Baltorussus Total Makeover: Rejuvenation and Sex Change in an Ancient Parasitoid Wasp Lineage

The Orussidae is a small and rare but phylogenetically important family of parasitoid wasps. The fossil record of the family is also very poor. Baltorussus velteni was described from Baltic amber from an allegedly female specimen. This and another recently discovered specimen are examined with microCT scanning and standard microscopy. We reveal that both the holotype and the new specimen are actually males. Furthermore, the results of the microCT scanning allow us to integrate the fossils in a morphological data set assembled for extant Orussidae. Phylogenetic analyses consistently retrieve Baltorussus as a separate basal lineage within the crown group, whereas two Cretaceous fossils are placed as stem group orussids and a Dominican amber fossil in an extant genus. Based on the positions of the fossils, we estimate that the extant Orussidae radiated in the mid-Cretaceous (approx. 100 Ma ago). This is considerably younger than a previously suggested Early Jurassic date (180 Ma ago), which was primarily based on biogeographic evidence.


Introduction
Fossils have traditionally played a crucial role when interpreting the evolutionary history of various groups of organisms. Vertebrate fossils (e.g., the proto-bird Archaeopteryx) have figured prominently in phylogenetic hypotheses ever since the publication of Darwin's The Origin of Species promoted broader acceptance of the occurrence of evolution in organisms over deep time. In contrast, it is only comparatively recently that sufficient information has accumulated to fully realize the potential of fossils to interpret insect evolution [1]. However, with the implementation of improved methods for handling fossils in phylogenetic analyses [2], fossils are now routinely employed to calibrate age estimates of splitting events within insect phylogenies. In this paper, we demonstrate how the incorporation of a recently described fossil in the phylogeny of a family of little known parasitoid wasps, the Orussidae, leads to drastic re-evaluation of the timing of its diversification.
The Orussidae is an obscure group of parasitoid wasps that nonetheless occupy a pivotal position in the early evolutionary history of the Hymenoptera, one of the most diverse and economically important insect orders. In most recent comprehensive treatments of hymenopteran phylogeny [3,4], the Orussidae are placed as the sister group to the Apocrita, the latter comprising the majority of the Hymenoptera including such well known groups as ants, bees and hornets as well as a multitude of parasitoid lineages. The Orussidae are parasitoids of woodboring insect larvae, typically of jewel beetles or longhorn beetles. This is a lifestyle that is probably very similar to that of the common ancestor of all carnivorous wasps, estimated to have lived 200+ Ma ago [2]. Orussidae is thus a crucial group when interpreting the transition from the herbivorous to the carnivorous lifestyle early in hymenopteran evolution [5]. Furthermore, the female orussids display remarkable adaptations to targeting host larvae living inside wood. The female apparently operates a vibrational sounding system to detect potential hosts. It generates vibrations in the wood by tapping it with the modified tips of the antennae; the vibrations are then picked up by the elongate basitarsus (the proximal part of the 'foot') of the fore leg and conveyed to an enlarged sensory unit, the subgenual organ, inside the swollen and subdivided tibia, where they are transduced into nerve impulses [6].
The Orussidae is a small taxon, currently with approx. 90 described extant species [7]. They occur in all major biogeographic regions, but are rarely encountered. Their fossil record is correspondingly poor, a total of four fossil taxa being unequivocally assigned to the family, all described from amber inclusions (age estimates from [1]): Mesorussus taimyrensis Rasnitsyn, 1977 Engel, 2008 ([11]; Dominican amber, 17-20 Ma). Vilhelmsen [12] analyzed the two Cretaceous fossils with the extant Orussidae. The fossils were invariably placed in the stem group of the family, although the relative positions of the fossils were unresolved. In contrast, Ophrynopus peritus was described in an extant genus [11]; this has been corroborated by later analyses [7]. Based on the placement of the Cretaceous fossils and in particular the biogeographic patterns within the Orussidae, it was concluded that the earliest splitting events within the extant lineages of the family probably took place at least 180 Ma ago [12]. Some Upper Jurassic compression fossils from the Karatau Formation, Kazakhstan (152-158 Ma) collectively referred to as the Paroryssidae have been considered to be 'ancestral' to the Orussidae [13,14]. These fossils share some wing reduction characters with extant orussids (and to some extent Apocrita), but they miss some of the characteristics of the Orussidae. Paroryssidae do not have an ocellar corona (a circle of small cuticular teeth on the top of the head) and they have a comparatively long external ovipositor whereas that of extant Orussidae is elongate but entirely concealed within the body when not in use [6]. The analyses of Ronquist et al. [2] included two paroryssid taxa, Praeoryssus and Paroryssus, as well as Mesorussus; these fossil taxa were retrieved as successive sister groups to Orussus, the only extant orussid included in the analyses.
Baltorussus velteni is the first known orussid from Baltic amber. It was recently described from a single, allegedly female specimen [10]. However, the description was done without reference to recent publications on orussid systematics [12,15], and no attempt was made to place the fossil in a phylogenetic context. Furthermore, the investigation of the fossil was carried out with only a dissection microscope. The holotype of Baltorussus is partly obscured by a whitish film covering most of the body and by cracks in the amber matrix surrounding the inclusion (Fig. 1).
Since the original description, a second specimen of Orussidae from Baltic amber has been discovered which we also identify as Baltorussus velteni. This specimen is better preserved (Fig. 2) and we use this and the more detailed investigation of the holotype to expand the original description and re-interpret some of the features previously reported. Furthermore, by combining observations from standard microscopy and micro-CT we are able to score enough characters for the data set initially assembled for extant Orussidae [15] to confidently place Baltorussus within the phylogeny of the family. Finally, we discuss the implications that the placement of Baltorussus has for the interpretation of the character evolution, biogeographic history and timing of the radiation of Orussidae.

Ethics statement
No permits were required for the described study, which complied with all relevant regulations.

Microscopy and imaging
The amber specimens were initially examined with a Leica M205C dissection microscope. They were imaged while being immersed in maple syrup. Images were captured with a Visionary Digital imaging setup with flash lighting and P-51 Camlift Driver v.2.6.1 to control the camera. Images were stored in Adobe Lightroom 2 and composite images were compiled from stacks with the software Zerene Stacker v. 1.04 by implementing the Pyramidal stacking method (PMax).

MicroCT-scanning
The holotype of Baltorussus velteni was imaged with an Xradia MicroXCT x-ray microtomography system (University of Vienna, Department of Theoretical Biology). The microCT data were reconstructed with 262 pixel binning to reduce noise and file size, and reconstructed volume images were exported as a PNG image stack which is available for download at Morphosource (http://  Evolutionary History of Orussidae morphosource.org) as media file 2371. The software Amira 5.4.3 was used for 3D-visualization and analysis of the data. For the surface renderings (e.g., Fig. 3B) the image stack was cropped with the crop editor to eliminate the surface of the amber piece as far as possible. Additionally the obliqueSlice function was used to blank disturbing surfaces temporarily. The surface renderings were generated with the isosurface function. For the volume renderings (e.g., male genitalia, see below) the area of the internal genitalia was segmented. A new volume dataset was generated from the original data and the label data by using the arithmetic function [16]. Subsequently the voltex function was used to visualize the structures. For the reconstructions (e.g., Fig. 3C, D) the structures were segmented manually with the brush tool and visualized with the surfaceGen and surfaceView functions.

Phylogenetic/dating analyses
The two Baltorussus specimens were scored for the morphological character set for Orussidae presented in [7], an update of the data matrix from Vilhelmsen [15]. The dataset was assembled in Mesquite [17]. Of the 169 characters included, Baltorussus could be scored for 98; eight were inapplicable and 63 had to be scored as unknown as the character in question could not be observed either because it was obscured by other body parts or because it was sex specific. In total, 58% (63% if inapplicable characters are included) of the characters could be scored for Baltorussus, a high percentage for a fossil hymenopteran taxon (compare with table 2 in [2]; this was a different data set which included a number of characters dealing with internal anatomy not observable in fossils). The data set is available from FigShare (http://dx.doi.org/10.6084/m9. figshare.951964). As was the case for Vilhelmsen [12], we did not include the Paroryssidae in the data set. Unlike the amber fossils included, Paroryssidae are comparatively poorly preserved compression fossils that can only be scored for very few of the characters included here, and it would be difficult to place them confidently in the phylogeny.
The data set was analyzed in TNT 1.  [19], the concavity constant K was set in turn to 3, 6, 8 and 10-20. For each weighting scheme, traditional analyses with 10 000 replications and TBR saving 100 trees per replication were Evolutionary History of Orussidae conducted. All analyses were run with collapsing rule 1. The root was Urocerus. To calculate the Bremer support values [20] suboptimal trees up to 15 steps longer were used (per step 10.000 replications and TBR saving 10 trees/rep.; time-out 15 min). Symmetric resampling [21] was performed with 10.000 replications and is displayed as Frequency and GC values.
It was also attempted to do a dating analysis in MrBayes 3.2 [22]. However, unlike the test case for total evidence dating [2], the current data set for Orussidae consists entirely of morphological characters. Even if the fossils included could be placed with reasonable confidence, it was not possible to get reliable age estimates; the variation for some of the deeper nodes was in the range of 300+ Ma in some analyses. There are too few fossils relative to the extant taxa and they are not dispersed evenly across the phylogeny, i.e., three out of four fossils are at or very close to the base of extant orussids, only one is placed more distally in the phylogeny. For these reasons, we do not consider the dating results any further.

Redescription of Baltorussus velteni Schedl, 2011
Body length 8.9 mm (holotype, Fig. 1), 7.8 mm (non-type, Fig. 2). Holotype covered with film partly obscuring sculpture and setation, making body colour appear whitish green; wing and wing venation appear similarly pale whitish, except for darkened area in fore wing distal to pterostigma; many cracks around specimen, e.g., around frons and along antennae and fore leg obscuring some features. Non-type also with film imparting golden sheen; wing venation dark, fore wing area distal to pterostigma maybe slightly infuscate; distal parts of fore and mid legs cut off through tibia during trimming of amber piece.
Wings. Fore wing vein 1r arises just over halfway from base of pterostigma (Fig. 8); vein 1r-Rs well developed, elongate, discal cell rectangular, proximal part not broader than distal part; vein cu-a inserts on Cu1 slightly distal to M. Hind wing venation cannot be observed.
Abdomen. Most of dorsal part of abdomen covered by wings in both specimens. Sculpture of abdomen obscured by artifactual film. No spiracle observed on tergum 8. Sternum 9 triangular in ventral view, with median longitudinal carina bordered by narrow groves extending for L of distance to apex (Fig. 9A, B); apex with prominent truncated projection extending beyond tip of abdomen; sternum 9 otherwise devoid of spines and tubercles. Male genitalia not visible externally. Cupula, gonoforceps and volsella closely integrated (Fig. 9C-E). Cupula strongly emarginate anterodorsally, thin and continuous ventrally, with well developed gonocondyle (apodeme) anteromedially. Gonoforceps grooved and distended anteroventrally, shorter than volsella. Volsella posteriorly differentiated into digitus and cuspis, contiguous to posterior margin of S9; paired penisvalvae approximately as long as gonoforceps.
Female. Not known.

Diagnosis
Baltorussus velteni is characterized by the following unique traits that have not been observed in any other living or extinct Orussidae: 1) frons with distinct sculpture in the middle, i.e., a column of transverse dorsally arched carinae forming washboardlike structure (Figs 3A, B, 4A); 2) mesoscutum with well developed mesoscutal sulcus extending from pronotum to transscutal articulation ( Fig. 7; some specimens of the extant species Orussella dentifrons (Philippi) have a sulcus developed for a short distance anteriorly, but never all the way back to the transscutal articulation); 3) the male sternum 9 with a median longitudinal carina extending for most of its length and distally terminating in a  Evolutionary History of Orussidae truncated projection (Fig. 9A, B). A similar projection is present in a number of genera within the 'ophrynopine' clade, see [7,23]. However, these taxa never have the median longitudinal carina and in addition to the posterior projection always have anteromedial and lateral spines or tubercles that are absent from Baltorussus.

Results of phylogenetic analyses
The analysis under equal weights yielded 20.503 trees with a length of 859 steps, a consistency index (CI) of 0.25, and a retention index (RI) of 0.766 (Fig. 10). The analyses under implied weighting with K 10-20 consistently produced three trees with a length of 860 and a relative fit ( = adjusted homoplasy) of 38.89 (Fig. 11). These trees differ only in the relationships within Orussus. In all analyses under equal and implied weights Baltorussus is placed as a separate lineage being sister to a large clade comprising all extant genera except Orussonia and Orussella. This position is moderately well supported under equal weights (Fig. 10) and well supported when homoplasious characters are down-weighted through implied weighting (Fig. 11).

Reinterpretation of the sex and head sculpture
The holotype of Baltorussus velteni was originally identified as a female, based on the perceived ovipositor apex protruding from the tip of the abdomen (a common artifact also in pinned extant specimens) and it was stated that the antennae 'probably' had ten segments and that the apical one was strongly tapered [10]. These statements fit well with the specimen being a female. However, they differ strongly from our observations in several points: a. The number of antennomeres is in fact 11. The distalmost antennomere is not reduced in size, but of similar width and longer than the penultimate antennomere (compare Fig. 5A with figs 1A-D in [6]), i.e., the antenna is not modified for tapping wood. b. The fore tibia (not observed in [10]) is not swollen and subdivided by a transverse furrow, and the tarsus is fivesegmented and does not have an elongate basitarsomere with a projecting spur distally (compare Fig. 5B with figs 2A, C in [6]), i.e., the fore leg is not modified for receiving and processing vibrations reflected from the wood. c. We did not observe a functional spiracle on abdominal tergum 8. Female orussids have functional spiracles on terga 1 and 8, males only on tergum 1. d. We cannot agree that the projection at the apex of the abdomen is the tip of an ovipositor. It is clearly continuous with the posteroventralmost sclerite in the abdomen (Fig. 9A,  B), which we interpret as sternum 9 (posteriormost ventral sclerite in the abdomen, absent in female Hymenoptera). The posteroventralmost sclerite in the female abdomen of Orussidae is tergum 9, which is subdivided ventromedially to allow the ovipositor to be exerted. The median longitudinal Evolutionary History of Orussidae structure observed on the posteroventralmost sclerite of the abdomen is a carina, not a furrow (Fig. 9A, B). e. The holotype of Baltorussus has what appear to be male genitalia inside the posterior end of the abdomen (Fig. 9C-E). All extant Orussidae have the male genitalia concealed when not in use. In the females, the concealed ovipositor apparatus extends all the way into the thorax. No trace of an internalized ovipositor was observed in the microCT scans.
In conclusion there can be no doubt that the holotype of Baltorussus velteni is a male. The newly discovered specimen resembles the holotype in all sex specific characters and is also a male.
The sculpture on the anterior part of the head (frons) between the eyes of the holotype of Baltorussus was not considered to belong to the specimen in the original description [10]. Instead, it was interpreted as an exuvium, the shed skin of a larve or a pupa of a not further identified insect. Indeed, in the holotype the frons is framed by a rectangle formed of cracks that might give the impression that this area is set apart from the rest of the specimen (Fig. 3A). However, in the microCT scannings performed by us it is evident that the spectacular sculpture is a part of the frons (Figs 3B, 4A). Furthermore, the finding of a second specimen with similar structure (Fig. 2A) confirms that it is indeed an integral part of the head.
The application of microCT scanning to the holotype of B. velteni as a supplement to examining the specimen in brightfield microscopy allows us not only to observe internal structures (e.g., the internal head skeleton, the male genitalia) but also to refine the interpretation of the external morphology. Many of the mistakes made in the original description [10] (most notably the misinterpretation of the anterior head sculpture) were apparently due to debris and cracks in the amber matrix and could be corrected by using microCT surface reconstructions. Still, traditional brightfield microscopy remains essential for the investigation of amber fossils as hairs and setae are often insufficiently resolved by microCT data (Fig. 9A, B).

Phylogenetic position of Baltorussus
Baltorussus displays a unique combination of plesiomorphic and apomorphic characters that decide its placement within the phylogeny of Orussidae. The plesiomorphic features are (character numbers refer to [7]): 1. The absence of ventral coronal teeth (char. 3:0; Figs 3A, B, 4A).
The ventralmost pair of cuticular teeth situated well below the median ocellus is present in most extant Orussidae, where they form the ventral part of the ocellar corona. However, these teeth are absent from the basalmost lineages Orussonia, Orussella, and Orussobaius as well as secondarily in Mocsarya and Pedicrista. This is not the case for Orussonia, Orussella, and Orussobaius, where the carina is less developed and the antennal bases are exposed. 3. Postocular and occipital carinae absent (chars 24:0, 26:0; Fig. 4B). Most orussid genera have these carinae, which are situated just behind the eye or on the back of the head, respectively, partly or fully developed (Ophrynon has secondarily lost the occipital carina), but not the genera Orussonia, Orussella, Orussobaius and Leptorussus. 4. Dorsal tentorial arms fully developed (Fig. 6). The tentorium forms the internal skeleton of the head and accommodates attachment sites for muscles moving the antennae and mouthparts. Usually, the tentorium has a characteristic Yshape in lateral view with the posterior tentorial arm forming the 'stem' and the diverging anterior and dorsal arms the 'arms' of the Y. This is the condition displayed by Baltorussus and is certainly a plesiomorphic trait. The presence or absence of the dorsal tentorial arms has not been scored across the Orussidae as it would require destructive dissection or expensive and time consuming scanning procedures. Therefore, the configuration of the tentorium has been documented only for two species of extant Orussidae [24,25], both within Orussus. In this derived genus, the dorsal tentorial arms are much reduced, hardly being developed at all. However, until the internal skeleton of the head of more orussid genera have been examined it cannot be determined when the reduction of the dorsal tentorial arms took place. 5. Median mesoscutal sulcus extending the length of mesoscutum (char. 65:2; Fig. 7). This is the most prominent plesiomorphic feature displayed by Baltorussus, as it is not known from any other extant or extinct orussids. This feature is present in most other basal hymenopteran lineages [26,27] and is probably a retained plesiomorphy that must have been lost independently at least twice among the extant Orussidae. Unfortunately, there is no information on this character from the two Cretaceous fossils, Mesorussus and Minyorussus. 6. Medioventral margins of hind coxa not angled (char. 96:0).
The fore wing venation of Baltorussus conforms completely with what can be inferred to be the ground plan of Orussidae, not displaying any features characteristic for some genera within the family; i.e., vein 2r is not displaced distally on the pterostigma, the discal cell is rectangular and vein 1r-Rs is elongate, and vein cu-a is not displaced significantly distally on Cu1.
On the other hand, there are some traits displayed by Baltorussus which are apomorphic relative to the ground plan of the Orussidae:   When analysed with the remainder of the Orussidae, Baltorussus is placed inside the extant members of the family as a separate, Evolutionary History of Orussidae basal lineage being the sister to a large clade comprising all the extant genera except Orussonia and Orussella (Figs 10, 11). This is in contrast with the two Cretaceous fossils Mesorussus and Minyorussus, which are stem group orussids, and Ophrynopus peritus from the Miocene, which is placed in an extant genus within the derived ophrynopine clade. There is thus good correlation between the respective ages of the four fossil orussid taxa and their placement in the phylogeny of the family.

Conclusion: The Timing of the Diversification of the Orussidae
Vilhelmsen [12] explored two sources of evidence when attempting to estimate the age of the radiation of the Orussidae: the direct evidence provided by the fossils and the indirect inferred from correlating splitting events in the family with major tectonic events in Earth history (i.e., biogeography). The age estimates from these two sources differed widely. Since only the two Cretaceous fossils (both from the northern hemisphere) were available for the analyses in Vilhelmsen [12] which placed them both as stem group orussids, they could only provide a minimum age (95 Ma) for the crown group. Consequently, the interpretation of the evolutionary history in Vilhelmsen [12] relied heavily on biogeography.
The four basalmost extant orussid lineages are all from the southern hemisphere in former parts of Gondwana: Orussonia (Australia), Orussella (Neotropics), Orussobaius (Australia), Leptorussus (Afrotropical). Based on this distribution, it was concluded that the earliest splitting events within the crown group probably occurred inside Gondwana [12]. Furthermore, the age of some of the splitting events basally in Orussidae between northern and southern hemisphere taxa were interpreted as possible vicariance events correlated with the initial separation between Laurasia and Gondwana (155 Ma ago), later splitting events perhaps corresponding to Gondwanan breakup. This was taken to indicate that the earliest splitting events within Orussidae occurred in the early Mesozoic, i.e., 180 Ma ago. This age estimate infers that the Cretaceous stem group fossils only occurred 80+ Ma after the initial radiation of the Orussidae and thus are poor indicators for the age of the family.
Since the publication of Vilhelmsen [12] the fossil record of Orussidae has been extended into the Paleogene by the discovery of Baltorussus and Ophrynopus peritus, both of which belong to the crown group. The different ages and phylogenetic placement of these two fossils allow for a better age estimate for the radiation of the family than from the stem group fossils alone. The geographic provenance of Baltorussus disproves that the earliest splitting events among extant Orussidae were restricted to southern hemisphere continents. Furthermore, the phylogenetic position and age of Baltorussus does not corroborate the biogeographical dating scenario. Having a comparatively young (44 Ma) fossil among the basal extant lineages in combination with the age of the stem group fossils indicates a much later age for the early radiation of Orussidae than indicated by the distributional history.
If the fossil orussid taxa are assumed to be not much younger than the splitting events that gave rise to them, the crown group Orussidae might not have started to radiate until well into the Cretaceous (Fig. 12). Some of the striking characteristics of the family (e.g., adaptations for vibrational sounding, the concealed ovipositor) may not have accumulated until comparatively shortly before that, i.e., still within the Cretaceous. This is corroborated by the apparent absence of these traits in the Paroryssidae (age up to 158 Ma), which are possibly also related to Orussidae (see Introduction). The sister group to the Orussidae and their closest fossil relatives are the Apocrita which have a fossil record extending back into the Early Jurassic, 180+ Ma ago [1]. Again, the temporal succession of fossil Apocrita, Paroryssidae and Cretaceous orussids is congruent with their phylogenetic placement relative to crown group Orussidae. Inferring that the radiation of extant orussids occurred not until well into the Cretaceous (e.g., after 100 Ma ago) is therefore in better accordance with the fossil record than an Early Jurassic date (180 Ma ago).
The revised sex of the holotype of Baltorussus velteni and the adjusted time estimate for the diversification of the Orussidae demonstrates the advances that can be made in understanding insect evolution both by the incorporation of new fossil discoveries in existing data sets, and the application of novel technologies (e.g., microCT-scanning) in the study of insect fossils.