Revision of the fossil species of Thaumatodryinus Perkins from Dominican amber, with a new combination and description of a new species (Hymenoptera, Dryinidae)

The fossil species of Thaumatodryinus from Dominican amber are studied, and the first revision is presented with a key to the known taxa. We recognize three species, T. miocenicus Olmi, 1995, T. priscus (Olmi, 1998), comb. nov. and T. fuscescens sp. nov. The current classification of the genus and relationships between fossil and living species are discussed. Comments on the host records for Thaumatodryinus


Introduction
In recent years, a significant number of species of Dryinidae have been described from different fossiliferous deposits Guglielmino et al. 2018;Perkovsky et al. 2019). Currently, 73 extinct species of Dryinidae are known from amber inclu-sions and rock impressions Guglielmino et al. 2018;Perkovsky et al. 2019Perkovsky et al. , 2020aMartins and Melo 2019;Martynova et al. 2020;Olmi et al. 2020). The Dominican amber is the third largest deposit in number of fossil dryinid species (Kachin Burmese amber in first with 33 species and Baltic amber in second with 22 species), with a total of 14 species of the following genera: Aphelopus Dalman, Dryinus Latreille, Harpactosphecion Haupt and Thaumatodryinus Perkovsky et al. 2019Perkovsky et al. , 2020aMartynova et al. 2019).
Thaumatodryininae is a small subfamily of Dryinidae, known to attack nymphs of the auchenorrhynchous Flatidae (Guglielmino et al. 2013). This subfamily has an almost worldwide distribution, being absent only from the Palearctic region. A single genus is recognized, Thaumatodryinus Perkins, with 32 described species, of which only T. miocenicus Olmi 1995 has been described from a fossil inclusion (Olmi and Virla 2014;Olmi and Xu 2015;Olmi et al. 2019). Olmi (1984) proposed the subfamily Thaumatodryininae for the type-genus Thaumatodryinus Perkins, but a few years later he transferred Thaumatodryinus to the subfamily Dryininae and sunk Thaumatodryininae under the latter (Olmi 1993). The resurrection of Thaumatodryininae was proposed by Tribull (2015), after her molecular phylogenetic studies having showed that Thaumatodryinus does not belong in Dryininae. Morphological support for this hypothesis was obtained more recently by Martins (2018). His phylogenetic hypothesis also corroborated an independent position for Thaumatodryinus outside of the clade composed of Dryininae plus Gonatopodinae.
In this study, we contribute for the knowledge of Dryinidae by reviewing the fossil species of Thaumatodryinus from Dominican amber and propose a new species for the genus. Furthermore, the status of Dryinus priscus Olmi, 1998 is reinterpreted based on our comparative morphological study.

Materials and methods
The amber pieces studied here are deposited in the Departament of Zoology, of the Universidade Federal do Paraná, Curitiba, Brazil (DZUP) and in the American Museum of Natural History, New York, USA (AMNH). The inclusions have been obtained from amber fossiliferous Miocene deposits in the Dominican Republic. Dominican amber has been estimated to be around 15-20 million years old, from the Burdigalian, in the Early Miocene (Iturralde-Vinent and MacPhee 2019).
Morphological terminology follows Olmi and Virla (2014); specific terms used for integumental sculpture, follows Olmi and Virla (2014) and Harris (1979); we adopted Brothers's (2011) terminology for the forewing venation. The term rhinaria (sensu Olmi 1984) is interpreted herein as equivalent to "ADOs", Antennal Dorsal Organs (sensu Riolo et al. 2016). The diagnosis for the previously described species was prepared based on the original descriptions and redescriptions by Olmi (1995Olmi ( , 1998 and Olmi and Virla (2014), complemented by notes taken by the first author during his visit to the AMNH.
In the description and diagnosis the following abbreviations were used: POL, refers to the minimum distance between the inner edges of the lateral ocelli; OL, refers to the minimum distance between the inner edges of the lateral ocellus and the median ocellus; OOL, refers to the minimum distance from the outer edge of the lateral ocellus to the eye inner margin; OPL, refers to the minimum distance from the posterior edge of a lateral ocellus to the occipital carina; and TL, refers to the minimum distance from the posterior edge of eye to the occipital carina. The measurements provided for ocellar ratio, antennomeres and for the fore leg articles represent relative values.
Study of the inclusions was carried out in a Leica M125 stereomicroscope. Color images of Thaumatodryinus fuscescens sp. nov. were obtained by a LEICA DFC295 digital camera, and those of T. miocenicus Olmi and Dryinus priscus Olmi by a Nikon DS-Ri2 using NIS-Elements D-Z-series 11*. Image stacking was carried using the software Zerene Stacker. The figure plates were prepared using Adobe Photoshop (version Cs6).
Diagnosis. Among living and fossil dryinids, the species of Thaumatodryinus can be recognized by the following combination of characters: mandible with four teeth in both sexes, teeth progressively larger from anterior to posterior ends along cutting edge; outer surface of mandible with a distinct basal sulcus, extending from the anterior to the posterior condyle; mid portion of anterior margin of clypeus straight or convex; occipital carina complete in fossil taxa and in males of most living species, females of many living taxa with incomplete carina; maxillary and labial palpomeres in proportion 6:3; antenna filiform in both sexes; flagellomeres 3-7 of female with single set of rhinaria on its mid length, flagellomere 8 with two sets of rhinaria along its length; rhinaria on flagellomeres 5-8 each with four long setae, two at each side; pronotum saddle-shaped and crossed by strong transverse impression at central portion, pronotal lobe reaching tegula; anterior margin of mesoscutellum with a foveate groove; fore wing with 2r-rs shorter than 3Rs&4Rs; female with tarsal claw simple and with basal expansion, male with bifid claw, without basal expansion; apex of 3 rd tarsomere of fore leg with variable number of thick bristles in female; chela with rudimentary claw, enlarged claw with one row of lamellae and two subapical teeth laterally on outer surface; female hind coxa with a distinct basal projection ventrally; dorsal and posterior surfaces of propodeum convex; tibial spur formula for both sexes 1/1/2.

Key to species of Thaumatodryinus from Dominican amber (females only)
1 Eyes hemispherical, about 1.1× longer than wide; maximum head width about 1.9× distance between inner margins of tegulae; antenna with 1 st flagellomere as long as 2 nd (Fig. 1 Diagnosis. Thaumatodryinus fuscescens sp. nov. is characterized by the body predominantly testaceous, except head and part of pronotum and remainder of mesosoma black; eyes hemispherical, about 1.1× longer than wide; maximum head width about 1.9× distance between inner margins of tegulae; frontal line present; mid portion of clypeus with anterior margin straight; antenna with 1 st flagellomere as long as 2 nd ; POL 3.2× OL; OOL about 6× OPL; occipital carina complete; notauli percurrent, extending from anterior to posterior margins of mesoscutum.
Description. Female holotype (Fig. 1A-F). Approximate body length: 4.1 mm. Color. Head black, except mandible, clypeus and antenna testaceous (Fig. 1A-D); pronotum black, except dorsal-posterior surface brown (Fig. 1B); mesosoma black; legs testaceous, except pro-and metafemur with middle portion darkened (Fig. 1); fore wing with one transverse darkened band; metasoma testaceous. Pubescence. Head with short setae (Fig. 1A-C); antenna with setae; eye glabrous (Fig. 1C); pronotum with fine and erect setae (Fig. 1B); mesosoma with conspicuous erect setae (Fig. 1B); propodeum with erect setae; metasoma with more conspicuous erect pilosity on last tergum. Integumental sculpture. Head finely granulated (Fig. 1A-C); frontal line in the form of a low carina and apparently complete, not visible near mid ocellus; malar sulcus well-marked; mandible with a marked groove at the base; pronotum granulate, except by rugosity on lateral portions of transverse impression; mesoscutum granulate; notauli complete; mesoscutellum rugulose, its anterior margin with a narrow row of small foveae, not with a groove; mesopleuron rugose; metapleuron reticulate-rugose; propodeum reticulate rugose. Structure and proportions. Vertex convex, except by ocellar triangle slightly elevated. Ocellar ratio: POL = 16; OL = 5; OOL = 25; OPL = 4; TL = 9. Antennomeres in following proportions: 35: 15: 45: 45: 60: 65: 55: 23: 27: 25. Eye hemispherical, about 1.1× longer than wide; maximum head width about 1.9× distance between inner margins of tegulae; occipital carina complete; mid portion of clypeus with anterior margin straight; maxillary palpomeres elongated; malar space about as long as basal width of mandible; notaulus complete (Fig. 1B). Head longer than pronotum (60:55); mesoscutum longer than mesoscutellum and metanotum shorter than propodeum. Fore wing with three cells closed by pigmented tubular veins; vein abscissa 3Rs&4Rs longer than 2r-rs (65:30). Dorsal surface of propodeum distinctly longer than posterior surface (50:45); mesopleuron with a groove in the central region. Articles of fore leg (Fig. 1E-F Remarks. Thaumatodryinus fuscescens sp. nov. is more similar to the extant Neotropical fauna than to the other two species found in the Dominican amber. It has in common with some of the living species a convex vertex and slightly elevated ocellar triangle, straight lower clypeal margin, complete occipital carina, percurrent notauli and anterior margin of mesoscutellum with a narrow row of small foveae. Among the fossil species studied here, T. fuscescens sp. nov. is more similar to T. miocenicus and differs from T. priscus by the head shape with eyes protuberant on the lateral side, frontal line present and ratio of OL short than TL. The new species differ from T. miocenicus and T. priscus by the granulated integument of the head; clypeus with lower margin straight; 1 st flagellomere as long as 2 nd ; hemispherical-shaped eyes; and percurrent notauli. The chela of T. fuscescens is closed and for this reason it is not possible to determine the number of lamellae in the enlarged claw and in the 5 th tarsomere.
Etimology. The species is named in reference to its overall dark coloration, from the Latin fuscus, dark, dusky, and -escens, beginning of, becoming.
Eye somewhat elongate, about 1.3× longer than wide; maximum head width about 1.7× distance between inner margins of tegulae. Pronotum shorter than head (16:37). Mesoscutum longer than pronotum (20:16). Fore wing with 2r-rs shorter than 3Rs&4Rs. Fore leg segments in following proportions: 30 (coxa): 15 (trochanter): 40 (femur): 43 (tibia): 22 (1 th tarsomere): 3 (2 nd tarsomere): 7 (3 rd tarsomere): 22 (4 th tarsomere): 36 (5 th tarsomere); 5 th tarsomere longer than enlarged claw and with two rows of at least 30 lamellae; distal apex with group of approximately three long lamellae; enlarged claw with two subapical teeth and one row of 29 lamellae. Propodeum short, distinctly shorter than high in lateral view, with dorsal surface shorter than posterior surface (45:50). Remarks. While in the original description Olmi (1995) mentioned that the mandibles are "not distinctly visible", Olmi and Virla (2004) stated that the mandibles are "tridentate". We could not clearly see the condition in the type material but consider more likely that T. miocenicus has quadridentate mandibles as in all other Thaumatodryinus. When the mandibles are closed, it is quite hard to observe the tooth closer to the clypeus, giving the impression that the mandibles are tridentate.
Extended diagnosis. Female holotype, 4.0 mm (see fig. 55 in Olmi 1998 (1 th tarsomere): 4 (2 nd tarsomere): 6 (3 th tarsomere): 21 (4 th tarsomere): 31 (5 th tarsomere); 5 th tarsomere slightly longer than enlarged claw (31:27) and with two rows of numerous lamellae (exact number hard to observe); enlarged claw apparently with one subapical tooth and with one row of numerous lamellae (exact number hard to observe). Propodeum distinctly long, about as long as high in lateral view, dorsal surface as long as posterior surface. Remarks. This species was originally described in the genus Dryinus by Olmi (1998) and redescribed as such by Olmi and Virla (2014), but this taxon belongs in Thaumatodryinus since it exhibits all the diagnostic features of the genus, including the characteristic long setae on flagellomeres (see Fig. 3). Furthemore, we have noticed that the specimen illustrated by Olmi and Virla (2014: 254) representing Dryinus priscus Olmi does not correspond to the holotype. The inclusion shown in their Plate 100 corresponds to a species of Harpactosphecion Haupt, which can be easily recognized by the long and distinctly curved, fore trochanter, and the thin flagella.
Examined material. Female holotype, in amber from Dominican Republic: amber from an unknown mine (probably El Valle mine) DR-14-341 (AMNH).

Discussion
All fossil species of Thaumatodryininae are currently known only from Dominican amber. Species described originally in Thaumatodryinus from Baltic amber by Brues (1923Brues ( , 1933 have been all transferred to Harpactosphecion, a distantly related fossil genus belonging to the Dryininae (Olmi 1984;Olmi and Bechly 2001). These two genera resemble each other in their general morphology, having elongated, slender bodies and long antennae and legs. Among many differences, they can be easily set apart by details of their antennae: while the flagella of Harpactosphecion lack conspicuous pilosity and the first two flagellomeres are distinctly longer than the 3 rd , Thaumatodryinus possesses distinctly long setae projecting from all rhinaria and their first two flagellomeres are about as long as the 3 rd .
We document here the presence of three species of Thaumatodryinus in Dominican amber. Two of them, T. fuscescens sp. nov. and T. miocenicus, resemble the extant species in their protuberant and hemispherical eyes, and a more compact propodeum. Thaumatodryinus priscus is more divergent, possessing more elongated eyes and a longer propodeum. In general, it has a more slim, elongated body and this might have contributed to its interpretation as a Dryinus by Olmi (1998). Also, the amber piece containing the holotype has been further polished after the original description, allowing for a better view of the inclusion, as can be seen in the photographs provided here.
As far as it is known, species of Thaumatodryinus attack only the fulgoroid family Flatidae and host records are known for four species of Thaumatodryinus from the Australian, Nearctic, Neotropical and Oriental regions (Guglielmino et al. 2013). They have been reared from six species belonging to five different genera of Flatidae (Guglielmino et al. 2013). The fossil record of Flatidae is scanty and the oldest records are from the Paleocene of Argentina and China (Szwedo and Stroiński 2013). Additional more recent records include a nymph preserved in Dominican amber and parasitized by a dryinid larva (Olmi 1995). This could represent a larva of Thaumatodryinus, although this cannot be stated with certainty considering that Flatidae are also attacked by three other dryinid genera, Gonatopus Ljungh and Neodryinus Perkins (Gonatopodinae), and Dryinus (Dryininae) (Guglielmino et al. 2013).