Schwarzfeld, M. D

Systematic Entomology (2016), 41, 191–206 DOI: 10.1111/syen.12152

Molecular phylogeny of the diverse parasitoid wasp genus Ophion Fabricius (Hymenoptera: Ichneumonidae: Ophioninae)

MARLA D. SCHWARZFELD1†, GAVIN R. BROAD2 andFELIX A. H. SPERLING1

1Department of Biological Sciences, University of Alberta, Edmonton, Canada and 2Department of Life Sciences, The Natural History Museum, London, U.K.

Abstract. Ophion Fabricius is a diverse genus of nocturnal ichneumonid wasps (Insecta: Hymenoptera) that is particularly species-rich in temperate areas, yet has received little taxonomic attention in the Holarctic region, where most species occur. While there have been some attempts to divide Ophion into monophyletic species groups, the vast majority of species have been lumped into a single, paraphyletic group, the O. luteus species group, which is defined only by the lack of characters specific to the other groups. The challenging morphology of this large catch-all group has limited attempts to subdivide it, and no phylogenetic hypothesis has been proposed for the genus as a whole. In this study, we use DNA sequence data [28S ribosomal RNA (28S), cytochrome oxidase 1 (COI) and internal transcribed spacer 2 (ITS2)] to present the first molecular phylogeny of Ophion. We also describe the secondary structure of ITS2 for the first time in Ichneumonidae, and explore its implications for phylogeny estimation. We define 13 species groups, nine of which were previously considered part ofthe O. luteus species group s.l. The included species groups are the O. minutus, O. areolaris, O. scutellaris, O. flavidus, O. parvulus, O. slossonae, O. nigrovarius, O. pteridis, O. luteus s.s., and O. obscuratus species groups, along with three groups lacking described species (Species group 1, New Zealand group, Madagascar group). This study provides a framework for future studies of this diverse and morphologically challenging genus.

Introduction there are relatively few tropical species and the majority of these are restricted to cooler climates at high elevation (Gauld The ichneumonid subfamily Ophioninae consists primarily & Mitchell, 1981; Gauld, 1988). In contrast, Ophion is by far of nocturnal orange or yellowish parasitoid wasps that are the most abundant and diverse ophionine genus in virtually all frequently observed at lights. Ophionines are koinobiont habitats across the Holarctic region (Gauld, 1980, 1988). endoparasitoids of holometabolous insect larvae, almost always Despite being a predominately temperate genus, Ophion in the Lepidoptera. There are currently 32 recognized genera within Holarctic region has received little taxonomic attention, as most Ophioninae (Yu et al., 2012), almost all of which are most revisions of Ophioninae have been for tropical regions (Gauld, diverse in tropical regions (Gauld, 1985). However, the genus 1977, 1988; Gauld & Mitchell, 1978, 1981; Fernández-Triana, Ophion Fabricius (Fig. 1) is one of the few exceptions to this 2005; Kim et al., 2009; Rousse & Van Noort, 2014). In the pattern. While the genus has a nearly worldwide distribution, Nearctic region, 17 species are currently described (Yu et al., 2012; Schwarzfeld & Sperling, 2014), although it has been Correspondence: Marla D. Schwarzfeld, 960 Carling Ave., Ottawa, estimated based on morphology that there are approximately 50 Ontario K1A 0C6, Canada. E-mail: [email protected] Nearctic species (Gauld, 1985) and recent molecular analyses suggest there are many more (Schwarzfeld & Sperling, 2015). †Present address: Canadian National Collection of Insects, Arachnids In the Palearctic region, Ophion is better known, with 79 and Nematodes, 960 Carling Ave., Ottawa, Ontario K1A 0C6, Canada. described species (Yu et al., 2012); however, there have been

Gauld, Sclerophion Gauld and Rhopalophion Seyrig), although none of these are found within the Holarctic region (Gauld, 1985). All Nearctic and Palearctic species have been placed within four, or possibly five, of Gauld’s eight species groups (Gauld, 1985). The O. areolaris group was previously classified as the genus Platophion Hellén (Gauld, 1977; Brock, 1982) and is characterized by the loss of the occipital carina, among other characters. The O. similis species group contains three Palearctic and two undescribed Nearctic species that have a stout body shape, short antennae, small eyes and ocelli, and appear to be diurnal. The O. dentatus species group is restricted to the eastern Palearctic region and is characterized by long slender mandibles and long claws. The O. bicarinatus species group is primarily Indo-Malayan and Australian but possibly contains the Palearctic O. minutus Kriechbaumer; it is characterized by having forewing vein Rs + 2r broadened and slightly curved before reaching the pterostigma. The O. luteus species group Fig. 1. Ophion sp. (Photograph: Andrea Jackson.) (sensu Gauld, 1985) is a large group that contains the vast majority of Palearctic and Neotropical species, and all of the no revisions of the genus across the Palearctic. Ophion was Nearctic species except for the two undescribed species from revised in the Netherlands by Oosterbroek (1978), and was the O. similis group. It was assumed by Gauld (1985) to be examined most comprehensively in the UK, with a number of paraphyletic with respect to the other species groups and was revisions over the years (Gauld, 1973, 1976, 1978) culminating only defined by the lack of characters that characterize the other in a thorough revision of the 16 known British Ophion species by species groups. Brock (1982). However, even among the relatively well-known Within the O. luteus group, Brock (1982) suggested some British fauna, molecular evidence suggests a host of undescribed sister-group relationships between British species, but these species (Schwarzfeld & Sperling, 2015). have not been tested with molecular characters and no hypoth- Ophion is a morphologically challenging genus, with lit- esis was proposed for the phylogeny of the genus as a whole. tle morphological divergence between most species and few Quicke et al. (2009) included nine species of Ophion as part of informative, discrete characters for identification (Gauld, 1980; their broad taxonomic sampling to obtain a phylogeny of Ichneu- Brock, 1982). Thus far, all attempts to divide Ophion into monidae based on 28S rDNA. Recently, Schwarzfeld & Sperling monophyletic species groups have been based on morphology (2014) defined one monophyletic species group (the O. scutel- and have been hampered by the lack of information regarding laris Thomson species group) from within the larger O. luteus the eastern Palearctic and Nearctic species (Gauld, 1985). A group, and revised the included Nearctic species. number of species groups and species complexes have been Two of the most commonly used molecular markers for informally referred to within Ophion; however, these have often hymenopteran phylogenies are the D2–D3 expansion region not been explicitly defined and were generally discussed in of 28S ribosomal RNA (28S) and the mitochondrial gene isolation, with little reference to the remainder of the genus (e.g. cytochrome oxidase 1 (COI) (e.g. Mardulyn & Whitfield, 1999; Gauld, 1973, 1977; Gauld & Mitchell, 1981; Brock, 1982). The Dowton & Austin, 2001; Whitfield et al., 2002; Michel-Salzat most comprehensive classification to date was by Gauld (1985), & Whitfield, 2004; Klopfstein et al., 2011). COI is often used at based in large part on the results of a previous morphologi- the species level, due to its relatively high mutation rate (Monti cal analysis (Gauld, 1980). Gauld (1985) outlined eight major et al., 2005; Li et al., 2010; Williams et al., 2012); however, it species groups, although there was no discussion of what consti- can also be informative at higher phylogenetic levels (Klopfstein tutes a major versus a minor species group. The majority of these et al., 2010; Quicke et al., 2012). In contrast, 28S is a highly groups were presumed to be monophyletic and contained a small conserved gene that has been widely used for higher-level insect number of morphologically distinctive species, whereas one phylogenies (Caterino et al., 2000; Heraty et al., 2011; Sharkey large group [the O. luteus (L) species group] served as a catch-all et al., 2012), although it has also proved useful in distinguishing for the remaining species. Ophion as a whole can be recognized species (Monaghan et al., 2005; Derycke et al., 2008; Raupach by the absence of a posterior mesosternal transverse carina, the et al., 2010). impressed clypeal apex, the long membranous flange on the fore The internal transcribed spacer 2 (ITS2) of nuclear ribosomal tibial spur, the presence of a distinct ramellus (nearly always), RNA is another rapidly evolving locus that has been most often and the straight, evenly slender forewing vein Rs + 2r (most used at the species-level in insects (e.g. Campbell et al., 1993; species), among other characters (Gauld, 1985). Ophion has also Gomez-Zurita et al., 2000; Wagener et al., 2006; Li et al., 2010; been hypothesized to be paraphyletic with respect to the other Schwarzfeld & Sperling, 2014). However, the highly conserved six genera within the Ophion genus group (Afrophion Gauld, secondary structure of this gene, with characteristic paired Agathophiona Westwood, Alophion Cushman, Xylophion domains and unpaired regions, also makes this an informative

© 2015 The Royal Entomological Society, Systematic Entomology, 41, 191–206 Phylogeny of Ophion 193 gene at higher phylogenetic levels (Coleman & Vacquier, 2002; included a nonophionine member of the higher ophioniformes Coleman, 2007). Incorporating secondary structure can improve (Anomalon Panzer) as a distant outgroup with which to root the alignment of rRNA, an otherwise problematic task due to phylogeny. the presence of numerous indels in the length variable loop The provenance and repositories of all specimens are listed regions (Kjer, 1995; Coleman & Vacquier, 2002). Also, since in Appendix S1. With the exception of sequences obtained the nonindependence of paired regions can mislead phylogenetic from GenBank or BOLD, all Nearctic taxa were identified by analyses, incorporating secondary structure can allow the use of M.D.S. and all Palearctic taxa were identified by G.R.B., with rRNA-specific models for phylogeny estimation (Savill et al., the exception of O. forticornis Morley, which was identified by 2001; Telford et al., 2005; Wolf et al., 2008; Letsch & Kjer, M. R. Shaw (National Museums Scotland, Edinburgh). 2011). The goal of this study is to obtain a preliminary molecular phylogeny of Ophion, with an emphasis on the O. luteus species Sequencing and alignment group sensu Gauld (1985), based on COI, 28S, and ITS2 DNA. The primary focus is on Nearctic species; however, representa- DNA was extracted from a single hind leg and COI and tives of most species known from the western Palearctic region ITS2 were sequenced using the primers and protocols described are included, as well as a few species from Costa Rica, Mada- in Schwarzfeld & Sperling (2014). For the D2–D3 region of gascar, Taiwan, Australia and New Zealand. 28S rDNA, the primers were: 5′-GCGAACAAGTACCGT GAG GG-3′ (forward) and 5′-TAG TTC ACC ATC TTT CGG GTC-3′ (reverse) (Laurenne et al., 2006), and PCR protocols Methods followed Schwarzfeld & Sperling (2014), except that cycling conditions were 94∘C for 2 min, 30 cycles of 96∘C for 15 s, Specimens 50∘C for 30 s, 72∘C for 30 s and a final extension of∘ 75 Cfor 7 min. Sequences are deposited in NCBI GenBank, and Genbank A total of 493 specimens of Nearctic Ophion were newly accession numbers are listed in Appendix S1. sequenced for this study. The majority of these specimens are Alignment of COI was unambiguous, and was confirmed by from Canada, with a smaller number from the US. Most were translating nucleotides to amino acids in mesquite (Maddison newly collected for this study; however, 25 specimens were & Maddison, 2011). A total of 681 COI sequences from sequenced from alcohol-preserved material on loan from the Ophion and 13 from outgroup taxa were included in the study Canadian National Collection of Insects, Arachnids and Nema- (including sequences from GenBank/BOLD). A few sequences todes. Specimens were selected for sequencing to represent the were not successfully sequenced in both directions, or were range of morphological variation observed in over 4000 spec- otherwise partially incomplete; these sequences were included imens from a variety of habitats. Eighty specimens from 15 with the gaps coded as missing data. For 28S, alignment was species of Palearctic Ophion were also newly sequenced, all of performed by eye in mesquite; there were occasional small which are from the UK, except for one specimen from Spain and indels, but generally alignment was unambiguous. The aligned one from France. sequences were 725 bp in length. The final dataset included In addition to these newly sequenced specimens, 93 Ophion 96 Ophion sequences (87 newly sequenced and nine from COI sequences were included from the Barcode of Life database Genbank), as well as 20 outgroup sequences. We were unable (BOLD). Finally, several Ophion sequences from GenBank to obtain complete ITS2 sequences for several specimens; as the were also included. This resulted in an additional 21 COI analysis of ITS2 secondary structure was sensitive to missing sequences (including eight sequences from Costa Rica, two data, we only included sequences that were complete and that from Madagascar and two from New Zealand) and nine 28S lacked ambiguous regions. The final dataset included 394 ITS2 sequences [including one sequence each from O. zerus Gauld sequences. The aligned sequences for each dataset are available (Australia) and O. bicarinatus Cameron (Taiwan)]. on the Dryad Data Repository: doi:10.5061/dryad.49g98. Outgroups were selected from across the range of genus groups within Ophioninae, as defined by Gauld (1985), although some of these genus groups are not well supported (Quicke ITS2 secondary structure et al., 2009). From the Ophion genus group, we included Afrophion Gauld, Alophophion Cushman, Rhopalophion The detailed protocol used to obtain the secondary structure of Seyrig and Xylophion Gauld. The other included genera were: ITS2 is described in Schwarzfeld & Sperling (2015). Briefly, we Sicophion Gauld, Janzophion Gauld (Sicophion group); Ere- first folded an arbitrarily chosen sequence using the RNAfold motylus Förster (Eremotylus group); Barytatocephalus Schulz, webserver (Hofacker, 2003) and default settings. We then Euryophion Cameron, Rhynchophion Enderlein, Thyreodon imported this structure into the ITS2 database, and used it as a Brullé (Thyreodon group); Dicamptus Szépligeti, Enicospilus template for folding the remaining sequences. If any sequences Stephens, Laticoleus Townes, Lepiscelus Townes, Leptophion had less than 90% similarity to the existing template, one Cameron, Ophiogastrella Brues, Pamophion Gauld, Simophion of the anomalous sequences was directly folded in RNAfold, Cushman, Stauropoctonus Brauns (Enicospilus group). As the and then added as an additional template to the database. phylogeny of Ophioninae in general is poorly known, we also Using this iterative process, the final dataset consisted of eight

© 2015 The Royal Entomological Society, Systematic Entomology, 41, 191–206 194 M. D. Schwarzfeld et al. sequences that were folded directly and 386 sequences that concatenated_all dataset, a supertree was constructed in clann were folded using homology. One sequence (from O. flavidus 3.0 (Creevey & McInerney, 2005), using the three individual NJ Brullé) appeared to be a pseudogene, and was thus excluded trees from paup* (COI, ITS2, and 28S) and the NJ function in from all phylogenetic analyses. The remaining sequences and clann. This tree was used as the starting tree for the concate- structures were aligned using 4sale, a program designed to nated_all ML analysis. The garli web service uses an adaptive synchronously align RNA sequences and structures (Seibel approach in which additional search replicates are based on the et al., 2006). The 28S and 5.8S flanking regions were not statistical probability that the best tree has been obtained with included in the phylogenetic analyses; the final alignment of 95% confidence. The ITS2, COI and concatenated_all analyses ITS2 was 1705 bp in length. each reached the maximum number of 1000 replicates, while the 28S and concatenated_common analyses were completed after ten replicates. One thousand bootstrap pseudoreplicates were Phylogeny estimation calculated for each dataset, using stepwise addition starting trees. Ophion phylogeny was estimated using maximum likelihood Finally, in order to incorporate the structural information of the (ML) and maximum parsimony (MP) analyses. Analyses were ITS2 sequence data, we calculated 100 NJ bootstrap replicates conducted on each dataset separately (COI, ITS2, and 28S), as using an ITS2 structure-specific rate matrix in profdists (Wolf well as on two concatenated datasets. The first dataset (‘con- et al., 2008). The rate matrix was obtained from profdists, catenated_all’) included all specimens and all loci. In order where it was calculated by converting the four bases into to investigate the effect of taxon sampling and missing data, a 12-letter alphabet, with each base in three possible states we also analysed a dataset consisting of only those specimens (unpaired, paired-right or paired-left), and then determining a that were successfully sequenced for all three loci (‘concate- GTR (ML)-corrected substitution model (Seibel et al., 2006; nated_common’). As no outgroup sequences were available for Wolf et al., 2008). ITS2, the ITS2 and concatenated_common trees were rooted with Ophion minutus Kriechbaumer, as this was consistently recovered as sister to the remaining ingroup species in the other Results three analyses, as well as having morphological characters that indicate it is sister to the remaining Ophion species. The concatenated_all and concatenated_common datasets were We conducted MP analyses using paup* 4.0 (Swofford, each partitioned by locus and codon position (CP), result- 2002). Analyses were run using heuristic searches with ten ing in five partitions. The concatenated_all dataset used CP1: replicates of random-addition starting trees (swap = tree– GTR+I+G; CP2: F81+G; CP3: GTR+I+G; ITS2: SYM+G; bisection–reconnection, saving no more than 10 000 trees per 28S: K80+I+G. Models for the concatenated_common dataset replicate), and a 50% majority-rule tree was constructed. For were the same except that CP1 and CP3 used GTR+G, and CP2 each analysis, 1000 bootstrap pseudoreplicates were conducted used F81. The individual analyses used the same models as the using the same search settings except that simple stepwise concatenated_all dataset. addition was used, and a maximum of 100 trees were saved per Thirteen major ingroup clades (= species groups) were recov- replicate. ered in this study, although not all clades were represented in all Models for the ML analyses were determined using parti- analyses (Figs 2, 3; Figures S1–S4; Table S1). In the COI and tionfinder version 1.0.1 (Lanfear et al., 2012). This program concatenated_all analyses, bootstrap support was much stronger considers all potential partitions within a dataset, and calculates in the MP analyses, compared with the ML analyses; however, the best partitioning scheme and best model(s) for each scheme. for the other analyses, support values were similar between the Only those models able to be implemented in the garli web ser- two methods (Table S1). The ITS2-specific NJ bootstrap values vice were included. Models were calculated for COI using the from profdists were almost identical to the MP bootstrap three codon positions as potential partitions, and the ‘greedy’ values for ITS2, and the analysis did not produce any divergent search algorithm. For the two concatenated analyses, each gene results. Bootstrap values are only shown for the major nodes in and the codon position for COI were included, with the ‘greedy’ Fig. 2; however, all bootstrap values are included in Figure S1. search algorithm. The 28S and ITS2 datasets were not parti- The trees from the individual analyses can be seen in Figures tioned, and all included models were assessed (‘search = all’). S2–S4, with bootstrap values shown for the major nodes. Best models were selected according to the Bayesian infor- The two concatenated datasets recovered the same major mation criterion (Schwarz, 1978), as recommended by Lanfear clades; however, bootstrap support was generally higher in the et al. (2012). concatenated_common analyses, with most major nodes having The ML analyses were conducted using the web ser- high bootstrap values (Fig. 3, Table S1). The same clades were vice for garli 2.1 (Zwickl, 2006), hosted at http://www. also recovered by the analyses of ITS2 and COI. However, molecularevolution.org (Bazinet et al., 2014). The ITS2 and the pattern of divergence differed between these two loci, with COI ML analyses used a neighbour-joining (NJ) starting tree ITS2 showing high sequence divergence between species groups constructed in paup* (p-distance), while the 28S and con- and very low sequence divergence within groups (Figure S3), catenated_common analyses used a stepwise addition starting while COI had a less distinct separation of within-group and tree. As there were taxa with no overlapping characters in the between-group divergence (Figure S2). The 28S tree was the

Fig. 2. (A–E) Maximum likelihood tree from the concatenated_all dataset, including 704 Ophion and 21 outgroup specimens and three loci [cytochrome oxidase 1 (COI), internal transcribed spacer 2 (ITS2) and 28S ribosomal RNA (28S)]. Maximum likelihood and maximum parsimony bootstrap values > 50% are given for the major nodes (MP/ML).

Fig. 2. Continued. least resolved of any of the analyses, with few clades having Fig. 2), and a clade containing all remaining Ophion as well strong bootstrap support (Figure S4). Summary statistics for all as Rhopalophion (labelled II in Fig. 2). However there was analyses are shown in Table S2. no bootstrap support for either of these clades. O. minutus At the root of the tree, a polytomy was recovered including and O. ventricosus formed a monophyletic clade, though with Alophophion, a clade containing Afrophion, Xylophion, O. zerus, weak bootstrap support (Fig. 2; Table S1). With the exception O. ventricosus Gravenhorst and O. minutus (labelled I in of the highly divergent O. flavidus sequence, O. minutus had

Fig. 3. Maximum likelihood tree from the concatenated_common dataset, including 67 Ophion specimens. Maximum likelihood and maximum parsimony bootstrap values > 50% are given for major nodes (MP/ML). the most atypically shaped ITS2, compared with other species concatenated_all tree, it also included the genus Rhopalophion, (Fig. 4H). We were not able to successfully sequence ITS2 for although only 28S was available for this genus. The O. areolaris O. ventricosus. group was represented by a single species in this study, and was Within clade II, there was little bootstrap support for the recovered as sister to the O. scutellaris group, with strong sup- deeper nodes in the concatenated_all tree; however, the con- port in the concatenated_common dataset (MP: 98, ML: 100), catenated_common tree was quite well resolved, with most but no support according to concatenated_all. nodes having strong bootstrap support (Fig. 3). The O. scutel- The O. slossonae Davis and O. parvulus Kriechbaumer laris group was recovered as monophyletic, with weak sup- species groups were also recovered as sister clades, with strong port according to the concatenated_all dataset, but very strong support in the concatenated_common analysis (Figs 2, 3). The support in the concatenated_common tree (Table S1). In the O. parvulus group was strongly supported by all analyses.

sequence ITS2 for one specimen within this group (O. flavidus). This sequence was highly divergent from all other species, indicating it may be a pseudogene, as it greatly exceeds the amount of expected variation in ITS2 (Kita & Ito, 2000; Alvarez & Wendel, 2003). The secondary structure was equally distinct; although the conserved regions of the structure were maintained, the unusually branched Helix III provides further evidence that this sequence may be a pseudogene (Fig. 4G). However, another indication of ITS2 pseudogenes (lower GC content) was not observed in this sequence (Buckler et al., 1997; Kita & Ito, 2000). Attempts to sequence two other specimens in this species group resulted in very small amounts of clean sequence that were clearly similar to that obtained for O. flavidus. Two COI sequences from New Zealand Ophion were included from GenBank. These two specimens formed a strongly supported monophyletic group (MP: 99, ML: 88), which was sister to (O. flavidus group + (O. parvulus group + O. slossonae group)), although without bootstrap support. The O. pteridis Kriechbaumer and O. nigrovarius Provancher species groups were recovered as sister groups, although with no support in the concatenated_all analyses, and only weak support in the concatenated_common analyses. Together, these two groups were sister to the O. luteus group, with strong support in the concatenated_common analyses (MP: 92, ML: 86), although again support was lacking in the concatenated_all analyses. The O. pteridis group was strongly supported in most analyses, although the ML analysis of the concatenated_all dataset had a bootstrap of < 50. The O. luteus s.s. group was strongly supported by both datasets, with bootstrap support ranging from 82 (concatenated_all ML) to 100 (concatenated_common MP, Fig. 4. Secondary structure of internal transcribed spacer 2 (ITS2) in ML). This group includes the Palearctic O. luteus (L) and Ophion spp., as determined by the free energy method in RNAfold. the Nearctic species O. elongatus Hooker. The O. nigrovarius (A) All other species groups; (B) O. scutellaris species group; (C) O. group includes two species: the Palearctic O. perkinsi Brock parvulus species group; (D) O. ocellaris species group; (E) O. slossonae and the Nearctic O. nigrovarius. It was strongly supported in species group A; (F) O. slossonae species group B; (G) O. flavidus the concatenated_all dataset (MP: 95, ML: 71); however, only species group; (H) O. minutus species group. O. nigrovarius was available for the concatenated_common dataset. Within this group, there was a deep divergence within O. parvu- Two COI sequences of Ophion from Madagascar were lus, indicating that two species are present (O. parvulus AandO. included from GenBank, forming a strongly supported mono- parvulus B, Fig. 2). The Nearctic specimens in this group were phyletic group (MP: 100, ML: 87). Species group 1 was also separated into two strongly supported clades (Fig. 2; Figure recovered with moderate bootstrap support in the concate- S1). In one of these clades, however, sequencing appears to have nated_all dataset (MP: 86, ML: 71). Support was very strong amplified a nuclear mitochondrial pseudogene (numt) (Gellissen in the concatenated_common dataset (MP and ML: 100), et al., 1983; Lopez et al., 1994), as there is a 2-bp deletion at the although this dataset only included two specimens from this same location in each sequence that would result in a frameshift group. O. longigena Thomson was recovered as sister to this in the coding mitochondrial gene. The O. slossonae group was group in the concatenated_all analyses, although with very strongly supported in the concatenated_common analysis (MP: low bootstrap support; only COI was available for this species. 100, ML: 96) and in the concatenated_all parsimony analysis The placement of the Madagascar group and species group 1 (MP: 90). However, it was recovered with less than 50% sup- varied between analyses. The concatenated_all MP analysis port in the concatenated_all ML analysis. This group includes recovered [Madagascar group + (Species group 1 + O. longi- O. slossonae, as well as morphospecies A from Schwarzfeld & gena)] as sister to the O. obscuratus group, as did the MP and Sperling (2015). ML analyses of the concatenated_common dataset (although In the concatenated_all analysis, the O. flavidus group was O. longigena and the Madagascar group were not included recovered with strong support (MP: 96, ML: 87). It was the sis- in this dataset). In comparison, the ML analysis of concate- ter group to (O. slossonae group + O. parvulus group), although nated_all recovered the Madagascar group as sister to (luteus with no bootstrap support. It was not included in the concate- group + (pteridis group + nigrovarius group)), while (species nated_common dataset, as we were only able to successfully group 1 + O. longigena) was sister to this combined clade.

The O. obscuratus Fabricius species group was recovered Within Ophion, the secondary structure of ITS2 was quite as monophyletic by most analyses. According to the concate- conserved among many taxa, and highly divergent between oth- nated_all ML analysis, two specimens (one sequence each of ers. The vast majority of specimens were successfully mod- O. mocsaryi and O. costatus) were not included in this clade; elled from a single template, with a minimum of 90% of the however, these were each only represented by 28S. Bootstrap base pairs assigned to the template structure. This includes all support for this group was very low, with the strongest support specimens from the obscuratus, luteus, pteridis and nigrovarius for this group coming from the concatenated_common dataset groups and species group 1 (Fig. 4A). The remaining species (Figs 2, 3; Table S1). The O. obscuratus group includes mor- groups each required a separate template. In most cases, these phospecies B from Schwarzfeld & Sperling (2015), as well as differed only slightly in shape, particularly in the number of a monophyletic clade consisting of several Palearctic species short helices between helix IIa and helix III, and in the branch- (O. crassicornis Brock, O. costatus Ratzeburg, O. mocsaryi ing pattern of helix III. The structures for O. minutus and O. Brauns, O. brevicornis Morley, O. forticornis Morley, and flavidus were the most distinct (Fig. 4G, H) The O. flavidus apparently two species within ‘O. obscuratus’) (Fig. 2). Within structure was particularly unusual as helix II was branched, a this Palearctic group, specimens identified as O. costatus and pattern rarely observed across Metazoa (Coleman, 2007). The O. mocsaryi were found in three clades: one of O. costatus, one slossonae group was the only species group to have two dis- of O. mocsaryi, and a third clade with two specimens identi- tinct secondary structures (Fig. 4E, F). One of these structures fied as mocsaryi and two as costatus. An additional specimen (Fig. 4F) was found in a single undescribed species (three spec- of O. mocsaryi was found separately from the above three imens) and differed from all other sequences by having a large clades, but as it was only represented by ITS2, this placement deletion in the centre of the sequence, thus lacking most of probably reflects the locus, rather than indicating an additional helix IIa. species. Two specimens identified as O. obscuratus (labelled O. obscuratus C; Fig. 2) were not recovered as part of this Palearc- tic clade, although they were part of the obscuratus group Discussion (Fig. 2). The single representative of O. bicarinatus from GenBank, This study presents the first phylogenetic hypothesis for the and two Nearctic specimens (MS10835 and UASM341138) genus Ophion. Gauld (1985) suggested that the genus Ophion were recovered as sister to the clade including species group 1 was paraphyletic with respect to the six other genera within the and the Madagascar, obscuratus, luteus, pteridis,andnigrovar- Ophion genus group. Four of these genera were included in this ius groups in the concatenated_all analysis, although without study (Afrophion, Alophophion, Rhopalophion and Xylophion), support. O. bicarinatus was only represented by 28S, which and our results support a close relationship between these was quite uninformative in this portion of the tree (Figure S4). species and Ophion. While more sampling is needed, these The other two specimens were also represented by COI and results suggest that by erecting a new genus for clade I (the ITS2; however, their placement on the tree varied depending on O. minutus group, O. zerus, Xylophion,andAfrophion)the which locus was examined (Figures S2 and S3). In the concate- remaining Ophion (possibly including Rhopalophion) would be nated_common tree, they were recovered as sister to the obscu- a monophyletic genus (= clade II in this study). ratus group, with weak support, possibly indicating they should This study could be regarded as a preliminary analysis due to be included in this species group (Fig. 3). limited geographical sampling; only three of the eight species groups outlined by Gauld (1985) are included in the analysis (the O. bicarinatus, O. areolaris and O. luteus groups), ITS2 secondary structure and the sampling is heavily biased towards Canadian and western Palearctic material. Delimitation of taxa above the The ITS2 of rDNA varied from 717 to 1056 bp in length, with species level is necessarily subjective. We delimited species all except two species having greater than 850 bp. The secondary groups primarily based on distinct sequencing gaps between structure was congruent with the conserved features of ITS2 well-supported clades. However, some species groups were less observed across a wide range of taxa (Schultz et al., 2005; supported and further sampling may fill in some of the sequenc- Coleman, 2007). In particular, all species have recognizable ing gaps, thus requiring a reanalysis of these species groups. helices II and III, the hallmark helices of almost all known Despite these limitations, this is the first attempt at resolving ITS2 sequences (Coleman, 2007). Helix II is relatively short, the ‘luteus anathema’ (Gauld, 1985, p. 125), and as such pro- almost always unbranched, and has a pyrimidine-pyrimidine vides a valuable framework for understanding the relationships mismatch bulge near its base. Helix III is long, often branched between the majority of Ophion species. While the vast major- and more highly variable, but with a conserved sequence ity of Ophion in this study are undescribed (Schwarzfeld & motif near the apex on the 5′ side. In all Ichneumonidae Sperling, 2015), there is at least one described species in all examined (M.D Schwarzfeld, unpublished data), this motif is except species group 1 and the Madagascar and New Zealand CGGTCGATCGAGTCC. In Ophion, there are an additional two groups. to four helices between helix II and helix III. The first of these Thirteen species groups are designated in this study, most of (tentatively labelled helix IIA) is present in all species and is which are newly coined here. With the exception of O. minutus, almost always very long (Fig. 4). O. zerus, O. bicarinatus (O. bicarinatus group), O. ocellaris

Fig. 5. Characters of selected Ophion species groups. Forewing: (A) O. minutus group (O. minutus), (B) O. parvulus group (undescribed sp.), (C) O. obscuratus group (undescribed sp.). Mid-tibial spurs: (D) O. scutellaris group (O. idoneus). Scutellum: (E) O. scutellaris group (O. keala), (F) O. pteridis group (undescribed sp.), (G) O. obscuratus group (undescribed sp.). First tergite: (H) O. parvulus group (undescribed sp.), (I) O. pteridis group (undescribed sp.). Hind trochantellus (Tr): (J) O. luteus group (O. luteus), (K) O. parvulus group (undescribed sp.). Mandibles: (L) O. luteus group (O. luteus), (M) O. pteridis group (undescribed sp.), IA, internal angle; face: (N) O. nigrovarius,(O):O. slossonae group (undescribed sp.). (Photographs: (A–C), (E), (H–M), Rylee Isitt.)

Ulbright (O. areolaris group), and the two species groups from having forewing vein Rs + 2r thickened and slightly curved Madagascar and New Zealand, the remaining species were all before reaching the pterostigma (Fig. 5A). Ophion ventricosus presumably included within the catch-all O. luteus species group Gravenhorst also has a slightly thickened Rs + 2r (Brock, 1982) (Gauld, 1985). but was not mentioned by Gauld (1985). Similarly thickened Rs + 2r veins are found in other ophionine genera, such as Eremotylus Förster, which lends support to these species being Ophion minutus Kriechbaumer species group sister to the remaining Ophion. In contrast, the O. luteus species group sensu Gauld (1985) has Rs + 2r straight and of equal According to Gauld (1985), the O. bicarinatus species group thickness along its length (Fig. 5B, C). includes O. bicarinatus, all of the Australian species (including While only weakly supported by the sequence data, the O. min- O. zerus), and possibly O. minutus; it is characterized by utus + ventricosus clade is further supported by biology, as both

© 2015 The Royal Entomological Society, Systematic Entomology, 41, 191–206 Phylogeny of Ophion 201 species have been reared from Geometridae, unlike the majority 28S sequence was included in this study, additional taxon of Ophion, which are parasitoids of Noctuoidea (Brock, 1982). and gene sampling is needed before a formal name change is In their analysis of 28S across Ichneumonidae, Quicke et al. proposed. (2009) recovered O. zerus and O. minutus as a clade, but did not recover O. ventricosus within Ophion. The 28S sequence from Genbank identified as Ophion bicarinatus was not recovered O. parvulus Kriechbaumer species group near O. minutus, but rather was near the O. obscuratus group and related species groups; it was similarly recovered by Quicke This group was very strongly supported by the sequence data. et al. (2009). This may indicate a misidentification of this Morphologically, this group can be recognized by the evenly species by Quicke et al. (2009), as it would be surprising to find arched forewing vein Rs (i.e. not sinuous) (Fig. 5B; cf. Fig. 5C) a species with a thickened Rs + 2r within a group with otherwise and by the unusually short first tergite, with the spiracles in line typical venation. In this study, we have included O. minutus with the ventral membrane (Fig. 5H). Brock (1982) referred to and O. ventricosus in the O. minutus species group but left O. O. parvulus as forming a ‘close species pair’ with O. pteridis, zerus unassociated, pending further study of O. bicarinatus and but our results show they are not in fact closely related. Thus of other members of the O. bicarinatus species group sensu far all Nearctic specimens in this group have been collected in Gauld (1985). O. minutus had the most atypical-shaped ITS2 late summer (August or September), although one of the two compared with other species, providing additional evidence Palearctic species was collected in May. that O. minutus is distinct from all Ophion species in clade II. Determining the ITS2 sequence and structure of other species in clade I would be informative for understanding the root of O. slossonae Davis species group the Ophion phylogeny. In most cases, we named each species group after the earliest-named species within it; however, as This species group is highly morphologically variable the term ‘O. minutus species group’ has previously been used and specimens have been collected from early June to in the literature (e.g. Gauld, 1977), we are retaining it for this mid-September (with one specimen from Florida in March). group. This is also the only species group to have two distinct ITS2 structures in different species (Fig. 4E, F). Additional morphological study of this group may reveal characters that are O. areolaris Brauns species group consistent within the group or, conversely, might suggest it should be further subdivided. The nominate species, O. slos- Due to the lack of the occipital carina, the O. areolaris sonae Davis, is by far the largest Nearctic species we have species group (Gauld, 1985) has been treated as a separate examined; however, other species in this group are of a more genus, Platophion Hellén, by some authors (Hellén, 1926; typical size. Brock, 1982); however, it was synonymized with Ophion by Gauld (1980). This study supports Gauld’s conclusion that the species group should be included within Ophion, probably as O. pteridis Kriechbaumer species group sister to the O. scutellaris species group. This group is also unusual within Ophion, as it has been reared from Thyatirinae This is a late-flying species group, with almost all collection (Drepanidae) (Brock, 1982), unlike the usual Noctuoidea hosts. records from July to September. Species in this group have As well as lacking the occipital carina, species of the O. strong scutellar carinae (Fig. 5F; cf. Fig. 5G) and a distinctly areolaris group are unusual in their rather square dorsum of sinuous first tergite (Fig. 5I). Again, there is no evidence that O. the mesoscutellum, black inter-ocellar area and, at least in some pteridis is closely related to O. parvulus, as was suggested by species, pronounced sexually dimorphic colour patterns (males Brock (1982). can be considerably darker).

O. luteus (L.) species group O. scutellaris Thomson species group Species from this group have been collected from May to The O. scutellaris group is treated in detail by Schwarzfeld September, with one species from Arizona in March or April & Sperling (2014). The members of this group are mostly (Appendix S1). While morphologically variable, most species dark reddish, early-season species and can be recognized by a can be recognized by the long hind trochantellus (Fig. 5J; cf. suite of morphological characters [e.g. subequal mid-tibial spurs Fig. 5K), relatively transverse face, and blunt-tipped mandibles (Fig. 5D), carinate scutellum (Fig. 5E)]. This group includes a with weak or absent internal angles (Fig. 5L; cf. Fig. 5M). minimum of seven Nearctic species and two Palearctic species. O. elongatus is a distinctly elongated Nearctic species which The recovery of Rhopalophion within this clade suggests that was recovered within this group, based on two COI sequences it should be transferred to Ophion. Quicke et al. (2009) also obtained from BOLD; we have not examined these specimens to recovered it as sister to O. scutellaris, using the same 28S determine whether this species also has the typical luteus group sequence as in the current study. However, as only a single characters.

O. nigrovarius Provancher species group is nonetheless a useful categorization pending further study. It is possible that three unplaced specimens (O. bicarinatus, Ophion nigrovarius and O. perkinsi are both buccate-headed UASM341138 and MS10835) also belong to this group; how- (i.e. have expanded genae; Fig. 5N; cf. Fig. 5O), which sup- ever, more evidence is needed before they can be conclusively ports the molecular data linking them. Based on morphol- placed. ogy, Brock (1982) believed O. perkinsi was closely related Brock (1982) states that O. mocsaryi and O. costatus are to O. brevicornis; however, in our study O. brevicornis was more closely related than any other pair of British Ophion, part of the Palearctic clade within the O. obscuratus group. and suggests they should be synonymized. Our study supports Ophion nigrovarius is a particularly interesting species as it this close relationship, but suggests that part of the difficulty has been recorded from Phyllophaga (Coleoptera: Scarabaei- in distinguishing these species may be because they form part dae), the only known record of a coleopteran host for Ophion of a species complex of more than two species. Brock’s other (Townes, 1971). No hosts have been recorded for O. perkinsi (Yu hypothesis, that the mocsaryi–costatus species pair is closely et al., 2012). related to the pteridis–parvulus pair, is not supported by our study, as the latter species are not closely related to each other or to the members of the obscuratus group. In the Nearctic Species group 1 region, this group includes an early-season, yellow-marked Ophion species that is by far the most abundant species in all This is a morphologically diverse clade; while COI had a Canadian localities that have been well sampled; however, the maximum sequence divergence of only 2.3% within the 12 precise boundaries of this species are still undetermined (M.D. included specimens, we predict based on morphology that the Schwarzfeld, unpublished data). In total, the O. obscuratus group contains a minimum of four species, none of which are group contained 56% of all COI sequences and 67% of ITS2 apparently described (M.D. Schwarzfeld, unpublished data). sequences in this study. Specimens have been collected from May to September. Some of these species are buccate-headed, though not as dramatically as in the O. nigrovarius group, while others are more typical. It is possible that this group should be expanded to include O. ITS2 secondary structure longigena, another buccate-headed species, however, additional specimens and genes need to be examined, as only a single COI The highly variable unpaired regions of ITS2, combined with sequence was included in this study and its inclusion was not the nonindependence of pairing regions, can potentially mis- supported by bootstrap analyses. lead phylogenetic inferences (Tillier & Collins, 1995; Telford et al., 2005; Wolf et al., 2008; Letsch & Kjer, 2011). However, using an ITS2-specific model that incorporates secondary struc- New Zealand and Madagascar groups ture had very little impact on the results of this study. This may be because the observed variation in ITS2, namely large differ- Gauld (1985) stated that all described New Zealand Ophion ences between species groups and low divergence within them, species were found in the O. peregrinus species group, based on provided a strong enough signal that the choice of model did not a number of morphological characters. It is therefore possible significantly affect the results. that the two New Zealand sequences included in this study As the ITS2 sequences were folded in silico, further studies are belong to this group. However, as the sequences from GenBank needed to confirm that these structures are accurate (Eddy, 2004; were not identified to the species level, and we have not Marinho et al., 2013). Overall, the ITS2 structures were consis- examined the specimens, we refrain from assuming that they are tent with the conserved patterns observed by Coleman (2007), part of this named species group. Similarly, the two sequences loosely corresponding to the four-domain model predicted for from Madagascar are clearly distinct from all other Ophion in ITS2 (Schultz et al., 2005). They did, however, differ in sig- this study; however, with such limited sampling from the area, nificant ways. Most dramatically, ITS2 in Ophion is consider- we are unable to draw any general conclusions regarding Ophion ably longer than in most other species that have been studied of Madagascar. (Schlötterer et al., 1994; Gomez-Zurita et al., 2000; Young & Coleman, 2004; Coleman, 2007; Marinho et al., 2011), including many other Ichneumonidae (Ashfaq et al., 2005; Wagener O. obscuratus Fabricius species group et al., 2006). This length contributed to difficulties in sequencing the gene, making it less optimal for phylogenetic analyses While this study provides much resolution of the species in Ophion compared with other taxa. Much of the additional groups within Ophion,theO. obscuratus group remains an length was found in the third helix between helix II and helix III. unresolved challenge. This species complex consists of many We have called this domain helix IIA, assuming it is equivalent apparently closely related species, some of which are morpho- to the IIA of other insect species (Coleman, 2007). However, logically distinct, while others are less so (M.D. Schwarzfeld, the lack of a variable helix IV in most Ophion species indi- unpublished data). While most evidence points to this group cates that perhaps this domain is equivalent to helix IV in other being monophyletic, bootstrap support is very low. This group taxa. It is biologically interesting that one species within the

O. slossonae group lacked this long helix, even though all other Supporting Information specimens within the species group had a more typical-shaped ITS2. The nearly identical structure found in species group 1 and Additional Supporting Information may be found in the online the pteridis, luteus, nigrovarius and obscuratus groups further version of this article under the DOI reference: support the monophyly of this clade. 10.1111/syen.12152 Figure S1. Maximum likelihood tree from the concate- nated_all dataset, including 704 Ophion and 21 outgroup Species delimitation and pseudogenes specimens and three loci (COI, ITS2, 28S). Maximum > The vast majority of Nearctic species in this study are likelihood and maximum parsimony bootstrap values 50% undescribed. However, the presence of several previously are presented (MP/ML). For clarity, occasional weak boot- unrecognized putative new species within the British material strap values for short branches that conflicted between anal- shows that even in well-studied areas there is much work to yses were excluded from the figure. be done at the species level within Ophion (Schwarzfeld & Figure S2. Maximum likelihood tree of 681 Ophion and 12 Sperling, 2015). This also suggests that using a single exemplar outgroup COI sequences. Maximum likelihood and maxi- per species for phylogenetic inference can potentially mislead mum parsimony bootstrap values > 50% are given for major the analysis if there is any doubt as to the identity of the nodes (MP/ML). species involved. Until Ophion are better resolved at the species level, we are limited in our understanding of their overall Figure S3. Maximum likelihood tree of 393 Ophion ITS2 phylogeny. sequences. Maximum likelihood and maximum parsimony The amplification of nuclear mitochondrial pseudogenes bootstrap values > 50% are given for major nodes (MP/ML). (nonfunctional copies of mitochondrial genes that have been Figure S4. Maximum likelihood tree of 96 Ophion and 20 incorporated into the nuclear genome, or numts) can mislead outgroup 28S D2-D3 sequences. Maximum likelihood and phylogenetic analyses based on mitochondrial DNA (Song > et al., 2008; Buhay, 2009; Moulton et al., 2010). In this study, maximum parsimony bootstrap values 50% are given for the amplification of a potential numt in some members of major nodes (MP/ML). the parvulus group means the results of the analyses should be Table S1. Summary of bootstrap support values for 12 viewed cautiously. As ITS2 and 28S also support the monophyly species groups of Nearctic and western Palearctic Ophion of this group, it appears that, at least at the species group level, (Only a single species of the O. areolaris species-group was the numts remain phylogenetically informative. Their presence, included in the analyses, and therefore this group is not however, limits the utility of COI for species delimitation. These included in the table.). nm, not monophyletic; all, all taxa numts were recognized due to a presumably homologous 2-bp and all loci included; common, only taxa with all three loci deletion in several otherwise clean sequences. However, not all present included. numts contain indels or stop codons, and may be difficult to rec- ognize (Song et al., 2008; Buhay, 2009; Moulton et al., 2010). Table S2. Summary of maximum likelihood (ML) and max- Similarly, the amplification of a possible ITS2 pseudogene in imum parsimony (MP) analyses of three genes of Ophion. the O. flavidus species group could also potentially mislead Appendix S1. List of all specimens included in this study. phylogenetic analyses. Both of these examples emphasize the importance of not relying on a single molecular marker for phylogenetic inference (Dupuis et al., 2012). Acknowledgements

Thank you to the collectors who provided specimens to Conclusion M.D.S., especially John Acorn, John Adams, Gary Anweiler, Charley Bird, Heather Bird, Brett Bodeux, Bryan Brunet, Jason This study provides a framework for future studies of the Dombroskie, Julian Dupuis, Jessica Edwards, Dave Holden, diverse and morphologically challenging genus Ophion.For Wes Hunting, Dave Langor, Nina Lany, Dave Lawrie, the first time, the ‘luteus anathema’ (Gauld, 1985, page 125) Lisa Lumley, Doug Macaulay, Boyd Mori, Ted Pike, Greg is divided into more manageable, discrete units that can be Pohl and Chris Schmidt. Thank you also to the numerous examined for morphological characters and biological informa- light-trappers who have sent specimens to G.R.B. for the Noc- tion. Many species are morphologically very similar between turnal Ichneumonoidea Recording Scheme. We are grateful species groups, and there are also highly divergent species to Jeremy deWaard for providing many specimens as well as within groups. With this phylogeny as a guide, it will now be nearly 100 COI sequences through BOLD. Thanks also to possible to conduct more targeted morphological analyses, with Andrew Bennett (Canadian National Collection) for the loan reanalysis of known morphological characters and discovery of of pinned and alcohol-preserved specimens, and to Mark Shaw new characters, to better characterize the species groups and to for sending a leg from O. forticornis. Matthias Wolf kindly facilitate the delimitation and description of species within each provided assistance with 4Sale and ProfDistS. Thank you group. to Rylee Isitt for help with photographing specimens and to

Andrea Jackson for a live shot of Ophion. Many thanks to Pascal Gauld, I.D. (1973) Notes on the British Ophionini (Hym., Ichneu- Rousse and two anonymous reviewers for their comments on an monidae) including a provisional key to species. Entomologist’s earlier version of this manuscript. This work was supported by Gazette, 24, 55–65. the Sustainable Forest Management Network; Alberta Conser- Gauld, I.D. (1976) Notes on the British Ophionini (Hym., Ichneu- vation Association Grants in Biodiversity; Natural Sciences and monidae). Part 3. The identity of the species described by Morley, Engineering Research Council (NSERC) and Alberta Ingenuity 1915 and Thomson, 1888 and a checklist of British species. Entomol- ogist’s Gazette, 27, 113–117. (Alberta Innovates) student scholarships to M.D.S.; and by an Gauld, I.D. (1977) A revision of the Ophioninae (Hymenoptera: Ichneu- NSERC Discovery grant to F.A.H.S. monidae) of Australia. Australian Journal of Zoology Supplementary Series, 49, 1–112. Gauld, I.D. (1978) Notes on the British Ophioninae (Hym., Ichneu- References monidae). Part 4. A revised key to the species of the genus Ophion Fabricius. Entomologist’s Gazette, 29, 145–149. Alvarez, J.M. & Wendel, J.F. (2003) Ribosomal ITS sequences and plant Gauld, I.D. (1980) An analysis of the classification of the Ophion phylogenetic inference. Molecular Phylogenetics and Evolution, 29, genus-group (Ichneumonidae). Systematic Entomology, 5, 59–82. 417–434. Gauld, I.D. (1985) The phylogeny, classification and evolution of para- Ashfaq, M., Erlandson, M.A. & Braun, L. (2005) Hyperparasitism by Mesochorus spp. (Hymenoptera: Ichneumonidae) in Peristenus sp. sitic wasps of the subfamily Ophioninae (Ichneumonidae). Bulletin of (Hymenoptera: Braconidae) and development of PCR primers for the British Museum of Natural History (Entomological Division), 51, hyperparasitoid detection. Biological Control, 32, 371–377. 61–185. Bazinet, A.L., Zwickl, D.J. & Cummings, M.P. (2014) A gateway for Gauld, I.D. (1988) A survey of the Ophioninae (Hymenoptera: Ichneu- phylogenetic analysis powered by grid computing featuring Garli 2.0. monidae) of tropical Mesoamerica with special reference to the fauna Systematic Biology, 63, 812–818. of Costa Rica. Bulletin of the British Museum of Natural History, Brock, J.P. (1982) A systematic study of the genus Ophion in Britain Entomology Series, 57, 1–309. (Hymenoptera, Ichneumonidae). Tijdschrift Voor Entomologie, 125, Gauld, I.D. & Mitchell, P.A. (1978) The Taxonomy, Distribution and 57–96. Host Preferences of African Parasitic Wasps of the Subfamily Ophion- Buckler, E.S., Ippolito, A. & Holtsford, T.P. (1997) The evolution of inae (Hymenoptera: Ichneumonidae). Commonwealth Institute of ribosomal DNA: divergent paralogues and phylogenetic implications. Entomology, Slough. Genetics, 145, 821–832. Gauld, I.D. & Mitchell, P.A. (1981) The Taxonomy, Distribution and Buhay, J.E. (2009) “COI-like” sequences are becoming problematic Host Preferences of Indo-Papuan Parasitic Wasps of the Subfamily in molecular systematic and DNA barcoding studies. Journal of Ophioninae (Hymenoptera: Ichneumonidae). Commonwealth Insti- Crustacean Biology, 29, 96–110. tute of Entomology, Slough. Campbell, B.C., Steffen-Campbell, J.D. & Werren, J.H. (1993) Phy- Gellissen, G., Bradfield, J.Y., White, B.N. & Wyatt, G.R. (1983) logeny of the Nasonia species complex (Hymenoptera: Pteromalidae) Mitochondrial DNA sequences in the nuclear genome of a locust. inferred from an internal transcribed spacer (ITS2) and 28S rDNA Nature, 24, 774–786. sequences. Insect Molecular Biology, 2, 225–237. Gomez-Zurita, J., Juan, C. & Petitpierre, E. (2000) Sequence, secondary Caterino, M.S., Cho, S. & Sperling, F.A.H. (2000) The current state structure and phylogenetic analyses of the ribosomal internal tran- of insect molecular systematics: a thriving tower of Babel. Annual scribed spacer 2 (ITS2) in the Timarcha leaf beetles (Coleoptera: Review of Entomology, 45, 1–54. Chrysomelidae). Insect Molecular Biology, 9, 591–604. Coleman, A.W. (2007) Pan-eukaryote ITS2 homologies revealed by Hellén, W. (1926) Beitrage zur Kenntnis der Ichneunioniden Finlands. RNA secondary structure. Nucleic Acids Research, 35, 3322–3329. 2 Subfam. Ophioninae und Anomaloninae. Acta Societatis pro Fauna Coleman, A.W. & Vacquier, V.D. (2002) Exploring the phylogenetic et Flora Fennica, 56, 3–27. utility of ITS sequences for animals: a test case for abalone (Haliotis). Heraty, J., Ronquist, F., Carpenter, J.M. et al. (2011) Evolution of the Journal of Molecular Evolution, 54, 246–257. hymenopteran megaradiation. Molecular Phylogenetics and Evolu- Creevey, C.J. & McInerney, J.O. (2005) Clann: investigaging phylo- tion, 60, 73–88. genetic information through supertree analyses. Bioinformatics, 21, Hofacker, I.L. (2003) Vienna RNA secondary structure server. Nucleic 390–392. Acids Research, 31, 3429–3431. Derycke, S., Fonseca, G., Vierstraete, A., Vanfleteren, J., Vincx, M. Kim, K.-B., Suh, K.-I. & Lee, J.-W. (2009) Taxonomical review of & Moens, T. (2008) Disentangling taxonomy within the Rhabditis the subfamily Ophioninae (Insecta: Hymenoptera: Ichneumonidae) (Pellioditis) marina (Nematoda, Rhabditidae) species complex using Ophion Korean Journal of Systematic molecular and morphological tools. Zoological Journal of the Linnean from Korea II. Genus Fabricius. Society, 152, 1–15. Zoology, 25,1–4. Dowton, M. & Austin, A.D. (2001) Simultaneous analysis of 16S, 28S, Kita, Y. & Ito, M. (2000) Nuclear ribosomal ITS sequences and phy- COI and morphology in the Hymenoptera: Apocrita – evolutionary logeny in East Asian Aconitum subgenus Aconitum (Ranunculaceae), transitions among parasitic wasps. Biological Journal of the Linnean with special reference to extensive polymorphism in individual plants. Society, 74, 87–111. Plant Systematics and Evolution, 225, 1–13. Dupuis, J.R., Roe, A.D. & Sperling, F.A.H. (2012) Multi-locus species Kjer, K.M. (1995) Use of rRNA secondary structure in phylogenetic delimitation in closely related animals and fungi: one marker is not studies to identify homologous positions: an example of alignment enough. Molecular Ecology, 21, 4422–4436. and data presentation from the frogs. Molecular Phylogenetics and Eddy, S.R. (2004) How do RNA folding algorithms work? Nature Evolution, 4, 314–330. Biotechnology, 22, 1457–1458. Klopfstein, S., Kropf, C. & Quicke, D.L.J. (2010) An evaluation of Fernández-Triana, J.L. (2005) The taxonomy and biogeography of phylogenetic informativeness profiles and the molecular phylogeny of Cuban Ophioninae (Hymenoptera: Ichneumonidae). Zootaxa, 1007, Diplazontinae (Hymenoptera, Ichneumonidae). Systematic Biology, 1–60. 59, 226–241.

Klopfstein, S., Quicke, D.J.J., Kropf, C. & Frick, H. (2011) Molecular Quicke, D.L.J., Smith, M.A., Janzen, D.H. et al. (2012) Utility of and morphological phylogeny of Diplazontinae (Hymenoptera, Ich- the DNA barcoding gene fragment for parasitic wasp phylogeny neumonidae). Zoologica Scripta, 40, 379–402. (Hymenoptera: Ichneumonoidea): data release and new mea- Lanfear, R., Calcott, B., Ho, S.Y.W. & Guindon, S. (2012) Partition- sure of taxonomic congruence. Molecular Ecology Resources, Finder: combined selection of partitioning schemes and substitution 12, 676–685. models for phylogenetic analyses. Molecular Biology and Evolution, Raupach, M.J., Astrin, J.J., Hannig, K., Peters, M.K., Stoeckle, M.Y. 29, 1695–1701. & Wägele, J.-W. (2010) Molecular species identification of Cen- Laurenne, N.M., Broad, G.R. & Quicke, D.L.J. (2006) Direct optimiza- tral European ground beetles (Coleoptera: Carabidae) using nuclear tion and multiple alignment of 28S D2-D3 rDNA sequences: prob- rDNA expansion segments and DNA barcodes. Frontiers in Zoology, lems with indels on the way to a molecular phylogeny of the cryptine 7, 26. ichneumon wasps (Insecta: Hymenoptera). Cladistics, 22, 442–473. Rousse, P. & van Noort, S. (2014) Afrotropical Ophioninae Letsch, H.O. & Kjer, K.M. (2011) Potential pitfalls of modelling (Hymenoptera, Ichneumonidae): an update of Gauld and Mitchell’s ribosomal RNA data in phylogenetic tree reconstruction: evidence revision, including two new species and an interactive matrix from case studies in the Metazoa. BMC Evolutionary Biology, 11, 146. identification key. ZooKeys, 456, 59–73. Li, Y., Zhou, X., Feng, G., Hu, H., Niu, L., Hebert, P.D.N. & Huang, Savill, N.J., Hoyle, D.C. & Higgs, P.G. (2001) RNA sequence evolution D. (2010) COI and ITS2 sequences delimit species, reveal cryptic with secondary structure constraints: comparison of substitution taxa and host specificity of fig-associated Sycophila (Hymenoptera, rate models using maximum-likelihood methods. Genetics, 157, Eurytomidae). Molecular Ecology Resources, 10, 31–40. 399–411. Lopez, J.V., Yuhki, N., Masuda, R., Modi, W. & O’Brien, S.J. (1994) Schlötterer, C., Hauser, M.T., von Haeseler, A. & Tautz, D. (1994) Com- Numt, a recent transfer and tandem amplification of mitochondrial parative evolutionary analysis of rDNA ITS regions in Drosophila. DNA to the nuclear genome of the domestic cat. Journal of Molecular Molecular Biology and Evolution, 11, 513–522. Evolution, 39, 174–190. Schultz, J., Maisel, S., Gerlach, D., Müller, T. & Wolf, M. (2005) A Maddison, W.P. & Maddison, D.R. (2011) Mesquite: A Modular System common core of secondary structure of the internal transcribed spacer for Evolutionary Analysis. Version 2.75 [WWW document]. URL 2 (ITS2) throughout the Eukaryota. RNA, 11, 361–364. http://mesquiteproject.org [accessed on 1 February 2011]. Schwarz, G. (1978) Estimating the dimension of a model. The Annals of Mardulyn, P. & Whitfield, J.B. (1999) Phylogenetic signal in the COI, Statistics, 6, 461–464. 16S, and 28S genes for inferring relationships among genera of Schwarzfeld, M.D. & Sperling, F.A.H. (2014) Species delimitation using Microgastrinae (Hymenoptera; Braconidae): evidence of a high diver- morphology, morphometrics, and molecules: definition of the Ophion sification rate in this group of parasitoids. Molecular Phylogenetics scutellaris Thomson species group, with descriptions of six new and Evolution, 12, 282–294. species (Hymenoptera, Ichneumonidae). ZooKeys, 462, 59–114. Marinho, M.A.T., Junqueira, A.C.M. & Azeredo-Espin, A.M.L. (2011) Schwarzfeld, M.D. & Sperling, F.A.H. (2015) Comparison of five meth- Evaluation of the internal transcribed spacer 2 (ITS2) as a molecular ods for delimiting species in Ophion, a diverse genus of parasitoid marker for phylogenetic inference using sequence and secondary wasps (Hymenoptera, Ichneumonidae). Molecular Phylogenetics and structure information in blow flies (Diptera: Calliphoridae). Genetica, Evolution, 93, 234–248. 139, 1189–1207. Seibel, P.N., Müller, T., Dandekar, T., Schultz, J. & Wolf, M. (2006) Marinho, M.A.T., Azeredo-Espin, A.M.L. & Zanchin, N.I.T. (2013) 4Sale – a tool for synchronous RNA sequence and secondary struc- Structural characterization of the internal transcribed spacer 2 (ITS2) ture alignment and editing. BMC Bioinformatics, 7, 498. of the ribosomal DNA (rDNA) cluster in Calyptratae (Diptera: Sharkey, M.J., Carpenter, J.M., Vilhemsen, L. et al. (2012) Phylogenetic Schizophora) and its implications for molecular phylogenetic anal- relationships among superfamilies of Hymenoptera. Cladistics, 28, yses. Journal of Molecular Evolution, 76, 158–171. 80–112. Michel-Salzat, A. & Whitfield, J.B. (2004) Preliminary evolutionary Song, H., Buhay, J.E., Whiting, M.F. & Crandall, K.A. (2008) Many relationships within the parasitoid wasp genus Cotesia (Hymenoptera: species in one: DNA barcoding overestimates the number of species Braconidae: Microgastrinae): combined analysis of four genes. Sys- when nuclear mitochondrial pseudogenes are coamplified. Proceed- tematic Entomology, 29, 371–382. ings of the National Academy of Sciences United States of America, Monaghan, M.T., Balke, N., Gregory, T.R. & Vogler, A.P. (2005) 105, 13486–13491. DNA-based species delineation in tropical beetles using mitochon- Swofford, D.L. (2002) PAUP*. Phylogenetic Analysis Using Parsimony drial and nuclear markers. Philosophical Transactions of the Royal (*and Other Methods). Version 5. Sinauer Associates, Sunderland, Society B, 360, 1925–1933. Massachusetts. Monti, M.M., Nappo, A.G. & Giorgini, M. (2005) Molecular charac- Telford, M.J., Wise, M.J. & Gowri-Shankar, V. (2005) Consideration terization of closely related species in the parasitic genus Encarsia of RNA secondary structure significantly improves likelihood-based (Hymenoptera: Aphelinidae) based on the mitochondrial cytochrome estimates of phylogeny: examples from the bilateria. Molecular oxidase I gene. Bulletin of Entomological Research, 95, 401–408. Biology and Evolution, 22, 1129–1136. Moulton, M.J., Song, H. & Whiting, M.F. (2010) Assessing the effects Tillier, E. & Collins, R. (1995) Neighbor-joining and maximum likeli- of primer specificity on eliminating numt coamplification inDNA hood with RNA sequences: addressing the inter-dependence of sites. barcoding: a case study from Orthoptera (Arthropoda: Insecta). Molecular Biology and Evolution, 12, 7–15. Molecular Ecology Resources, 10, 615–627. Townes, H. (1971) The genera of Ichneumonidae, part 4. Memoirs of the Oosterbroek, P.(1978) Dutch Ophionini (Hymenoptera, Ichneumonidae, American Entomological Institute, 17, 1–372. Ophioninae). Entomologische Berichten, 38, 103–112. Wagener, B., Reineke, A., Lohr, B. & Zebitz, C.P.W. (2006) Phylo- Quicke, D.L.J., Laurenne, N.M., Fitton, M.G. & Broad, G.R. (2009) A genetic study of Diadegma species (Hymenoptera: Ichneumonidae) thousand and one wasps: a 28S rDNA and morphological phylogeny inferred from analysis of mitochondrial and nuclear DNA sequences. of the Ichneumonidae (Insecta Hymenoptera) with an investigation Biological Control, 37, 131–140. into alignment parameter space and elision. Journal of Natural Whitfield, J.B., Mardulyn, P., Austin, A.D. & Dowton, M. (2002) Phy- History, 43, 1305–1421. logenetic relationships among microgastrine braconid wasp genera

based on data from the 16S, COI and 28S genes and morphology. Yu, D.S.K., van Achterberg, C. & Horstmann, K. (2012) Taxapad Systematic Entomology, 27, 337–359. 2012, Ichneumonoidea 2011, Database on USB, Ottawa. [WWW Williams, P.H., Brown, M.J.F., Carolan, J.C. et al. (2012) Unveiling document]. URL www.taxapad.com [accessed on 8 June 2012]. cryptic species of the bumblebee subgenus Bombus s. str. worldwide Zwickl, D.J. (2006) Genetic algorithm approaches for the phylogenetic with COI barcodes (Hymenoptera: Apidae). Systematics and Biodi- analysis of large biological sequence datasets under the maximum versity, 10, 21–56. likelihood criterion. PhD Dissertation, The University of Texas, Wolf, M., Ruderisch, B., Dandekar, T., Schultz, J. & Müller, T. (2008) Austin, Texas. ProfDisS: (profile-) distance based phylogeny on sequence-structure alignments. Bioinformatics, 24, 2401–2402. Accepted 4 September 2015 Young, I. & Coleman, A.W. (2004) The advantages of the ITS2 region First published online 30 October 2015 of the nuclear rDNA cistron for analysis of phylogenetic relationships of insects: a Drosophila example. Molecular Phylogenetics and Evolution, 30, 236–242.