A new megaspilid wasp from Eocene Baltic amber (Hymenoptera: Ceraphronoidea), with notes on two non-ceraphronoid families: Radiophronidae and Stigmaphronidae

1. Compact mesosoma.

In Ceraphronoidea, the pronotum, mesopectus, metapectus, first abdominal tergum, and the metanotum comprise a single, compact sclerite (Fig. 1A). This modification is only found in some wingless hymenopterans, as the presence of the conjunctivae that allows mobility between the above-mentioned mesosomal regions is less important in flightless hymenopterans (Reid, 1941; Keller, Peeters & Beldade, 2014). The functional consequences of mesosomal compactness in ceraphronoids have never been explored. Most ceraphronoids are able to fly; therefore, mesosomal compactness most likely evolved for a different reason than that of flightless hymenopterans with a similarly compact mesosoma.

2. Orientation of the toruli and antennal bases.

The ventral position of the antennal insertion sites (toruli) is a well-known characteristic of Ceraphronoidea (Masner, 1993) and it is shared by some other apocritan taxa (e.g., Megalyroidea and Platygastroidea; Sharkey et al. (2012)). The orientations of the torular surface and the antennal base are, however, often overlooked traits of the superfamily. In Ceraphronoidea, the lateral torular margin is elevated relative to the median margin and therefore the antennal foramina, and the antennal scapes are oriented medially in a resting position (Figs. 1B, 1C).

3. Articulation between pronotum and mesoscutum.

In Ceraphronoidea, the pronotum and the mesoscutum have a unique, ball-and-socket type articulation that corresponds to sharp anterolateral edges on both the mesoscutum and pronotum (Fig. 1A) (Mikó & Deans, 2009). The notauli arise from these articulation (Figs. 2A, 2B) in Ceraphronoidea. While distinct anterolateral edges on the mesoscutum are present in Megalyridae, they never correspond with ball-and-socket articulations.

4. Wing venation.

All winged Ceraphronoidea have a stigmal vein that originates from the pterostigma or from the distal portion of the marginal vein posterior to the costal notch (Masner, 1993). Along with this, ceraphronoids have a single vein extending along the anterior margin of the fore wing. This vein is equipped with unique triangular elements (Figs. 1D, 1E), whose functions and origins remain unknown. A single vein on the anterior wing margin is present in numerous other hymenopterans (e.g., aphidiine braconids) but the triangular elements are seemingly specific to Ceraphronoidea (I Mikó, pers. obs., 2018). In many Chrysidoidea the two wing veins on the anterior wing margin (costal and subcostal veins) are adjacent and may superficially look “fused”, but they are always separated by a faint line (Olmi, 1994; Richards, 1939).

The two type specimens of the present paper are embedded in an approximately 33–55 million-year-old (Gillung & Winterton, 2018; Larsson, 1978; Penney, 2010; Sadowski et al., 2017; Wolfe et al., 2016), somewhat cloudy piece of Baltic amber (Figs. 3C, 3D and 4A–4D). These specimens belong in Ceraphronoidea based on the the lyre-shaped notaulus and the orientation of the torulus and antenna (traits that are not obscured). As ceraphronoid experts, we examined the specimens with a light microscope at 230× magnification, and based on their general habitus (Gestalt) and the presence of a bifurcated anteromedian projection of the metanoto-propodeo-metapectal complex (Fig. 4A) we first classified them as Megaspilus, a genus that has never been recorded from any fossil deposit. However, after looking at the high resolution 3D data (see File S1), we were able to observe otherwise-obscured traits and determine that the new species actually belongs in Conostigmus, as part of a species group that shares some key features with Megaspilus.

While small body size, reduced wing venation, and the presence of a pterostigma are often used to determine specimens as Ceraphronoidea, these characters occur in almost all hymenopteran superfamilies. Despite the above-listed clear and obvious ceraphronoid synapomorphies, it is common to find smaller Ichneumonoidea and Aculeata, especially Bethylidae and Dryinidae, misidentified as Ceraphronoidea in collections. Perhaps the most commonly misidentified are small specimens of the dryinid genus Aphelopus (Fig. 2C). One factor that may contribute to these misidentifications is the use of low-power microscopes for sorting and identification of specimens. Even the most distinct external traits specific to ceraphronoids are obscure with lower magnification and inadequate lighting. In such poor conditions, only those who have trained their eyes by looking at hundreds of ceraphronoid specimens are able to identify these taxa correctly. In some cases, the examination of fossil specimens is similar to studying specimens of recent taxa with a low quality microscope. Morphological traits are usually obscured by artifacts or debris, making them difficult to properly observe (if they can be seen at all).

The discovery of these Conostigmus specimens encouraged us to review fossil ceraphronoids and revise the classification of two hymenopteran families that are exclusively represented by fossils and currently classified in Ceraphronoidea. Radiophronidae and Stigmaphronidae are minute, winged wasps that have a distinct pterostigma and, in some cases, two protibial spurs. Besides these traits, however, these two families do not share any other characteristics with Ceraphronoidea. Here, we remove them from the superfamily, leaving them incertae sedis, and provide some guidance for their future classification.

Conostigmus

The new species belongs in Conostigmus based on the presence of a distinct sternaulus (ste: Fig. 5A), a shorter posterior ocellar line (POL) than ocular-ocellar line (OOL) (Figs. 1C, 5B, 5C) and the presence of a preoccipital carina (Figs. 1C, 5A–5C). These traits are absent from Dendrocerus, Trichosteresis, and Platyceraphron; clavate female antenna and F1 length (proximodistal anatomical line) shorter than the combined length of F2 and F3 are traits that are present in Conostigmus and absent from Megaspilus.

Locality of type specimens

Gulf of Gdańsk (Baltic amber)

Etymology

The new species is named after Elijah Talamas (Florida State Collection of Arthropods), who drew our attention to these unique fossils.

Megaspilus vs. Conostigmus

Based on the latest phylogenetic analysis (Mikó et al., 2013), and preliminary phylogenomic data from an ongoing molecular study using ultra-conserved elements (UCEs) (B Blaimer, pers. comm., 2018) Conostigmus is polyphyletic and includes Megaspilus. Until Dessart’s revisions (1972; 1981) of Nearctic and Palaearctic species, Megaspilus was a broad taxonomic concept that essentially included all larger megaspiline species with an acute ocellar triangle (Conostigmus-type, in contrast to a Dendrocerus-type obtuse ocellar triangle), well-defined sternaulus, distinct posterior orbital carina and/or some rugulose sculpture on the frons. Dessart (1972) narrowed the generic concept of Megaspilus to include only those species that have an acute ocellar triangle, sternaulus, bifurcated anteromedian projection, and elongate female first flagellomere. The ocellar triangule shape and presence of the sternaulus are shared with numerous Conostigmus species, leaving the bifurcated anteromedian projection and the elongate female proximal flagellomeres as diagnostic features for Megaspilus.

The bifurcated anteromedian projection of Megaspilus is clearly derived from the elevated and medially-projected lateral propodeal carinae. This bifurcated condition can be found in certain Ceraphron and Dendrocerus (Alekseev, 1978) species, and so far has never been reported from any Conostigmus (Mikó et al. 2016; Dessart 1997; CT personal communictaion). The median portion of the lateral propodeal carina is elevated and forms a bifurcated projection in C. talamasi that is smaller than that of Megaspilus and most likely represents an intermediate state.

With the discovery of C. talamasi, the only diagnostic characters for separating Megaspilus from Conostigmus remain the elongate proximal female flagellomeres. The first female flagellomere is more than two times as long as the pedicel in Megaspilus, while in other megaspilids, including Conostigmus, it is less than 1.5 times as long as the pedicel. The length of the male and female flagellomeres correlate to each other and are important in species-level diagnosis in Megaspilinae (Mikó et al., 2016). The correlation of flagellomere length between different sexes might be related to their courtship behavior as males extensively antennate and repeatedly touch the female antenna during mating (Liebscher, 1972). Consequently, the first male flagellomere is much longer than the scape in Megaspilus in contrast with other male megaspilids where the first flagellomere is either shorter or slightly longer than the scape (Mikó et al., 2016; Dessart, 1972; Dessart, 1981; Dessart, 1974; Dessart, 1995; Dessart, 1997; Dessart, 1999; Dessart, 2001).

Ceraphronoid and non-ceraphronoid fossils

To date, there are 18 fossil specimens that share key characteristics with recent Ceraphronoidea (see Fig. S1), including the compact mesosoma, lyre-shaped notaulus, and the orientation of the antennal bases. Based on their visible morphological characters, these fossils represent (mostly) megaspilid wasps from the Late Cretaceous Santonian to the Early Miocene. There is also a single ceraphronid specimen from the Early Miocene (Fig. S1).

Wing venation characters are perhaps the most distinctive features shared between these fossils and recent ceraphronoids. In these taxa, a single wing vein is present along the anterior fore wing margin and the stigmal vein is never angled (an angle is present in numerous other taxa in the stigmal vein, marking its intersection with 2RS or r-m veins Mikó et al. (2014)), never tangential to the pterostigma and arising from or anterior to the midpoint of the pterostigma (Fig. 1D, Masner 1993). A single fore wing vein can be found in numerous hymenopteran families, but unlike in these fossils, it is always well separated from the anterior margin.

The remaining fossil hymenopterans currently classified as ceraphronoids (Fig. S1) lack key ceraphronoid characteristics listed in the introduction. Although they do each possess a pterostigma, the remaining wing venation traits are inconsistent with the superfamily. These taxa might not even be closely related to Ceraphronoidea.

Is Ceraphronoidea the most structurally diverse hymenopteran superfamily?

Superfamilies, among the highest taxonomic ranks (family-level) that are regulated by the ICZN (1999), serve as important taxa for communicating about Hymenoptera evolution (see Sharkey et al., 2012). They represent the highest functional and pragmatic taxa, defined, in part, by their natural history and a set of distinct morphological characteristics. It is critical, therefore, that they remain monophyletic and free from extraneous, unrelated taxa.

Recent and fossil Ceraphronidae and Megaspilidae, including C. talamasi, comprise a monophyletic group that can be clearly defined using numerous apomorphic traits (Masner, 1993). The inclusion of Stigmaphronidae, Radiophronidae and, until recently, the stephanoid Aptenoperissidae and the trigonaloid Maimetshidae (Zhang et al., 2018; Perrichot et al., 2011) within Ceraphronoidea, however, results in a polyphyletic morass that cannot be confidently diagnosed. The resulting taxon would have a variable number of protibial spurs and flagellomeres, the presence and absence of a wasp waist, tarsal plantulae, and cenchri, and either a compact or unabridged mesosoma, i.e., traits that are characteristic of other, well defined superfamilies. Ceraphronoidea has clearly been treated as a waste bin for minute fossil taxa with a pterostigma. Based on the characteristics of those fossils we remove these taxa from Ceraphronoidea.