The First Fossil Representatives of the Sawfly Genera Emphytus and Empria from the upper Miocene of France (Hymenoptera, Tenthredinidae)

Simple Summary Two sawflies belonging to the tenthredinid genera Emphytus and Empria are described from upper Miocene diatomite paleolakes from Southern France. They are compared with all their extant and fossil relative genera. The previously described fossil representatives of these two genera are discussed. These fossils are of great interest for dating in future phylogenetic analyses of the Tenthredinidae. Abstract Emphytusmiocenicus sp. nov., first fossil representative of this genus, is described from the upper Miocene of the diatom paleolake of Montagne d’Andance (Ardèche, France). Its placement is ensured by an in-depth comparison with all the extant and fossil genera of the subfamily Allantinae. The representatives of Emphytus are distributed in the Palearctic, Nearctic, and Oriental regions. Empria sammuti sp. nov., second representative of the latter genus, is described from the latest Miocene of the diatom paleolake of Sainte-Reine (Cantal, France). The placement of this new species is based on a detailed comparison with the extant genera of the tribe Empriini. The larvae of the extant Emphytus and Empria spp. are known to be phytophagous on angiosperm leaves of several families, all present as fossils in the taphocenoses of la Montagne d’Andance and Sainte-Reine. Emphytus miocenicus sp. nov. represents the oldest record of this genus and of its crown group, corroborating the estimate of a middle Eocene–middle Oligocene age for its stem group. Throughout our study, it appears that the first described fossil of the genus Empria, E. oligocaenica, from the Oligocene of Germany, needs to be revised and redescribed. It should preferably be treated until the revision as incertae sedis in Allantinae sensu lato.


Introduction
The family Tenthredinidae is the most speciose extant sawfly family [1]. Nevertheless, its fossil record remains scarce, especially for the Cretaceous, even if recent phylogenetic analyses have suggested a Cretaceous age for the diversification of the family into its modern subfamilies, especially for the Allantinae [2,3]. Described fossil Allantinae comprise

Material and Methods
The holotype of Emphytus miocenicus sp. nov. originates from the Miocene Konservat-Laggerstätte of the Montagne d'Andance (Saint-Bauzile, Ardèche, France). It was found ca. 35 years ago. The diatomitic unit from which the specimen originated was deposited in a lake formed by a maar crater [5] under a warm and humid climate [6], with possibly seasonal character. This lake was probably calm and shallow [7], surrounded by an immense heterogeneous forest with an undergrowth of ferns [8,9]. Over time, this lake was filled with diatomite, forming a rock strata up to 30 m thick in some localities [10]. This diatomite is a soft siliceous rock in which was preserved the insects that drowned in the lakes at the time. These stratigraphic layers of diatomite represent a reducing environment that has allowed a remarkable conservation of organic matter that is mummified. The age of this deposit corresponds to the Tortonian-Messinian boundary, ca. 7.30 ± 0.15 Ma [11].
The holotype of Empria sammuti sp. nov. originates from the Miocene Konservat-Laggerstätte of Sainte-Reine (Fourfouilloux, Virargues village), near Murat (Cantal, France). The diatomitic unit from which the specimen originated was deposited in a lake probably formed by a maar crater, under a rather high altitude (ca. 1000 m). The age of this deposit corresponds to the latest Miocene, ca. 5.60 ± 0.3 Ma [12]. After Roiron [13], the flora indicates an important cooling for the period.
The specimens were studied using a stereomicroscope Nikon SMZ25 and the photographs were taken with a Nikon D800 mounted on the previously mentioned stereomicroscope or using a Canon 50D with an attached Canon 65 MPE camera lens and mounted on an automated stacking rail (StackShot). The photographs were treated with graphic software. All images are digitally stacked photomicrographic composites of several individual focal planes, which were obtained using Helicon Focus 6.7. The figures were composed with Adobe Illustrator CC2019 and Photoshop CC2019.
We follow the tenthredinid classification proposed by Wei and Nie [14] rather than Taeger et al. [1] for the treatment of the genus Emphytus. Niu et al. [2] recently separated a subfamily Megabelesinae from the Allantinae on the basis of a molecular phylogenetic analysis. We follow the wing venation nomenclature of Goulet and Huber [15].

Results
Systematic paleontology.     Taxonomic remarks Following the key to tenthredinid subfamilies of Goulet [16], the new fossil falls in the Allantinae because of the following characters: "fore and hind wings as long as or longer than length of abdomen"; "veins lm-cu and Cu, of forewing meeting at an angle of 140°"; "forewing with main axis of vein M and that of 1m-cu subparallel"; "vein R of forewing not deviated between junctions of M and Sc"; "vein 2A and 3A of forewing complete and markedly constricted"; "crossvein of anal cell of forewing developed and long"; "vein Rs+M of forewing near junction with R straight". Following Haris's [17] key to tenthredinid subfamilies, the Nematinae and the Heterarthrinae are excluded because of the urn:lsid:zoobank.org:act:1697F94A-5445-4E15-8D7F-13D4C6835E71 Etymology. Named after the Miocene period. Material. Holotype specimen IGR-PAL-2859 (a compression of a nearly complete adult, but with head poorly preserved), housed in the Geological Department and Museum of the University of Rennes, France (IGR).
Diagnosis. A rather large species (compared to the extant European species), with body length ca. 11.0 mm; wings uniformly infuscate; pterostigma unicolored, dark brown; abdomen uniformly black; all femora, tibiae, and tarsi uniformly black; forewing vein 3A reaching posterior wing margin.
Color. As the fossil is a compression mummy, the differences in the color pattern have some significance because they reflect the original differences in coloration. Body black with faint traces of brown on the abdomen; compound eyes brown; legs black; wings infuscate with apical third slightly darker than basal two-thirds.
Head poorly preserved, ca. 2.0 mm long, 2.0 mm wide; compound eyes 0.8 mm apart; mouthparts not preserved; seven antennomeres preserved but antennae not complete; third antennomere as long as fourth. Thorax ca. 3.3 mm long, 2.4 mm wide, mainly smooth and shining, covered dorsally with very small punctuations. Abdomen shining without surface sculpture, 2.4 mm wide, but apex missing. Mid legs with femur 1.5 mm long, tibia 1.55 mm long, tarsus 2.0 mm long, longer than tibia; inner tibial spur 0.3 mm long; claw with an inner tooth 0.2 mm long, apical tooth 0.4 mm long, basal lobe small but distinct, rounded; hind legs with femur 1.9 mm long, tibia 2.7 mm long, tarsi not preserved. Forewing 7.7 mm long, 2.7 mm wide; pterostigma 1.7 mm long, 0.4 mm wide; no first free sector of Rs between the cells 1R1 and Rs; cell 1R1 + Rs 1.9 mm long; cell 2Rs 1.2 mm long; nervulus (1cu-a) situated just beside middle of cell 1M; vein a oblique and elongate; vein 3A reaching posterior wing margin. Hind wing 5.6 mm long, 1.7 mm wide; cells Rs and M opened; apex of cell 1A 0.4 mm beside apex of cell Cu; vein 3A very faint but reaching posterior wing margin.
Taxonomic remarks: Following the key to tenthredinid subfamilies of Goulet [16], the new fossil falls in the Allantinae because of the following characters: "fore and hind wings as long as or longer than length of abdomen"; "veins lm-cu and Cu, of forewing meeting at an angle of 140 • "; "forewing with main axis of vein M and that of 1m-cu subparallel"; "vein R of forewing not deviated between junctions of M and Sc"; "vein 2A and 3A of forewing complete and markedly constricted"; "crossvein of anal cell of forewing developed and long"; "vein Rs+M of forewing near junction with R straight". Following Haris's [17] key to tenthredinid subfamilies, the Nematinae and the Heterarthrinae are excluded because of the presence of the basal part of vein 2r-rs in forewing. Affinities with the Susaninae are also excluded because of the very different shape of the anal veins and base of Rs at point of fusion of M with R [18]. The Dolerinae are excluded because the new fossil has a vein 2r-m separating the cell 1R1 + Rs (vein Rs missing) from cell 2Rs. The subcosta straight, without widened area in front of pterostigma, excludes the Tenthredininae. The Selandriinae are excluded because of the presence of an anal crossvein a, and the base of Rs+M near Sc+R straight [19]. The Blennocampinae are excluded because the anal cell of the forewing is complete, without stem, and the crossvein of the anal cell is elongate in the new fossil. Lacourt [20] proposed a concept of the Blennocampinae that comprises several genera listed in the Allantinae by Taeger et al. [1]. Thus, it appears that there is no real consensus on the classification of the Tenthredinidae and that new phylogenomic analyses will help to clarify the systematics of the family.
The forewing 2A + 3A separated from 1A by a distinct anal vein, the veins M and lm-cu subparallel, and the forewing veins lm-cu and Cu1 forming an angle of 140 • , present on the specimen studied here, are characteristic of the Allantinae present in the new fossil and corroborate the keys used to tentatively place this fossil.
Within the Allantinae, affinities with the Eriocampini (Eriocampa Hartig, 1837, Pseudosiobla Ashmead, 1898, Dimorphopteryx Ashmead, 1898, Armitarsus Malaise, 1931, Corymbas Konow, 1903, etc.) are excluded because of the veins M and Rs+M meeting Sc+R at the same point, plus the mesoscutellum and the mesonotum without deep punctures in the new fossil, and the affinities with the Acidiophorini (Acidiophora Konow, 1899) are excluded because the pterostigma is rather broad and the tarsal claws are bifid [21,22]. The Adamasini (genus Dinax Konow, 1897 = Adamas Malaise, 1945) are excluded because of their simple claw, the short crossvein a, and the presence of the first free sector of Rs in forewing [23,24].
The main differences between the Empriini and Allantini concern the clypeus, the mandibles, and the thorax, which are not or poorly preserved in the new fossil. Thus, we proposed an in-depth comparison with all the genera encompassed by these two tribes.
The two characteristics "forewing first free sector of Rs missing between 1r-rs and M+Rs" and "vein 3A relatively developed reaching posterior wing margin" exclude affinities with Indostegia Malaise, 1934 [29]. Haymatus Smith, 1979 has a hind wing cell M and a forewing first free sector of Rs present, unlike the new fossil [30]. Somanica Smith, 1979 Ermolenko, 1968) have also a forewing first free sector of Rs [62][63][64][65]. Monostegia and Monsoma further have a hind wing cell M [30,66], preventing attribution of the new specimen to these genera. Harpiphorus Hartig, 1837 has the forewing cell M with a distinct dorsal petiole and the cell M of hind wing closed [24]. Conobeleses Wei, 1997a has also a forewing cell M with a distinct dorsal petiole [25]. The genera Athalia Leach, 1817 and Hypsathalia Benson, 1962 generally have also such petiole, a forewing first free sector of Rs, and hind wing cell Rs and M [67]. Hennedyella Forsius, 1935 and Hennedyia Cameron, 1891 have a small inner tooth of the tarsal claws, unlike the new fossil [67]. Ametastegia Costa, 1882 (including Ocla Malaise, 1957) differs from the new fossil because of the vein a clearly less oblique, and a very long third antennomere [30]. The presence vs. absence of the forewing first free sector of Rs is variable in this genus. Ocla has the hind wing without closed cell M, but its nervulus (1cu-a) is curved and the basal lobe of claw is more pronounced than in the new fossil [32].
The species of the genus Empria Lepeletier, 1828 have the forewing first free sector of Rs absent or present and the hind wing cell M opened or closed, tarsal claws with or without an inner tooth, but they have paired opalescent white spots on abdominal terga 2 to 5 or more [30], while the new fossil has none.
The two genera Tritobrachia Enderlein, 1920 and Nagamasaia Togashi, 1988 of the subtribe Nagamacina have claws without basal lobe, unlike the new fossil. Tritobrachia has no forewing first free sector of Rs, while Nagamasaia has one [72,84]. Empronus Malaise, 1935 has also claws without basal lobe, plus a closed hind wing cell M [85]. Mallachiella Malaise, 1934 has also a very minute basal lobe of claw [86]. Hemkuntus  has no basal lobe of claw [87]. Hemocla Wei, 1995 shares with the new fossil the absence of the forewing first free sector of Rs, and the opened hind wing cell M, but it also has no basal lobe of the claw [88].
Macremphytus MacGillivray, 1908, Asiemphytus Malaise, 1947, Parabeleses Wei & Nie, 1998, Filixungulia Wei, 1997, and Plumalaminia Wei, 1997 share with the new fossil the absence of the forewing first free sector of Rs, but they have a closed hind wing cell M [30,47,55,[89][90][91][92]. Apethymus Benson, 1939 also shares with the new fossil the absence of the forewing first free sector of Rs, but it has the teeth of its tarsal claws nearly of the same lengths [93,94].
The two genera Athlophorus Burmeister, 1847 and Hemathlophorus Malaise, 1945 of the subtribe Athlophorina sensu Wei & Nie [49] (=tribe Athlophorini sensu Lacourt [22]) have no forewing first free sector of Rs, and hind wing without closed cell M. Athlophorus differs from the new fossil in the forewing veins M and Rs + M not meeting Sc + R at the same point [94][95][96]. Hemathlophorus differs from the new fossil in a venation similar to that of Athlophorus and the subapical tooth of the claw longer and stronger than the apical one [88,97]. Stenempria Wei, 1997 and Fernophytus Wei, 1997 have also no forewing first free sector of Rs and hind wing without closed cell M, but the former differs from the new fossil in the claw without inner tooth and basal lobe, and the later in the claw without basal lobe and inner tooth very small and remote from apical tooth [90,98].
Paralinomorpha Koch, 1988a has no forewing first free sector of Rs, but differs from the new fossil in the very large basal lobe of the claw and the hind wing cell subclosed [99]. Yushengliua Wei & Nie, 1999b also has a larger basal lobe of the claw than in the new fossil [71,100]. Nervobeleses Wei et al., 2006 (new name for Doleroides Nie & Wei, 1998) also has no forewing first free sector of Rs but differs from the new fossil in the two teeth of the claw of the same lengths and a strong basal lobe [101].
Allantus  [103] has shown that some specimens of Emphytus can have this vein), the hind wing cells Rs and M opened, and the basal lobe of the claw small and rounded [104]. Allantus differs from the new fossil in the nervulus 1cu-a of the forewing close to the base of vein M [71,104]. Emphytus Klug, 1815 has this vein in a more distal position, meeting the cell 1M at basal third, as in the new fossil. Both Allantus and Emphytus have acute basal lobes of the claw, unlike Maghrebiella. The shape of the claw in the new fossil strongly resembles that in Emphytus and strongly differs from those of Allantus and Maghrebiella [104]. Taxonemphytus and Mimathlophorus have an inner tooth stronger and longer than the apical one, unlike the new fossil [88]; also, the apex of the hind wing cell 2A is touching that of cell 1A in these two genera, unlike in the new fossil. Linomorpha has a very particular cell 2Rs with a very short crossvein 2r-m.
Apethymus and Hemiphytus Malaise, 1947 have claws and a venation very similar to that of the new fossil, but the forewing vein 3A is absent in Apethymus, and the length of cell 1R1+Rs is twice as long as cell 2Rs in Hemiphytus, while it is only 1.75 times as long as cell 2Rs in the new fossil [88,105]. Taxonemphytus has the inner tooth of claw longer than apical one [88]. Parallantus has the two teeth of the claws of the same length and the forewing anal cell without crossvein, unlike the new fossil [49].
Stenemphytus  has no forewing first free sector of Rs, the hind wing cells Rs and M opened, but it differs from the new fossil in the cell1R1 + Rs two and a half times as long as the cell 2Rs, while they are only 1.75 times as long as cell 2Rs in the new fossil [79].
The latest Oligocene genus Hylotomites Meunier, 1914 was originally placed near the argid genus Hylotoma, but Taeger et al. [1] placed it in the Allantinae without giving arguments. Therefore, as far as we know, this genus has never been revised. It differs from the new fossil in the forewing cell M with a distinct dorsal petiole ("cellule radiale distinctement appendiculée") [106]. The fossil genus Palaeotaxonus Brues, 1908 (lowermost Oligocene of Florissant Fm., Colorado, USA) has a forewing first free sector of Rs [107]. Palaeathalia Zhang, 1985 (Early Cretaceous, Laiyang Fm., Liaoning, China) was originally considered as a Tenthredinidae [108], placed in the Allantinae by Abe and Smith [109] and Taeger et al. [1], but considered as a Tenthredinoidea by Ronquist et al. [110]. Palaeathalia also differs from the new fossil in the presence of a forewing first free sector of Rs [108].
The new fossil shares with Emphytus a majority of its preserved characters to the exclusion of the other allantine genera. Taeger et al. [1] considered that Emphytus is a subgenus of Allantus, but Lacourt [104] separated the two taxa Allantus and Emphytus, followed by Magis [111], Pesarini [112], and Wei and Nie [14]. Here, we follow this treatment and propose to include it in the genus Emphytus.
Tribe Etymology. Named after Claude Sammut, who allowed us to study the type specimen and donated to the University of Rennes. It is to be treated as a noun in the genitive case.
Material. Holotype specimen IGR-PAL-2860 (a compression of a nearly complete adult), housed in the Geological Department and Museum of the University of Rennes, France (IGR).
Diagnosis. Body black except brown lateral parts of abdominal tergites; wing infuscate with costal margin black; hind femur brown, hind tibia with basal half brown and apical one black, hind tarsus black; anterior margin of clypeus has two weak lateral bumps separated by a very weak concavity; forewing first free sector of Rs between cells 1R1 and 1Rs present.
Description. Body length 10.2 mm. Color. As for above, the differences in the color of the different parts of the body reflect the original differences because it is a compression mummy. Body black with lateral traces of brown on abdomen; compound eyes black; legs brown with hind tarsus darker than femur and tibia; wings infuscate with costal margin darker.  Head well preserved, 1.3 mm long, 1.8 mm wide; compound eyes large, 0.5 mm apart; labrum 0.1 mm long, covering the mandibles; anterior margin of clypeus has two weak lateral bumps separated by a very weak concavity; antenna 3.2 mm long, with nine antennomeres preserved, third as long as fourth, longer than more distal ones. Thorax 4.0 mm long, 2.5 mm wide, mainly smooth and shining, covered dorsally with very small punctuations; medial length of metapostnotum is 0.5 of its maximal length. Abdomen shining without surface sculpture, 5.4 mm long, 2.7 mm wide; median part of tergites sclerotized and black, lateral sides brown and less sclerotized; short ovipositor present but poorly preserved. Only hind legs visible, femur 1.4 mm long, tibia 1.4 mm long, tarsus 2.7 mm long, longer than tibia; claws not preserved. Forewing 7.5 mm long, 2.8 mm wide; pterostigma 1.4 mm long, 0.5 mm wide; first free sector of Rs between cells 1R1 and 1Rs present; cell 1R1 0.5 mm long; cell 1Rs 1.0 mm long; cell 2Rs 1.3 mm long; nervulus (1cu-a) situated at middle of cell 1M; vein a oblique and elongate. Hind wing 5.0 mm long; cells Rs and M opened; cells 1A and Cu not visible.
Taxonomic remarks: Following the key to tenthredinid subfamilies of Goulet [16], the new fossil falls in the Allantinae because of the following characteristics: "fore and hind wings as long as or longer than length of abdomen"; "veins lm-cu and Cu, of forewing meeting at an angle of 120 • -150 • "; "forewing with main axis of vein M and that of 1m-cu subparallel"; "vein R of forewing not deviated between junctions of M and Sc"; "vein 2A and 3A of forewing complete and markedly constricted"; "crossvein of anal cell of forewing developed and long"; "vein Rs + M of forewing near junction with R straight". After this key, the only character that precludes affinities with the Allantinae and rather suggests affinities with the Selandriinae is "veins lm-cu and Cu, of forewing meeting at an angle of 120 • ", while it would be between 140 • and 150 • in Allantinae. However, this character is not stable in the Allantinae (ca. 120 • in Dinax Konow, 1897) [29]. The Selandriinae are excluded because of the presence of an anal crossvein a, and the base of Rs + M near Sc + R straight [19]. Within the Allantinae, affinities with the Eriocampini are excluded because of the veins M and Rs + M meeting Sc + R at the same point, plus the mesoscutellum and the mesonotum without deep punctures in the new fossil. Affinities with the Acidiophorini (Acidiophora Konow, 1899) are because the pterostigma is rather broad. The Adamasini are excluded because of their short crossvein a [23,113]. After Goulet [16], the Allantini are characterized by "medial length of postnotum of metathorax 0.8 or more of maximal length of postnotum, about halfway between side and midline" and "anterior margin of clypeus widely concave" vs. "medial length of postnotum of metathorax 0.7 or less of maximal length of postnotum, halfway between side and midline" and "anterior margin of clypeus less widely concave than above" in the Empriini. In the new fossil, the medial length of metapostnotum is 0.5 of its maximal length, and the anterior margin of clypeus has two weak lateral bumps separated by a very weak concavity, in accordance with an attribution to the Empriini.
Within this tribe, affinities with the genera Monosoma MacGillivray 1908, Monostegia Costa 1859, Phrontosoma MacGillivray 1908, Plumalaminia , Harpiphorus Hartig, 1837, and Haymatus Smith, 1979 are excluded because the hind wing cell M is opened in the new fossil [24,30,90]. Stenempria Wei, 1997 has cells 1R1 and 1Rs fused [90]. Blennallantus Wei, 1998 and Allanempria Wei, 1998 have a flat clypeus as in the new fossil, but a metapostnotum flat, medially not constricted, unlike the new fossil [26]. Dasmithius Xiao, 1987 has a very shallowly and broadly circularly emarginated clypeus, and cell 2Rs as long as cell 1Rs, unlike the new fossil [35]. Ungulia Malaise, 1961 has the forewing cell 2Rs as long as cell 1Rs and the anal vein nearly perpendicular; Oralia Malaise, 1961 (with junior synonym Himindica  has also a nearly perpendicular anal vein, and the fourth antennomere as long as fifth, forewing vein 3A relatively developed reaching posterior wing margin, and the hind wing vein 3A well developed and elongate [33]. Somanica Smith, 1979 and Aphilodyctium Ashmead 1898 have a more pronounced concavity of the clypeus than in the new fossil [30]. Protemphytus Rohwer 1909 has the forewing cells 1R1 and 1Rs combined [114]. Hemiphytus Malaise, 1947 and Fernophytus Wei, 1997g also have the forewing cells 1R1 and 1Rs combined and the third antennomere shorter than the fourth [88,98]. Hemocla Wei, 1995 has the forewing cells 1R1 and 1Rs combined and the third antennomere longer than the fourth [87]. Heptapotamius Malaise 1935 has the third antennomere longer than the fourth [31]. Kattakumia Zhelokhovtsev, 1964 (Mongolempria Ermolenko 1968) has a venation rather similar to that of the new fossil, but with a distinctly shorter forewing cell 2Rs [61].
The new fossil has a clypeus and a wing venation quite similar to those of Ametastegia Costa 1882, but its third antennomere is as long as the fourth, while it is distinctly longer in the latter. In addition, the forewing anal crossvein is distinctly less oblique in Ametastegia than in the new fossil [30]. Phrontosoma MacGillivray 1908 has a clypeus very shallowly emarginated as in the new fossil, but a third antennomere subequal in length to fourth plus fifth antennomeres, and a closed hind wing cell M. Allantunicus Smith in Smith & Schiefer, 1997 and Gulingia Wei et al., 1997 have a third antennomere longer than fourth [43,115].
Some species of Empria Lepeletier & Serville 1828 have a clypeus very similar to that of the new fossil; also, the third antennomere is of the same length as the fourth or nearly so, and the wing venation of the new fossil fits well with that in this genus. The Empria spp. have paired opalescent white spots on abdominal tergites 2 to 5 or more, a character that allows an easy attribution of sawflies to this genus [30]. In the case of the new fossil, the presence of a dark, sclerotinized zone on the median part of the tergites together with desclerotinized lateral parts is in accordance with the presence of such opalescent spots.
Thus, we consider that this new fossil can be attributed to the genus Empria.
The numerous extant species of Empria can be separated using the shape of the claws, head, ovipositor, and colorations, characters hard to determine in the new fossil. Nevertheless, the absence of the hind wing cell M would suggest affinities with the species of the "candidata-group" (=subgenus Parataxonus comprising E. candidata and E. multicolor) [27,30,65].
The unique fossil currently listed in the genus Empria by Taeger et al. [1] and Prous [27] is Empria oligocaenica (Meunier, 1923) from the Oligocene of Rott (Germany) [116]. It was originally attributed to the genus Dolerus under the name of Dolerus oligocaenicus. Statz [117] excluded this fossil from Dolerus and attributed it to Leucempria Enslin, 1913, a subgenus of Empria after Ross [118], on the basis of wing venation characters only, which are not sufficient to exclude several other empriine genera, while nothing is known on the antenna and the clypeus. The current location of the type specimen is unknown to us. The original description of Meunier [116] does not give any character that could be diagnostic of Empria ss. The original photograph of Meunier does not allow to give a precise attribution, except for the possible presence of white spots on the abdominal tergites, suggesting affinities with Empria. Statz [117] gave a reconstruction drawing but it is unclear if he has revised or see the type specimen. Therefore, we treat it as an incertae sedis within Allantinae sensu lato.

Conclusions
Taeger et al. [1] listed in their catalog of the Symphyta seven species of Allantus ss, 33 species of Emphytus, plus five species not assigned to a subgenus. The extant species are separated based on body characters such as the color of labrum, palps, and head, shape of ovipositor, etc. [111,[119][120][121]. As these structures are not preserved in the new fossil, it is not possible to compare it with all the extant species. Nevertheless, as it is the first known fossil of this genus, we prefer to name it rather than leaving the specimen in an open nomenclature. Additionally, the age of the specimen (Miocene) is potentially a sufficient justification to prevent affinities with any extant species of the genus. In addition, its forewing vein 3A reaching posterior wing margin seems to be infrequent among the extant species and could constitute an important difference.
The attribution of the fossil Empria to a species different from the extant one is more complicated to justify because of its younger age; nevertheless, we prefer to do so for the reason stated above.
Niu et al. [2] dated the separation between Allantus and Hemathlophorus to the middle Eocene-middle Oligocene. Therefore, it is surprising that no older Allantus or Emphytus fossil has been describe before our study; thus, it is necessary to further explore the tenthredinid fossil record. In fact, numerous specimens are known from fossil imprints or compressions but the fastidious comparisons and sometimes the uncertainties for their placement prevent paleoentomologists to begin their description.
Extant representatives of the genera Allantus and Emphytus are distributed in the Palearctic, Nearctic, and Oriental regions. Empria is also known from the Palearctic, Nearctic, and Oriental regions, but two species are Neotropical. Both genera are known to live on angiosperms (Rosaceae, Salicaceae, Betulaceae, Fagaceae, Polygonaceae, etc.) [111,119,122,123]. Fossils of these families are known in the paleoflora of la Montagne d'Andance and Sainte-Reine [8,13], supporting the putative phytophagy on such angiosperms for Emphytus miocenicus sp. nov. and Empria sammuti sp. Nov. since the Miocene period.