0.10 level by . AU = approximately unbiased test; SH = Shimodaira–Hasegawa test. Hypothesized clade

ln L

AU

SH

Unconstrained phylogeny Opomyzoidea (excl. Margo) Above—Xenasteiidae Above—Clusioinea Above—Opomyzoinea Above—Periscelis Asteioinea Above—Xenasteiidae Above—Periscelis Asteioinea grp. 1 Above—Neminidae Above—Neurochaetidae Above—Aulacigaster (= Periscelididae s. lat.) Agromyzoinea Opomyzoinea Clusioinea

18,239.3 18,322.0 18,291.9 18,287.8 18,257.1 18,247.9 18,305.8 18,272.3 18,263.3 18,271.0 18,255.7 18,249.2 18,257.3 18,247.3 18,257.9 18,249.2

N/A <0.001** 0.047** 0.051*

N/A 0.012** 0.065* 0.093*

0.281 0.402 <0.001** 0.075* 0.154 0.117 0.243 0.523 0.128 0.363 0.325 0.346

0.544 0.688 0.005** 0.118 0.243 0.194 0.447 0.620 0.405 0.375 0.302 0.334

ations at which point the convergence statistic (standard deviation of split frequencies between independent runs) was approximately 0.04. Alternative topologies representing a priori hypotheses of relationships were tested to see if they differed significantly from the ML tree based on the combined data. For some hypotheses of monophyly (e.g. Opomyzoidea), additional hypotheses representing sequential modification of the original hypothesis were also tested (i.e. subtracting families; see table 5). Hypothetical clades tested were Opomyzoidea exclusive of Marginidae, Asteioinea, subgroup 1 of Asteioinea (see Fig. 1), Agromyzoinea, Anthomyzoinea, Clusioinea, and Periscelididae sensu lato (including Stenomicra, Cyamops, and Planinasus). To accomplish this, ML analyses were first repeated in Garli (20 runs) with the hypothesized group constrained as monophyletic. Site likelihoods were obtained separately for these topologies and for the unconstrained ML topology in PAUP* (Swofford, 2001) using the model parameters and branch lengths estimated by Garli for each. These site likelihoods served as input for the program CONSEL (Shimodaira and Hasegawa, 2001), which performed the Shimodaira–Hasegawa (SH) test (Shimodaira and Hasegawa, 1999) and the Approximately Unbiased (AU) test (Shimodaira, 2002). P-Values resulting from these tests (Table 5) represent the probability that a single constrained phylogeny represents the true optimal tree given the provided set of topologies and bootstrapped data sets generated by the program. Topologies representing related, nested hypotheses were tested simultaneously; testing these independently changed the Pvalues slightly, but had no effect on the final inferences. 3. Results Although containing some common elements, trees from each analysis and gene partition differed substantially from each other. Also, exploratory analyses with different alignments, inclusion sets, and taxon sampling resulted in different topologies, though sometimes only slightly so. Each of the 11 families represented by two or more species (except for Periscelididae sensu lato) were generally recovered as monophyletic, except for Psilidae, which was recovered only by the combined ML and BI analyses. Opomyzoidea and the four included suprafamilies were not monophyletic in any of the analyses. Divergence levels for 28S and for first and second positions of CAD were low to moderate (P distance < 0.25; Fig. 2) and the plot of P distance vs. ML distance showed little evidence of saturation. Third positions of CAD were, however, highly divergent (P distance

752

I.S. Winkler et al. / Molecular Phylogenetics and Evolution 54 (2010) 746–762

mostly > 0.4) and the plot showed strong evidence of saturation past the most recent divergences (Fig. 2). Base frequencies at third positions of CAD were exceptionally variable (9–59% G + C; see Table 4), but frequencies were not significantly variable in the included portions of both genes, as calculated in PAUP* (ignoring phylogenetic correlation). However, G + C content does vary in slight but potentially consequential ways when only the included portions of 28S and CAD are considered. For example, the families Anthomyzidae, Sphaeroceridae, Muscidae, and Carnidae share a slightly elevated (3% higher) G + C content in 28S compared to the average value, and these families grouped together in the 28S ML analysis (Fig. 5). Parsimony analysis of the combined data set resulted in two most parsimonious trees, the consensus of which (Fig. 3) was fully resolved except for a single node. ML analysis resulted in a much

different tree (Fig. 4). Bootstrapping showed that, with few exceptions, only relationships uniting families were strongly supported by either of these analyses. Fig. 5 shows the Bayesian majority rule consensus tree, which showed levels of support generally comparable to the ML analysis, but with moderate to strong support for three additional relationships recovered with weak support in the ML analysis (see below). Although Bayesian values may be subject to artificial inflation (Suzuki et al., 2002), this difference may also reflect the desirable effect of modeling genes separately. Although 28S and CAD had approximately the same number of parsimony informative sites, CAD performed much more poorly as measured by the number of families recovered in both the MP and ML individual analyses (four of eight for CAD and 10 of 11 for 28S). Because of this, the CAD only analyses were not further considered. The 28S MP analysis resulted in 40 MP trees, the consensus of

Fig. 3. Strict consensus cladogram of the two most parsimonious trees obtained from analysis of the combined dataset (length 2928, CI = 0.3914, RI = 0.3953). Bootstrap values (500 reps.) are noted if greater than 50%. Taxa included by McAlpine (1989) in Opomyzoidea are in bold type.

I.S. Winkler et al. / Molecular Phylogenetics and Evolution 54 (2010) 746–762

753

Fig. 4. Maximum likelihood phylogeny from the combined 28S and CAD data set (ln L = 18,239.3). Bootstrap values (500 reps.) are noted if greater than 50%. Taxa included by McAlpine (1989) in Opomyzoidea are in bold type.

which (not shown) was largely compatible with the 28S ML tree (Fig. 6) and Bayesian consensus tree (Fig. 5). A consistent result, found in all analyses of the 28S and combined data sets, was the placement of Cyamops, Stenomicra, and Planinasus in a clade with Aulacigaster, as suggested by Hennig (1958, 1969, 1971). This clade received strong support (100% PP) in the combined BI analysis, though support was weak in the MP and ML bootstrap analyses. Upon further inspection of the individual trees for each bootstrap replicate (not shown), however, one or more of the three genera mentioned above

was found to be sister group to Aulacigaster in most trees for the 28S ML analysis (90%) and combined ML analysis (80%). Within this clade (Aulacigastridae s.lat.), Cyamops and Aulacigaster grouped together with low support, and Planinasus and Stenomicra were recovered as sister taxa with moderate to strong support. Xenasteia Hardy and Australimyza Harrison grouped together consistently, with strong support values comparable to those uniting members of the same family. A sister-group relationship of Fergusoninidae and Asteiidae was observed in all 28S and combined analyses. This relationship was moderately

754

I.S. Winkler et al. / Molecular Phylogenetics and Evolution 54 (2010) 746–762

Fig. 5. Majority rule consensus tree resulting from Bayesian analysis (107 generations) of the combined data set, with posterior probabilities labeled above branches. Taxa included by McAlpine (1989) in Opomyzoidea are in bold type.

supported (0.87 PP) in the BI analysis, but poorly supported (<50%) in remaining analyses. A number of previously hypothesized sister-group relationships in ‘‘Opomyzoidea” (McAlpine, 1989) were never recovered by our data, though we cannot statistically disprove monophyly of these putative clades. For example, Opomyzidae and Anthomyzidae were distant from each other on every observed topology. Three small families, Acartophthalmidae, Neminidae, and Fergusoninidae, did

not show close relationship with the families in which they were originally included (Clusiidae, Aulacigastridae, and Agromyzidae, respectively), despite the retention of these hypothesized relationships by later authors after they were given family status. Finally, the position of Periscelis was variable, but it was never placed near Aulacigaster nor did it ever group with Cyamops, Stenomicra, or Planinasus. It is notable also that the Carnoidea of McAlpine and later authors (Brake, 2000; Buck, 2006) was not supported, as Carnidae,

I.S. Winkler et al. / Molecular Phylogenetics and Evolution 54 (2010) 746–762

755

Fig. 6. Maximum likelihood phylogeny from analysis of the 28S data set (ln L = 12,508.7). Taxa included by McAlpine (1989) in Opomyzoidea are in bold type. Bootstrap values (500 reps.) are noted if greater than 50%, for both ML (above branches) and MP (below branches) analyses. Nodes also recovered in the 28S parsimony analysis, but supported by less than 50% bootstrap values are marked by an asterisk .

Australimyzidae and Acartophthalmidae never appeared near Milichiidae + Chloropidae. The topology tests in CONSEL (Table 5) resulted in rejection of monophyly for several hypothesized clades. At the conservative level of a = 0.05, monophyly of only the Opomyzoidea exclusive of Marginidae (including or excluding Xenasteia) and Asteioinea (including Xenasteia) could be rejected using the Approximately Unbiased test. However, with a less conservative criterion

(a = 0.10), two additional groups were rejected: Opomyzoidea (excluding Xenasteia and Clusioinea), and Asteioinea (excluding Xenasteia). The SH test is known to be overly conservative, especially when multiple topologies are compared (Shimodaira and Hasegawa, 1999), and rejected only Opomyzoidea (including Xenasteia), Asteioinea (including Xenasteia), and (with a = 0.10) Opomyzoidea (excluding Xenasteia) and Opomyzoidea (excluding Xenasteia and Clusioinea).

756

I.S. Winkler et al. / Molecular Phylogenetics and Evolution 54 (2010) 746–762

4. Discussion 4.1. Phylogenetic conclusions Relationships among ‘‘acalyptrate” flies are notably difficult because of their large diversity and the high level of homoplasy observed in most studied character systems. DNA sequence data have proven to be no exception to this latter difficulty. However, our results permit a few definitive conclusions. Most importantly, Opomyzoidea sensu McAlpine and variants thereof proposed by other authors are decidedly non-monophyletic, even if Xenasteia is removed. In our recovered topologies, Opomyzoidea is not simply paraphyletic or diphyletic, but lacks cohesion in any sense. This failure of monophyly cannot be attributed simply to lack of phylogenetic resolution because it is statistically supported by the results of the topology tests, and recovery of monophyletic families through analysis of our data suggests the presence of some useful phylogenetic signal. This result should not be wholly surprising, given the previously ambiguous morphological support for this group and the earlier status of many component families as ‘‘unplaced”. In the case of McAlpine’s (1989) opomyzoid suprafamilies, our results are not as decisive. Non-monophyly for these groups was statistically significant according to the topology tests only in the case of Asteioinea; monophyly for the other three suprafamilies (Opomyzoinea, Agromyzoinea, Clusioinea) cannot be ruled out. Furthermore, alternative arrangements for families in these suprafamilies were weakly supported and inconsistent among analyses. Morphological evidence has been ambivalent at best in support for these groups. For example, Acartophthalmidae (Clusioinea) is now generally included in the superfamily Carnoidea (Brake, 2000; Buck, 2006), and both Hennig (1971, 1973) and Griffiths (1972) considered Fergusoninidae (Agromyzoinea) as too highly derived to place anywhere. Opomyzoinea represents a special enigma. The close relationship of Opomyzidae and Anthomyzidae has been accepted by most recent authors (Hennig, 1971; McAlpine, 1989; Rohácˇek, 1998a, 2006; but see Colless and McAlpine, 1991) and is supported by similar larval biologies and a similarly modified structure of the male aedeagal complex, among other characters. Our failure to recover Opomyzoinea should thus be considered suspect, especially in light of the observed 28S base composition bias in Anthomyzidae, which may have resulted in spurious placement. However, we note that a close relationship between Anthomyzidae and other families (notably members of ‘‘Heteromyzidae” sensu McAlpine, 2007) has been suggested before based on morphology (Speight, 1969; Griffiths, 1972). Additionally, in a study of the female internal genital tract of the Opomyzidae, Kotrba and Baptista (2002) failed to find any putative synapomorphies between Opomyzidae and Anthomyzidae. Instead, the authors advanced a few putative synapomorphies that may unite each of these two families with different groups within the Opomyzoidea. In light of these considerations, our results cast further doubt upon the validity of McAlpine’s (1989) opomyzoid suprafamilies, despite lack of statistical support. The genera Stenomicra, Cyamops, and Planinasus evidently are not closely related to Periscelis, but instead show close affinity with Aulacigaster, as proposed by Hennig (1958, 1969, 1971). We therefore follow Hennig (1971) and McAlpine (1989) in considering these genera (as well as the related Procyamops and Stenocyamops) as belonging to Aulacigastridae. Although support for this expanded concept of Aulacigastridae was poor in the ML and MP analyses, its strong (100% PP) support in the BI analysis and the consistent grouping of these genera in the ML analyses are sufficient in our view to uphold Hennig’s classification. Inclusion of these genera in Periscelididae was based primarily on antennal

characters (McAlpine, 1978, 1983; see below) which, in view of our results, must be interpreted as convergence. Additionally, our data do not support a close relationship of Neminidae with Aulacigaster, as suggested by McAlpine (1983). In conclusion, based on our results we favor an expanded concept of Aulacigastridae that includes Stenomicra, Cyamops, and Planinasus, in addition to Stenocyamops and Curiosimusca (not sampled in this analysis), and excludes Neminidae. Morphological support for Aulacigastridae, as defined here, might depend on the exploration of alternative character systems. The sister-group relationship between Aulacigaster and Curiosimusca is supported by a handful of putative synapomorphies (Rung et al., 2005). In contrast, all synapomorphies advanced by Hennig (1971) to support an expanded Aulacigastridae are highly homoplasious and none of them is present in all aulacigastrids as the family is here defined. The most important of these characters (McAlpine, 1989), ‘‘ocellar setae weak or absent,” has intermediate states amongst undescribed species of Aulacigaster, and the ocellar seta is fully developed in Curiosimusca. Within an expanded Aulacigastridae, we did not find the genera Cyamops, Stenomicra, and Planinasus to form a monophyletic group. Instead, Cyamops appeared more closely related to Aulacigaster than to Stenomicra and Planinasus, though with low support. Although Cyamops, Stenomicra, and Planinasus are now generally acknowledged to be closely related, the presumption that they form a monophyletic group is not universal. For instance, Stenomicra is morphologically quite reduced, and its relationship with Cyamops and Planinasus was not initially apparent to Hennig (1958, 1959) or Griffiths (1972). Hennig (1971) suggested, with some reservations, placement in Aulacigastridae as the sister group to Cyamops, Planinasus, and Aulacigaster. On the other hand, Baptista and Mathis (1994), while retaining the concept of Stenomicrinae within Periscelididae, thought it likely that Planinasus was in fact more closely related to Periscelidinae, some genera of which show a strong resemblance to Planinasus. From the above review, we conclude that relationships recovered by the current study within Aulacigastridae, while not strongly supported, are plausible considering the morphological evidence. The enigmatic genus Periscelis does not group with other Asteioinea in any analysis, and its position varies widely among the separate analyses. The hypothesis of a close relationship between Periscelis, Stenomicrinae, and Neurochaetidae is based primarily on a distinctive antennal form in which the pedicel is cap-like with a dorsal seam; the first flagellomere (third antennal segment or basal flagellomere) is geniculate, or reflexed downward; and the arista has alternate, long rays (McAlpine, 1978). These taxa also have a unique type of articulation between the pedicel and first flagellomere in which a basal process of the first flagellomere is inserted into a cavity in the pedicel (McAlpine, 1978; Rung, Mathis, and Mitter, unpublished); in most other Schizophora, a membranous extension of the pedicel is instead inserted into a cavity in the flagellum. The placement of Aulacigastridae in this group by McAlpine (Asteioinea group 1, see Fig. 1) implies a reversal of these antennal characters in this family. Because of this, Grimaldi and Mathis (1993) suggest instead a sister-group relationship between Periscelididae (including Stenomicrinae) and Neurochaetidae. Despite these striking similarities in antennal structure, it is not clear if the antennae of Periscelis are substantially different from members of the Diopsoidea (including part of ‘‘Nothyboidea” of Griffiths, 1972) and some other acalyptrates in these structural details. For example, McAlpine (1997) considers the diopsoid groundplan to include a deeply concave pedicel with a dorsal seam, deflexion of the first flagellomere, and articulation of the first flagellomere by means of a dorsobasal process inserted into a cavity on the pedicel. Similarly, the antennal similarity between Periscelididae sensu stricto and Cyamops, Stenomicra, and Planinasus may be due to convergence.

I.S. Winkler et al. / Molecular Phylogenetics and Evolution 54 (2010) 746–762

One strongly supported result from our data is that Xenasteiidae is closely related to Australimyzidae and does not belong with Asteiidae or any other ‘‘opomyzoid” family. These small, enigmatic families share a common habitat (seashore), and proposed characters uniting the Xenasteiidae and Asteiidae (Hardy, 1980), most notably in wing venation, can probably be attributed to morphological reduction associated with reduction in size. This result also confirms the distinctness of Australimyza from Carnidae, as first proposed by Griffiths (1972). Although later authors (McAlpine, 1989; Buck, 2006) supported its inclusion as a separate family in Carnoidea, Griffiths found it to be unrelated to any other muscoid lineage and accordingly designated a separate prefamily for Australimyza. Furthermore, the concept of Carnoidea fares little better than Opomyzoidea in our analyses and morphological support for this group is similarly weak (Buck, 2006). More systematic work including more characters and additional taxa is desirable. 4.2. Causes and implications of phylogenetic uncertainty The differences between the topologies presented here and previous morphological hypotheses may be explained partly by a high degree of phylogenetic ‘‘noise” (i.e. large fraction of highly homoplastic characters) in both molecular and morphological data sets. Because both molecular and previous morphological hypotheses are subject to this random error and generally poorly supported, it is difficult to evaluate the accuracy of many relationships presented here. In our exploratory analyses (not shown), we tried various means of maximizing the signal contained in our data, including attempting to extract signal from the excluded hypervariable regions (Lutzoni et al., 2000; Wheeler et al., 2006) and implicitly downweighting homoplastic characters (Goloboff, 1997). None of these methods produced demonstrably better results than the approach presented here of excluding rapidly evolving regions and positions that are more likely to be saturated and biased. In our experience during this study, using Gblocks (Castresana, 2002) to delimit the hypervariable regions of 28S seems to facilitate a good balance between including too many ‘‘noisy” sites and excluding too many informative sites, depending on the parameters used. However, even with hypervariable regions included, a 28S ML analysis (not shown) resulted in a substantially similar tree, with all but one (Psilidae) of the included families recovered and including the clade Stenomicrinae + Aulacigastridae. Despite this encouraging observation, inclusion of these regions may exacerbate base composition bias, and they are more likely to be saturated, and thus misleading, at deeper levels in the phylogeny. Systematic biases may also contribute to lack of phylogenetic resolution among families included here. For example, these may take the form of base composition bias or long branch artifacts in the molecular data or repeated parallel reduction in morphological characters. Systematic biases are more problematic than random error because they cannot be resolved by the inclusion of more characters. Furthermore, detecting systematic error in phylogeny construction is generally difficult (Bergsten, 2005), and we do not know which, if any, results of our study may be attributable to systematic error. Even in cases where non-stationarity of base frequencies is observed, we are unsure whether similar base frequencies represent convergence or common inheritance. More realistic models of sequence evolution are one possible solution to this problem. For example, Baurain et al. (2007) and Lartillot et al. (2007) found that using a model (CAT model; Lartillot and Philippe, 2004) that accounts for site-specific variation in base frequencies corrected systematic biases in inferring animal phylogeny. Our exploratory analyses implementing the CAT model in PhyloBayes v. 2.3 (Lartillot et al., 2004) for our 28S data (results not shown) found largely the same result as the combined analysis

757

in Mr. Bayes (Fig. 5). Models that account for taxon-specific base frequency variation also exist (Blanquart and Lartillot, 2006; Gowri-Shankar and Rattray, 2007), but their application has not yet been well-developed. In the future, such improved models will probably be important in more confidently resolving relationships in the Schizophora and other phylogenetically difficult groups. Cladistic (and non-cladistic) analysis of morphological characters faces similar problems, and the highly homoplastic nature of many characters used in acalyptrate higher classification has been acknowledged by previous workers (Hennig, 1971; McAlpine, 1989) and demonstrated in work by the second author (Rung, Mathis, and Mitter, unpublished). Debates in the literature thus largely center on the a priori polarity of characters and on the subjective weight that should be given to various characters or character systems (Yeates and Wiegmann, 1999). Taxa of the ‘‘Opomyzoidea” may be especially prone to morphological bias due to frequent reduction in size, with accompanied morphological changes. This is especially true with characters of the wing venation, but may also apply to others such as thoracic structure, chaetotaxy, and even genitalic structures. Other characters of the male and female postabdomen may vary rapidly, and a correlation with mating system can be expected. In general, it is agreed that novel character systems, including molecular sequence data, will be important for substantial future progress in schizophoran phylogenetics (Yeates et al., 2007). Regardless of the quality of data available, deeper relationships of Schizophora may prove difficult because this represents an ancient, rapid radiation comparable to those of similar age in other insect orders (e.g. Ditrysian moths, parasitic Hymenoptera; Whitfield and Kjer, 2008). Although the hypothesis of rapid radiation may be invoked more often than warranted for intransigent data sets, there are reasons to believe that the lack of resolution we observed reflects, in part, a true rapid radiation of schizophoran flies. To begin with, the markers used by us have been successful in resolving dipteran phylogeny at several different levels (Moulton and Wiegmann, 2007; Bertone et al., 2008; Scheffer et al., 2007; Winterton et al., 2006; Petersen et al. 2007), and observed divergence levels do not suggest that saturation is an important problem in our data (Fig. 2; except for third positions of CAD). The conjecture of a rapid Tertiary radiation of Schizophora is also supported by the fossil record (Blagoderov et al., 2002; Grimaldi and Engel, 2005). Despite a few possible earlier records (McAlpine, 1970; Grimaldi et al., 1989), most extant family-level diversity is first encountered in the extensively studied Baltic amber deposits (Eocene, 50 mya; Hennig, 1965; Evenhuis, 1994; von Tschirnhaus and Hoffeins, 2009). This is true of many families treated here, including Clusiidae, Acartophthalmidae, Odiniidae, Anthomyzidae, Aulacigastridae, Neurochaetidae, and Carnidae. If this picture of schizophoran evolution is accurate, it may be difficult to resolve the phylogeny of Schizophora even with a large number of characters. However, this degree of pessimism is not yet warranted, as lack of resolution due to rapid radiation, systematic biases, or use of markers with inappropriate evolutionary rates may be overcome with addition of more taxa or genes (e.g. Ortiz-Rivas et al., 2004; Baurain et al., 2007; Jian et al., 2008). Molecular phylogenetics of the ‘‘Acalyptratae” is still in its infancy and is now primarily limited to work within individual genera or families, for example, Drosophilidae (Remsen and O’Grady, 2002), Coelopidae (Meier and Wiegmann, 2002), Agromyzidae (Scheffer et al., 2007), Diopsidae (Meier and Baker, 2002), Sepsidae (Su et al., 2008), and Tephritidae (Han and McPheron, 1997). Although calyptrate flies have been slightly better studied (Bernasconi et al., 2000; Nirmala et al., 2001; Petersen et al., 2007; Kutty et al., 2008), to our knowledge only one other published molecular phylogenetic study has investigated relationships between acalyptrate families: that of Han and Ro

758

I.S. Winkler et al. / Molecular Phylogenetics and Evolution 54 (2010) 746–762

(2005) on Tephritoidea. Molecular phylogenetic studies of other groups face greater obstacles, including less robust initial morphological hypotheses of internal relationships, and more importantly, weaker evidence for monophyly of clades to begin with. As a result, it is difficult to obtain a reliable initial frame of reference from which to design taxon sampling and identify egregious examples of systematic biases. Our study highlights these problems, which we tried to alleviate by using extensive sampling of possible outgroups, including one non-schizophoran, and by evaluating alternative analyses based on recovery of recognized families. Although some groups appear to be based on firm morphlogical grounds (e.g. Ephydroidea, Sciomyzoidea), a number of schizophoran superfamilies proposed by Hennig (1971) and others were originally based on weak evidence and intended as tentative working hypotheses. Some of these hypotheses have been given more weight than originally intended as they have been repeated in the literature. Given the rampant polyphyly of the Opomyzoidea observed in our study, the existence of a paraphyletic residue of ‘‘orphan” lineages within Schizophora is possible, and dipterists should be wary of accepting the monophyly of other proposed acalyptrate superfamilies without a careful review of the evidence. Consequently, future studies of schizophoran phylogeny should include as broad a taxon sample as possible, at least until monophyletic groups are identified with a greater degree of confidence. Increased character sampling, including more protein-coding genes, is also desirable, though the poor performance of CAD here suggests this may not be a straightforward solution.

Although we were not able to definitively resolve any clades above the family-level (except for Xenasteiidae + Australimyzidae), our results are valuable in corroborating some previous hypotheses and contradicting others. Most notably, Hennig’s (1971) placement of Stenomicra, Cyamops, and Planinasus within Aulacigastridae is confirmed by molecular data. Additionally, some novel hypotheses emerged from analysis of our data which deserve further investigation. Accordingly, we view this study as an important initial step in using molecular data to evaluate the plethora of morphological hypotheses available for higher relationships of the Schizophora. As schizophoran flies represent a major clade including important model organisms and agricultural pests, further work to resolve the phylogenetic history of Schizophora is warranted and will yield rich fruit for comparative evolutionary biology. Acknowledgments We thank Charlie Mitter and the Department of Entomology at the University of Maryland for encouraging and assisting us to begin this project. Matthew Lewis (USDA) was very helpful in providing assistance and instruction in the laboratory. Brian Wiegmann supported the final stages of the study under the NSF-funded FLYTREE project (EF-03-34948). Special thanks are extended to those who donated specimens: Amnon Freidberg, Wayne Mathis, Peter Kerr, Brian Wiegmann, Marion Kotrba, Gary Taylor, and Riley Nelson. Owen Lonsdale, Bob Kula, Wayne Mathis, and two anonymous reviewers gave helpful comments on the manuscript.

Appendix A Exemplar specimens used in the phylogenetic analysis. Family

Species

Agromyzidae

Agromyza ambrosivora Spencer Liriomyza huidobrensis (Blanchard)

Collection data

Genbank No. (28S/CAD)

USA, Maryland, Prince George’s Co., SE of Clinton, IX.2000, S.J. Scheffer South Africa, Western Cape Province, Lambert’s Bay, XI.1999, D. Visser, ex. potato Cerodontha dorsalis (Loew) USA, Colorado: Gunnison Co., Irwin, VI.1997, S.J. Scheffer

EF104822

EF104735

EF104876

EF104788

EF104859

EF104771

Acartophthalmidae

Acartophthalmus bicolor Oldenberg

UK, Gwynedd Co., Afon Eden, Coed y Brenin Park, 19.VI.2004, I. Winkler

GU299253

GU299254

Anthomyzidae

Anthomyza gracilis Fallén

Germany, Berlin, Charlottenburg, GU299220 1.VII.2002, M. Kotrba & A. Rung USA, Maryland, Montgomery Co., GU299251 Beltsville, 7.VIII.2002, I. Winkler & A Rung

—

Switzerland, Genève, Jussy Prés-deGU299218 Villette, 2.VIII.2002, B. Merz USA, Washington, Pend Oreille Co., NE of GU299232 Sullivan Lake, 7.VI.2003, I. Winkler

GU299255

Ecuador, Puerto Orellana, Rio Tiputini, GU299237 12–26.VIII.1999, W.N. Mathis, A. Rung, M. Kotrba Germany, Berlin, Tiergarten, 28.VI.1999, GU299238 M. Kotrba USA, Maryland, Prince George’s Co., GU299224 Greenbelt, 27.VII.2002, A. Rung

GU299256

Mumetopia sp. Asteiidae

Asteia amoena Meigen Leiomyza curvinervis Zetterstedt

Aulacigastridae s.str.

Aulacigaster ecuadoriensis (Hennig) Aulacigaster leucopeza (Meigen) Aulacigaster neoleucopeza Mathis and Freidberg

GU299270

GU299268

GU299257 GU299258

759

I.S. Winkler et al. / Molecular Phylogenetics and Evolution 54 (2010) 746–762 Appendix A (continued)

Family

Species

Collection data

Clusiidae

Clusia czernyi Johnson

USA, Maryland, Montgomery Co., GU299246 Plummers Island, 7–15.V.2001, A. Rung USA, Maryland, Prince George’s Co., GU299240 College Park, 25.VIII.2002, I. Winkler & A. Rung

GU299263

Fergusonina purcelli Taylor (FC24) Fergusonina sp. (FC73)

Australia, Queensland, Daintree Ferry, 76-MC-F3, ex. Melaleuca cajuputi Australia, WINC2004, ex. Eucalyptus intertexta

GU299241

GU299265

GU299242

GU299266

Marginidae

Margo clausa McAlpine

Madagascar, Andasibe, Analamazaotra Forest, 31.X-4.XI.2007, A. Freidberg

GU299230

—

Neminidae

Nemula longarista Freidberg Madagascar, Andasibe, Analamazaotra Forest, 31.X-4.XI.2007, A. Freidberg

GU299223

—

Neurochaetidae

Neurochaeta inversa McAlpine

GU299217

GU299271

Odiniidae

Odinia boletina (Zetterstedt) UK, Hertfordshire, Aldbury Common, GU299247 14.VI.1997 Neoalticomerus seamansi UTAH, Utah Co., American Fork Cyn., 1– GU299229 Shewell 15.VII.2005, C.R. Nelson et al.

GU299272

Geomyza tripunctata Fallén Switzerland, Genève, Bernex ChanteMerle 16.VIII.2002, B. Merz Opomyza sp. Germany, Bradenburg, Kloster Chorin, VII.2000, M. Kotrba & A. Rung

GU299222

GU299267

GU299221

GU299275

Periscelididae s.str.

Periscelis flinti (Malloch)

USA, Maryland, Prince George’s Co., Greenbelt, 29.VII.2002, A. Rung

GU299228

GU299274

Stenomicrinae (included in Periscelididae or Aulacigastridae)

Cyamops sp.

Australia, Queensland, Josephine Falls, 23.IX.2002, W.N. & D. Mathis

GU299236

—

Cyamops nebulosus Melander Stenomicra angustata Coquillett

USA, Maryland, Montgomery Co., GU299226 Pautuxent Refuge, S.J. Scheffer et al. USA, Maryland, Montgomery Co., GU299244 Beltsville, 7.VIII.2002, I. Winkler & A. Rung Brazil, Rio de Janeiro, Floresta da Tijuca, GU299243 VIII.1998, A. Rung, R. Baptista

—

Clusiodes johnsoni Malloch

Fergusoninidae

Opomyzidae

Planinasus sp.

Australia, Upper Cedar Cr., nr. Sanford, 5.X.2002, M. Kotrba

Genbank No. (28S/CAD)

GU299264

—

GU299277

—

Teratomyzidae

Teratomyza (Vitila) sp.

Australia, Queensland, Millaa Millaa, 24.IX.2002, W.N. & D. Mathis

GU299227

GU299279

Xenasteiidae

Xenasteia shalam (Freidberg)

Israel, Enot, 3.IX.2007, A. Freidberg

GU299225

—

Australimyzidae (Carnoidea)

Australimyza salicorniae Harrison

New Zealand, N. Isl., WN, Waikanae Beach, 6.I.2004, W.N. Mathis

GU299219

GU299259

Carnidae (Carnoidea)

Meoneura sp.

USA, California, Inyo Co., South Fork Bishop Cr., 4-Jeffrey campground, 3.III.2003, I. Winkler

GU299239

GU299260

Chloropidae (Carnoidea)

Chaetochlorops inquilinus (Coquillett) Thaumatomyia sp.

USA, Maryland, Montgomery Co., Plummers Island, 7-15.V.2001, P. Kerr USA, Maryland, Prince George’s Co., Greenbelt, 2.VII.2002, I. Winkler & A. Rung

GU299234

GU299261

GU299235

GU299280

Aphaniosoma sp.

USA, Colorado, Adams Co., Barr Lake St. Park, 3.III.2004, I. Winkler

GU299231

—

Outgroups

Chyromyidae (Sphaeroceroidea)

(continued on next page)

760

I.S. Winkler et al. / Molecular Phylogenetics and Evolution 54 (2010) 746–762

Appendix A (continued)

Family

Species

Collection data

Genbank No. (28S/CAD)

Gymnochiromyia sp.

USA, Maryland, Prince George’s Co., Adelphi, 19.IV.2004, I. Winkler

GU299250

—

Milichiidae (Carnoidea)

Paramyia nitens (Loew)

USA, Maryland, Prince George’s Co., Adelphi, 3.VIII.2002, I. Winkler

GU299233

GU299273

Muscidae (Calyptratae)

Musca domestica Linnaeus

(Genbank)

AF503004 AF503025 AY123358

AY280689

Psilidae (Diopsoidea)

Chyliza sp.

USA, Maryland, Prince George’s Co., Greenbelt, 29.VIII.2002, A. Rung USA, Maryland, Montgomery Co., Plummers Island, 7–15.V.2001, P. Kerr

GU299249

GU299262

GU299245

GU299269

Loxocera sp. Sphaeroceridae (Sphaeroceroidea) Pterogramma sp.

USA, Maryland, Montgomery Co., Plummers Island, 4–18.VI.2001, P. Kerr

GU299248

GU299276

Syrphidae (Aschiza)

Rhingia nasica Say

(Genbank)

AF502977 AF502998 AF503019

AY280697

Tephritidae (Tephritoidea)

Procecidochares utilis Stone Australia, Queensland, 2001, A.L. Norrbom, ex. Ageratina adenophora

GU299252

GU299278

References Baptista, A.R., Mathis, W.N., 1994. A revision of New World Cyamops Melander (Diptera, Periscelididae). Smithson. Contrib. Zool. 563, 1–28. Barraclough, D.A., McAlpine, D.K., 2006. Natalimyzidae, a new African family of acalyptrate flies (Diptera: Schizophora: Sciomyzoidea). Afr. Invertebr. 47, 117– 134. Baurain, D., Brinkmann, H., Philippe, H., 2007. Lack of resolution in the animal phylogeny: closely spaced cladogeneses or undetected systematic errors? Mol. Biol. Evol. 24, 6–9. Bergsten, J., 2005. A review of long branch attraction. Cladistics 21, 163–193. Bernasconi, M.V., Valsangiacomo, C., Piffaretti, J.C., Ward, P.I., 2000. Phylogenetic relationships among Muscoidea (Diptera: Calyptratae) based on mitochondrial DNA sequences. Insect Mol. Biol. 9, 67–74. Bertone, M.A., Courtney, G.W., Wiegmann, B.M., 2008. Phylogenetics and temporal diversification of the earliest true flies (Insecta: Diptera) based on multiple nuclear genes. Syst. Entomol. 33, 668–687. Blagoderov, V.A., Lukashevich, E.D., Mostovski, M.B., 2002. Order Diptera Linné, 1758. The true flies (= Muscida Laicharting, 1781). In: Rasnitsyn, A.P., Quicke, D.L.J. (Eds.), History of Insects. Kluwer Academic Publishers, Dordrecht, pp. 227– 240. Blanquart, S., Lartillot, N., 2006. A Bayesian compound stochastic process for modeling non-stationary and nonhomogeneous sequence evolution. Mol. Biol. Evol. 23, 2058–2071. Brake, I., 2000. Phylogenetic systematics of the Milichiidae (Diptera, Schizophora). Entomol. Scand. 57, 1–58. Brake, I., Mathis, W.N., 2007. Revision of the genus Australimyza Harrison (Diptera: Australimyzidae). Syst. Entomol. 32, 252–275. Britton, D., 2009. Rare – or just hidden? Australian Museum Online. Available from: (accessed Jan. 2010). Brunel, É., 1998. Family Opomyzidae. In: Soós, A., Papp, L. (Eds.), Catalogue of Palearctic Diptera. Elsevier Science Publishers and Akadémiai Kiadó, Budapest, pp. 259–266. Buck, M., 2006. A new family and genus of acalypterate flies from the Neotropical region, with a phylogenetic analysis of Carnoidea family relationships (Diptera, Schizophora). Syst. Entomol. 31, 377–404. Castresana, J., 2000. Selection of conserved blocks from multiple alignments for their use in phylogenetic analysis. Mol. Biol. Evol. 17, 540–552. Castresana, J., 2002. Gblocks v.0.91b. Available from: . Colless, D.H., McAlpine, D.K., 1991. Diptera (flies), second ed.. In: Naumann, I.D. (Ed.), The Insects of Australia. A Textbook for Students and Research Workers, vol. II Melbourne University Press, Carlton, Victoria, Australia, pp. 717–786. Currie, G.A., 1937. Galls on Eucalyptus trees: a new type of association between flies and nematodes. Proc. Linn. Soc. N. S. W. 62, 147–174. Edgar, R.C., 2004. MUSCLE: multiple sequence alignment with high accuracy and high throughput. Nucleic Acids Res. 32, 1792–1797. Available from: .

Evenhuis, N.L., 1994. Catalogue of the Fossil Flies of the World (Insecta: Diptera). Available from: (accessed Oct. 2008). Evenhuis, N.L., Pape, T., Pont, A.C., Thompson, F.C. (Eds.), 2008. Biosystematic Database of World Diptera, Working Version. Available from: (accessed Dec. 2008). Fish, D., 1983. Phytotelmata: Flora and Fauna. In: Frank, J.H., Lounibos, L.P. (Eds.), Phytotelmata: Terrestrial Plants as Hosts for Aquatic Insect Communities. Plexus, Medford, New Jersey, pp. 1–27. Freidberg, A., 1994. Nemula, a new genus of Neminidae (Diptera) from Madagascar. Proc. Entomol. Soc. Wash. 96, 471–482. Freidberg, A., 2002. Recent advances in the study of Asteiidae. In: Abstracts Volume, Fifth International Congress of Dipterology, Brisbane, Australia. University of Queensland, Brisbane, p. 73. Gaimari, S.D., Mathis, W.N., in press. World catalog and conspectus of the family Odiniidae (Diptera: Schizophora). Myia. Goloboff, P.A., 1997. Self-weighted optimization: tree searches and character state reconstructions under implied transformation costs. Cladistics 13, 225– 245. Gowri-Shankar, V., Rattray, M., 2007. A reversible jump method for Bayesian phylogenetic inference with a nonhomogeneous substitution model. Mol. Biol. Evol. 24, 1286–1299. Griffiths, G.C.D., 1972. The phylogenetic classification of Diptera Cyclorrhapha with special reference to the structure of the male postabdomen. Ser. Entomol. 8, 1– 340. The Hague. Grimaldi, D.A., Engel, M.S., 2005. Evolution of the Insects. Cambridge University Press, Cambridge, UK. Grimaldi, D.A., Mathis, W.N., 1993. Fossil Periscelididae (Diptera). Proc. Entomol. Soc. Wash. 95, 383–403. Grimaldi, D.A., Beck, C.W., Boon, J.J., 1989. Occurrence, chemical characteristics, and paleontology of the fossil resins from New Jersey. Am. Mus. Novit. 2948, 1–28. Han, H.Y., McPheron, B.A., 1997. Molecular phylogenetic study of Tephritidae (Insecta: Diptera: Tephritidae) using partial sequences of the mitochondrial 16S ribosomal DNA. Mol. Phylogenet. Evol. 7, 17–32. Han, H.Y., Ro, K.E., 2005. Molecular phylogeny of the superfamily Tephritoidea (Insecta: Diptera): new evidence from the mitochondrial 12S, 16S, and COII genes. Mol. Phylogenet. Evol. 34, 416–430. Hardy, D.E., 1980. Xenasteiidae, a new family of Schizophora from the Pacific and Indian oceans. Proc. Hawaii Entomol. Soc. 23, 205–225. Hennig, W., 1958. Die Familien der Diptera Schizophora und ihre phylogenetischen Verwandtschaftsbeziehungen. Beitr. Entomol. 8, 505–688. Hennig, W., 1965. Die Acalyptraten des Baltischer Bernsteins. Stuttgarter Beitr. Naturkd. 145, 1–215. Hennig, W., 1969. Neue Gattungen und Arten der Acalypteratae. Can. Entomol. 101, 589–633. Hennig, W., 1971. Neue Untersuchungen über die Familien der Diptera Schizophora (Diptera: Cyclorrhapha). Stuttgarter Beitr. Naturkd. 226, 1–76. Hennig, W., 1973. Diptera. In: Kukenthal, W. (Ed.), Handbuch der Zoologie, IV: Arthropoda, Section 2, Part 2, vol. 31. De Gruyter, New York, pp. 1–337.

I.S. Winkler et al. / Molecular Phylogenetics and Evolution 54 (2010) 746–762 Hoffeins, C., Rung, A., 2005. Procyamops succini, a new genus and new species of Periscelididae (Diptera, Brachycera) from Baltic amber (Tertiary, Eocene). Stud. Dipterol. 12, 23–27. Jian, S., Soltis, P.S., Gitzendanner, M.A., Moore, M.J., Li, R., Hendry, T.A., Qiu, Y.-L., Dhingra, A., Bell, C.D., Soltis, D.E., 2008. Resolving an ancient rapid radiation in Saxifragales. Syst. Biol. 57, 38–57. Khoo, K.C., Sabrosky, C.W., 1989. Family Stenomicridae. In: Evenhuis, N.L. (Ed.), Catalog of the Diptera of the Australasian and Oceanian Regions. B.P. Bishop Museum Special Publication No. 86.. E.J. Brill and Bishop Museum Press, Honolulu. p. 551. Kotrba, M., Baptista, A.R., 2002. The internal reproductive tract of the Opomyzidae (Diptera, Schizophora). Stud. Dipterol. 9, 57–71. Kutty, S.N., Pape, T., Pont, A., Wiegmann, B.M., Meier, R., 2008. The Muscoidea (Diptera: Calyptratae) is paraphyletic: evidence from four mitochondrial and four nuclear genes. Mol. Phylogenet. Evol. 49, 639–652. Lartillot, N., Philippe, H., 2004. A Bayesian mixture model for across-site heterogeneities in the amino-acid replacement process. Mol. Biol. Evol. 21, 1095–1109. Lartillot, N., Blanquart, S., Lepage, T., 2004. PhyloBayes 2.3: a Bayesian software for phylogenetic reconstruction using mixture models. Available from: . Lartillot, N., Brinkmann, H., Philippe, H., 2007. Suppression of long-branch attraction artefacts in the animal phylogeny using a site-heterogeneous model. BMC Evol. Biol. 7 (Suppl. 1), S4. Lonsdale, O., Marshall, S.A., 2006. Redefinition of the Clusiinae and Clusiodinae, description of the new subfamily Sobarocephalinae, revision of the genus Chaetoclusia and a description of Procerosoma gen. n. (Diptera: Clusiidae). Eur. J. Entomol. 103, 163–182. Lonsdale, O., Marshall, S.A., 2007. Redefinition of the genera Clusiodes and Hendelia (Diptera: Clusiidae: Clusiodinae) with a review of Clusiodes. Stud. Dipterol. 14, 117–159. Lonsdale, O., Marshall, S.A., 2008. Synonymy within Clusia and description of the new genus Melanoclusia (Diptera: Clusiidae: Clusiinae). Ann. Entomol. Soc. Am. 101, 327–330. Lutzoni, F., Wagner, P., Reeb, V., Zoller, S., 2000. Integrating ambiguously aligned regions of DNA sequences in phylogenetic analyses without violating positional homology. Syst. Biol. 49, 628–651. Mathis, W.N., in press. World Catalog and Conspectus on the Family Periscelididae (Diptera: Schizophora). Myia. Mathis, W.N., Papp, L., 1992. A new genus of Periscelididae (Diptera) from the Neotropics. Proc. Biol. Soc. Wash. 105, 366–372. Mathis, W.N., Papp, L., 1998. Family Periscelididae. In: Papp, L., Darvas, B. (Eds.), Contributions to a Manual of Palaearctic Diptera, Volume 3: Higher Brachycera. Science Herald, Budapest, pp. 285–294. Mathis, W.N., Rung, A., 2004. Redescription of the genus Diopsosoma Malloch (Diptera, Periscelididae). Rev. Bras. Entomol. 48, 303–309. McAlpine, J.F., 1970. First record of calypterate flies in the Mesozoic era (Diptera: Caliphoridae). Can. Entomol. 102, 342–346. McAlpine, D.K., 1978. Description and biology of a new genus of flies related to Anthoclusia and representing a new family (Diptera, Schizophora, Neurochaetidae). Ann. Natal Mus. 23, 273–295. McAlpine, D.K., 1983. A new subfamily of Aulacigastridae (Diptera: Schizophora), with a discussion of aulacigastrid classification. Aust. J. Zool. 31, 55–78. McAlpine, D.K., 1988a. Studies in upside-down flies (Diptera: Neurochaetidae). Part I. Systematics and phylogeny. Proc. Linn. Soc. N. S. W 110, 31–58. McAlpine, D.K., 1988b. Studies in upside-down flies (Diptera: Neurochaetidae). Part II. Biology, adaptations and specific mating mechanisms. Proc. Linn. Soc. N. S. W 110, 59–82. McAlpine, J.F., 1989. Phylogeny and classification of the Muscomorpha. In: McAlpine, J.F., Wood, D.M. (Eds.), Manual of Nearctic Diptera, Volume 3: Research Branch, Agriculture Canada, Monograph 32, pp. 1397–1518. McAlpine, D.K., 1991. Marginidae, a new Afrotropical family of Diptera (Schizophora: Opomyzoidea). Ann. Natal Mus. 32, 167–177. McAlpine, D.K., 1993. Review of the upside-down flies (Diptera: Neurochaetidae) of Madagascar and Africa, and evolution of neurochaetid host plant associations. Rec. Aust. Mus. 45, 221–239. McAlpine, D.K., 1997. Gobryidae, a new family of acalyptrate flies (Diptera: Diopsoidea), and a discussion of relationships of the diopsoid families. Rec. Aust. Mus. 49, 167–194. McAlpine, D.K., 1998. Neminidae (Nobody flies). In: Oosterbroek, P. (Ed.), The Families of Diptera of the Malay Archipelago. Fauna Malesiana Handbook 1. Brill, Leiden, pp. 123–124. McAlpine, D.K., 2007. Review of the Borboroidini or wombat flies (Diptera: Heteromyzidae), with reconsideration of the status of families Heleomyzidae and Sphaeroceridae, and descriptions of femoral gland-baskets. Rec. Aust. Mus. 59, 143–219. McAlpine, D.K., de Keyzer, R.G., 1994. Generic classification of the fern flies (Diptera: Teratomyzidae) with a larval description. Syst. Entomol. 19, 305–326. Meier, R., 2005. Role of dipterology in phylogenetic systematics: the insight of Willi Hennig. In: Yeates, D.K., Wiegmann, B.M. (Eds.), The Evoutionary Biology of Flies. Columbia University Press, New York, pp. 45–62. Meier, R., Baker, R., 2002. A cladistic analysis of Diopsidae (Diptera) based on morphological and DNA sequence data. Insect Syst. Evol. 33, 325–336. Meier, R., Wiegmann, B.M., 2002. A phylogenetic analysis of Coelopidae (Diptera) based on morphological and DNA sequence data. Mol. Phylogenet. Evol. 25, 393–407.

761

Moulton, J.K., Wiegmann, B.M., 2004. Evolution and phylogenetic utility of CAD (rudimentary) among Mesozoic-aged eremoneuran Diptera (Insecta). Mol. Phylogenet. Evol. 31, 363–378. Moulton, J.K., Wiegmann, B.M., 2007. The phylogenetic relationships of flies in the superfamily Empidoidea (Insecta: Diptera). Mol. Phylogenet. Evol. 43, 701–713. Murphy, D.H., 1990. The natural history of insect herbivory on mangrove trees in and near Singapore. Raffles Bull. Zool. 38, 119–203. Nirmala, X., Hypsa, V., Zurovec, M., 2001. Molecular phylogeny of Calyptratae (Diptera: Brachycera): the evolution of 18S and 16S ribosomal rDNAs in higher dipterans and their use in phylogenetic inference. Insect Mol. Biol. 10, 475–485. Ortiz-Rivas, B., Moya, A., Martínez-Torres, D., 2004. Molecular systematics of aphids (Homoptera: Aphididae): new insights from the long-wavelength opsin gene. Mol. Phylogenet. Evol. 30, 24–37. Papp, L., 1980. New taxa of acalyptrate flies (Diptera: Tunisimyiidae fam. n., Risidae, Ephydridae: Nannodastiinae subfam. n.). Acta Zool. Acad. Sci. Hung. 26, 415–431. Papp, L., 1984. Family Aulacigastridae. In: Soós, A., Papp, L. (Eds.), Catalogue of Palearctic Diptera. Elsevier Science Publishers and Akadémiai Kiadó, Budapest, pp. 60–61. Papp, L., 1998a. Family Aulacigastridae. In: Papp, L., Darvas, B. (Eds.), Contributions to a Manual of Palaearctic Diptera, Volume 3: Higher Brachycera. Science Herald, Budapest, pp. 279–284. Papp, L., 1998b. Family Asteiidae. In: Papp, L., Darvas, B. (Eds.), Contributions to a Manual of Palaearctic Diptera, Volume 3: Higher Brachycera. Science Herald, Budapest, pp. 295–303. Papp, L., 1998c. Family Xenasteiidae. In: Papp, L., Darvas, B. (Eds.), Contributions to a Manual of Palaearctic Diptera, Volume 3: Higher Brachycera. Science Herald, Budapest, pp. 305–308. Papp, L., Ozerov, A.L., 1998. Family Acartophthalmidae. In: Papp, L., Darvas, B. (Eds.), Contributions to a Manual of Palaearctic Diptera, Volume 3: Higher Brachycera. Science Herald, Budapest, pp. 227–232. Papp, L., Merz, B., Földvári, M., 2006. Diptera of Thailand: a summary of the families and genera with references to the species representations. Acta Zool. Acad. Sci. Hung. 52, 97–268. Petersen, F.T., Meier, R., Kutty, S.N., Wiegmann, B.M., 2007. The phylogeny and evolution of host choice in the Hippoboscoidea (Diptera) as reconstructed using four molecular markers. Mol. Phylogenet. Evol. 45, 111–122. Remsen, J., O’Grady, P.M., 2002. Phylogeny of Drosophilidae (Diptera), with comments on combined analysis and character support. Mol. Phylogenet. Evol. 24, 248–263. Rohácˇek, J., 1998a. Taxonomic limits, phylogeny and higher classification of Anthomyzidae (Diptera), with special regard to fossil record. Eur. J. Entomol. 95, 141–177. Rohácˇek, J., 1998b. Family Anthomyzidae. In: Papp, L., Darvas, B. (Eds.), Contributions to a Manual of Palaearctic Diptera, Volume 3: Higher Brachycera. Science Herald, Budapest, pp. 267–278. ˇ as. Rohácˇek, J., 2006. A monograph of Palaearctic Anthomyzidae (Diptera) part 1. C Slez. Muz. Opava (A) 55 (suppl. 1), 1–328. Ronquist, F., Huelsenbeck, J.P., 2003. MrBayes 3: Bayesian phylogenetic inference under mixed models. Bioinformatics 19, 1572–1574. Rung, A., Mathis, W.N., Papp, L., 2005. Curiosimusca, gen. nov., and three new species in the family Aulacigastridae from the Oriental Region (Diptera: Opomyzoidea). Zootaxa 1009, 21–36. Scheffer, S.J., Giblin-Davis, R.M., Davies, K.A., Taylor, G.S., Purcell, M., Lewis, M.L., Goolsby, J., Center, T.D., 2004. Phylogenetic relationships, species limits, and host specificity of gall-forming Fergusonina flies (Diptera: Fergusoninidae) feeding on Melaleuca (Myrtaceae). Ann. Entomol. Soc. Am. 97, 1216–1221. Scheffer, S.J., Winkler, I.S., Wiegmann, B.M., 2007. Phylogenetic relationships within the leaf-mining flies (Diptera: Agromyzidae) inferred from sequence data from multiple genes. Mol. Phylogenet. Evol. 42, 756–775. Shatalkin, A.I., 1993. On the taxonomy of the flies of the family Strongylophthalmyiidae (Diptera). Zool. Zh. 72, 124–131. Shimodaira, H., 2002. An approximately unbiased test of phylogenetic tree selection. Syst. Biol. 51, 492–508. Shimodaira, H., Hasegawa, M., 1999. Multiple comparisons of log-likelihoods with applications to phylogenetic inference. Mol. Biol. Evol. 16, 1114–1116. Shimodaira, H., Hasegawa, M., 2001. CONSEL: for assessing the confidence of phylogenetic tree selection. Bioinformatics 17, 1246–1247. Available from: . Speight, M.C.D., 1969. The prothoracic morphology of Acalyptrates (Diptera) and its use in systematics. Trans. R. Entomol. Soc. Lond. 121, 325–421. Spencer, K.A., 1990. Host specialization in the world Agromyzidae (Diptera). Ser. Entomol. 45, 1–444. Kluwer Academic Publishers, Dordrecht. Su, F.-Y.K., Kutty, S., Meier, R., 2008. Morphology versus molecules: the phylogenetic relationships of Sepsidae (Diptera: Cyclorrhapha) based on morphology and DNA sequence data from ten genes. Cladistics 24, 902–916. Suzuki, Y., Glazko, G.V., Nei, M., 2002. Overcredibility of molecular phylogenies obtained by Bayesian phylogenetics. Proc. Natl. Acad. Sci. USA 99, 16138–16143. Swofford, D.L., 2001. PAUP* 4.0 beta 10. Phylogenetic Analysis Using Parsimony (*and Other Methods). Sinauer Associates, Massachussets. Tautz, D., Hancock, J.M., Webb, D.A., Tautz, C., Dover, G.A., 1988. Complete sequences of the rRNA genes of Drosophila melanogaster. Mol. Biol. Evol. 5, 366–376. Taylor, G.S., Davies, K.A., Giblin-Davis, R.M., 2004. Species-richness in gall-flies (Diptera: Fergusoninidae), nematodes (Nematoda: Sphaerulariidae) and associated parasitoids and inquilines on Myrtaceae. Rec. S. Aust. Mus. Monogr. Ser. 7, 249–256.

762

I.S. Winkler et al. / Molecular Phylogenetics and Evolution 54 (2010) 746–762

Taylor, G.S., Head, E.R., Davies, K.A., 2005. Gall flies (Diptera: Fergusoninidae) on Myrtaceae: a mutualistic association between flies and nematodes. In: Raman, A., Schaefer, C.W., Withers, T.M. (Eds.), Biology, Ecology, and Evolution of Gallinducing Arthropods. Science Publishers, Enfield, New Hampshire, pp. 643–671. Thompson, J.D., Higgins, D.G., Gibson, T.J., 1994. Clustal W: improving the sensitivity of progressive multiple sequence alignment through sequence weighting, position-specific gap penalties and weight matrix choice. Nucleic Acids Res. 22, 4673–4680. Vockeroth, J.R., 1987. Family Opomyzidae. In: McAlpine, J.F., Wood, D.M. (Eds), Manual of Nearctic Diptera, Volume 2: Research Branch. Agriculture Canada, Monograph 32, pp. 881–885. von Tschirnhaus, M., Hoffeins, C., 2009. Fossil flies in Baltic amber—insights in the diversity of Tertiary Acalyptratae (Diptera, Schizophora), with new morphological characters and a key based on 1000 collected inclusions. Denisia 26, 171–212. Wheeler, T.A., 2000. Systematics of Mallochianamyia Santos-Neto (= Gayomyia Malloch), an unplaced genus of Neotropical acalyptrate flies (Diptera: Schizophora). Insect Syst. Evol. 31, 209–224. Wheeler, W.C., Aagesen, L., Arango, C.P., Faivovich, J., Grant, T., D’Haese, C., Janies, D., Smith, W.L., Varón, A., Giribet, G., 2006. Dynamic Homology and Phylogenetic

Systematics: A Unified Approach Using POY. American Museum of Natural History, New York. pp 365. Whitfield, J.B., Kjer, K.M., 2008. Ancient rapid radiations of insects: challenges for phylogenetic analysis. Annu. Rev. Entomol. 53, 449–472. Wiegmann, B.M., Tsaur, S.C., Webb, D.W., Yeates, D.K., Cassel, B.K., 2000. Monophyly and relationships of the Tabanomorpha (Diptera: Brachycera) based on 28S ribosomal gene sequences. Ann. Entomol. Soc. Am. 93, 1031–1038. Winterton, S.L., Wiegmann, B.M., Schlinger, E.I., 2006. Phylogeny and Bayesian divergence time estimations of small-headed flies (Diptera: Acroceridae) using multiple molecular markers. Mol. Phylogenet. Evol. 43, 808–832. Yeates, D.K., Wiegmann, B.M., 1999. Congruence and controversy: Toward a higherlevel phylogeny of Diptera. Annu. Rev. Entomol. 44, 397–428. Yeates, D.K., Wiegmann, B.M., Courtney, G.W., Meier, R., Lambkin, C., Pape, T., 2007. Phylogeny and systematics of Diptera: two decades of progress and prospects. Zootaxa 1668, 565–590. Zwickl, D.J., 2006. Genetic algorithm approaches for the phylogneetic analysis of large biological sequence datasets under the maximum likelihood criterion. Ph.D. Dissertation, the University of Texas at Austin. Available from: .