Institutional abbreviations:

NHMUK Natural History Museum, London, UK;

ZISP Zoological Institute of the Russian Academy of Sciences, St Petersburg, Russia;

NHRS Swedish Museum of Natural History, Stockholm, Sweden.

The English specimen (NHMUK014425411) was collected at a white sheet with a 15W actinic bulb. The right hind leg was removed from the dry, card-pointed specimen for sequencing. Photos were taken with a Canon SLR EOS 5DSR with either a 65 mm macro lens or a Mitutoyo 10 × lens in combination with a 70–130 mm macro lens, mounted on a copy stand with an automated Z-stepper; images were aligned using Helicon Focus software version 6.6.1.

Morphological terminology follows van Achterberg (1993).

DNA was extracted from the hind leg using the Thermo Labsystems KingFisher extraction robot at the NHRS laboratory facility. The standard DNA barcode fragment of the mitochondrial cytochrome oxidase I (CO1) gene was obtained using primers by Folmer et al. (1994). Ready-To-Go PCR beads were used (Amersham Pharmacia Biotech, Amersham, UK) on the following program: 5 min 94 °C hot- start; 40 cycles: denature 94 °C for 15 s, anneal 46 °C for 15 s, extend 72 °C for 15 s; final extension 72 °C for 10 min. This gene has been used in previous studies of braconid phylogenetics (Belshaw et al. 2000; Belshaw and Quicke 2002; Dowton et al. 2002; Zaldívar-Riverón et al. 2006; Sharanowski et al. 2011; Stigenberg and Ronquist 2011), but latterly mainly in combination with other, particular nuclear, genes. In addition to the CO1 gene we also utilised the nuclear 28S D2 fragment (see Stigenberg and Ronquist 2011). PCR products were purified with EXO1 and FastAP. The product was sequenced using both the forward and reverse primers. Sequences were assembled and edited using Geneious Pro v.9.1.8. The Voseq v.1.7.4 (Peña and Malm 2012) database was used for storing voucher and DNA sequence data. BLAST analysis was performed to identify similar sequences within the NCBI database.

Bayesian Inference analyses were performed using MrBayes 3.2.7a (Ronquist et al. 2012). All BI analyses were performed through the Cipres web portal for phylogenetic analysis (Miller et al. 2010) (http://www.phylo.org). The protein coding CO1 was partitioned after codon positions 1+2 and 3, the 28S dataset was not partitioned. The Bayesian analysis was performed by running four MCMC chains for 10 million generations, sampling post-burn-in trees every 1000 generations. A CO1 analysis including 60 taxa as well as a combined CO1+28S analysis including 65 taxa were run. The datasets included an assortment of members of the Euphorinae tribe Perilitini with an ichneumonid and distant as well as closer braconids as outgroups. GenBank accession numbers for the Orionis species are shown in Table 1.

Bortoni et al. (2016) revised the species of Orionis known to them, which included four Neotropical species and one from Thailand. Including the species transferred here to Orionis, the genus is now represented by equal numbers of described species in the Old and New Worlds. None has a recorded host.

While this work was in preparation, van Achterberg et al. (2020) reported the occurrence of a very similar specimen in Amsterdam, the Netherlands, which they identified as Perilitus erratus. Subsequent posts on Facebook showed that similar specimens have been found in Belgium and Germany recently and a further UK specimen came to light on Facebook (England, Hampshire, New Forest, October 2021; Colin Easton). We strongly suspect that all these specimens belong to the same species and that, as van Achterberg et al. (2020) surmised, this is a recent introduction from the Far East. However, we offer a different perspective on the identification. The Dutch specimen illustrated by van Achterberg et al. (2020) has curved dorsal striae on the hind coxa, identifying this as Orionis coxator. Given that two of the authors also described Perilitus erratus as a new species (Chen and van Achterberg 1997), perhaps the hind coxa is variable within the species and these names are synonymous. We have not examined the holotypes of Perilitus erratus or Perilitus coxator and are relying on the (very informative) original descriptions; should these names prove to be synonymous, Orionis coxator has priority.

Whether Orionis coxator is native to Europe is probably unknowable, although circumstantial evidence, mainly that these distinctive wasps have never been found in Europe before, strongly suggests that there has been an accidental introduction and perhaps a rapid range expansion within Europe. However, other ichneumonoids have been discovered in Europe following their description in the Russian Far East and these massive range discontinuities have been assumed to represent spot samples from an extensive and undocumented range; for example, the diplazontine ichneumonid, Episemura diodon Kasparyan & Manukyan, 1987 (Klopfstein 2014). The distribution of Orionis species is rather similar to that of another ichneumonoid genus that was poorly known until recently; Rodrigama Gauld, 1991 (Ichneumonidae, Poemeniinae) comprises a few known species, with the type species described from Costa Rica (Gauld 1991) then additional species being discovered in the Oriental and Eastern Palaearctic regions (Matsumoto and Broad 2011; Choi and Lee 2020) and one species in the Western Palaearctic (Broad and Kuslitzky 2019). There are similar examples of apparently huge range discontinuities in Braconidae, for example Proclithrophorus Tobias & Belokobylskij, 1981, with the type species (Proclithrophorus mandibularis Tobias & Belokobylskij, 1981) described from the southern Russian Far East (Tobias and Belokobylskij 1981) and a second species (P. genalis Vikberg & Koponen, 2001), of very similar morphology, subsequently described from Finland (Vikberg & Koponen 2001). Or Colastes (Pseudophanomeris) pilosus Belokobylskij, 1984, known from the Palaearctic Far East, recently being found in Ukraine and the Czech Republic (Belokobylskij 2019). Some parasitoid species suddenly become numerous and then suddenly disappear again. A good example in Britain is Phrudus badensis Hilpert, 1987, described from Germany (Hilpert 1987), collected in Britain in good numbers (Shaw 1991) and then apparently disappearing again (M.R. Shaw pers. comm.). Phrudus badensis closely resembles Phrudus longius Chiu, 1987, described from Taiwan (Chiu and Wong 1987). The lack of known hosts for most of these ichneumonoids hinders interpretation of their distribution.

Orionis eximius is associated with unknown hosts on Lantana camara L. (Verbenaceae) (Shaw 1987). Lantana camara does not survive outdoors in the cool temperate climate of NW Europe, although other ornamental Verbenaceae do. The as yet unknown hosts of Orionis are presumably Coleoptera, as with other Perilitini (Shaw 1985; Stigenberg et al. 2015). The large eyes of Orionis coxator, and the fact some specimens have been caught at light (as have some O. flavifacies; Belokobylskij 2000a), suggest it is at least partly nocturnal.

Orionis is diagnosed by a combination of characters, more than one of which can be present in species classified in Perilitus. For example, Perilitus mylloceri (Wilkinson, 1929) (originally described in Dinocampus Förster, 1863) shares a narrow, ventrally fused first metasomal segment, strongly convex propodeum (Fig. 4) and bulging, ventrally convergent eyes with Orionis. However, we consider this species best placed in Perilitus, as the spiracles of the first metasomal tergite are at the posterior third and this segment is virtually straight until the postpetiole. An unidentified or undescribed African species of Perilitus in NHMUK has a strongly angled propodeum, very similar to some Orionis, and ventrally convergent eyes (Fig. 5). In this case, the first metasomal segment is more typical of Perilitus, being only 2.5 × as long as posteriorly wide, with the spiracles at the posterior third of the tergite and the sides of the first tergite ventrally distinctly separated. Nevertheless, Perilitus as currently defined is probably not monophyletic (Stigenberg et al. 2015; Suppl. material 1, 2) and some of these species might be better classified elsewhere, perhaps in an expanded concept of Orionis. Figure 6 compares the propodeum and first metasomal segment of O. coxator and Perilitus rutilus (Nees, 1811), the type species of Perilitus.

Figure 4. 

Perilitus mylloceri paratype female, propodeum and first metasomal segment.

Figure 5. 

Unidentified Perilitus species from Cameroon (NHMUK014425412). Scale bar: 1 mm.

Figure 6. 

Comparison of propodeum and first metasomal segment of a Orionis coxator and b Perilitus rutilus.

Chen and van Achterberg (1997) described erratus in the genus Meteorus, despite the lack of vein rs-m (i.e., the second submarginal cell is open). GRB’s first thought was that the English specimen of O. coxator could be an aberrant Meteorus. The structure of the first metasomal segment and the strongly convergent eyes are both more similar to some Meteorus species than they are to Perilitus. The first metasomal segment of these Old World species is not as long and slender as in the type species of Orionis, O. eximius (Muesebeck), which has hindered the correct identification of the Old World species. Bortoni et al. (2016) had recognised Orionis as occurring in the Old World, with the description of O. orientalis from Thailand, and pointed out the close relationship to certain ‘Perilitus’ species in Stigenberg et al.’s (2015) phylogeny. It took a chance discovery of O. coxator on a different continent to join the dots and more accurately describe the diversity of Orionis species.