Compression fossil Mymaridae (Hymenoptera) from Kishenehn oil shales, with description of two new genera and review of Tertiary amber genera

Abstract Compression fossils of three genera and six species of Mymaridae (Hymenoptera: Chalcidoidea) are described from 46 million year old Kishenehn oil shales in Montana, USA. Two new genera are described: Eoeustochus Huber, gen. n., with two included species, Eoeustochus kishenehn Huber (type species) and Eoeustochus borchersi Huber, sp. n., and Eoanaphes, gen. n., with Eoanaphes stethynioides Huber, sp. n. Three new species of Gonatocerus are also described, Gonatocerus greenwalti Huber, sp. n. , Gonatocerus kootenai Huber, sp. n., and Gonatocerus rasnitsyni Huber, sp. n. Previously described amber fossil genera are discussed and five genera in Baltic amber are tentatively recorded as fossils: Anagroidea, Camptoptera, Dorya, Eustochus, and Mimalaptus.


introduction
The family Mymaridae (Hymenoptera: Chalcidoidea) is represented by 103 genera and about 1400 nominal extant species in all terrestrial habitats and a few fresh water habitats (Huber 1986, Noyes 2010. Fossil Mymaridae have been described from amber inclusions from various localities, including five extinct genera and species from Cretaceous amber (Yoshimoto 1975, Huber andPoinar 2011) and eight genera (seven still extant, one extinct) from Tertiary amber, mostly from the Samland Peninsula, Kaliningrad district, Russia, and Chiapas, Mexico. Probably because of their small size, no compression fossils of Mymaridae have been discovered until now. Here we describe the first ones, representing three genera and six species. These and the Tertiary amber fossil genera are discussed and compared with extant genera.

Methods
A total of 17 specimens of fossil Mymaridae, all females, were collected in 2009 and 2010 at six sites (e.g. W113°42.173', N48°23.476') along the Middle Fork of the Flathead River in northwestern Montana between Paola and Coal Creeks in accordance with USFS Authorization HUN281. Fossil specimens were collected from the middle sequence of the Coal Creek member of the Kishenehn Formation, which has been estimated to be 46.2 +/-0.4 myo (Lutetian) by 40 Ar/ 39 Ar analysis and 43.5 +/-4.9 myo by fission-track analysis (Constenius 1996). Although the paper/oil shales of the middle sequence are thin (<1 mm to several mm), they were often easily split into even thinner pieces so as to expose unweathered surfaces on which the fossil insects reside.
The compression fossils were immersed in 95% ethanol for examination and photography. For figures 1-10 and 13-18, specimens were photographed using a Zeiss AxioCam MRc5 digital CCD camera mounted on a Zeiss Discovery V20 microscope and Zeiss AxioVision EDF software in Ottawa, ON. Measurements were generated using Zeiss AxioVision software. For figures 11 and 12, specimens were photographed using an Olympus SZX12 microscope, DP-25 camera and DPM imaging software in Washington, DC, and measurements were taken with the DP2-BSW software. All measurements are in micrometers (mm). If a measurement could not be made (e.g., scape length) it is represented by a dash.
Abbreviations used: fl x = funicle segment x. Measurements were taken as accurately as possible but given that the beginning and end points of a structure were not always clear or were hidden, the measurements may not be accurate. Appendage measurements are the most accurate except the wing bases cannot be clearly determined. Consequently, wing lengths were taken from the visible edge of the mesosoma and therefore they and length/width ratios are slightly smaller than they should be.
Fossils examined are in the following institutions: AMNH American Museum of Natural History. Description. Female. Body length 718-1133 (seven specimens in total, only four described and named to species). Head. Normal in shape, wider than long and about ¾ as high as wide. Face slightly convex in lateral view; vertex flat and slightly sloping anteriorly, forming a moderately sharp angle with occiput; back of head slightly concave. Eye higher than wide, about 2/3 head height; malar space about 1/3 eye height. Antenna. Funicle 6-segmented, with each funicle segment longer than wide; clava 3-segmented with the claval sutures perpendicular to claval length. Wings. Fore wing wide, symmetrical, with evenly rounded apex (shape resembling that of the extant genus Eustochus); marginal fringe shorter than fore wing width. Venation about 0.4× fore wing length, with long marginal vein and short but distinct stigmal vein. Hind wing narrow; marginal fringe much longer than wing width. Mesosoma. Shorter than gaster. Pronotum length at most about half length of mesoscutum. Mesoscutum length subequal to scutellum. Scutellum with frenum apparently entire, not divided longitudinally. Metanotum much shorter than scutellum. Metasoma. Constricted at base, probably with short petiole. Gastral terga similar in length. Ovipositor moderately short, probably arising near midpoint of gaster and its apex not or barely exserted beyond gastral apex.
Type species. Eoeustochus kishenehn Huber, sp. n. Derivation of generic name. The name is from eo-, Greek for early, and Eustochus, an extant genus. The gender is masculine.
Discussion. Two species are described, each from two specimens. Three additional specimens (Kishenehn #30,356,40,410,and 40,023) belong to Eoeustochus but are not included in the type series of either species because they are not in as good a condition.

Eoeustochus kishenehn
Derivation of species name. Named after Harold Borchers, professor of entomology and early mentor to the junior author at Bemidji State University, Bemidji, Minnesota.

Gonatocerus Nees Figs 9-16
Gonatocerus is a worldwide group with numerous described extant species classified in several subgenera (Triapitsyn et al. 2010). Its members are often the most commonly collected Mymaridae in almost any habitat so it seems surprising that it has been recorded only once as a fossil, in Baltic amber by Meunier (1905).
The three species described below definitely belong to Gonatocerus based on the entire clava, 8-segmented funicle, similar sized gastral segments, short and probably narrow petiole, and 5-segmented tarsi (at least in the one species where they can be counted). Because they all appear to have a rhomboidal dorsellum they would be classified either in G. (Gonatocerus) or in G. (Cosmocomoidea), but not G. (Lymaenon), the most common extant subgenus because it has a narrow, strap-shaped dorsellum or G. (Gastrogonatocerus) because the ovipositor does not project forward under the mesosoma.
Comments. Gonatocerus kootenai differs from the following two species by the relatively narrower wings and shorter funicle segments. I tentatively place it in G.
(Gonatocerus) because of the relatively narrow fore wing and fairly long and relatively slender fl 3 -fl 5. The reverse side of the shale piece contains a specimen of Trichoptera.
Derivation of species name. Named after the Kootenai tribe of the Flathead Nation in northwestern Montana, site of the Kishenehn shales. Comments. Gonatocerus rasnitsyni differs from G. kootenai by the relatively smaller size, shorter and wider fore wing, and colour of the antennal pedicel and legs. I tentatively place it in G. (Cosmocomoidea) because of the relatively wider fl 3 (on one antenna at least, suggesting it bears multiporous plate sensilla) compared to the slightly narrower fl 2 and fl 4. This resembles several extant members of this subgenus that also have alternately wider and narrower basal funicle segments.

Gonatocerus rasnitsyni
Derivation of species name. Named in honour of A. P. Rasnitsyn, the world's foremost Hymenoptera palaeontologist, on the occasion of his 75 th birthday. Description. Female. Colour dark brown except apex of pedicel and legs beyond coxae lighter (yellowish). Holotype (Fig. 14) measurements as follows. Body length 926. Antenna (Fig. 15)  Comments. G. greenwalti differs from G. kootenai and G. rasnitsyni by the wider fore wing and thicker funicle segments. The apparent absence of microtrichia behind the venation, the wide fore wing, and fairly uniformly thick funicle segments suggest that G. greenwalti should be classified in G. (Cosmocomoidea).The shale fragment in which the fossil occurs contains several aquatic insects (e.g., Notonectidae), an indication of the lacustrine environment in which the mymarid lived (though it is not aquatic itself ).
Derivation of species name. Named by the senior author in honour of the junior author, Dale Greenwalt, who collected and curated the insect fossils from Kishenehn shale. Description. Female. Body about 700. Head. About as wide as high (measured in frontal view). Malar space long, almost eye height. Antenna. Funicle 6-segmented, with fl 1 very short and remaining segments longer than wide; clava 3-segmented with the claval sutures almost perpendicular to claval length. Wings. Fore wing wide, slightly asymmetrical, with somewhat truncate apex; marginal fringe much shorter than fore wing width. Venation almost 1/3 fore wing length, with marginal vein fairly short, about as long as stigmal vein. Hind wing narrow; marginal fringe much longer than wing width and along posterior margin extending to base of membrane. Mesosoma. About 0.75 x gaster length. Mesoscutum length shorter than scutellum. Metanotum much shorter than scutellum, apparently with strap-like dorsellum. Metasoma. Apparently slightly constricted at base, probably with fairly wide, short petiole. Gastral terga similar in length. Ovipositor moderately short, apparently arising near base of gaster and its apex not exserted beyond gastral apex.
Type species. Eoanaphes stethynioides Huber, sp. n. Derivation of generic name. The name is based on the extant genus Anaphes, which is also known from one extinct species in Baltic amber. The gender is masculine.
Discussion. Eoanaphes appears to be related to two genera, Anaphes and Stethynium. Four features suggest Anaphes: 1) in lateral view there is a distinct constriction dorsally between the mesosoma and metasoma, indicating that the propodeum slopes strongly down relative to the horizontal scutellum and dorsellum (in Stethynium, the entire dorsal margin of the mesosoma is horizontal and in line with the base of the metasoma, without a depression between the two parts); 2) the very short fl 1 ; 3) the apparent absence of a distinct rounded lobe on the posterior margin of the fore wing opposite the marginal + stigmal veins (one wing appears to have a rounded lobe but this is due to the membrane being partly folded over on itself ); 4) the wing surface behind the marginal vein seems to be bare except for a few setae behind the stigmal vein and a faint suggestion of a hair line separating medial from marginal space. Two features suggest Stethynium: 1) the clearly 3-segmened clava; 2) the strap-like dorsellum (rhomboidal in Anaphes). Description. Female. Colour dark brown; pedicel, basal two funicle segments, and tarsi except perhaps apical tarsomere yellowish. Holotype (Fig. 17)

Discussion
Compression fossils in sedimentary rock are less informative morphologically than amber fossils. Fine details such as body setation and sculpture are not visible (wing setae may be visible), and body flattening during fossilization usually results in distortion. Measurements of body parts may therefore not be accurate though they can sometimes be more easily measured because they are flat. Appendages are the best source of useful diagnostic characters because some (legs, antenna) are less and some (wings) are not distorted. Because of the preservation method, compression fossils cannot be easily compared to amber fossils or extant genera and species. Nonetheless, they can be di-agnosed moderately well and differentiated with reasonable certainty from each other and from other extinct and extant genera.
Among the 17 Kishenehn mymarids found, Eoeustochus is the most common, followed by Gonatocerus and Eoanaphes. It is surprising that no males were found; in extant genera males are less common than females but they are rarely almost completely absent. The compression fossils are of middle Eocene (Lutetian) age, as indicated above, and so are most Baltic amber fossils (Weitschat and Wichard 2010), i.e., about 41-49 my old. Incidentally, Meunier (1901), quoting A. Jentzsch, had noted that their age was lower Eocene, not lower Oligocene (yet the latter was stated again by Meunier 1909 andDoutt 1973). The Eocene had a warm humid climate, significantly warmer than the present (Zachos et al. 2001), allowing for development of subtropical and tropical rain forests to which the Baltic amber forests belong (Weitschat and Wichard 2010). Baltic amber forests were apparently rich in lentic water, flood plains, ponds, lakes with littoral habitats and temporary micro waters. Aquatic insects make up 25% of Baltic amber inclusions (Weitschat and Wichard 2010). The Middle sequence of the Coal Creek member of the Kishenehn Formation consisted of lacustrine sediments of a large permanent and shallow lake and associated lakeside and/or marsh, and has a high proportion of aquatic insect fossils (e.g., > 60% of all the fossil insects are Corixidae and Chironomidae). Because the climate and habitat of both Baltic and Kishenehn fossils was apparently similar it is not surprising that the genera of Mymaridae found in each are similar. Meunier (1901) described most Baltic amber fossils of Mymaridae. Doutt (1973) described a few species from Mexican amber (Simojovel, Chiapas), dated as 15-20 my, and Thuróczy (1983) described one species. There is no indication from morphology that any of these or the mymarid compression fossils form a link between Cretaceous and Tertiary/Quaternary genera and species. Instead, they add support to the analysis by Rasnitsyn and Kulicka (1990) that showed Hymenoptera assemblages from Baltic amber seem to be more similar to the extant fauna than to the Late Cretaceous fauna.
One clear morphological tendency from Cretaceous to Tertiary and Quaternary (present) species can be seen -a reduction in the number of funicle or claval segments in females. A 3-segmented clava occurs in three of five (60%) Cretaceous genera, four of ten (40%) previously reported Tertiary genera from amber and shale (reported above), and 13 of 103 (about 8%) currently recognized extant genera. An 8-segmented funicle occurs in three of five (60%) Cretaceous genera, one of the 10 (10%) previously reported Tertiary genera, and five of the 103 (about 5%) extant genera. Except for Gonatocerus, which has eight funicle segments but an entire clava, all other Tertiary fossils have six (rarely fewer) funicle segments in females. Spahr (1987) catalogued the literature and listed 11 mymarid genera from amber, three of them Cretaceous and eight Tertiary. He followed previous authors by including genera and species now correctly classified in Mymarommatidae (Gibson et al. 2007). Witsack (1986) described Palaeopatasson for one species from Dominican amber.
The senior author examined 16 Baltic amber pieces (SVT, UCRC) in 2005, 13 pieces (AMNH) in 2011, and three pieces (IPMGö., ZMUC) in about 2005. Only a few specimens could be referred confidently to an extant genus. These were four specimens of Gonatocerus, and (less confidently) three specimens of Anaphes and two of Stethynium. If correctly identified, five more genera, all known from the extant fauna, are reported here for the first time as amber fossils. Having examined these 32 additional amber specimens and realizing that most of them cannot definitely be classified in an extant genus, the senior author has doubts about the correct generic placement of at least some of the specimens studied by past workers. None of those specimens were examined, however. Each Tertiary fossil genus is discussed briefly below.
Anagroidea (1 female, C.V. Henningsen, B-1 1956, ZMUC, examined) is no longer known from extant species in Europe and is rare in the Holarctic region (eastern Asia and south east USA).
Alaptus is known from at least two species in Miocene amber (Doutt 1973) that are probably correctly identified to genus.
Anaphes was known from three specimens in Baltic amber. Anaphes splendens Meunier is probably incorrectly classified. According to Meunier the ovipositor extends appreciably beyond the apex of the metasoma, unlike any extant Anaphes. The fore wing with a slight but distinct ventral lobe, its overall shape, the venation with a distinct stigmal vein, and the antenna with fl 1 almost as long as fl 2 also do not fit the genus. Anaphes schellwieniens Meunier is described from a male so it is impossible to determine its placement. Doutt (1973) examined a male from Chiapas amber and tentatively placed it in Anaphes. Three specimens examined here are fairly confidently placed in Anaphes (1 female, UCRC; 2 females, SVT). Additional specimens (2 females, SVT) examined are tentatively placed in Anaphes.
Arescon is known from two species in amber (Meunier 1901(Meunier , 1905 that are probably correctly placed because of their 5-segmented funicle, though fore wing venation length, which is important information for correct generic placement, was not given by Meunier. ?Camptoptera (1 female, SVT), examined in 2005, if correctly classified, would be the first fossil specimen for this worldwide genus.
?Dorya (1 female, UCRC) is known only from extant species in Australia and New Zealand.
?Eustochus (1 female, AMNH). The gaster appears to be petiolate. This suggests a Eustochus with a distinct 3-segmented clava (most have a 2-segmented clava) though the wing shape and antenna suggests Eoeustochus.
Gonatocerus was known from only one fossil specimen, Gonatocerus henneberti (Meunier 1905), which appears to belong to either G. (Gonatocerus) or G. (Lymaenon) as suggested by the narrow wings. However, the "écusson semilunaire" [semilunate shield], which I interpret from Meunier's drawing as being the dorsellum, appears to be rhomboidal in shape, which would eliminate G. (Lymaenon) as the correct subgenus. The small size of the specimen, only 1/3 mm long, is unusual; it is much smaller than the smallest extant Gonatocerus I have seen. A specimen of Gonatocerus (Cosmocomoidea) is reported here (1 male, Poinar collection, OSU) from Dominican amber, 15-20 my.
Litus is represented by two fossil species. Litus elegans Meunier (1901) may or may not be correctly classified because the description and illustration are insufficient to place it. It apparently differs considerably from Litus mexicanus Doutt (1973), which is likely correctly classified. Extant Litus species tend to have a bi-geniculate funicle, unlike that illustrated by Meunier. Litus beneficus Meunier from copal (recent) from Madagascar also needs to be re-examined; it is unlikely to be classified correctly.
Malfatia molitorae Meunier (1901) is based on a male and from the description alone it is impossible to determine how it relates to extant genera. Its status remains uncertain.
Palaeopatasson grollei Witsack (1986) is based on a female and may be related to Anaphes.
Polynemoidea mexicana Doutt (1973) from Chiapas amber was stated to be similar to P. domestica Girault, but the latter species is likely incorrectly placed in Polynemoidea (Lin et al. 2010). Consequently, I suspect that P. mexicana is also incorrectly classified; it does not have a strongly exserted ovipositor as in the type species of Polynemoidea.

Conclusions
1. Tertiary fossils of Mymaridae are clearly more related to the extant fauna than to the Cretaceous fauna. 2. Except for Gonatocerus and the two new genera described above, tertiary fossils of Mymaridae (except perhaps Anaphes and Stethynium) are for the most part doubtfully assigned to extant genera. 3. Several Eocene amber fossils, if correctly identified to genus, represent genera that no longer occur in Europe (Anagroidea, Dorya, Mimalaptus) or are very poorly represented in most of the Holarctic region (Stethynium). These genera are now distributed mostly in the tropics or southern hemisphere, supporting observations that the Eocene climate (in parts of the Holarctic region at least) was considerably warmer than at present. 4. If correctly identified, Anagroidea, Eustochus, and perhaps Eoeustochus would be the first fossil records of genera belonging to the extant tribe Mymarini sensu Annecke and Doutt (1961), likely the most derived lineage in Mymaridae.
collection for study of the senior author. These were then purchased by AMNH, where they will be deposited.