The geological record and phylogeny of spider wasps (Hymenoptera: Pompilidae): A revision of fossil species and their phylogenetic placement

Accurate fossil identification has become increasingly relevant with the widespread use of phylogenetic divergence time estimation methods, which rely on fossil data to determine clade hard-minimum ages. Here we revise, diagnose and illustrate known spider wasp (Hymenoptera: Pompilidae) fossil species and place them within the latest Pompilidae phylogenetic hypothesis. Ceropalites infelix Cockerell, from the Florissant Fossil Beds (Priabonian), is no longer recognized as Pompilidae, but as Aulacidae. Agenioideus saxigenus (Cockerell) comb. nov., Deuteragenia wettweri (Statz) comb. nov., Caputelus scudderi (Cockerell, 1906) comb. nov., Pepsinites avitula (Cockerell, 1941) comb. nov., Pepsinites contentus (Theobald, 1937) comb. nov., Pepsinites florissantensis (Cockerell, 1906) comb. nov., Pepsinites laminarum (Rohwer, 1909) comb. nov., Pepsinites scelerosus (Meunier, 1919) comb. nov., Pepsinites cockerellae (Rohwer, 1909) comb. nov., Pompilinites coquandi (Theobald, 1937) comb. nov., Pompilinites depressus (Statz, 1936) comb. nov., Pompilites incertus (Theobald, 1937) comb. nov., Pompilites induratus (Heer, 1849) comb. nov., Pompilites fasciatus (Theobald, 1937) comb. nov., and Pompilites senex comb. nov. are new combinations. Twenty-three fossil species of spider wasps are now recognized in 13 genera. Four new genera are proposed: Caputelus Waichert & Pitts gen. nov., Pompilites Rodriguez gen. nov., Pompilinites Rodriguez & Waichert gen. nov., and Pepsinites Rodriguez & Waichert gen. nov., of which the three latter are collective-group names for fossils with taxonomic uncertainty. One species of fossil spider wasp is described: Deuteragenia catalunyia Rodriguez, Waichert & Pitts sp. nov., from the Bellver deposits in Catalonia, Spain. Five of the 23 known species can be used to determine hard-minimum age for calibrations of genera stem-groups (Agenioideus, Anoplius, Cryptocheilus, Deuteragenia, Priocnemis). The fossil belonging to the stem-group of the tribe Ageniellini (Chubutholites) is not recommended for calibration because of the high uncertainty in its age and taxonomy. The remaining taxa can be assigned to the lineage comprising Pompilinae + Pepsinae (12 species) or crown-group Pompilidae (four species).

Introduction Fossil species identification is becoming more relevant with the widespread use of molecular data for phylogenetics and the possibility of producing time-calibrated trees. Unfortunately, many fossilized specimens lack relevant preserved structures necessary for classification at the genus and species level, which makes taxonomically difficult groups even more challenging. Spider wasps (Hymenoptera: Pompilidae), for instance, have a puzzling taxonomy, because of morphological uniformity. Pompilid fossil taxa have an additional issue and it is the lack of appropriate preserved structures, and descriptions which mostly date from the early 1900s, where the majority of images accompanying species descriptions are poor or non-existent.
Fossil spider wasps have not been revised as a whole. The most recent work described Dominican amber species and excluded a Cretaceous fossil species from the family [1]. This paleontological study corroborated molecular phylogenetic analyses, showing that the oldest Pompilidae are from the Eocene and not the Cretaceous, as was previously proposed [1,2] Here we provide a revision of the existing spider wasp fossils and place them as potential calibration points on the most recent phylogeny of the group [2].

Taxonomic revision
Compression and amber fossils from various natural history collections were studied. The compression fossils studied belong to seven main deposits: the Florissant Fossil Beds (Florissant, Colorado, USA), the Oeningen deposits (Baden-Württemberg, Germany), the Rott deposits (Rhineland-Palatinate, Germany), the Aix-en-Provence deposits (Bouches-du-Rhône, France), the Terrains sannoisiens du Gard deposits (Gard, France), Camoins-les-Bains (Marseille, France) and the Bellver deposits (Lleyda, Spain). The Dominican amber fossils studied derive from deposits found in mines between the cities of Santiago and Puerto Plata (Dominican Republic). One of the Baltic amber fossils derives from the Kaliningrad region (Russia). The locality of the second Baltic amber fossil is unknown.
Abbreviations used are the same as those by Wasbauer & Kimsey [3]. They are defined as follows: LA3, length of third antennal segment and WA3, width of third antennal segment. Measurements of the clypeus are as follow: WC, width of clypeus, measured between the widest points; and LC, maximum height of clypeus. Wing venation terminology follows that of Huber & Sharkey [4]. Pictures were either provided by curators or taken with a Jenoptik camera coupled to a Leica Mz7.5 microscope; line drawings were processed in Adobe Illustrator. The kind of material studied is summarized in Table 1.
The species treated here were assigned to the family Pompilidae based mainly on wing venation features, which are relatively uniform for the family [5]. All of the specimens studied have a preserved forewing, and most of them have the hindwing also preserved ( Table 1). They were placed in the family Pompilidae based on a combination of wing venation character states [5][6][7]: forewing with ten closed cells, C vein present, 1Rs not directly joining pterostigma, abscissa distad Rs + M from M vein present, veins 2rs-m and 3rs-m present, 2m-cu present, 1cu-a closer to vein 1M than to its junction with Cu vein, 2-Rs vein present; hindwing with distinct claval lobe absent, first abscissa of Cu present, 1A vein absent, 1rs-m crossvein absent, second abscissa of M present, and the veins C+Sc+R+Rs fused basally. Pompilidae shares most of Sharkey & Roy's [6] wing venation character states with Tiphiidae and Sapygidae, but Tiphiidae have a distinct claval lobe [7] and Sapygidae have a vein 1Rs that joins or is closer to the pterostigma [6]. Additionally, Tiphiidae usually have a cilindrical metasoma and males with spined hypopygium, and in both Tiphiidae and Sapygidae the hind leg femur does not surpass the metasoma.
The acronyms for the collections used in this study are as follows:

Nomenclatural acts
The electronic edition of this article conforms to the requirements of the amended International Code of Zoological Nomenclature, and hence the new names contained herein are available under that Code from the electronic edition of this article. 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: urn:lsid:zoobank.org:pub: 260A7C4F-2B6E-44FC-87E0-033297134EF4. The electronic edition of this work was published in a journal with an ISSN, and has been archived and is available from the following digital repositories: PubMed Central and LOCKSS.

Phylogenetic placement
After identification, fossils were placed onto the most recent molecular phylogeny [2] by morphologically matching the lowest reliable Linnaean taxonomic category with a monophyletic group and locating it on the stem group. The topology aims to provide and clarify the current data on pompilid fossils and phylogenetic relationships for future studies that might require chronological analyses. A conservative placement was used because none of the species can be matched to extant species, and therefore using them as crown-group calibrations could lead to error. It has been demonstrated that applying fossils to crown groups can inflate inferred ages [8]. We also provided in the topology hard minimum age values (lowest age of oldest fossil) to be used as priors in divergence time estimation analyses.

Phylogenetic placement
We revised 23 species, of which 18 are compression fossils, one is an ichnofossil (i.e. geological record of biological activity consisting on impressions or fossilized organic material) and four are preserved in amber (Table 1). Age information from these fossils was used to determine hard minimum age for future calibration studies. We plotted these bounds on the topology from Waichert et al. [2]. All fossil date ranges fell within estimated age error bars (Fig 1). Five fossils were useful in calibrating nodes at the genus stem-group level: Agenioideus saxigenus (Cockerell, 1908 (Theobald, 1937) comb. nov. and Pompilites incertus (Theobald, 1937) comb. nov. This age, however, is younger than the suggested hard-minimum age of the most recent common ancestor of Pompilinae and Pepsinae, and therefore is not recommended as a calibration point.  Type stratum. Colorado Florissant Fossil Beds (Priabonian). The formation is a heterolithic accumulation of shale, tuffaceous mudstone and siltstone, tuff, and arkosic, volcaniclastic sandstone and conglomerate [11]. It is dated from the Priabonian, Eocene. Epis & Chapin [12] dated the formation from 34.9 Ma. Later, Evanoff et al. [11] analyses yielded a range of ages from 34.3 to 33.5 Ma. The most recent study [13]  Diagnosis. Forewing hyaline, with two bands that cover 1Rs and 2Rs cells; maximum width 0.31X its length; 1Rs cell almost as long as 2Rs; 2m-cu vein slightly curved, meeting 2Rs cell 0.95X distance from base to apex of cell; and 2M cell with an inflection at the base of Cu.
Remarks. Two specimens collected from the same locality are found in the type series of this species. Some species of Agenia are currently classified under Deuteragenia Šustera, 1912. Agenia saxigena was placed in Deuteragenia, probably due to the double dark bands on the wings, which entirely cover the 1Rs and 2Rs cells (Figs 2A, 2B, 3A and 3B). Nevertheless, the inflection in the vein Cu at the base of the 2M cell identifies this specimen as Pompilinae. We place this species in Agenioideus because the 2m-cu vein of the forewing arises on the Cu vein much more than half the distance from the origin of 2M to the outer wing margin ( , 1902). This specimen differs from Priocnemis and Balboana by the presence of an inflection at the base of the Cu vein in 2M cell and by having long pronotum, which is short and angulate in the latter; and from Tachypompilus by the absence of irregular contours on the propodeum. Finally, this specimen has a large stigma, whereas in Aplochares the stigma is reduced. Because of its preserved morphological characters, and consequently, its taxonomic accuracy, this fossil can be used as a calibration point for stem-group Agenioideus (Fig 1,  Type stratum. Dominican Amber (Burdigalian to Langhian). The age of Dominican Amber was established by Iturralde-Vinent [14] at "around" 16 Ma [1], based on the Midmiocene Climatic Optimum. This event has a possible age range from 17-15 Ma [15], and therefore we treat this as the age uncertainty for Dominican Amber.
Diagnosis. Antennae elongate, third antennal segment 1.8X as long as second segment; wings hyaline; maximum width 0.18X its length; 2Rs cell as long as 1Rs; 2m-cu vein slightly curved, meeting 2Rs cell 0.70X distance from base to apex of cell; and 2M cell with an inflection at the base of the vein Cu; base of metasoma paler, apical portion darker.
Remarks. This is the only Anoplius from Dominican amber.   (Cockerell, 1908)  Type stratum. Dominican amber (Burdigalian to Langhian). Diagnosis. Antennae slightly crenulated; wing hyaline; maximum width 0.31X its length; 2Rs cell as long as 1Rs; 2m-cu vein curved, meeting 2Rs cell 0.55X distance from its base to apex of cell; 2M cell with an inflection at the base of Cu vein; propodeal disc flat in profile view, posterior surface sloping abruptly at an angle.
Remarks. This is the only described species of Tainopompilus. It was placed in the subfamily Pompilinae by Rodriguez et al. [1] based on the presence of an inflection at the base of the vein Cu on 2M cell (Rodriguez et al. [1], fig 2b). It is morphologically similar to Priochilus, but the spine-like setae of uniform length on the metatibia and the inflection at the base of the vein Cu on 2M cell separate the two genera. Morphological convergence has been observed in several distantly related groups of Pompilidae that hunt in similar niches [2,16]. This may result from the ecological specificity in host use, which is phylogenetically conserved [17]. For these reasons, we abstain from assigning T. argentum to any stem group.
Diagnosis. Forewing with two dark bands, maximum width 0.30X its length. The wing venation characters are not visible.
Remarks. The holotype of this specimen could not be located. The original description of this species does not mention its location and none of the museums contacted claimed to have it in their collection. Our only source of information is the species description and a blurry photograph included in the original publication. Meunier [18] mentioned the presence of this specimen in the Natural History Museum of Marseille. However, the curator did not respond to our inquiries concerning this specimen. Theobald [22] suggested a resemblance of this species to the extant Pompilus maculipes Smith, 1870, which is now placed in Anoplius [23]. We have no morphological evidence to place this species in Anoplius, and therefore we keep its original name.

Pompilinae incertae sedis
urn:lsid:zoobank.org:act:F5B5ABEC-55DC-4279-B46E-051E6BFAE746 Diagnosis. Wings hyaline with or without transverse dark spots; 2M cell with an inflection at the base of the Cu vein; legs with long spine-like setae set on grooves, spines uneven in length; metasoma without a transverse groove in the second sternite.
Description. Body length variable; forewing length variable. Head. Head wide, TFD > FD (Figs 2C and 3C); ocelli and mandible usually inconspicuous in fossils; flagellum elongate. Mesosoma. Pronotum elongated or not (Figs 2C and 3C); notauli present; propodeum variable; wing elongated; 2M cell with downward inflection at base of Cu vein; mid and fore femur with long setae in the apical margin; mid and fore tibia with long, uneven spine-like setae, setae set on grooves. Metasoma. Metasoma usually large, total length > total width.
Etymology. From Pompilin[ae] the latin suffix-ites "nature of, quality of". This suffix is traditionally used for the generic epithet in fossils. The gender is masculine.
Remarks. This name is established as a collective-group name for all Pompilinae fossil species for which the generic position is unclear because of lack of diagnostic characters in preserved specimens. Pompilinae are promptly classified by the presence of long spine-like setae on metatibia and the Cu1 vein deflected downward at the base, with exceptions among species of Priochilus and Balboana. As a collective group, no description or diagnosis is provided. A type species is not required for collective groups ( [24], art. 13.3.2, 42.3.1, 66, 67.14). Because of the uncertainty in their lower-level classification, all Pompilinites fossils can only be assigned to stem-group Pompilinae (Fig 1, divergence between Pompilinae and Pepsinae).
Pompilinites coquandi (Theobald, 1937)  Type stratum. Aix-en-Provence deposits (Oligocene). Diagnosis. Antennae long, third antennal segment 2.40X as long as wide; wing hyaline, maximum width 0.37X its length; 2Rs cell as long as 1Rs; 2m-cu vein straight, meeting 2Rs cell 0.50X distance from base to apex of cell; and 2M cell with an inflection at the base of the Cu vein.
Remarks. This species is unquestionably Pompilinae due to the presence of an inflection at the base of the Cu vein on the 2M cell ( Fig 4B). The low quality of specimen preservation hinders an accurate identification to genus level.
Pompilinites depressus (Statz, 1936)  . The matrix of this deposit consists mainly of fine-grained paper shales, and thus contains very well-preserved fossils. Its age is controversial, ranging from the Chattian, Oligocene to the Aquitanian, Miocene [26]. More recent publications establish the Chattian as the correct age for this deposit [27].
Diagnosis. Wing hyaline; maximum width 0.31X its length; 1Rs cell 1.10X as long as 2Rs; 2m-cu slightly curved, meeting 2Rs cell 0.40X distance from base to apex of cell; and Cu vein of 2M cell with an inflection at its base.
Remarks. The generic position of this species is equivocal. We have only examined the wing venation illustrations in Statz [25] (pl. 12, fig 34), and the characters observed are dubious for determining its taxonomic status. Wing venation characters are not enough to determine the genus of this specimen. An inflection at the base of the Cu vein in the 2M cell is present, which places the species in the subfamily Pompilinae. Further inference would lead to an inaccurate taxonomic placement. The original description of this species does not indicate the location of the holotype and none of the natural history museums contacted claimed to have it in their collection.
Subfamily Type stratum. Sarmiento Formation (Priabonian). This formation is composed of alluvial rocks.
Remarks. Chubutolithes was described based on an ichnofossil composed of a mud nest. The structure of the pellets observed in the fossil suggests its affinity to Auplopus and its placement in Pompilidae [29][30][31]. Further analysis of this fossil suggests its formation on plant material and subsequent alluvial transport [29]. Simpson [32] studied the age of the entire Gaiman formation and concluded that the stratum where the holotype was found is Casamayoran (early Eocene) and not Patagonian (Miocene) as previously proposed by Schiller [33]. The stratigraphic distribution of the species is wide, ranging from Colhuehaupian (~20 Ma, [34]) to Casamayoran (55.8-48 Ma, [32]). According to Genise & Cladera [29] the only reliable fossils are from the same stratigraphical level as the holotype. The only other specimen from a younger stratum within the Sarmiento Formation (Colhuehaupian) shows signs of abrasion, which has been proposed as evidence of transport from older strata (i.e. Sarmiento Formation, [29]). If the age of Chubulolithes falls in the early Eocene, it would constitute the oldest Pompilidae fossil to date, which would have repercussions for the divergence-date estimates for the family, because the age is outside of the confidence interval obtained for Ageniellini (Fig 1). Nonetheless, recent analyses of fossil mammals from the Casamayoran revealed a much younger age in the late Eocene, Priabonian (36.6-35.3 Ma, [35]). The oldest sediments from the Sarmiento formation have been found to be no older than middle Eocene [36], rather than early Eocene as proposed previously. Therefore, the uncertainty in the age of Casamayoran fossils would impose a younger age range (36.6-35.3 Ma) on this calibration. Moreover, the fact that this is an ichnofossil adds taxonomic uncertainty, and even Evans & Shimizu [31]-to whom the confirmation of the assignment to Pompilidae has been attributed-have regarded it only as a "probable" Auplopus-looking nest. This uncertainty leaves open the possibility that this fossil derives from a much older ancestor within or even outside Pompilidae. Therefore, we do not recommend using this fossil in divergence-time analyses.
Genus Type stratum. Lacustrine deposits of Bellver, Spain (Messinian; [37]). Diagnosis. Wing hyaline with two bands covering cells 1Rs, 2Rs, 2M and most of 2R1; maximum width 0.31X its length; 2Rs cell as long as 1Rs; 2m-cu vein almost straight, meeting 2Rs cell 0.40X distance from base to apex of cell; and 2M cell without an inflection at the base of Cu vein.
Etymology. This species is named after the autonomous community of Catalonia (Catalunyia in Catalan), where the fossil was found. The epithet is placed as a noun in apposition (ICZN [24], art. 31.2.1).
Remarks. This species was first studied by Arillo [37], who described it, but did not name it. We compared the holotype with the extant and extinct species of Dipogon, and name the new species herein. Because this species was found in Miocene deposits (younger than 11 Myr), the specimen was compared to all extant Dipogon (Deuteragenia) species [38,39] to confirm that it is a new species. Characters from wing venation were compared against all the other species. The most conspicuous difference found was the width of the 2R1 cell. Dipogon catalunyia sp. nov. has a 2R1 cell that is 4X as long as wide (Figs 2D and 3D), while in all other species it is less than 3X as long as wide (Arillo [37], fig 2). There are also differences in the distance from the beginning of the 1Rs where the 1m-cu vein is received at the base of 1Rs. Most of the described species do not receive this vein at 0.4X from the beginning of the 1Rs. Also, the distance where the 2m-cu vein is received by 2Rs is not 0.3X its length from the beginning of the cell in most species, as it is in D. catalunyia (e.g., 0.2X in D. vechti Day, 1979). Moreover, the extent to which the 2R1 cell is covered by dark banding, which in D. catalunyia is covered by 0.9X its length, is less in most of other species (e.g. 0.3X its length in D. monticolus Wahis, 1972, 1.0X its length in D. subintermedius [Maggretti, 1886]). Finally, the 1cu-a vein of the forewing meets the M+Cu slightly beyond the origin of the M, while in some species it meets M+Cu well beyond the origin of M (e. g., D. austriacus Wolf, 1964). This species can be distinguished from Priocnemis species, because in Dipogon the dark region covers the 1Rs and 2Rs cells completely, whereas in Priocnemis these are covered only partially [37]. Because of confidence in its taxonomic placement, this fossil species can be used as calibration point for Deuteragenia stem-group.
Deuteragenia wettweri (Statz, 1938)  Diagnosis. Wing with two bands, apex of hind wing darkened; maximum width 0.39X its length; 2Rs 1.2X longer than 1Rs; 2m-cu vein straight, meeting 2Rs cell 0.45X distance from base to apex of cell; and 2M cell without an inflection at the base of the Cu vein; metasoma 1.4X as long as mesosoma.
Remarks. The wing venation (Fig 4C) of this specimen resembles that of Deuteragenia species rather than Priocnemis. The 3r-m vein is curved in Deuteragenia (Fig 4C), whereas in Priocnemis it is straight; and the maximum width of 2Rs cell is greater than 2X its maximum length. Therefore, we transfer this species to Deuteragenia.
Genus  2a1); maximum width 0.45X its length; cells short and rounded; 2Rs cell about the same size as 1Rs; 2m-cu vein slightly curved, meeting 2Rs cell 0.5X distance from base to apex of cell; 2R1 ending on apex of the forewing instead of anterior margin; and 2M cell without an inflection at the base of the Cu vein.
Remarks. Rodriguez et al. 2016 suggested that this species was probably a cleptoparasitoid, because of the short antennal segments with thick conspicuous setae (Rodriguez et al. [1], fig  2b1).
Genus Priocnemis Schiodte, 1837 Priocnemis aertsi Statz, 1936  Diagnosis. Wing hyaline with two dark bands, apex of wing darkened; maximum width 0.27X its length; 2Rs cell 1.20X longer than 1Rs; 2m-cu vein rounded, meeting 2Rs cell 0.33X distance from base to apex of cell; and 2M cell without an inflection at the base of the Cu vein.
Remarks. The original description of this species contained drawings of the wing venation and the entire specimen (Statz [25], pl. 12, figs [33][34]. This species was placed in Priocnemis probably due to the presence of the 1cu-a vein extending beyond the M vein by about 0.70 to 1.30 of its length (Fig 4D). In other related Pepsini genera, this vein is closer to the M vein. Nevertheless, other Pepsini genera, such as Entypus Dahlbom, 1843, Pepsis, and Calopompilus Ashmead, 1900, also possess this character. The wing venation in Priocnemis, however, is distinguished from Pepsis by having the 2R1 cell ending straight, not separated apically from the costal margin of the wing, and from Calopompilus and Entypus by having the 2r-m vein slightly curved. In addition, it can be separated from Dipogon by the straight 3r-m vein, which is curved in Dipogon. Schoberlin [42] mentioned the presence of Priocnemis in the Oeningen deposits. However, the author did not provide a species description, or the location of the specimen mentioned. This fossil can be used as a calibration point for the stem-group Priocnemis (Fig 1).
Description. Body length 12.00 mm. Forewing 7.00 mm; maximum wing width 2.00 mm. Head. Head wide, TFD 1.16X FH, vertex extended posteriorly (Figs 5A and 6A); ocelli in acute angle, lateral ocelli closer to each other than to compound eyes; first four antennal segments in the ratio 0.30: 0.34: 0.28: 0.24. Mesosoma. Mesosoma short, 0.75X as long as wide. Pronotum short, anterior margin straight, lateral edges rounded. Notauli present. Propodeum large, width 0.63X its length; propodeal posterior and lateral carina present. Forewing narrow (Figs 5A and 6A); maximum width 0.28X its length; length of 2R1 cell 0.60X the distance from its apex to wing margin; 2Rs cell 1.28X longer than 1Rs; two dark bands present. Leg without even spines distally; mid and hind tibia with scale-like spines; fore spur as long as antennal scape. Metasoma. Metasoma 2.94X as long as wide; 1.75X as long as mesosoma.
Etymology. The generic epithet comes from the Latin, caput-, which means head, and-elus, which refers to the suffix widely used for Pompilidae taxa. The name denotes the distinctive head shape of this taxon, which is extended in the vertex. The gender is masculine.
Remarks. Caputelus gen. nov. has the overall shape of the head, pronotum, and propodeum similar to those in extant species of Ctenocerinae, an Afrotropical subfamily. Additionally, the head has two dark spots, which could indicate it possessed antennal pits (Figs 5A and 6A), as diagnosed in Ctenocerinae. Extant genera of the New World, such as Abernessia Arlé, 1947 and Lepidocnemis Haupt, 1930, share similar morphological characteristics (see Waichert & Pitts with Caputelus gen. nov. [16,44]). Abernessia and Lepidocnemis were considered representatives of Ctenocerinae in the Neotropical Region until Waichert et al. [2] transferred these genera to Pepsinae based on molecular phylogenetic analysis. Waichert et al. [2] suggested that similarities between these Neotropical genera resulted from morphological convergence. The prolonged vertex and the antennal pit on the head, the large and carinate propodeum, and the absence of spines on legs suggest that Caputelus gen. nov. is morphologically related to Abernessia and Lepidocnemis, but differs from them by having the head with vertex well-prolonged; the pronotum short and angulate (it is long and rounded in the other genera); the 2R1 cell is large; and the forewing is banded.
Caputelus gen. nov. is here classified as Pepsinae, regardless of its similarities with Ctenocerinae, because it is more likely that these morphological features are the result of convergence, as it is for Abernessia and Lepidocnemis, than to consider that the Tertiary distribution of Ctenocerinae extended to the Northern Hemisphere. Geological record and phylogeny of spider wasps (Hymenoptera: Pompilidae) Type material. Amber fossil. Holotype not available. Type stratum. Baltic amber (Lutetian to Priabonian). The exact locality is unknown. Baltic amber deposits have been obtained for more than 100 years, and their age is controversial. Microfaunistic dating of the deposits containing the largest amount of amber suggest they are from the Priabonian, Eocene (37.7 Ma) (Kaplan et al. [41]), whereas radiometrically dated glauconite as Lutetian, Eocene (47.0 to 44.1 Ma) [46]. Perkovsky et al. [47] considered the Ritzkowski [46] data insufficient to disprove Kaplan [41], because the former was based on two samples and the latter on seven samples. More recent data indicate that the age of Baltic Amber can be narrowed to 34 to 38 Ma based on palynological data [48,49].
Diagnosis. Wing hyaline, maximum width 0.26X its length; 2Rs cell 1.29X longer than 1Rs; 2m-cu vein straight, meeting 2Rs cell 0.40X distance from base to apex of cell; and 2M cell without an inflection at the base of Cu vein.
Remarks. This is a very small specimen from Baltic amber that does not have an inflection at the base of the Cu vein of the 2M cell, which excludes it from Pompilinae. The description and drawings provided by Meunier ([50], figs 1-3) suggest that this specimen should be placed in Pepsinae. The location of the holotype of this species was not mentioned in the original description and none of the museums contacted claimed to have it in their collection. Because we could not locate the holotype, and the published drawing is inadequate, we cannot make further taxonomic conclusions about this taxon.
Pepsinites avitula (Cockerell, 1941)  Remarks. Vardy [52] studied this specimen, and mentioned a number of characters that differentiate it from Pepsis. He also mentioned the lack of Pepsis species that possess banded wings, and suggested the proximity of this species to Chirodamus Haliday, 1837, thus establishing the combination. We did not find enough wing venation characters to justify its placement in Chirodamus (Fig 4G). Therefore, we place this species in the collective group Pepsinites.
Pepsinites florissantensis (Cockerell, 1906)  Diagnosis. Wing hyaline with two dark bands, apex darkened; maximum width 0.20X its length; 2Rs cell slightly shorter than 1Rs; 2m-cu vein slightly curved, meeting 2Rs cell 0.70X distance from base to apex of cell; 2M cell without an inflection at the base of Cu; and hindwing with cu-a ending distinctly before the juncture of M with CU.
Remarks. The specimen was preserved with wings overlapping (Figs 5D and 6D), which precluded accurate description and illustration of wing venation characters. Nevertheless, Cockerell [43] provided a good description and several measurements of hindwing cells. The placement of C. florissantensis in Hemipogonius Cockerell, 1906 was not justified by Cockerell [43]. The same author later placed the species within Cryptocheilus [45]. The generic position of this species cannot be certain because wing venation characters are not entirely visible. It certainly belongs to Pepsinae, because of the absence of an inflection at the base of the Cu vein in 2M cell. (Rohwer, 1909)  Diagnosis. Wing hyaline with apex darkened; maximum width 0.26X its length; 2Rs cell almost as long as 1Rs; and 2m-cu and 2M cell not observable.

Pepsinites laminarum
Remarks. Cockerell [45] placed this species in Cryptocheilus. The wing venation is not well preserved, but the robust body, flat metasoma, and the absence on an inflection at the base of the Cu vein of 2M cell likely places it in the tribe Pepsini (Figs 5E and 6E). Characters that might assign it to a particular genus within the tribe are not visible.
Pepsinites contentus (Theobald, 1937)  Type stratum. Terrains sannoisiens du Gard, France (Priabonian). The age of this deposit has been addressed in various recent publications that refer to the Theobald specimens [54,55], but the source of the age justification is not addressed and is, so far, unknown to us. The only age range has been mentioned as 37.2-33.9 Ma [54], falling within the Priabonian. Collomb et al. [55] set the date as "Late Priabonian, 35 Mya, Gard, France" and Perrard et al. [56] suggest a "Late Eocene" date based on the French Geological Survey (Bureau de Recherches Géologiques et Minières) geological map 1/50 000, number 912. There is no other mention of these deposits and their age in recent literature.
Diagnosis. Wing hyaline; maximum width 0.31X its length; 2Rs cell 1.5X longer than 1Rs; 2m-cu vein straight, meeting 2Rs cell 0.33X distance from base to apex of cell; and 2M cell without an inflection at the base of the Cu vein.
Remarks. This species was described as "Criptochilus contentus", with a misspelling in the generic name, and was not mentioned by Engel & Grimaldi [57] in their revision. The venation of the hindwing was not illustrated by Theobald [22], but based on the forewing it belongs unquestionably to Pepsinae. However, the lack of information on the hindwing venation precludes its accurate generic placement.
Pepsinites cockerellae (Rohwer, 1909)  Remarks. This species was placed in the subgenus Deuteragenia based on the wing-banding patterns (Figs 5F and 6F). Many pompilid genera in the two most species-rich subfamilies (e.g. Priochilus and Ageniella Banks, 1912), possess dark areas on the wings, which makes it an ambiguous character. We conclude that the current generic position of this species is equivocal, and the preservation of the fossil does not allow us to draw any further conclusions. The main wing venation characters of Pepsinae, however, are present. Therefore, we place this species in Pepsinites.
Etymology. From Pompil[idae] the latin suffix-ites "nature of, quality of". This suffix is traditionally used for the generic epithet in fossils. The gender is masculine.
Remarks. This name is established as a collective-group name for all Pompilidae fossil species for which the subfamilial and generic position is unclear because of lack of diagnostic characters in preserved specimens. As a collective group, a type species is not required for collective groups ( [24], art. 13.3.2, 42.3.1, 66, 67.14). Because of lack of evidence to place them in extant subfamilies, these are currently assigned to crown-group pompilidae, but not recommended for in phylogenetic time-calibrations (Fig 1).
Pompilites induratus (Heer, 1849)  Type stratum. Oeningen fossil beds (Langhian to Serravallian). This region had first been reported as belonging to Switzerland. Cockerell [43] corrected it and located the region in Baden-Württemberg, Germany. Oeningen is one of the richest insect fossil deposits known in the world. It is composed of freshwater limestone deposits that date from the Langhian to the Serravallian , based on stratigraphic data and terrestrial insect sediment analysis [60][61][62]27].
Remarks. We are doubtful about the subfamilial and generic classification of this species, because of the poor preservation of the specimen (Figs 7A and 8A). Rohwer [53] transferred this species to Anoplius, but made no comments on the reasons for this decision. Characters that would place it in the subfamily Pompilinae, such as the inflection on the base of Cu vein in the 2M cell or the spines of different lengths on the apex of the metatibia, cannot be observed, nor can diagnostic characters of other subfamilies.
Pompilites fasciatus (Theobald, 1937)  Type stratum. Aix-en-Provence fossil deposits (Chattian). Diagnosis. Antennae short, antennal segments wide; wing hyaline, darkened towards the apex in about 1/3 of the total length, maximum width 0.35X its length; 2Rs cell 1.20X longer than 1Rs; 2m-cu vein slightly curved, meeting 2Rs cell 0.60X distance from base to apex of cell; and 2M cell with an inflection at the base of Cu vein.
Remarks. The general morphology of the specimen is very similar to the description of extant cleptoparasitic Pompilidae by Shimizu [63]. Pompilidae cleptoparasites typically have shortened antennal segments (Figs 7B and 8B), such as in Aridestus Banks, 1947, or Poecilagenia Haupt, 1927 [63]. Nevertheless, cleptoparasites are known in three subfamilies, and the accurate generic placement of this specimen is not possible due to the lack of detail in preserved structures.
Pompilites incertus (Theobald, 1937)  Another issue arises because of the misuse of stratigraphic information, which may increase uncertainty in age estimates. This becomes more apparent in younger groups like Pompilidae, where the most accurate fossil identifications are from amber deposits, for which the age is controversial with wide time ranges.
Pompilidae is one of the least taxonomically studied Hymenoptera families. One of the main reasons for the paucity of systematics research is the difficulty in identification stemming from morphological uniformity even between subfamilies. Recent molecular and paleontological studies place the origin of the family in the Eocene [1,2], with origin of crown-group subfamilies no later than the Oligocene. The morphological homogeneity, together with limited amount of preserved structures, makes fossil Pompilidae identification highly inaccurate, especially for compression fossils.
Our study is the first attempt at clarifying the classification of fossil spider wasps. From the 17 fossil species described, only eight can be determined to genus with confidence. The remaining species have been placed in collective-group genera that indicate the uncertainty in taxonomic placement and avoid further confusion in fossil Pompilidae ages.