A new species of Ophrella Middlekauff, 1985 (Hymenoptera, Orussidae) from French Guiana

Ophrella seagi sp. n. is described and illustrated from a female and a male collected in French Guiana. Additional Orussidae from this country are listed in an Appendix. The new species differs from other members of Ophrella in a number of features, and the diagnosis of the genus is revised accordingly. A key to Ophrella species is included. The new species is included in a continuously updated morphological data set assembled for the Orussidae and its phylogenetic position discussed.


Introduction
The Orussidae are a small family of parasitoid wasps, currently with about 90 extant described species.They have a worldwide distribution (Vilhelmsen 2003(Vilhelmsen , 2004), but they are very rarely collected, many species being known from only one or a few specimens.Orussidae have a fossil record spanning the Cretaceous and Tertiary, the radiation of the extant members of the family probably occurred in the mid-Cretaceous, approx.100 Ma ago (Vilhelmsen and Zimmermann 2014).Detailed biological information is lacking for most species, but they are generally associated with dead wood, targeting woodboring insect larvae (e.g., Buprestidae, Cerambycidae) and completing their pre-adult development inside the wood (Vilhelmsen et al. 2001).
The genus Ophrella Middlekauff, 1985 is one of the smallest genera of Orussidae and until recently one of the least frequently collected.The first species was described from a single specimen from Panama as Ophrella lingulata by Middlekauff (1985).Smith (1988) transferred Oryssus amazonicus Westwood, 1874 (also based on a single specimen, from Brazil) to Ophrella as Ophrella amazonica (Westwood, 1874), citing the shape of the frontal setae and the position of fore wing cross wing cu-a relative to the discoidal cell as evidence for this taxonomic placement.Vilhelmsen (2003) in his cladistic treatment of the Orussidae confirmed this and added the presence of a longitudinal furrow on top of the head separating the posteriormost coronal teeth as further support for the monophyly of Ophrella.Vilhelmsen et al. (2013) cited the presence of only a single short hind tibial apical spur as additional corroboration for the monophyly of Ophrella.
Ophrella belongs to the ophrynopine clade (Vilhelmsen and Smith 2002), an assemblage currently comprising six genera and 33 extant species.The monophyly of the clade is usually retrieved as well supported in cladistics analyses of Orussidae, prominent putative autapomorphies being the presence of a triangular protrusion on the posteroventral corner of the hind femur (reversed in Kulcania Benson, 1935), the fore wing vein 2r inserting distally on the pterostigma, and the presence of spines or pegs on sternum 9 in the male.More than half the species occur in the Neotropical Region, but ophrynopines are also represented in the Nearctic, Australasian (Australia, New Guinea), Oceanic (Fiji, New Zealand), Oriental (Southeast Asia) and eastern Palaearctic (Japan, South Korea) regions.The ophrynopine clade probably radiated in the Tertiary (Vilhelmsen et al. 2013), but the biographic history of the clade is highly complex and likely involves multiple dispersal events (Vilhelmsen 2004).
The Société Entomologique Antilles Guyane (SEAG; see http://insectafgseag. myspecies.info/en)has for the past decade been inventorying the insect fauna in French Guiana.Various passive and active collecting techniques have been implemented in a series of projects in different localities around the country.These intensive collecting efforts aim, among other things, to produce and update checklists for the country (Brûlé and Touroult 2014).The SEAG efforts have also generated material of otherwise rarely collected insect taxa, including Orussidae (see Appendix and Vilhelmsen et al. 2013).Each of the orussid specimens collected by the SEAG surveys probably represents several thousand trapping hours.Only by an effort of this magnitude is it possible to get a more comprehensive impression of the true diversity of such rare taxa in an area.
Including the material reported in the present paper, about ten specimens of Ophrella species have been collected in French Guiana until now.This has led to substantial changes in the taxonomy of the genus.Vilhelmsen et al. (2013) described Ophrella eldorado Vilhelmsen as new and placed O. lingulata in synonymy with O. amazonica.Here, I describe a new species of Ophrella.The new species differs substantially from the previously recognized species and the diagnosis of the genus is adjusted accordingly.

Material and methods
All the material treated in the present paper (see also Appendix) is stored in the Natural History Museum of Denmark, University of Copenhagen (NHMD).
Specimens were examined with a Leica M205C dissection microscope both before (in ethanol) and after mounting and scored for the characters in the data set presented in Blank and Vilhelmsen (2016).
The following characters were added to the data set of Blank and Vilhelmsen (2016).175.Shape of female antennomere 10: less than three times as long as wide = 0; at least three times as long as wide = 1.
Digital images were produced with a Visionary Digital imaging setup with flash lightning and P-51 Camlift Driver ver.2.6.1 to control the camera.A cylinder of semitransparent plastic and a cone of semitransparent paper were placed around the specimen to disperse the light.Images were stored in Adobe Lightroom 2 and composite images were compiled from stacks with the software Zerene Stacker ver.1.04by implementing the Pyramidal stacking method (PMax).

Phylogenetics
The phylogenetic analyses under various weighting schemes produced variable results, the shortest/most fit trees being retrieved in most in less than 10% of the replications.
The results of the implied weights analyses with concavity constant k set to 7, 10 and 20 are shown in Figs 1-3.In the following, only the results relevant to evaluating the monophyly of Ophrella and its placement within the ophrynopine clade will be discussed.
In all the analyses where the male and female of O. seagi were included as separate terminals, they were retrieved as a monophylum, corroborated by the presence of a prominent transverse carina dorsally on the pronotum with a median incurvation (char.176:2).Ophrella is retrieved as monophyletic in the implied weighting analyses with k settings 1-15, usually with Ophrella seagi as sister to the two other Ophrella species (Figs 1, 2).When k = 1, O. seagi is the sister to O. eldorado; this relationship is supported by the presence of a transverse carina dorsally on the pronotum (char.176:1/2).Ophrella amazonica + O. eldorado is corroborated by the presence of dense pilosity posterior to the eyes (char.23:1) and on the mesoscutum (char.69:1) and having only one short hind tibial apical spur (chars 108:1; 109:1).The monophyly of Ophrella is supported by the presence of a longitudinal groove on the top of the head between the posteriormost coronal teeth (char.7:1), the insertion of fore wing cross vein cu-a on Cu at least 0.3 the length of the discal cell distal to vein M (char.123:1) and antennomere 10 in the female being at least three times longer than wide (char.175:1).When k = 20, 25 or 30, O. seagi is the sister to a monophyletic Ophrynopus Konow, 1897 (Fig. 3), and under equal weights, O. seagi is sister to all Ophrynopus species except Ophrynopus carinatus Vilhelmsen & Smith, 2002.When Ophrella is not monophyletic, O. amazonica + O. eldorado are sister to Argentophrynopus (Fig. 3).
The position of Ophrella within the ophrynopine clade is somewhat unstable.When k = 1, 3, or 7-9 (Fig. 1), it is sister to Ophrynopus.When k = 2, 4-6 or 10-15 (Fig. 2), Ophrella is inside Ophrynopus, being sister to Ophrynopus carinatus; this is corroborated by the absence of the lateral longitudinal frontal carina (char.12:0) and the pronotum being of equal length medially and laterally in dorsal view (char.48:0), but these characters are variable within the ophrynopine clade.O. seagi shares some characters with all or some members of Ophrynopus, e.g., the presence of a mesepisternal carina (char.85:1) and the presence of a projection posteriorly on the male sternum 9 (char.157:2) that are absent in other Ophrella species but has the effect of pulling the entire genus inside Ophrynopus.
Even with the inclusion of the somewhat aberrant O. seagi, Ophrella remains well supported and diagnosable (see below for adjustments).Vilhelmsen et al. (2013) synonymized Stirocorsia Know, 1897 with Ophrynopus, but the latter remains poorly supported and difficult to circumscribe.Some subgroups within Ophrynopus are well defined and at some point it might be useful to subdivide the genus, but based on the findings of this study this is premature.Ophrella seagi can be accommodated in the current generic classification of the Orussidae without necessitating major adjustments.Description.Female.Body length 9.2 mm, fore wing length 5.9 mm.Body predominantly black (Fig. 4A).Small brownish spot situated laterally on frons, just median to lower part of eye (Fig. 4B).Antenna and mouthparts dark brown to black, Antennomeres 9-10 slightly paler (Fig. 4C).Fore femur black, fore tibia and tarsus dark brown; mid leg dark brown to black throughout, except for brown trochanters; hind coxa brown, hind femur laterally with large brown area, remainder of hind leg black (Fig. 5B).Fore wing predominantly heavily infuscated, with dense covering of short coarse dark hairs (Fig. 4A); pterostigma with basal half pale; small hyaline spot situated between basal parts of M+Cu and anal veins, narrowing hyaline band extending from pterostigma proximally of vein 2r to hind margin, apex of wing hyaline; venation predominantly dark brown, except for vein M clear in median hyaline band.Hind wing weakly infuscated in anterior and distal parts, otherwise more or less hyaline, venation dark to light brown (Fig. 4A).
Male.Body length 4.8 mm, fore wing length 3.2 mm.Body uniformly dark brown to black, even more so than female (Fig. 6A).Appendages and mouthparts dark brown to black, legs slightly paler towards apex, hind trochanters light brown.Fore wing infuscated almost throughout, not as heavily as in female, infuscation fades towards apex (Fig. 6C); pterostigma predominantly pale, with brown medial spot in distal half; venation dark brown proximally and anteriorly, light brown distally and posteriorly, vein M proximally hyaline.
Less pilose on top of head and on hind coxa than in female.Mesoscutellar sulcus not interrupted medially.Hind tibia with 22-23 pegs in two rows dorsally.Fore wing vein 2r arises 0.63 from base of pterostigma (Fig. 6C).Sternum 9 with three posteriorly directed spines, one anteromedially and two posterolaterally, sternum 9 terminating in stubby projection (Fig. 6A).
Etymology.Named to acknowledge the contributions of Société Entomologique Antilles Guyane (SEAG) to further the exploration of the diversity of Orussidae in the Neotropics.
Comments.The female and male that have been assigned to Ophrella seagi were collected in the same locality, albeit almost four months apart.There are some differences between the two specimens in the coloration of the body and appendages, and in the degree and pattern of infuscation of the fore wing (compare Figs 4-6), but this is within the degree of variation observed in other known species of Orussidae, especially between sexes (e.g., Vilhelmsen andSmith 2002, Blank et al. 2010).
Ophrella seagi has a unique combination of characters that differs somewhat from the other members of Ophrella.The generic placement is based on the presence of a median longitudinal furrow between the posteriormost coronal teeth (Figs 4B, 6B; less developed in O. seagi than in O. amazonica and O. eldorado), the presence of an elongate antennomere 10 (at least three times as long as broad; Fig. 4C), and the position of the fore wing vein cu-a on Cu1 some distance from vein M (Figs 4A, 6C); all these characters are unique within the ophrynopine clade.Previously diagnostic features suggested for Ophrella, e.g., the presence of flattened, leaf-shaped setae (Middlekauff 1985) and the presence of only one hind tibial apical spur (Vilhelmsen et al. 2013) are not observed in O. seagi and cannot be upheld as potential autapomorphies for Ophrella.Nevertheless, the monophyly of the genus, including O. seagi, is well supported, and it is still possible to identify O. seagi correctly to Ophrella in the genus key in Vilhelmsen et al. (2013).
Ophrella seagi is a very distinct species, especially when compared to the other two species in Ophrella.The most distinctive feature is the prominent, medially subdivided transverse carina on the dorsal part of the pronotum (Figs 4C, 5A, 6A); O. eldorado also has a transverse carina, but it is less developed and not subdivided medially (Vilhelmsen et al. 2013, fig. 7d).Like many other morphological features observed in Hymenoptera pupating in wood, the carina might help the wasp escaping from the wood after eclosion (see Vilhelmsen and Turrisi 2011), probably acting as a brace when the wasp is digging its escape tunnel with the mandibles.A possible analogue occurs in several species of Aulacidae, another family of woodliving parasitoid wasps.Some species of Pristaulacus Kieffer, 1900 have a prominent, medially interrupted transverse crest anteriorly on the mesoscutum (Turrisi and Vilhelmsen 2010, fig. 14).Topologically it is in a similar position, i.e., anterodorsally on the thorax, indicating a similar function; morphologically it is developed on a different part (mesoscutum in the Aulacidae, pronotum in Orussidae), perhaps because the pronotum is weakly developed medially in most Aulacidae (Turrisi et al. 2009;char. 25).

Figure 1 .
Figure 1.Consensus tree of 9 trees of fit 51,39931 produced by implied weighting analysis with k = 7.Only crown group Orussidae shown; genera outside the ophrynopine clade have been collapsed to single terminals.

Figure 2 .
Figure 2. Consensus tree of 9 trees of fit 41,51167 produced by implied weighting analysis with k = 10.Only crown group Orussidae shown; genera outside the ophrynopine clade have been collapsed to single terminals.

Figure 3 .
Figure 3. Consensus tree of 9 trees of fit 25,58786 produced by implied weighting analysis with k = 20.Only crown group Orussidae shown; genera outside the ophrynopine clade have been collapsed to single terminals.