A new lineage of Cretaceous jewel wasps (Chalcidoidea: Diversinitidae)

Jewel wasps (Hymenoptera: Chalcidoidea) are extremely species-rich today, but have a sparse fossil record from the Cretaceous, the period of their early diversification. Three genera and three species, Diversinitus attenboroughi gen. & sp. n., Burminata caputaeria gen. & sp. n. and Glabiala barbata gen. & sp. n. are described in the family Diversinitidae fam. n., from Lower Cretaceous Burmese amber. Placement in Chalcidoidea is supported by the presence of multiporous plate sensilla on the antennal flagellum and a laterally exposed prepectus. The new taxa can be excluded from all extant family level chalcidoid lineages by the presence of multiporous plate sensilla on the first flagellomere in both sexes and lack of any synapomorphies. Accordingly, a new family is proposed for the fossils and its probable phylogenetic position within Chalcidoidea is discussed. Morphological cladistic analyses of the new fossils within the Heraty et al. (2013) dataset did not resolve the phylogenetic placement of Diversinitidae, but indicated its monophyly. Phylogenetically relevant morphological characters of the new fossils are discussed with reference to Cretaceous and extant chalcidoid taxa. Along with mymarid fossils and a few species of uncertain phylogenetic placement, the newly described members of Diversinitidae are among the earliest known chalcidoids and advance our knowledge of their Cretaceous diversity.


INTRODUCTION
Jewel wasps (Hymenoptera: Chalcidoidea) are estimated to constitute one of the most species-rich insect lineages. Estimated numbers range from 100,000 to 500,000 species, which may comprise 10% of insect diversity, though only about 22,000 species have been described to date (Noyes, 1978;Noyes, 2000;Noyes, 2017;Heraty & Gates, 2001). Their evolutionary success is mirrored by and likely results from their varied biological life styles. Jewel wasps develop mainly as parasitoids of 13 different insect orders, as well as some nematodes, pseudoscorpions and arachnids, and thus are essential beneficial regulators, while some species are phytophages or even obligate pollinators of figs (Ficus) (Gibson, Heraty & Woolley, 1999;Weiblen, 2002;Heraty, 2009). Despite recent progress (Munro et al., 2011;Heraty et al., 2013;Peters et al., 2018), the relationships among most chalcidoid taxa as well as their evolutionary history still remain unresolved. The role of fossils in a phylogenetic framework is pivotal in understanding some of the evolutionary processes that led to chalcidoid megadiversity and provide valuable information on morphological character evolution (Donoghue et al., 1989;Peters et al., 2018). Reliably placed fossils can shed light on the minimum age of taxa and allow calibrations of molecular phylogenies to resolve timing and patterns of biological shifts (Ware & Barden, 2016;Gunkel et al., 2017;Slater, Harmon & Alfaro, 2012).
The fossil Minutoma yathribi Kaddumi, 2005 is currently the oldest described putative chalcidoid wasp from Jordanian amber, dated about 115 million years old (myo) (Kaddumi, 2005). It was placed in Mymaridae, which is considered to be the sister group to all other chalcidoid families (Heraty et al., 2013). Heraty et al. (2013), however, commented that the photo of M. yathribi rather suggests affiliation with Bouceklytinae, an extinct subfamily of uncertain placement. Kaddumi (2005) also mentioned a putative eupelmid fossil, which was not formally described. The family assignment of the concerned fossil is questionable as the metasomal and wing venational characters depicted in Kaddumi, (2005, figs. 95-97) are characteristic for Scelioninae (Platygastridae) (own observation). Myanmymar aresconoides Poinar & Huber, 2011 represents the oldest verified fossil record of Chalcidoidea, dating back to the Early Upper Cretaceous, approximately 99 mya (Shi et al., 2012). Although there are some reports of Eulophidae and Chalcididae from the transition between the Upper and Lower Cretaceous, no information concerning their validity is available (Penney, 2010). Schmidt et al. (2010) reported Eulophidae, Trichogrammatidae and Mymaridae from Ethiopian amber, which they dated through chemical and spectroscopic methods to an Upper Cretaceous origin (around 94 mya). Though the family identifications might be right, doubt was raised concerning the age of Ethiopian amber. Coty, Lebon & Nel (2016) described a myrmecine ant from the same deposit, which could readily be described in the tribe Crematogastrini, suggesting through phylogenetic dating that the specimen cannot be of Cretaceous age. Subsequent revised gas chromatography and infrared spectroscopy analyses showed, that indeed, though not completely unequivocal, evidence strongly suggested that Ethiopian amber is of Cenozoic origin, probably at least 50 million years younger than formerly suspected (Coty, Lebon & Nel, 2016). Currently, the oldest verified record of the families Trichogrammatidae and Aphelinidae are from Baltic amber, approximately 44 myo (Burks et al., 2015).
A putative member of Pteromalidae, Parviformosus wohlrabeae Barling, Heads & Martill, 2013, was described from limestone originating from the Crato formation, dated to the Aptian period, about 110 mya. Because of its age, it might be considered as one of the oldest known fossils of Chalcidoidea, but evidence for its placement is ambiguous because none of the diagnostic features of Chalcidoidea was preserved (Barling, Heads & Martill, 2013;Farache et al., 2016). It was placed within Pteromalidae only because of a putative habitus resemblance to Sycophaginae (now Agaonidae sensu Heraty et al., 2013). The limited morphological characters of P. wohlrabeae need to be reassessed before phylogenetic conclusions can be drawn from this fossil. The original placement of P. wohlrabeae in Pteromalidae is in this case highly problematic, because the family, in its current concept, is indicated to be polyphyletic (Campbell et al., 2000;Krogmann & Vilhelmsen, 2006;Heraty et al., 2013).
We here contribute to the scarce Cretaceous fossil record of Chalcidoidea by describing three new fossil genera and species within a new family. These fossils lack synapomorphies with any of the currently described chalcidoid families, but possess many putatively plesiomorphic features, suggesting a basal position within Chalcidoidea.

Specimens
Four specimens in four different pieces of Burmese amber were examined. Burmese amber is of Upper Cretaceous origin, approximately 99 my old (Shi et al., 2012). Additional information about the geographical origin of the individual pieces is not known. All pieces are deposited in the amber collection of the State Museum of Natural History, Stuttgart, Germany (SMNS).

Imaging
Imaging was done, using a MZ 16 APO Leica R microscope, with an attached DXM 1200 Leica R camera. The images were generated by stacking single images using the

Terminology
Terminology follows the Hymenoptera Anatomy Ontology (HAO) (Yoder et al., 2010). Abbreviations listed in Table 1 are used throughout the text and illustrations.

Cladistic analysis
Morphological cladistic analyses were performed using the 233 characters from Heraty et al. (2013). Their comprehensive matrix, encompassing 19 families, 78 subfamilies, 268 genera and 283 species of Chalcidoidea was used as basis for the here conducted phylogenetic analysis. Due to preservation and inaccessibility, some characters could not be scored for the fossils without reasonable doubt and were marked as unknown ''?'' (Appendix S1). Analyses were conducted using the program TNT ver. 1.5 (Goloboff, Farris & Nixon, 2008) following Heraty et al. (2013) in analysis setup. A sectorial search, with equally weighted characters, under New Technology methods was performed, using a ratchet weighting probability of 5% with 50 iterations, tree-drifting of 50 cycles, tree-fusing of five rounds and a best score hit of 10 times. New Technology searches in TNT provide refined algorithms more effective than simple branch swapping techniques applied in traditional searches, leading to shorter analyze times, especially in large datasets (Goloboff, Farris & Nixon, 2008). Nevertheless, traditional searches with and without implied weighting were conducted as well to test consistency of the results. Dependent on the used concavity constant (k), implied weighting aims to decrease the phylogenetic impact of supposed homoplasious characters, in comparison to equal weighting, (Congreve & Lamsdell, 2016).

Nomenclature
The electronic version of this article in Portable Document Format (PDF) will represent a published work according to the International Commission on Zoological Nomenclature (ICZN), and hence the new names contained in the electronic version are effectively published under that Code from the electronic edition alone. This published work and the nomenclatural acts it contains have been registered in ZooBank, the online registration system for the ICZN. The ZooBank LSIDs (Life Science Identifiers) can be resolved and the associated information viewed through any standard web browser by appending the LSID to the prefix http://zoobank.org/. The LSID for this publication is: LSID urn:lsid:zoobank.org:pub:B936D52D-7165-47CE-9C3E-0B79A17AC5AC. The online version of this work is archived and available from the following digital repositories: PeerJ, PubMed Central and CLOCKSS.  . Fore wing completely pilose, i.e., speculum absent ( Fig. 1E), basal vein only anteriorly pigmented. Gaster distinctly petiolate (Fig. 1D). Etymology. The generic name Diversinitus is composed of two parts. The first being ''Divers-'', originating from the Latin adjective ''diversus'', meaning diverse or different. The second part, ''-initus'', is the Latin noun ''initus'' translating to ''origin'' or ''start''. Together the two parts can be translated to ''origin of diversity'', referring to the age of the fossil and the diversity which evolved since its appearance in the Upper Cretaceous. The generic name is masculine in gender.  Habitus dorsal. Scale bars: 0.5 mm. Abbreviations: ax, axilla; bv, basal vein; cer, cercus; cx 1/2/3 , pro-/meso-/metacoxa; F1/11, funicular 1/11; frn, frenum; msc, mesoscutum; Mt 2 , metasomal tergum 2; Mt 8+9 , syntergum; no 1 , pronotum; not, notaulus; pl 1 , propleuron; pre, prepectus; prp, propodeum; ptl, petiole; sctl, scutellum; tgl, tegula. Drawings by M Haas.

Systematic Paleontology
Full-size DOI: 10.7717/peerj.4633/ fig-3 margin straight. Mandible at least two times as long as broad with slight curvature and few short setae on outer surface. Maxillary palps with at least three segments. Labial palps with at least two segments. Malar space shorter than 1/3 length of an eye.
Wings. Fore wing hyaline, immaculate, entirely pilose; humeral plate with at least three setae; basal vein apically pigmented and angled relative to submarginal vein at about 10-15 • ; marginal vein slightly thickened relative to postmarginal vein; stigmal vein about 0.5 times length of marginal vein; uncus bent at angle of about 95-100 • in direction of postmarginal vein, almost reaching it; postmarginal vein not reaching apex of wing, 1.5 times as long as marginal vein. Hind wing apical 2/3 pilose, rest relatively bare; posterior marginal fringe moderately long.
Legs. Pro-and metacoxa larger than mesocoxa; metacoxa dorsally bare, except few hairs posteriorly. Protibial setae inconspicuous and short. Basitarsal comb not visible. Metatibia laterally flattened, bearing two spurs, one robust, the other short and more slender.
Metasoma. Petiole (Mt 1 ) cylindrical distinct and reticulate, length 0.09 mm (h), breadth 0.06 mm (h). Gaster of holotype 0.66 mm in length, lanceolate; terga smooth and bare except of Mt 6 -Mt 8+9 with longitudinal rugosity and lateral setae, hindmargins straight, length of terga of holotype: Mt 2 : 0.24 mm, Mt 3 : 0.07 mm, Mt 4 : 0.07 mm, Mt 5 : 0.07 mm, Mt 6 : 0.11 mm, Mt 7 : 0.06 mm, Mt 8+9 : 0.04 mm. Cerci peg-like with few long setae. Female. Unknown. Specimen examined. Male holo-(SMNS Bu-4) and paratype (SMNS Bu-5) deposited in the SMNS. The amber piece hosting the holotype also includes syninclusions: three complete Diptera and three further Diptera, which are preserved only in part. Additionally, a Serphitidae (Hymenoptera) is included in the same piece. The amber piece including the paratype also hosts a Platygastridae: Scelioninae (Hymenoptera). Etymology. Named after the well renowned British broadcaster and naturalist Sir David Frederick Attenborough for his inspiring enthusiasm and devotion to natural sciences. This species was dedicated to Sir Attenborough during his visit to the SMNS on the occasion of his 91st birthday.
Wings. Fore wing hyaline, immaculate, speculum present, basal cell bare, costal cell pilose throughout; humeral plate with at least two setae; basal vein almost completely pigmented, Legs. Pro-and mesocoxa about same size, metacoxa slightly larger, dorsally completely bare. Protibial setae inconspicuous and short. Basitarsal comb not visible. Metatibia hardly flattened, bearing two equally short and robust spurs.
Glabiala gen. nov. LSID urn:lsid:zoobank.org:act:10644623-4534-4848-B961-1E608CBB773B Type species. Glabiala barbata sp. n. Diagnosis. Head densely pilose, with mouth margin surrounded by especially long setae (Figs. 4D and 5B). Clypeus quadrate. Toruli situated at about center of face, closer to each other than to margin of eyes. All funiculars rather thistle shaped (Fig. 4E). Pronotum and mesonotum with dense, short pilosity (Fig. 4F). Pronotum about 1/3 the length of the mesoscutum (Figs. 4G and 5B). Axillae advanced about 1/4 the length of the mesoscutum (Fig. 4G). Frenum large, delimited by deep frenal groove (Fig. 4F). Lateral propodeal callus with dense pilosity. Fore wing with speculum (Fig. 4D); basal cell pilose, basal vein only anteriorly pigmented. Metacoxa dorsally with short pilosity. Ovipositor hardly protruding apex of gaster (Fig. 4D). Etymology. The name consists of two parts originating from the Latin words for ''hairless'' (glabellus) and ''wing'' (ala), referring to the distinct speculum on the wing of the specimen. The generic name is feminine in gender. Head. Frontal view largely blocked, appearing trapezoid, finely pilose, except quite long pilosity on gena and mouth margin, about as broad as body, actual breadth and height not measurable. Foramen magnum situated higher than half height of head. Eye length 0.23 mm, height 0.27 mm, distance between eyes not measurable. Transfacial sulcus not discernable. Antennal scrobes absent. Clypeus quadrate with subparallel sides, apically truncate, tentorial pits absent, dorsal margin straight. Mandible not measurable, appearing broad and straight, with numerous longer setae on its outer surface. Maxillary palps probably with four segments. Labial palps with at least two segments. Malar space about 1/3 length of an eye.
Wings. Fore wing hyaline, immaculate, speculum present, basal cell pilose, costal cell pilose throughout; humeral plate with at least two setae; basal vein apically pigmented and angled relative to submarginal vein at about 9 • ; marginal vein slightly thickened relative to postmarginal vein; stigmal vein about 0.4 times length of marginal vein; uncus bent at angle of about 95 • in direction of postmarginal vein, almost reaching it; postmarginal vein not reaching apex of wing, 1.6 times as long as marginal vein. Hind wing apical 1/2 densely pilose, the rest relatively bare; posterior marginal fringe short.

Taxonomic remarks
It may seem counterintuitive to place the only two known males of Diversinitidae in a separate genus than the two females, especially since sexual dimorphism is widely spread in Chalcidoidea, most notably in Agaonidae and Eupelmidae resulting in a separation of sexes in morphological analysis of females and males, when coded separately (Krogmann & Vilhelmsen, 2006;Heraty et al., 2013). In most other chalcidoids however, those modifications do not include severe changes to the body plan and are often confined to body size (Hurlbutt, 1987) and antennal characters (Barlin & Vinson, 1981). Males of D. attenboroughi differ from both known females of Diversinitidae by the absence of a speculum on the forewing (versus presence of speculum), an elongate petiole (versus a transverse petiole) and an antennal insertion in the lower 1/3 of the face (versus an insertion near center of face). In addition, they also lack each of the diagnostic characters of the other two females (see below) so that a separate generic placement seems to be justified. Furthermore, we consider the two females as not congeneric based on significant morphological differences: Glabiala barbata differs from B. caputaeria in having the foramen magnum situated higher than half the height of the head (versus lower third of head), a pronotum only 1/3 length of mesoscutum (versus slightly shorter than mesoscutum), distinctly advanced axillae (versus slightly advanced), a large and clearly anteriorly delimited frenum (versus short and shallowly delimited) and a pilose basal cell on the forewing (versus a bare basal cell).

Results of cladistics analyses
The new technology analysis in TNT found 39 most parsimonious trees (5,395 steps) with the strict consensus tree being 5,861 steps long. The general topology of Heraty et al. (2013) could largely be retrieved (Fig. 6). As in Heraty et al. (2013) the following families appeared as monophyletic: Agaonidae, Chalcididae, Encyrtidae, Eurytomidae, Leucospidae, Mymaridae, Rotoitidae, Signiphoridae, Torymidae (including Megastigminae) and Trichogrammatidae. Contrary to Heraty et al. (2013), Aphelinidae and Eucharitidae could be retrieved as monophyletic as well. In the unweighted new technology analysis Mymarommatoidea was nested within Chalcidoidea as part of a larger clade containing the chalcidoid families Aphelinidae, Mymaridae, Rotoitidae and Signiphoridae, as well as few members of Tetracampidae and Eulophidae. Leucospidae were recovered as sistergroup to all other Chalcidoidea, including Mymarommatoidea. The fossils were recovered as a monophyletic group with Micradelus rotundus Walker, 1834 as sister taxon, nested within a large polytomy. Monophyly of the fossils could be retrieved in all analyses, however general tree topology changed considerably between different analyses. Using a traditional search without implied weighting (Appendix S2), Diversinitidae were recovered as sistergroup of all other Chalcidoidea with the inclusion of Mymarommatoidea. Mymaridae as well as Rotoitidae clustered in deeper clades far from the base of the tree. Using a traditional search with implied weights (Appendix S2), Mymarommatoidea were almost always recovered as sistergroup of Chalcidoidea (except k = 45), but topology changed drastically with increasing k value, as did the position of the fossils within the tree. In most analyses with k values below 30, the fossils were closely affiliated with the pteromalid genera Habritys brevicornis (Ratzeburg, 1844), Cheiropachus quadrum (Fabricius, 1787) and other interchanging groups. Above a k of 30, M. rotundus was recovered as a sistertaxon (k = 35 and 55) or only Cheiropachus quadrum (k = 40), Diversinitidae were sister to all Chalcidoidea including Mymarommatoidea (k = 45) or they were recovered close to Platynocheilus cuprifrons (Nees, 1834) and some Ormocerinae (k = 50 and 60).

DISCUSSION
The placement of Diversinitidae within Chalcidoidea is well supported by several morphological synapomorphies. One of the key autapomorphies of Chalcidoidea are the structurally unique multiporous plate sensilla (mps) on the antennal funicle, with their apices free of their surrounding antennal cuticle, the lack of an encircling groove around the sensillum and elevation of the multiporous plate above the antennal cuticular level (Barlin & Vinson, 1981;Gibson, 1986;Basibuyuk & Quicke, 1999). Evidently, Diversinitidae have modified sensilla (Figs. 1C, 4C and 4E), which are raised above the antennal surface and have their apices not completely surrounded by the antennal cuticle. Some mps, although not all, even protrude slightly over the funicular apices, as seen with backlighting under high magnification. The lack of an encircling groove cannot be unequivocally confirmed, but overall resemblance to mps of other Chalcidoidea is apparent. Within those groups Figure 6 Phylogenetic placement of Diversinitidae within Chalcidoidea based on morphological characters. Strict consensus tree calculated from 39 trees (tree length = 5861, CI = 0.077, RI = 0.567, 232 characters and 304 taxa, equal weights, new technology search). Yellow box highlights described fossils. Mymarommatoidea, potential sistergroup to all Chalcidoidea, collapsed and highlighted in blue. Green names indicate monophyletic and therefore collapsed families. Red names indicate monophyletic and therefore collapsed pteromalid subfamilies. Grey names indicate single taxa. For more information on the dataset of extant taxa refer to Heraty et al. (2013).
Full-size DOI: 10.7717/peerj.4633/ fig-6 of Proctotrupomorpha that are most closely related to Chalcidoidea (Peters et al., 2017), few possess mps on their antennae. Only Cynipoidea and the family Pelecinidae within Proctotrupoidea share this feature, but show a quite different sensillar morphology with their sensillae usually only slightly raised above the antennal surface and possessing a groove surrounding the multiporous plate (Basibuyuk & Quicke, 1999). Other Proctotrupoidea, Ceraphronoidea, Platygastroidea and Diaprioidea possess setiform multiporous sensilla sharing little resemblance with the morphology of chalcidoid mps (Gibson, 1986;Basibuyuk & Quicke, 1999). Even Mymarommatidae, the putative sister group of Chalcidoidea, lack mps (Gibson, 1986;Munro et al., 2011;Heraty et al., 2013). Another diagnostic feature of Chalcidoidea is the presence of a free, externally visible prepectus between the pronotum and mesopleuron, which separates the pronotum from the tegula (Gibson, 1985;Gibson, 1999;Gibson, Heraty & Woolley, 1999). Diversinitidae have a large triangular prepectus, neither fused to the pronotum or mesopleuron nor hidden beneath its lateral margin (Figs. 3A, 4A, 5A and 5B). Additionally, like in other chalcidoids, the mesothoracic spiracle is situated between the lateral margin of the mesoscutum and the pronotum directly adjacent to the anterodorsal edge of the prepectus, another autapomorphy of Chalcidoidea that is correlated with its external prepectus. Gibson (1999) hypothesized the more dorsal position of the spiracle compared to other hymenopterans as a derived state. Other hymenopterans having a concealed prepectus or a prepectus that is fused either to the pronotum or mesopleuron have the spiracle originating somewhat more ventrally below the level of the mesoscutum between the pronotum and mesepisternum. In Rotoitidae and Mymaridae, the spiracle is situated between the lateral margin of the mesoscutum and the pronotum, but in Rotoitidae and some Mymaridae the prepectus is slender and more or less concealed under the pronotum. Mymaridae and Rotoitidae are hypothesized as basalmost clades within Chalcidoidea (Gibson, 1986;Munro et al., 2011;Heraty et al., 2013;Peters et al., 2018) and their prepectal structure may represent a transitional state (Gibson, 1999).
Assignment of the fossils to extant chalcidoid families is not possible due to the lack of synapomorphies. The most prominent characteristic of Diversinitidae separating them from all other chalcidoid families, except for some Mymaridae, is the possession of mps on the first flagellomere (F1) in both sexes. Mps on F1 is found in Chalcidoidea only in very few cases. In Mymaridae, most males possess mps on their first flagellomere and also females of very few species (e.g., within the genera Eustochomorpha Girault 1915 andYoshimotoana Huber, 2015) have them (Heraty et al., 2013;Huber, 2015;Huber, 2017). Some Aphelininae (Aphelinidae) and Eucharitidae also seemingly possess mps on their apparent F1, but this is only because the first two flagellomeres are fused (Heraty et al., 2013). In Diversinitidae, the first visible flagellomere is undoubtedly F1 in both sexes. A well-developed F1 that has mps is hypothesized as plesiomorphic for Chalcidoidea (Heraty et al., 2013), suggesting a basal position of Diversinitidae within Chalcidoidea. During the evolution of Chalcidoidea, the first funicular likely secondarily lost mps in association with the segment being reduced in length to a ring-like segment (anellus) as is suggested by some chalcidoids that have additional funiculars reduced to anelli-like segments that lack mps. In those, comparatively few chalcidoids with F1 lacking mps but being reduced in size, F1 is hypothesized to have been secondarily lengthened (see character 11 in Gibson, 2003).
The labrum of Diversinitidae can be described as free, semicircular or rectangular, flap-like and broadly continuous with the clypeal margin. Darling (1988) postulated, that the ground plan structure of the labrum for Chalcidoidea is flap-like, with many evenly distributed setae. Darling (1988) referred to the labrum of Chalcididae as ''remarkably uniform and (. . . ) similar to that hypothesized as the ground plan for Apocrita'', being heavily sclerotized and contiguous with the margin of the clypeus, bearing long, tapered setae on the entire surface, arising from distinct sockets. In Pteromalidae, the plesiomorphic state of the labrum is found in Cleonyminae, and the labrum is also exposed in Spalangiinae, Asaphinae, Eunotinae and others, which bear in comparison to Cleonyminae setae only along their apical margin (Darling, 1988). Some Mymaridae also possess an exposed labrum (Heraty et al., 2013;. In Diversinitidae, the setal pattern is difficult to assess due to refractions within the amber in conjunction with the small size of the specimens. Setae are at least situated along the apical margin in Diversinitidae, but whether they are also found on the surface remains uncertain. If so, the labrum might also be putatively plesiomorphic for Diversinitidae. Diversinitidae possess a bidentate mandible, which is widely distributed in Chalcidoidea, although a three or more dentate mandible appears to be more common (Bouček & Noyes, 1987;Woolley, 1988;Dzhanokmen, 1996;Gibson, Heraty & Woolley, 1999;Gibson & Huber, 2000;Heraty et al., 2013). The plesiomorphic state for this character is not known and has so far not been discussed for Chalcidoidea comprehensively so that the evolutionary patterns are difficult to assess. Putatively basal chalcidoid families already exhibit varied states of mandible dentation, with Rotoitidae having bidentate mandibles, of which Chiloe micropteron (Gibson & Huber, 2000) has the upper tooth finely serrated (Bouček & Noyes, 1987;Gibson & Huber, 2000). Denticulation in Mymaridae varies greatly, with taxa lacking mandibular teeth (Erythmelus rosascostai Ogloblin, 1934) to taxa with many fine denticles (Eubroncus spp.) (Heraty et al., 2013;Jin & Li, 2014). The mymarid genera Triadomerus (Yoshimoto, 1975) (extinct), Macalpinia (Yoshimoto, 1975) (extinct) and Neotriadomerus (Huber, 2017) (extant) are considered to be the most basal taxa in this family (Huber, 2017). In those early groups mandibular dentation is already differing, with bidentate mandibles in Triadomerus and Macalpinia and four uneven teeth in Neotriadomerus, hampering phylogenetic implications. Outgroup comparisons with Mymarommatoidea and other Proctotrupomorpha (sensu Peters et al., 2017) reveal that also in those groups, mandibular dentation is highly variable (Naumann & Masner, 1985), not permitting a stable hypothesis about the groundplan state for Chalcidoidea. However, Diversinitidae as putative basal group within Chalcidoidea might indicate that bidentate mandibles could be plesiomorphic for at least a smaller subset of chalcidoid taxa.
A frenum is found in Diversinitidae, which is likely a plesiomorphic character state for Chalcidoidea (Krogmann & Vilhelmsen, 2006). Presence is observed in many chalcidoid families and in closely related groups, such as Mymarommatidae, Diapriidae and Platygastridae: Scelioninae (Heraty et al., 2013), suggesting that it is probably part of the ground plan structure for a subgroup of Proctotrupomorpha. Frenal morphology is used in species and subfamily distinction of Torymidae and Pteromalidae (Graham, 1969;Graham & Gijswijt, 1998;Gibson, 2003). The morphological variation of the frenum led to frequent discussions about its homology between different taxonomic groups (Grissell, 1995;Gibson, Heraty & Woolley, 1999;Vilhelmsen & Krogmann, 2006).
Diversinitidae possess peg-like cerci, which are more or less spatulate. This character state has been considered as plesiomorphic in contrast to a button-like cercus (Gibson, 2003) or, alternatively, as an apomorphic character state, which has independently evolved in different chalcidoid groups (Grissell, 1995). Grissell (1995) postulated that though peg-like cerci are found in Agaonidae sensu lato, Eulophidae (Entia Hedqvist, 1974), Pteromalidae (Cea Walker, 1837 andChromeurytoma Cameron, 1912), Torymidae and Megastigmidae, evolution of this character must have been convergent because positioning of the cerci is different in those groups. On the other hand, Gibson (2003) stated that many other groups have peg-like cerci as well, though most often not as prominent as those listed above, and therefore he considered exerted, basally articulated cerci as plesiomorphic relative to more reduced, plate-like cerci. In Heraty et al. (2013) many taxa were also coded as possessing exerted cerci to various degrees, such as Perilampidae (Brachyelatus sp.), Tetracampidae (Platynocheilus sp.), Signiphoridae (Signiphora sp.), Mymaridae (Borneomymar sp.) and Tanaostigmatidae (Protanaostigma sp.). Outgroup comparison for this character in Heraty et al. (2013) is however not conclusive due to sparse taxon sampling. Mymarommatidae (Mymaromella sp.) was coded as not possessing exerted cerci, compared to Scelioninae (Archaeoteleia mellea Masner, 1968), which show slightly exerted cerci and Diapriidae (Belyta sp.) without coding for this character. The wide distribution of peg-like cerci within Chalcidoidea and its appearance in Mymaridae and Diversinitidae supports the hypothesis that they represent the plesiomorphic state over button-like cerci.
Presenting a solid phylogenetic placement of Diversinitidae within Chalcidoidea is not unequivocally possible. All cladistic analyses provided evidence for monophyly of Diversinitidae, but did not resolve further relationships within Chalcidoidea, because placement of the fossils and general tree topology remained highly variable between different analysis. Although Micradelus rotundus was recovered as sister taxon of Diversinitidae in the new technology analysis and few traditional searches with implied weighting, a true relationship is highly doubtful. Micradelus rotundus belongs to the pteromalid subfamily Pireninae. This subfamily is characterized, though not only, by a reduced number of antennal segments and at least one annellus (Bouček, 1988), which is also the most prominent difference to Diversinitidae, sharing little resemblance to M. rotundus aside from morphologically variable characters like a bidentate mandible, lack of pronotal collar, deep notauli or exposed labrum. Additionally placement of M. rotundus was inconsistent over the different analyses and it behaved like a rogue taxon, jumping between several clades. However, high inconsistencies in the analyses were expected, because the morphology-only analysis in Heraty et al. (2013) was also poorly resolved. Due to the expected high rate of homoplasious characters in morphological datasets of Chalcidoidea (Krogmann & Vilhelmsen, 2006;Heraty et al., 2013), especially the results from analyses with and without implied weighting differed considerably. With increasing k values, the base of the phylogenetic tree was mostly relatively well resolved. Mymmarommatoidea were the sistergroup to Chalcidoidea and Rotoitidae and Mymaridae were retrieved as basal lineages within the superfamily. However, changes in topology of higher relationships were substantial. Through weighing down putative homoplasious characters, implied weighting is capable of better resolving polytomies . This can lead to trees with more correctly resolved clades, but also higher risks of erroneous placements and more inconsistent topologies as demonstrated by Congreve & Lamsdell (2016). Implied weighting can therefore be considered as less conservative over equal weighing of characters. There are conflicting views on whether parsimony analyses (Goloboff, Torres & Arias, 2017), as conducted in this study, or likelihood analyses (O'Reilly et al., 2018) perform better with morphological datasets. A comparison between likelihood and parsimony methods performed by Heraty et al. (2013) on the original dataset, however, resulted in a generally congruent tree with equally poor resolution of taxa. Additionally, probabilistic methods infer an evolutionary model on the data, based on subjective decisions and previous knowledge (Goloboff, Torres & Arias, 2017). We therefore favored the conservative equal weight parsimony analysis over implied weighting and likelihood analyses.
Unfortunately, there is no evident autapomorphic character of Diversinitidae, which would support its monophyly and all characters that exclude this group from existing families are seemingly plesiomorphic (see above). However, based on the unique combination of morphological characters (see diagnosis) and the preliminary results from the cladistic analyses (Fig. 6), we decided to place the new fossils into their own family rather than leaving them unplaced within Chalcidoidea.
Morphologically, Diversinitidae appear to be an early lineage of Chalcidoidea, possessing many putatively plesiomorphic characters (see discussion above). Mymaridae are thought to form the sister group to all remaining Chalcidoidea and can be traced back at least to the mid-Cretaceous (Gibson, 1986;Munro et al., 2011;Heraty et al., 2013). Resemblance between Diversinitidae and Mymaridae is not obvious and they only possess few putatively symplesiomorphic characters, such as an exposed labrum and mps on the true F1 in males and some females. In general, the mymarid body plan is characterized by a number of derived autapomorphies that have not changed much since the Mid Cretaceous (Poinar & Huber, 2011). The phylogenetic position of Diversinitidae can therefore not be established with certainty and several hypotheses are possible. Firstly, Diversinitidae could represent the sister group to all remaining chalcidoids, since they show a multitude of plesiomorphic characters, foremost mps on F1. During chalcidoid evolution mps on F1 might have been lost at first in females (as in most Mymaridae) and subsequently also in males (as in all remaining Chalcidoidea). This would imply, that the prepectus in Diversinitidae was either secondarily enlarged or that Mymaridae and Rotoitidae reduced the prepectal size during their evolution. Diversinitidae might also represent a sistergroup to a smaller subset of Chalcidoidea, suggesting that mps on F1 were independently lost at least twice, once in most females in Mymaridae and once in all other Chalcidoidea. Prepectal size might therefore have been increased in other Chalcidoidea relative to the prepectus in Mymaridae and Rotoitidae.
Biological implications of the new fossils are difficult to draw, because their phylogenetic position is not fully resolved. Egg parasitoidism is hypothesized to be the putative ground plan biology of Chalcidoidea (Heraty et al., 2013;Peters et al., 2018). Diversinitidae share a relatively small body size, which unites nearly all egg parasitizing taxa, but does not necessarily exclude ectoparasitoid groups. Body shape is not indicative, because both ectoand endoparasitoids can be very diverse in this regard. The length of the ovipositor and its saw-like tip might be indicative for concealed hosts inside plant material.

CONCLUSION
With the newly described fossils we reduce a significant fossil gap of Chalcidoidea from the Cretaceous. The wasp species described herein provide important new information of chalcidoid evolution because they are early representatives of a parasitoid lineage that was just beginning to evolve. One hundred million years later we merely start to fully appreciate the great morphological diversity and ecological significance of these ''green myriads in the peopled grass'' (Walker, 1839), which still rank among the least known of all insects. Further Cretaceous fossils will hopefully reduce the fossil gap even further to help us to understand how chalcidoid wasps have evolved and shaped the evolution of their arthropod host groups and associated plant species, as one of the most diverse and influential insect groups that life has ever seen.