On the identity of the fossil aquatic beetles from the Tertiary localities in the southern part of the Upper Rhine Graben (Coleoptera, Hydrophilidae, Dytiscidae)

Abstract This study focuses on the fossil beetles assigned previously to the family Hydrophilidae described from the localities in the southern part of the Upper Rhine Graben: Brunstatt (France, Alsace) and Kleinkems (Germany, Baden-Württemberg) (both dated ca. to Eocene-Oligocene boundary, 34 Ma). The identity of Escheria convexa Förster, 1891 is fixed by the designation of its neotype, the species is redescribed, illustrated, transferred from the hydrophilid genus Hydrobius Leach, 1815 to the genus Copelatus Erichson, 1832 (Coleoptera: Adephaga: Dytiscidae) and compared with other fossil representatives of Copelatus. The identity of the remaining three species, Hydrobius crassipunctatus (Förster, 1891), Hydrobius dimidiatus (Förster, 1891) and Hydrobius punctulatus (Förster, 1891),is briefly evaluated on the basis of the original descriptions and illustrations only, because their types were lost or destroyed during World War II; all three species are removed from the fossil record of the Hydrophiloidea and placed into Polyphaga incertae sedis. The geology and stratigraphy of Brunnstatt and Kleinkems are discussed briefly.


Introduction
In his study of the insects from the Tertiary outcrop of Brunstatt (Alsace, France), Förster (1891) described four species of the family Hydrophilidae, assigning all of them to the fossil genus Escheria Heer, 1847. Later, Th éobald (1937) recorded two of these species from the locality of Kleinkems (spelled incorrectly as 'Kleinkembs' by the latter author), which is situated 16 km from Brunstatt and is believed to be of the same age and origin (see below for details). Based on the preserved characters, Th éobald (1937) transferred all four of Förster's (1891) species to the recent hydrophilid genus Hydrobius Leach, 1815. After 1937, the specimens were never re-examined and their identity remained unclear. Unfortunately, the Förster collection was either lost or destroyed during World War II (see Material and methods for details), which further complicates research.
For this study, we have re-examined the specimens from the locality of Kleinkems studied previously by Th éobald (1937). In order to resolve the complicated situation concerning the four hydrophilid taxa, a neotype is designated for Escheria convexa Förster, 1891 and its taxonomic position is thus illuminated. Th e remaining fossils are excluded from the Hydrophiloidea due to the absence of any hydrophiloid apomorphy. Hence, our study supports the opinion by Fikáček et al. (2010) that various middle-sized Tertiary beetles were assigned into the hydrophilid genus Hydrobius irrespectively to their real taxonomic position.

Geology and stratigraphy of the fossil sites
Th e Tertiary outcrops of Brunstatt and Kleinkems were located in the south-west of central Europe and no longer exist today. Brunstatt was situated south of the city of Mulhouse in France (47°41'N 7°31'E); Kleinkems was situated in Germany (47°43'N 7°19'E) northwest of the city of Basel (Switzerland), about 16 km from Brunstatt. Several hundred fossils in total were collected at these localities (Wappler et al. 2005).
According to the reconstruction of the sedimentation history, the limnic sediments of Brunstatt and Kleinkems were deposited on the shore of a very large shallow saline lake (with an area of several hundred square kilometres) with episodical intrusion of fresh water (Lutz 1997). Th e landscape in this area originated during the formation of the Upper Rhine Graben (URG) which forms the central part of the Cenozoic Central European Rift System. Increased rifting during the late Middle Eocene to Early Oligocene led to the formation of the Mulhouse Potash Salt Basin (also called Potash Basin or Potassic Basin) which is located in the narrowest part of the graben and fl anked by the highest of the Vosges Mountains and Black Forest Mountains (Hinsken et al. 2007). A detailed overview of the development of the URG and the Potash Basin is provided by Berger et al. (2005 a, b) and Hinsken et al. (2007). Lutz (1997) and Mai (1995) assign the age of Brunstatt and Kleinkems to the Lower Oligocene, but Mai (1995) also allocated Brunstatt to the Mammal Reference Level MP20 which corresponds to the Priabonian (37.2-33.9 Ma) in the latest part of the Eocene. Th us, the stratigraphical position of Brunstatt and Kleinkems seems to be close to the Eocene-Oligocene boundary, as is the case of the similar fossil site of Altkirch in France (Wappler et al. 2005). According to Mai (1995) [based on Lakowitz (1895)], the palaeoclimate of Brunstatt was characterised by an average annual temperature of 18°C and abundant rainfall.
It seems very likely that the sediments of Brunstatt and nearby Kleinkems are nearly identical in age and genesis. Lutz (1995) even combined both localities in his study reconstructing their paleoenvironment, and according to Mai (1995), plant fossils from Brunstatt and Kleinkems are both deposited in the same layers of laminated clay ('plattiger Steinmergel').

Material and methods
Only the fossils from the locality of Kleinkems mentioned by Th éobald (1937) deposited currently in the Naturhistorisches Museum in Basel, Switzerland (NHMB) were studied for this paper. Th e material originally examined by Förster (1891) was deposited at the 'Service de la Carte Geologique de Strasbourg' (Th éobald 1937) and is considered to have been lost or destroyed during World War II on the basis of information we received from Jean Claude Horrenberger (École et Observatoire de la Terre, Strasbourg, France) as well as two letters sent to Volker Püthz, a specialist on Staphylinidae, by Marguerite Wolf (Université Louis Pasteur, Institut de Géologie, Strasbourg, France) in July 1967and September 1971(Püthz, pers. comm. 2010. Th e identity of species missing from the Kleinkems material is only discussed briefl y on the basis of the original drawings by Förster (1891).
Fossils were examined using the Olympus SZ61 binocular microscope. Photographs were taken using the Canon MP-E 65 mm macro lens attached to the Canon EOS 550D camera. Drawings were traced from photographs. Abbreviations used in descriptions are: EL -length of the elytron; TL -total length, a single measurement of length from front of head to apex of elytra; TL-h -total length minus head length, length of body from anterior margin of pronotum to apex of elytra; TW -maximum width of body measured at right angles to TL.
Fossils whose family placement and hence also generic placement remains unclear are cited using the original combination of the name, placing the respective genus name in quotation marks. Redescription. Body oblong-oval, broadest in basal third of elytra. Head relatively broad; compound eyes large, not exceeding body outline; clypeus rounded. Pronotum broadest between posterior angles, lateral margins regularly, moderately curved. Mesoscutellar shield well preserved, broadly triangular. Base of elytra as broad as pronotal base; lateral margins of elytra moderately curved. Only mesocoxae, part of metathoracic anepisternum, and probably part of apical abdominal ventrite perceptible from ventral part of body (Figs 1-4).
Surface sculpture. Pronotum with distinct longitudinal median impression, and poorly perceptible short longitudinal striolae on disc. Elytra with 12 moderately impressed longitudinal striae.
Measurements. TL: 6.3 mm, TL-h: 5.6 mm, TW: 3.2 mm. EL: 5.2 mm. Notes on the type material. Th e neotype corresponds well with the original description and drawings by Förster (1891) in the following characters: (i) general body shape; (ii) shape of the pronotum with projecting anterior angles; (iii) elytra with large number of longitudinal striae [preserved only in posterior portion of elytra in the holotype and their number is therefore estimated by Förster (1891) to be at least 10; 12 striae are present in the specimen from Kleinkems]; (iv) shape and proportion of the scutellar shield [much wider than long]; (v) body size [TL = 6.5 mm, EL = 4.3 mm, TW = 3.8 mm for the specimen from Brunstatt according to Förster (1891)]. Th e body proportions diff er slightly between both specimens (the specimen from Brunstatt is relatively wider), but this may easily have been caused by deformation during the fossilization process or by the inaccuracy of the drawings by Förster (1891) which is quite usual for historic authors (e.g., compare the drawings by Th éobald (1937) in Figs 6 and 12 with the actual appearance of the respective fossils). Moreover, Förster (1891) mentioned that his fossil resembles the dytiscid genus Agabus in most characters and did not assign it to the Dytiscidae merely because of its ventral morphology which was reconstructed by him as resembling that of Polyphaga (Fig. 5). It seems that Th éobald (1937) examined Förster's types as he mentioned certain details which are absent in Förster's (1891) original publication, and his opinion about the conspecifi city therefore also has to be considered as a strong argument.
Th e reasons provided above together with the same age, geological origin and geographical proximity of both outcrops (Brunnstatt and Kleinkems) provide strong support for the conspecifi city of both specimens mentioned by Th éobald (1937). As the specimen from Brunstatt (i.e., the holotype) is lost, we consider it adequate to designate the specimen from Kleinkems as the neotype.
Generic attribution. Th e preserved characters of the ventral morphology, i.e. the narrow metathoracic anepisternum arising from the median coxal cavity and the anepisterno-metaventral suture directed lateroposteriad correspond closely with the ventral morphology of the Dytiscidae (see, e.g., Fig. 7.6.1 in Balke (2005)). Th e hydrodynamic body shape, large eyes, broad mesoscutellar shield, medium body size and distinct elytral striae enable us to classify the specimen without any doubt as belonging to the genus Copelatus Erichson, 1832 of the family Dytiscidae.
Copelatus is currently pantropical in its distribution and contains more than 400 described species (Nilsson 2001). Most species of Copelatus are characterised by longitudinal elytral striae whose number has been used to group the species into species groups (Sharp 1882); only a few species have smooth elytra (e.g., Hájek et al. 2010). Although the presence and number of elytral striae provides only limited evidence of phylogeny (Balke et al. 2004), the species groups delimited by number and position of elytral striae are frequently used as a tool for better orientation within the genus (e.g., Guignot 1961;Guéorguiev 1968;Nilsson et al. 1997). Th e European species previously classifi ed in Copelatus have elytra without striae and have been transferred to the genus Liopterus Dejean, 1833 by Balke et al. (2004); they are not closely related to the fossil dealt with in this study.
Copelatus convexus diff ers from all known fossil and extant species of the genus in the presence of 12 longitudinal striae on each elytron. Sharp (1882) erected a group characterized by 12 discal striae (group 7) for a single species Copelatus interruptus Sharp, 1882 which is, however, currently classifi ed in the genus Exocelina Broun, 1886 (Nilsson 2007). In contrast to the fossil C. convexus, this recent species has elytra with numerous short striolae rather than 'true' striae, see, e.g. Wewalka et al. (2010). Th erefore, C. convexus might be provisionally classifi ed in a separate C. convexus-group. However, it is necessary to point out that the counting of the precise number of lateral elytral striae is problematic in compressed fossils as the imprint of the submarginal stria may coincide with the lateral margin of the body or with the epipleuron. Th erefore, we cannot rule out that a short submarginal stria was present in C. convexus although it is not perceptible in the fossil. In this case, C. convexus would belong to the C. simoni-group sensu Nilsson (2001).

Coleoptera: Polyphaga Family incertae sedis
'Escheria' crassipunctata Förster, 1891 Fig. 7 Escheria crassipunctata Förster 1891: 364, plate XI, Fig. 11 (1937) transferred Escheria crassipunctata to the hydrophilid genus Hydrobius and assigned fossil specimen no. R 707 from the locality of Kleinkems (deposited in NHMB) to this species. We have examined the specimen from Kleinkems for this study (Figs 10-11) but we cannot confi rm that it is conspecifi c with Escheria crassipunctata for the following reasons: (i) the elytra are slightly constricted sub-basally in the specimen from Kleinkems, but evenly rounded laterally in E. crassipunctata; (ii) the body outline is distinctly interrupted between the pronotum and the elytra, but it is uninterrupted in E. crassipunctata, (iii) eyes are large and globular in the specimen from Kleinkems, but relatively small in E. crassipunctata. A more detailed comparison is impossible as the holotype of E. crassipunctata is lost and was moreover preserved in dorsal view based on the drawing by Förster (1891), whereas the specimen from Kleinkems is preserved in ventral view. Based on the original drawing by Förster (1891), Escheria crassipunctata does not bear any synapomorphy of the Hydrophiloidea. For this reason, the species is removed from the fossil record of the Hydrophiloidea and is placed in Polyphaga incertae sedis. .3-6.5 mm 12 discal C. convexus-group C. fossilis Miocene 6.1-6.5 mm 10 discal + 1 submarginal C. erichsoni-group C. ponomarenkoi Miocene 5.5-5.6 mm 6 discal + 1 submarginal C. irinus-group C. predaveterus Miocene 3.8-4.4 mm 11 discal + 1 submarginal C. trilobatus-group C. stavropolitanus Miocene 5.1 mm 11 discal C. nigrolineatus-group Specimen no. R707 from Kleinkems does not bear any synapomorphy of the Hydrophiloidea, and moreover bears a combination of characters which excludes its assignment to the Hydrophiloidea: (i) prosternal process wide, (ii) mesocoxal cavities rather wide apart, and (iii) eyes large and globular. Th e preserved characters of this specimen do not allow an unambiguous family assignment (see Lawrence et al. 1999 (1891), the morphology of Escheria dimidiata agrees with that of Hydrophilidae: Hydrophilinae in many aspects: (i) mesocoxal cavities transverse, narrowly isolated from each other, (ii) metanepisternum rather narrow, (iii) epipleuron narrow but reaching elytral apex, and (iv) elytron with 10 longitudinal punctural series. None of these characters or their combination is, however, unique for the Hydrophiloidea and may be found in other beetle families as well (see e.g. Lawrence et al. 1999). Moreover, the medium body size (EL: 9 mm according to Förster (1891)) would indicate that the fossil should belong to the subtribes Hydrobiusina or Hydrophilina, whose representatives are characterized by a relatively large and well developed triangular mesoscutellar shield; in contrast, the scutellar shield is very small or reduced in Escheria dimidiata. Moreover, the re-examination of the fossil is impossible as the holotype was lost or destroyed. For all these reasons, Escheria dimidiata is removed from the fossil record of the Hydrophiloidea and is placed in Polyphaga incertae sedis.  (1891), the ventral morphology of this species agrees with that of Hydrophilidae: Hydrophilinae in many characters: (i) mesocoxae transverse and very narrowly separated, (ii) mesepimeron well separated, triangular, (iii) metanepisternum rather narrow; (iv) abdomen with fi ve ventrites.
However, none of these characters or their combination is unique for the Hydrophiloidea and may be found in other beetle families as well (see Lawrence et al. 1999). Moreover, two characters illustrated on the drawing and/or mentioned in the original description exclude the placement of E. punctulata in the Hydrophiloidea: (i) elytra bear only 6 deeply impressed striae [9-11 striae are present in all Hydrophiloidea with striate elytra, only rarely is the number of series higher but in such cases they are never impressed to striae]; (ii) mesoventrite fused with mesepisternal (i.e. not separated from them by sutures) [in Hydrophiloidea, the mesoventrite is fused to mesepisterna only in derived groups of the Sphaeridiinae which are characterized by a highly elevated median portion of the mesoventrite; the elevated median elevation is missing from the fossil]. For these reasons, Escheria punctulata is removed from the fossil record of Figures 7-12. 7 'Escheria' crassipunctata Förster, 1891, original illustration of the holotype 8 'Escheria' dimidiata Förster, 1891, original illustration of the holotype 9 'Escheria' punctulata Förster, 1891, original illustration of the holotype 10-12 specimen NHMB R707 (10 photograph 11 drawing 12 original drawing by Th éobald (1937)).
the Hydrophiloidea and is placed into Polyphaga incertae sedis; its family placement remains unclear.