Notes on rhopalosomatid wasps of Dominican and Mexican amber (Hymenoptera: Rhopalosomatidae) with a description of the first fossil species of Rhopalosoma Cresson, 1865

. Rhopalosomatidae are a family of aculeate wasps that are ectoparasitoids of crickets as larvae and are predominantly distributed pantropically. The published fossil record of the family is scarce. Here, we report three new fossil rhopalosomatid wasp specimens from Dominican and Mexican amber. Rhopalosoma hispaniola Lohrmann sp. nov. is described and documented from Dominican amber by two separate inclusions – one of each sex. An additional fossil female Rhopalosoma is described and documented from Mexican amber but is not named due to the insufﬁcient preservation of the fossil. The


Introduction
The Recent New World fauna of Rhopalosomatidae is currently comprised of 35 formally described species assigned to three genera: Liosphex Townes, 1977, with 13 species, Olixon Cameron, 1887, with 5 species, and Rhopalosoma Cresson, 1865, with 17 species (Townes, 1977;Lohrmann and Ohl, 2010;Lohrmann et al., 2012). Their distribution ranges from southeastern Canada and the northeastern USA to northern Argentina and southern Brazil, with the vast ma-jority occurring in the tropic zone. As far as it is known, female rhopalosomatids hunt for crickets (Orthoptera) that ultimately serve as hosts for their ectoparasitoid larvae (Perkins, 1908;Hood, 1913;Gurney, 1953;Lohrmann et al., 2014;Blaschke et al., unpublished results;see also Fig. 1a-b).
The recent discovery of an ectoparasitoid rhopalosomatid wasp larva attached to its cricket host in mid-Cretaceous Burmese amber revealed a remarkable constancy of the biology of Rhopalosomatidae over the last 100 million years (Lohrmann and Engel, 2017). In some way, this extraordinary fossil hides the fact that we still know little about the origin and evolution of this peculiar family of aculeate wasps. Two of the four fossil taxa previously assigned to the family, Mesorhopalosoma Darling (in Darling and Sharkey, 1990) and Paleorhopalosoma Nel et al., 2010, have been transferred to the sphecoid wasps (Osten, 2007;Archibald et al., 2018). Additionally, Propalosoma Dlussky and Rasnitsyn, 1999, has been transferred to the myrmeciine ants (Archibald et al., 2018). This leaves Eorhopalosoma Engel, 2008, which is another Cretaceous Burmese amber fossil, as the sole fossil species in the family so far.
Due to their general appearance, members of the rhopalosomatid wasp genus Rhopalosoma (see Fig. 1c), as well as the Old World Paniscomima Townes, 1977, are often mistaken for nocturnal ichneumonids, e.g., Ophion spp. or Netelia spp. (McGown, 1998), and one is strongly encouraged to go through collections of unidentified subtropical or tropical nocturnal ichneumonids when searching for material of the genus. This is also true for the fossil record. During our surveys of fossil inclusions, three adult rhopalosomatid wasps have been discovered in Dominican and Mexican amber, and one of the fossils described herein was indeed initially identified and advertised as a "Large Ichneumonid Wasp".
All three of the new fossils belong to the Recent genus Rhopalosoma. The aim of the present study is to describe and illustrate these fossils, which are not only the first fossil records of the genus but are also the first of the family from Dominican and Mexican amber and from the Miocene as a whole.

Thermochemolysis-gas chromatography-mass spectrometry
Of the three new rhopalosomatid fossils reported herein, each is preserved in a separate amber piece. Two of these fossils originate from the La Bucara mine in the Dominican Republic, and one comes from a mine in Mexico (Chiapas). Because the Mexican amber piece was obtained from an online auction, without an indication of the exact provenance in Chiapas, we performed a chemical analysis to ascertain its amber vs. copal nature and its affinities with genuine Chiapas amber. Thermochemolysis-gas chromatography-mass Lab PY-2020iD pyrolyzer coupled with a Shimadzu GCMS-QP2010 Plus system operating with a split ratio of 100. Separation was achieved using a capillary column SLB-5ms (60 m × 0.25 mm inner diameter -ID, 0.25 µm film thickness) with the carrier gas He with a flow of 1.1 mL min −1 . The operating conditions were as follows. The initial temperature was held at 50 • C for 2 min and was increased to 310 • C at a rate of 4 • C min −1 for 34 min. Individual compounds were identified based on comparison of (i) MS data with the NIST 2014 library and literature (Anderson, 1994(Anderson, , 1995 and (ii) retention time with the Baltic amber. The molecules were referenced according to Anderson (1995). The relative distribution of identified compounds was determined by measuring the area of a specific fragment, denoted as m/z integration in Appendix A. The peak area of the selected m/z for each compound was integrated and corrected by a mass spectra factor (Appendix A) calculated as the reciprocal of the proportion of the fragment used for the integration and the entire fragmentogram. The molecular ratios were calculated using those corrected areas that allow an approximation of areas on the total ion chromatogram, while preventing from the simultaneous analysis of co-eluting compounds.

Age of the amber
There has been a wide range of proposed ages for Dominican amber, from 45 ± 9 to 15 Ma (Lambert et al., 1985;Schlee, 1990;Landis and Snee, 1991;Grimaldi, 1995). However, most recent studies reviewing its geological dating indicate a late Early to early Middle Miocene age (about 15-20 Ma; Iturralde-Vinent and MacPhee, 1996;Iturralde-Vinent, 2001; and Penney, 2010 for a review). Mexican amber is of similar age (15-20 Ma;Solórzano Kraemer, 2007, but see Serrano-Sánchez et al., 2015, for a slightly older Miocene dating), which is supported by the similarities of the organismal inclusions of these two amber deposits. Indeed, many insect and arachnid genera are found in both Dominican and Mexican amber (Solórzano Kraemer, 2010). Yet, the authors are not aware of a single species in common described or recorded in both ambers.

Preparation and study of fossils
All three amber pieces were manually polished using a Buehler MetaServ 3000 polisher and Buehler CarbiMet silicon carbide papers to obtain flat surfaces for optimal observation and imaging of the insect inclusions.
The inclusions were studied with a Leica MZ12 and a Zeiss SteREO Discovery.V20 stereomicroscope. Measurements were taken with an ocular micrometer. Extended-focus images were taken with a Leica DFC 490 digital camera with a Leica Z16 Apo A system and a Canon 5D Mark II digital camera attached to the Leica MZ APO stereomicroscope or with the BK PLUS Lab system (Dun Inc.) with a Canon MPE 65 mm lens mounted on a Canon 6D camera. Pictures were aligned and stacked using the software package Auto-Montage Essentials by Syncroscopy (Version 5.03.0061 ES), Helicon Focus software (Helicon Soft Ltd., version 6.7.2), or Zerene Stacker under the PMax value.
The general morphological terminology is adopted from Huber and Sharkey (1993), with additions from Mason (1986Mason ( , 1990, and the following morphological abbreviations and modifications are used in the text. Cu 2 is the section of fore wing cubitus separating the 2Cu and 2M (not closed) cells (Fig. 4f). IOD is the interocellar distance, or the shortest distance between the lateral ocelli. LOD is the maximum diameter of a lateral ocellus. MOD is the maximum diameter of the median ocellus. M 1 is the section of fore wing media separating the R and 1M cells (Fig. 4f). M 2 is the section of fore wing media separating the 1M and 1Rs cells (Fig. 4e). OOD is the ocellocular distance, or the shortest distance between the lateral ocellus and eye. Rs 1 is the section of fore wing radial sector separating the 1R1 and 1Rs cells (Fig. 4e). Rs * is the section of hind wing radial sector separating the R and R1 cells (Fig. 5e). TL is the temple length, or the shortest distance between the posterior eye margin and occipital carina.
All three specimens are deposited in the amber collection of the Museum für Naturkunde in Berlin, Germany (inventory numbers: MB.I 5915, MB.I 6046, and MB.I 6549).

Nomenclatural acts
This published work and the nomenclatural acts it contains have been registered in ZooBank (http://www.zoobank.org/, last access: 1 October 2018), with the following LSID 3 Results

Chemical analysis
The THM-GC-MS analysis of the Mexican resin sample ( Fig. 2; see also Appendix B) revealed characteristic bicyclic products derived from polylabdanoid structures including carboxylic acids (XIa-XIVa), analyzed as their methyl ester; alcohols (XIIIb-XIVb) partially methylated (XIIId-XIVd); and hydrocarbons (XIc, XIIIc and XIVc). Moreover the THM of this sample also produced 1-methylbicyclic hydrocarbons (A-K) that are believed to come from A-ring defunctionalization (Anderson, 1995). The shift in the retention times of methylic ether XIIId and XIVd between the Baltic amber and the Mexican resin ( Fig. 3) indicated an enantio configuration for the latter. Consequently, it falls into the class Ic as defined by Anderson et al. (1992). The distribution of bicyclic products derived from the polylabdanoid macromolecular structure of the Mexican sample was dominated by 1-dimethylbicyclic and 1methylbicyclic hydrocarbons (50 % and 28 %, respectively), while bicyclic acids and alcohols represented only 6 % and 17 % of those compounds, respectively (Table 1). This predominance of 1-dimethylbicyclic hydrocarbons was already observed in Sicilian amber (van der Werf et al., 2016).
The composition was compared to those of known class Ic ambers from the Eocene of Oise (France; Nohra et al., 2015) and the Miocene of the neotropics (the Dominican Republic and Mexico, from the collection of Vincent Perrichot, Univ. Rennes, unpublished results). The relative proportions of carboxylic acids, alcohols, and hydrocarbons among bicyclic products derived from polylabdanoid structures were similar in the Mexican sample and the two neotropical ambers, while the Oise amber was richer in carboxylic acids ( Table 1). The proportion of 1-methylbicyclic hydrocarbons was proposed as a maturity proxy, since they are especially observed in samples of moderate thermal maturity (Anderson, 1995). The present Mexican sample (28 %) falls between Oise (19 %) and Dominican (22 %) ambers on one side and Mexican amber (41 %) on the other side, thus attesting to its amber rather than copal nature as well as its affinity with neotropical ambers.
With the exception of the earliest record of a class Ic amber from the Carboniferous (Bray and Anderson, 2009), all other known Ic ambers are Cenozoic and are derived from angiosperms of the legume family Fabaceae (Lambert et al., 2008;Nohra et al., 2015;van der Werf et al., 2016), particularly the genus Hymenaea Linnaeus that produced neotropical Miocene ambers from the Dominican Republic and Mexico (Anderson et al., 1992;Anderson, 1995).

Diagnosis
Among extant and fossil rhopalosomatids, species of Rhopalosoma are characterized by the following character combination. The wings are fully developed (brachypterous in Olixon); the occipital carina is present (absent in Liosphex); the apical section of the fore wing cubitus (Cu 2 in Fig. 4) is at least slightly sinuate (arched, but indented at the level of junction with CuA in Paniscomima, and almost straight in Eorhopalosoma); female pretarsal claws have a preapical tooth (as in Figs. 4b, 6c); female tarsomeres II-IV have apicolateral tarsal fenestrae (Fig. 6b); and the second preapical process of the male penis valve is narrow to broad, distad from the head of the penis valve, and is not overlapped at its apex by the third preapical process ( Fig. 1d; broadly triangular, near head of penis valve, and overlapped in Paniscomima).
(Males and females of the family can be differentiated by the number of flagellomeres -males have 11, and females have 10, by the number of visible metasomal tergites -males have 7, and females have 6, and by the form of tarsomeres II-IV -cylindrical for males and flattened for females. For a more detailed morphological description of the genus and all extant species, including a species-level identification key, the family revision of Townes (1977), should be consulted.)

Distribution
Distribution is predominantly in the tropics and subtropics of the Americas (eastern United States to southern Brazil and northern Argentina), including some of the Caribbean countries (Bahamas, Cayman Island, Cuba, and Haiti; Townes, 1977).

Material
Holotype: female, MB.I 5915, in Miocene amber from the Dominican Republic; the specimen is deposited in the amber collection of the Museum für Naturkunde Berlin. The specimen is almost completely preserved without any significant distortions. However, large parts of the legs are missing. The following syninclusions have been observed: one Collembola and one Mymaridae.
Paratype: male, MB.I 6046, in Miocene amber from the La Bucara mine in the Dominican Republic; the specimen is deposited in the amber collection of the Museum für Naturkunde Berlin. The specimen is completely preserved without any significant distortions. In order to access crucial characters of the fossil the amber piece has been cut into two fragments. The following syninclusions have been observed: two Hymenoptera (Mymaridae), four Diptera, one Isoptera, one large Isoptera wing, one unidentified Insecta, and one Araneae.
Description Female (Fig. 4). Measurements. Total body length (head, without antenna, plus mesosoma plus metasoma) in lateral view 12.0 mm. Maximum head width in dorsal view 2.0 mm. Flagellar length (measured on left antenna with the last segment added from the right antenna) 11.4 mm (including pedicel and scapus 12.1 mm). Mesosomal length in lateral view 3.2 mm. Fore wing length 8.9 mm. Hind wing length 6.5 mm. Metasomal length in lateral view 7.8 mm.
Head. (Base of right flagellum and last flagellomere of left flagellum not preserved.) FI-FV each with a pair of apical bristles. FI shorter than FII, which is about as long as flagellomeres III-V. FVI-FX becoming sequentially shorter from base to apex. Inner margin of compound eye emarginate. Depth of eye notch about the same size as the width of toru-lus. FI 5.0 ×, FII 8.2 ×, and FVII 9.5 × as long as wide. TL 0.6 ×, OOD 0.25 ×, and MOD 1.0 × LOD.
Mesosoma. Scutellum about half as long as mesoscutum and approximately 3 times as long as metanotum. Mesosternal lobes present, separated from mesosternum by slight constriction.
Fore wing. With eight enclosed cells: C, R, 1Cu, 1R1, 2R1, 1Rs, 1M, and 2Cu. Costal cell very narrow, apically slightly wider, over complete distance narrower than bordering veins. Cell 1Rs about 1.05 × as long as cell 1R1. Vein 1cu-a slightly bowed, distad of M 1 by about 1.0 × its length. Distance between 1cu-a and M 1 about half the distance between 1cu-a and m-cu. Vein Rs 1 straight, about 0.8 × length of M 2 . Pigmented traces of 2m-cu present. Anal cell without longitudinal spurious vein at its center (present in some Recent Asian Paniscomima). Pterostigma narrow.
Hind wing. Rs * straight, recurrent, its anterior section meeting Sc + R at an angle of about 55 • . M diverging far beyond cu-a. With two clusters of hamuli. Basal hamuli straight. With 13 distal hamuli, all of same size and curvedand/or hook-like.
Legs. (The following parts of the legs are not preserved: tarsomeres III and following of right fore leg, tarsomere II and pretarsus of left fore leg, distal two-thirds of tarsomere I and tarsomere II of right mid leg, tarsomeres II-IV of left mid leg, distal half of tarsomere I and following of both hind legs.) Fore leg with one tibial spur. Mid leg and hind leg each with two tibial spurs, inner spur of hind tibia at base with dorsal tuft of bristles (calcar). Tarsomeres II-IV with apicolateral fenestrae. Pretarsal claws toothed medially (Fig. 4c). Arolia large.
Pilosity. Body and wing membranes covered with regular distributed fine, short setae. Male (Fig. 5). Measurements. Total body length (head, without antenna, plus mesosoma plus metasoma) about 7.3 mm. Maximum head width in oblique ventral view about 1.0 mm. Flagellar length (measured on left antenna in ventral view) 5.7 mm (including pedicel and scapus: 6.2 mm). M 2 , ratio of the length of fore wing cells 1Rs and 1R1) seems to be slightly higher than in the Recent species of Rhopalosoma. However, we chose a conservative approach by assigning both specimens to the same species rather than describing two species based on only a single specimen of each sex.
( Fig. 6) Material Female, MB.I 6549, in Miocene amber from Chiapas, Mexico; the specimen is deposited in the amber collection of the Museum für Naturkunde Berlin. The specimen is almost completely preserved, except for part of the antennae, but it is largely distorted, except for the wings and legs. In order to access the crucial characters of the fossil the amber piece has been cut into three fragments. The following syninclusions have been observed: one Hymenoptera (Braconidae), six Diptera, one Heteroptera, and one Thysanoptera.

Description
Head, mesosoma, and metasoma as in R. hispaniola except the following. Measurements. Due to the heavy distortion of the specimen all measurements except for those of the wings, should be treated rather as an estimate of the general size. Body length (head, without antenna, plus mesosoma plus metasoma) in dorsal view 10.4 mm. Head width in dorsal view 1.4 mm. Mesosomal length in lateral view 2.9 mm. Fore wing length 8.2 mm. Hind wing length 6.3 mm.
Fore wing. 1Rs cell about 0.85 × as long as cell 1R1. 1cua slightly curved, distad of M 1 by about 1.0 × the length of 1cu-a. Distance between 1cu-a and M 1 about 0.7 × the distance between 1cu-a and m-cu.

Note
This specimen resembles the female from Dominican amber. Unfortunately, many important characters, such as the size and position of the ocelli, form of the inner eye margin, etc., cannot be examined due to the insufficient preservation of the fossil. However, we do not include this specimen in the type series of R. hispaniola to emphasize its different geological origin and the slightly different wing venation as an indication of a separate distinct species which might be properly described and named after new material becomes available.
Currently, the genus is represented in Mexico only by Rhopalosoma simile Brues, 1943, and R. obliquum Townes, 1977(Townes, 1977. However, neither of these species has a similar hind wing Rs * .

Discussion
Based on the presence of a median tooth on the tarsal claws, females of Rhopalosoma are easily identified, and there is no doubt about the generic placement of the new species described herein. Townes (1977) divided Rhopalosoma into three species groups -the isopus, poeyi, and lanceolatum groups (comprising two, 14, and one species, respectively). Whereas the identification of Rhopalosoma at the species level remains a challenge, the new species likely belongs to the poeyi group based on the characters provided by Townes: front spur of mid tibia and hind tibia without narrow basal neck (as present in the lanceolatum group), and the form of tarsomere V, which is gradually tapered at its apical 0.65 (parallel-sided in the isopus group). Nevertheless, the new species has some affinities with a member of the isopus group, and with R. minus in particular, by having the hind wing Rs * straight and recurrent. But because this state is also found in Eorhopalosoma gorgyra, all Liosphex, and several Paniscomima (Lohrmann and Ohl, 2010;Lohrmann, 2011;Volker Lohrmann, personal observation, 2014), it can be assumed that this is the plesiomorphic character state compared to the angular or strongly bowed form found in many Recent species of Rhopalosoma. Townes (1977) regarded the isopus group as the "primitive species group". His assumption is based on the similarities of the morphology of the fifth tarsomeres with those found in Paniscomima, a genus he regarded as "definitely more primitive". Thus, the fossils likely do not represent the most archaic forms in the genus but represent either basal members of the poeyi group or an intermediate link between the isopus and the poeyi species group.

Note
The current contribution includes revised parts of a chapter used earlier in the PhD thesis of Volker Lohrmann, i.e., the description and discussion of the female of R. hispaniola (including Figs. 1d and 4b). Because the thesis includes a statement disclaiming all relevant nomenclatural acts from taxonomic use, the new taxon is formally described here to make its name available according to the ICZN (1999