Redescription and revision of Georges Demoulin’s Eocene Baltic amber Heptageniidae (Insecta: Ephemeroptera)

Fossils provide a window into the past, offering insights into prehistoric aquatic ecosystems, elucidating evolutionary processes at different taxonomic scales, and clarifying relationships between extinct and extant taxa, as well as their temporal and spatial distribution.

The family Heptageniidae Needham, 1901 represents Arctogaean elements with a presumed Laurasian origin^1,2. Extant Heptageniidae are markedly more diverse in the Palaearctic realm, suggesting an origin in this region with successive radiation. Their diversification in the Palaearctic is more than twice as extensive as in the Nearctic, supporting the hypothesis of vicariant dispersal with subsequent introgression into other bioregions, including the Asian refugium². Fossil records of the family remain comparatively scarce across areas formerly comprising Laurasia, but its diversity is consistently higher in the Palaearctic^{3,4,5,6,7,8,9,10,11,12,13}.

Most extant Heptageniidae are rheophilic, inhabiting running waters, e.g. Rhithrogena Eaton, 1881 and Heptagenia Walsh, 1863, although a few species, such as Heptagenia longicauda (Stephens, 1836), also occur along lake shores. Sedimentation in dynamic palaeoenvironments such as running waters is problematic, and fossil heptageniid larvae are therefore extremely rare, particularly in compression deposits. This rarity is further explained by taphonomic factors, notably the fragility of their bodies, which makes them highly susceptible to damage or destruction prior to burial¹⁴.

The fossil records of Heptageniidae comprise 25 taxa based on adults (all of them preserved as amber inclusions), and 5 larvae mostly represented by compressed fossils¹⁵. Only one of these species was described as larval inclusion from the Eocene Baltic amber by Demoulin⁴.

The earliest records of extinct Heptageniidae date to the Upper Cretaceous of North America, namely from the latest Campanian of the Wapiti Formation, Alberta, Canada¹⁶, the Redmond Formation, Labrador, Canada¹⁷, and the Turonian amber deposits of New Jersey, USA¹⁴. Two mid-Cretaceous records from Burmese amber (Kachin, northern Myanmar) remain questionable: The monotypic genus Burmaheptagenia Chen & Zheng, 2023, recently established based on a single male imago from mid-Cretaceous Burmese amber, is in fact a member of the family Vietnamellidae, as reported by Godunko et al.¹⁸. The systematic position of the second Burmese genus, Longiheptagenia Yang, Zhao & Ren, 2023, comprising two species known from both sexes, requires clarification to establish its precise subfamily placement¹⁹. Finally, a male imago of an as yet undescribed species from Burmese amber has been assigned to the subfamily Ecdyonurinae Ulmer, 1920²⁰.

Most Heptageniidae from the Eocene comprise adult specimens of the subfamily Heptageniinae Needham, 1901 (twelve nominal taxa), with fewer representatives from Rhithrogeniinae Lestage, 1917 (five taxa) and Ecdyonurinae, including other undescribed species (five taxa in total)^{3,4,5,6,8,10,11,12}. In addition, a larval fossil impression from the Klondike Mountain Formation, Republic, Washington, USA, suggests a broad distribution of Heptageniinae during the Eocene²¹.

In the Miocene, compression fossils of Heptageniinae larvae have been documented, including one species from Belyi Yar, Russia⁷, and two species from the Shanwang Formation, Shandong, China (Burdigalian)⁹. Larval specimens from Faufoilloux, France, dating to the Upper Miocene, were described as the monotypic genus Pseudocageronia Masselot & Nel, 1999¹³. The extant genus Maccaffertium Bednarik, 1979 is also recorded in fossil form from Miocene (Burdigalian) Mexican amber²².

Representatives of only three of the five subfamilies of Heptageniidae have been identified in Eocene Baltic amber: Heptageniinae^4,5, Ecdyonurinae^10,11, and Rhithrogeninae Lestage, 1917⁵. The remaining subfamilies are either absent or generally poorly represented in the fossil record. The psammophilous and monospecific Pseudironinae are strictly Nearctic without any known fossil record²³, while the limnophilous Arthropleinae, which are Holarctic, include the dubious, extinct species Electrogenia dewalschei Demoulin, 1956²⁴. Arthroplea congener (Bengtsson, 1908), currently occurring in Central Europe, is regarded as an Ice Age relict^25,26.

The study of the amber-preserved mayfly fauna is of great importance, providing valuable information on the morphology of winged stages, which dominate as inclusions in fossil resins. In contrast to compression fossils, which are often deformed or damaged during burial, extraction, or preparation, specimens preserved in amber are typically well preserved, allowing detailed morphological analysis. Traditionally, characters used to describe alate stages have included linear body measurements, wing venation, leg structure and proportions (particularly tarsal segmentation), and the shape and structure of male genitalia^3,4,5,6. More recently, the thoracic structure of adult specimens has been used to establish systematic placement and describe extinct species^8,27,28,29. The thorax, as the most sclerotised part of the Ephemeroptera adult body, is frequently preserved in fossil resins with minimal alteration in shape or proportions. Accordingly, thoracic sutures, particularly of the mesothorax, can provide more reliable criteria for assessing the systematic position of extinct taxa at family and genus level. Redescription of fossil species using thoracic characters can also clarify the affinities and potential origins of Eocene taxa.

Georges Demoulin (*1919–†1994) (Fig. 1D) examined and revised numerous adult Heptageniidae fossils from Baltic amber, among other specimens from several German collections. Most of these specimens (26 inclusions in total) are housed in the Museum für Naturkunde, Berlin, forming part of the historical G.C. Berendt and W. Simon collections, derived from ambers originating in Gdańsk and Kaliningrad (then Danzig and Königsberg, former Prussia)^5,30. Although F.J. Pictet-Baraban and H. Hagen³¹ published the first descriptions of Baltic amber Ephemeroptera (see Staniczek & Godunko³⁰), Heptageniidae were not included. Over a century later, Demoulin revisited Berendt’s collection, redescribing previously known species and describing new Ephemeroptera taxa, including adults of Heptageniidae⁵. In his series of publications from 1956 to 1970 (see also Supplementary Table S1), Demoulin defined the systematic position of fossil taxa using the standard characters known at the time, following, for example, the key provided by Ulmer³².

In 2004, 2008, and 2015, one of the authors of the present study (RJG) re-examined all mayfly collections in Baltic amber housed at the Museum für Naturkunde Berlin, including the type series of Heptageniidae species originally described by Georges Demoulin. Given that the generic placement of some of these species remained uncertain, additional morphological analyses were undertaken, focusing on thoracic structure and using characters previously proposed by Kluge^29,33,34,35. In the present contribution, we reinvestigate these eleven adult specimens and propose several important changes in the systematic positions of fossil representatives from the subfamilies Heptageniinae and Rhithrogeniinae.

Material

This contribution is the result of a long-term collaboration among all the authors. The analysis of the material presented herein is based on multiple visits to the collections of the Museum für Naturkunde, Berlin [MNB], the Staatliches Museum für Naturkunde, Stuttgart [SMNS], the Bayerische Staatssammlung für Paläontologie und Geologie, Munich [BSPG] and Geowissenschaftliches Zentrum, Universität Göttingen (Germany) [GZG], National Museums Collection Centre, National Museums Scotland, Edinburgh [NMS], conducted in 2004, 2008, 2014–2016, 2019, and 2024 (Supplementary Table S1). We also had the possibility to investigate specimens from the private Baltic amber collections of C. Gröhn, Glinde, Germany, and W. Krzemiński, Krakow, Poland.

The MNB collection of amber inclusions is one of the oldest in Europe and worldwide, comprising more than 600,000 specimens, the majority of which are Baltic Eocene amber³⁶. Through acquisitions and donations from various private collectors, the MNB collection was established during the first half of the nineteenth century. Its principal and oldest component is the collection of Georg Carl Berendt (*1790–†1850), a physician from Danzig [Gdańsk]. Berendt was among the first to study and collaborate with other researchers in describing organic inclusions in Baltic amber, including the first descriptions of the fossil mayflies (see, for example, Berendt³⁷; Pictet-Baraban & Hagen³¹; Demoulin⁵; Staniczek & Godunko^30,38; Godunko et al.³⁹; Fig. 1A).

Another significant part of the MNB holdings is the “Ostpreußischer Bernsteinschrank” [East Prussian Amber Cabinet] of Walter Simon (*1857–†1920). Professor W. Simon, a royal banker and city councilor in Tübingen, acted as a patron of science by financing the preparation of several cabinets containing carefully selected inclusions. These were subsequently donated to the University of Tübingen, the Berliner Museum für Meereskunde (founded in 1906), and the MNB. Although the amber cabinet in the Berliner Museum für Meereskunde was destroyed during the Second World War, nearly identical cabinets in the palaeontological collection of the University of Tübingen and in the MNB have survived^40,41.

The material donated by W. Simon to the MNB was divided into a didactic section and a collection of approximately 300 inclusions prepared using the “Königsberg method”. This technique, developed by Gotthold Künow (*1840–†1909), conservator at the Königsberg Zoological Institute, and later refined by Richard Klebs (*1850–†1911)^{36,37,38,39,40,41}, involves embedding cut and polished amber pieces in gum dammar (also referred to as dammar resin or dammar varnish). The specimens were subsequently re-ground, repolished, mounted on microscope slides in Canada balsam, and sealed with cover glasses (Fig. 1B,C).

Gum dammar is a natural resin obtained primarily from trees of the Dipterocarpaceae family, which are common in Southeast Asia, particularly the Philippines, Indonesia, and India. Owing to its excellent optical clarity, hardness, and ability to take a polish, it has long been used in microscopy and specimen preparation as an embedding medium, as well as in art conservation as a transparent varnish. Its properties make it an effective substitute for, or complement to, Canada balsam⁴². Remarkably, this method of preparation has proven so durable that no visible signs of ageing or degradation are apparent in the treated amber pieces more than a century later^40,41.

Although synthetic materials are increasingly common in modern museum practice (see e.g., Godunko et al.⁴³), the use of natural dammar resin exemplifies the ingenuity of earlier conservators in ensuring the long-term stability and scientific accessibility of fossil arthropods preserved in amber. Consequently, specimens from the Simon collection, sorted and prepared more than 115 years ago, serve as an exemplary model for long-term amber conservation (Neumann^36,40; Fig. 1A–C).

In 1966, the Berlin collection containing inclusions of mayflies was transferred for study to Georges Demoulin at the Musée des Sciences Naturelles de Belgique [part of the Institut Royal des Sciences naturelles de Belgique (IRSNB), located in Brussels (Belgium)], as recorded in the inventory book of the MNB by curator Dr. H. Jaeger. Among other specimens later described by Demoulin⁵, belonging to several families (such as Ameletidae, Ameletopsidae, Coloburiscidae, Ephemeridae, Ephemerellidae, Metretopodidae, Leptophlebiidae, and Siphlonuridae), six slides prepared using the “Königsberg method” from the Simon collection contain adult specimens of Heptageniidae. This material formed the basis of our study (see below).

In addition to the Berlin material, Demoulin’s studies also included two inclusions of a male and a female Heptageniidae from the BSPG, which are part of the famous and extensive collection of the industrialist and private scholar Adolf Bachofen-Echt (*1864–†1947). These specimens were provided to Demoulin by the then director of the Bavarian State Collection, Prof. Richard Dehm (*1907–†1996). They are currently stored in Munich, where they were examined by us in 2015.

Finally, two specimens were subsequently designated by Demoulin⁵, a male imago as the holotype of Heptagenia ligata and a female imago as “Rhithrogena? sp.”, are now part of the large historical Friedrich Wilhelm Stantien and Moritz Becker collection, currently housed at GZG. The material was initially housed in the collection of SMNS; subsequently, part of the amber inclusions was transferred to Göttingen. Additionally to fossil specimens, comparative material of Heptageniidae housed in the Institute of Entomology, Biology Centre of the Czech Academy of Sciences was used.

Specimen processing and imaging

Line drawings of the material housed at the MNB were produced using a Leica WILD M3Z microscope equipped with a WILD 308,700 camera lucida in 2004 and 2008. The same equipment had previously been used by us for the description and redescription of mayfly taxa from the Simon collection^44,45. Initial photographs of the type material from the MNB were taken during the same period using a Leica DC200 camera mounted on a Leica MZ FL III microscope and were digitally processed using Adobe Photoshop 5.0.

Subsequently, all specimens discussed herein were loaned to the SMNS and re-examined in 2015–2016, 2019, and 2024. Observations were conducted using Leica M205 C (Leica Microsystems, Wetzlar, Germany) and Olympus SZX7 (Olympus Corporation, Tokyo, Japan) stereomicroscopes. A Leica Z16 APO macroscope, equipped with a Leica DFC450 digital camera and operated via Leica Application Suite v. 3.1.8, was used to capture stacked series of images at varying focal planes. These image stacks were processed using Helicon Focus Pro 8.3.0 to generate composite images with extended depth of field. Final photographs were sharpened and adjusted for contrast and tonality in Adobe Photoshop™ (Adobe Systems Inc., San Jose, CA, USA). Additional line drawings were prepared using a camera lucida mounted on a Leica M205 C stereomicroscope. All graphical material was processed using native graphic tools in Windows 10 and 11.

Terminology

Additionally to a set of adult characters used by Demoulin in original descriptions, thoracic features were studied and described in detail. General anatomical terminology is based on Kluge²⁹ and Bauernfeind & Soldán⁴⁶. Morphological terms and abbreviations to describe thorax morphology used herein follow Kluge^29,34 and Kluge et al.⁴⁷. Abbreviations for wing veins used throughout the text follow Kluge²⁹, with some modifications by Bauernfeind & Soldán⁴⁶, Staniczek et al.⁴⁸, and Godunko et al.^18,43. Terminology of male genitalia follows Stocker et al.⁴⁹. The measurements of individual body parts were inferred from the specimens embedded in the fossil resins, taken with a calibration scale or ocular grid.

Systematic notes

Several different approaches exist regarding the systematics of the genus Heptagenia. Beginning with the earliest works of N.J. Kluge^8,50 and continuing through his later revisions of the family Heptageniidae^{28,29,33,35,51}, the genus is treated in a broad sense. Thus, in addition to nominal Heptagenia s.str., the concept also includes Dacnogenia, Kageronia and the Stenonema generic complex as subgenera (see also Lewis⁵²; Bednarik⁵³; Bednarik & McCafferty⁵⁴). In terms of taxonomic circumscription, Heptagenia s.l. in this classification approximately corresponds to the subfamily Heptageniinae as understood by several other authors (see below). Additionally, Cinygma Eaton, 1885 was placed by Kluge³⁵ in a separate tribe Cinygmini Kluge, 1988 within Heptageniinae, whereas other classifications consider the tribe Cinygmatini within Rhithrogeninae¹. In any case, according to current consensus, the small genus Cinygma is included in Heptageniinae (see e.g. Tomka and Zurwerra⁵⁵; Tshernova et al.⁵⁶; Bauernfeind & Soldán⁴⁶).

Further developing the broad concept of Heptagenia s.l., Kluge²⁹ formalized this interpretation by proposing the non-hierarchical taxon Heptagenia/f6 = g5 as a convenient grouping of Heptageniinae taxa (excluding Cinygma).

Bauernfeind & Soldán⁴⁶ noted that Heptagenia is usually subdivided into three subgenera, which are sometimes treated as separate genera. The authors emphasized that, in comparison with current usage in other heptageniid genera, the subgeneric arrangement within Heptagenia is probably more appropriate. Accordingly, they recognize Dacnogenia, Kageronia and Heptagenia s.str. as subgenera (same in Haybach⁵⁷). A similar approach to the systematics of Western Palaearctic genera of Heptageniidae can be seen, for example, in Ecdyonurus, which is divided into two subgenera⁴⁶. For other groups of Heptageniidae, even for such a diverse genus as Rhithrogena, it is common practice to designate numerous species-groups based on larval and adult characters (Jacob⁵⁸; Sowa⁵⁹; Soldán & Landa⁶⁰; Vuataz et al.⁶¹; Bauernfeind & Soldán⁴⁶).

In contrast to the above concept of Heptagenia s.l., many other authors treat Dacnogenia, Kageronia and several related Heptageniinae lineages as separate genera^{1,58,62,63,64,65,66,67,68,69}.

Among extant Heptageniinae, delimiting several genera and subgenera remains a challenging task. This is well illustrated by the Stenonema generic complex, which shows major difficulties in distinguishing its constituent taxa based on larval and adult morphology⁶⁶. Additionally, multigene molecular studies (mitochondrial + nuclear markers) focused on part of this North American complex (Stenonema, Maccaffertium, Macdunnoa and Stenacron) reveal a tight cluster that does not always form monophyletic groups. Published results recover Maccaffertium as paraphyletic with Stenonema, either nested within or closely adjacent to it⁶⁶. These results clearly highlight problematic generic boundaries that require further taxonomic work (either redefining genera or reassigning species), using both larval and adult characters. Finally, mitogenomic analyses⁷⁰ revealed conserved gene-order features uniting Maccaffertium, Stenacron and Stenonema, supporting the existence of a closely related subclade of genera or subgenera within Heptageniinae.

Extinct taxa often preserve only fragmentary morphological features, resulting in a much more limited character set compared to extant groups. Even in resin inclusions, identification of reliable diagnostic features suitable for establishing systematic position can be difficult or impossible. We previously demonstrated that examining thoracic structures of adult Ephemeroptera can significantly advance the systematic interpretation of fossil taxa (See Introduction). The same approach is applied here to establish the subfamilial placement of the fossil species within Heptageniidae.

However, even well-preserved amber specimens may not provide sufficient information to determine the systematic position of extinct adult Ephemeroptera with confidence. This is especially evident in Heptageniidae, where reliable diagnoses of most genera are based on larval morphology^29,35,50. Adult stages can show substantial morphological variation in wing venation, leg proportions, and genital structure within a single genus, as illustrated above by the recent Stenonema generic complex.

Therefore, the systematics of the fossil taxa described by Demoulin^5,6 presents certain difficulties. Some species, as shown in more detail below, unquestionably belong to Heptagenia s.l. (or Heptageniinae), yet cannot be associated with particular subgenera according to the classifications of Kluge^29,35 or with genera as understood by other authors. For example, while H. bachofeni may belong to Dacnogenia or Heptagenia s.str., both H. atypica and H. demoulini sp. nov. can only be placed in Heptagenia s.l. (that is, the non-hierarchical taxon Heptagenia/f6 = g5 sensu Kluge²⁹), because important diagnostic characters are either insufficiently preserved or require verification using additional investigative tools. Thus, a placement in Heptagenia s.l. rather represents a conservative nomenclatural assignment and should not be interpreted as a phylogenetic hypothesis implying a placement within Heptagenia s.str.

Given the general difficulty of delimiting genera and subgenera within Heptageniinae, particularly in fossil material where key adult structures (e.g. male genitalia) are incompletely preserved and conspecific larvae remain unknown, the present study follows the broad taxonomic concept of Heptagenia as employed by Kluge^29,35 and Bauernfeind & Soldán⁴⁶. Fossil species that cannot be confidently assigned to existing subgenera are provisionally placed in Heptagenia s.l., corresponding in the circumscription of the taxon Heptagenia/f6 = g5 sensu Kluge²⁹. This temporary approach will remain in place until problematic fossil species can be reassessed using additional techniques (e.g. µCT analysis) or until newly discovered, well-preserved material becomes available. Such a framework ensures systematic stability and a consistent nomenclatural structure for Heptageniinae.

Nomenclatural acts

The electronic version of this article conforms to the requirements of the International Code of Zoological Nomenclature, and hence the new names and combinations 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:9089A11A-6506-40C3-87B2-9284A9F69A7D. A new species is registered under urn:lsid:zoobank.org:act:E4F3C251-7C46-4291-A604-B38D96793F5C. The online version of this work is archived and available from the following digital repository: https://www.nature.com/srep/.

Systematic palaeontology

Class Insecta Linnaeus, 1758.

Subclass Pterygota Lang, 1888.

Order Ephemeroptera Hyatt & Arms, 1891.

Family Heptageniidae Needham, 1901.

Subfamily Rhithrogeniinae Lestage, 1917.

Genus Rhithrogena Eaton 1881.

Rhithrogena Eaton, 1881; Entomol. Mon. Mag. 18: 23.

Type species Baetis semicolorata Curtis, 1834; Phil. Mag., 1834: 121 [original designation].

Included species More than 150 extant species with Holarctic distribution, 75 of them in the Palaearctic Region, and a few species in the Oriental Region. Except for a single Eocene species assigned to this genus by Demoulin⁵ and discussed below, the monotypic genus Burshtynogena fereci Godunko & Sontag, 2004, established based on a female imago from Baltic amber¹², is attributed together with R. sepulta to the subfamily Rhithrogeninae in the current contribution.

Rhithrogena sepulta Demoulin, 1968

(Figures 2, 3 and 4; Table 1).

Rhithrogena sepulta Demoulin, 1968; Dtsch. Entomol. Z., 15: 261, fig. 27a, b [original description]

Rhithrogena sepulta Demoulin, 1968: Keilbach⁷¹: Dtsch. Entomol. Z., 29: 208 [list of fossil mayfly taxa]; Hubbard⁷²: Fossilium Catalogus I: Animalia, 129: 50 [list of fossil mayfly taxa]; Spahr⁷³: Stuttgarter Beitr. Naturk. Ser. B, 182: 8 [list of fossil mayfly taxa]; Wichard & Weitschat⁷⁴: Entomol. Mitt. Löbbecke-Museum + Aquazoo Beiheft, 4: 25, fig. 5 [list of fossil mayfly taxa]; Weitschat & Wichard⁷⁵: Atlas of Plants and Animals in Baltic Amber: 92, fig. 43 [list of fossil mayfly taxa]; Godunko & Neumann⁴⁴: Ann. Zool., 56(1): 177 [list of fossil mayfly taxa]; Wichard et al.⁷⁶: Aquatic Insects in Baltic Amber 291 [list of fossil mayfly taxa]; Varela-Hernández et al.⁷⁷: Can. J. Earth Sci., 59: 483 [tabular list of fossil mayfly taxa]

“Rhithrogena” sepulta Demoulin, 1968: Kluge⁸: Paleontol. J., 2: 112 [comparative characters]

sepulta Demoulin 1968 [Rhithrogena]: Kluge²⁹: The Phylogenetic System of Ephemeroptera: 208 [list of fossil Heptageniidae]

Rhithrogena sepulta: Yang et al.¹⁹: Cretac. Res., 147: 7 [tabular record of fossil Heptageniidae]

Type material Holotype: male imago, Eocene, Baltic amber, originally labelled: “Holotype Fam. Ephemeridae 5”; “G. Demoulin det. 1966 Rhithrogena sepulta sp. n. ♂ imago Demoulin, 1968, fig. 27”; inventory number MB.I.2243; housed in MNB, Simon collection (Fig. 2A).

Type locality and horizon Eocene, Priabonian, 37.7–34 Ma, following the revised dating of Baltic amber by Ross et al.⁷⁸; the same age assignment is applied to all material examined in the present study. The piece of amber with the embedded holotype of R. sepulta originates from Eocene deposits of Baltic amber in the Königsberg Region [Kaliningrad]. The piece of amber is additionally enclosed in translucent resin to seal the specimen from oxygen. Due to its fragility, it is mounted on a slide to prevent mechanical damage (Fig. 2A).

Diagnosis Male imago Rhithrogena sepulta is a representative of the subfamily Rhithrogeninae due to the combination of following characters: (a) thorax of Rhithrogeninae type: mesonotum with transverse suture; lateroparapsidal sutures laterally curved, not connected to medioparapsidal sutures; prosternum lacking transverse ridge; mesosternal median impression between furcasternal protuberances, narrowing anteriorly; (b) cubital field of forewings with two pairs of long intercalary veins; (c) first tarsomere of all legs distinctly articulated with tibia.

It is distinguished from other extant and fossil Rhithrogena by the combination of following characters: (i) body length 5.24 mm; forewing length 4.92/5.16 mm [all as preserved]; (ii) upper compound eyes large, medially contiguous; (iii) hind wing crossveins developed only between Sc and RA; (iv) hind-wing costal process well developed, apically rounded; (v) penial lobes short; apices truncate, divergent, not laterally expanded or outwardly curved; apical teeth absent (ventral view); (vi) titillators broad-based, apices indistinct.

Redescription of the holotype [male imago] (Figs. 2, 3 and 4). Specimen almost complete, poorly visible in dull, only partly translucent amber; resin with scattered, small whitish-greyish spots or cracks; piece of amber enclosed within artificial, translucent resin of unknown composition and mounted on microscopic slide (Fig. 2A–C). Surfaces of head, thorax, and abdomen dorsally and ventrally darkened, riddled by cracks rendering the epidermal morphological features nearly indiscernible. All wings well preserved, stretched laterally; details of venation generally well visible (Fig. 2B–D); some wing parts covered by cracks and darkened, especially outer margin of right fore- and hind wings. The left hind leg missing (Fig. 2C).

Colouration Preserved colour of specimen yellow ochre, with brown to dark brown fissured texture on the surface of head, thorax and abdomen. Eyes yellow ochre, with brown to dark brown maculations; facial keel brown to light brown. Mesonotum light brown, with dark brown maculae centrally between two lateroparapsidal sutures, and from mesonotal suture to scutoscutellar impression. Fore- and hind wings covered with dirty brown colour, cracks with dark brown sides. Legs almost translucent; forelegs lighter than middle ones, even more than hind leg coloured in brown to dark brown. Abdomen with completely dark opaque segments VII–XI; segments I–VI paler, with dark brown markings along lateral margins of abdominal segments (Fig. 2B,C).

Measurements Body length 4.56 mm; length of forewings 5.16/5.08 mm; length of hind wings 1.60/1.66 mm; hind/forewing length ratio 0.31–0.33; length of left cercus 1.25 mm (damaged). See Table 1 for other measurements.

Head General colour dirty yellow to dark brown. Compound eyes with dark fissured patches, right side of right compound eye covered by a blackish cloud. Compound eyes rounded and large, contiguous medially; border between upper and lower portions of compound eyes indiscernible; ocelli poorly visible, relatively large. Facial keel not prominent; antennae translucent, scape and pedicel darker than flagellum (Fig. 3A,C).

Thorax Mesonotal suture [MNs] transverse, poorly visible; medioparapsidal suture [MPs] straight, not recognizable posteriorly; lateroparapsidal suture [LPs] clearly visible, moderately curved laterally, not joining medioparapsidal suture (Fig. 3A,B). Prosternum without transverse ridge; median impression of furcasternum of mesothorax narrowed anteriorly; furcasternal protuberances [FSp] oval-shaped (Fig. 3C).

Wings Wings opaque. Fore- and hind wings opaque; venation typical of Heptageniidae, i.e. two pairs of intercalaries present in cubital field of forewings, and hind wings with RS, MA and MP triads; crossvenation of forewings well-developed. Hind wing crossvenation poorly developed, with fewer transversal veins; without basally connected veins along posterior margin of wing (Fig. 3D,E).

Forewings Longitudinal venation light brown to brown; crossvenation well-developed, light brown; pterostigma with 7 simple crossveins; C and Sc uniformly light brown; RS forked near wing base, at ca. 1/4 of its length; iRS long, connected with RSp by 7 crossveins, proximally equidistant to RSa₁ and RSp; MA fork slightly asymmetrical, forked after ca. 0.55 of its length; iMA long, connected with MA₁ and MA₂ by 1–2 crossveins; MP asymmetrical, forked after 0.35 of its length, MP₁ and MP₂ basally connected by a single crossvein; iMP relatively short, connected with MP₁ and MP₂ by 2 crossveins from each side; CuA closely approximated to CuP; CuP smoothly curved toward wing base; only single cua–cup crossvein between CuA and CuP apically, no cup-a₁ crossvein between CuP and A₁; two pairs of simple intercalary veins iCu₁,–iCu₂ and iCu₃–iCu₄ in cubital field, each reaching basitornal margin; cubital intercalaries not arising from CuA or CuP; A₂ not arising from A₁; no crossveins in anal field; intercalary veins iRSa₁, iRSa, iRS and iMP connected to longitudinal veins by 3–7 crossveins; two small, basally free marginal intercalaries in MA (one) and Cu (one) fields, and between CuP and A₁ (one); one marginal intercalary vein in cubital field and one such vein basally linked to iCuA₄ (Table 1; Fig. 2D).

Hind wings opaque, relatively narrow, widely rounded distally; venation light brow; crossvenation poorly developed, present mainly between Sc and RA; no basally connected veins along of posterior margin of wing; one crossvein between C and Sc; ten crossveins between Sc and RA; 2–3 crossveins between RA and RSa; Rsp connected apically to RSa by one crossvein, and by another further to RA; iRS short, not connected to other veins; MA forked, longer than MP; MP forked; CuA relatively long; CuP present; one free small marginal intercalary in cubital field. Costal process protruding, smoothly pointed apically, situated close to wing base (Table 1; Fig. 3D,E).

Legs Legs dirty yellow stained with dark brown coloration (Fig. 2B,C), especially middle legs and even more right hind leg, which is almost completely darkened [coloration probably due to fossilization]; femora yellow to brown, tibiae and tarsi paler, yellowish with a few light brown spots (Fig. 4D); right foreleg femur-tibia-tarsus [tarsal segments] length ratio: 0.98 : 1.48 : 1.94 [0.18 : 0.60 : 0.63 : 0.38 : 0.15]; left foreleg length ratio: 0.95 : 1.53 : 1.94 [0.18 : 0.63 : 0.63 : 0.35 : 0.15]; pretarsal claws dissimilar, with one pointed and one blunt claw.

Abdominal segments All preserved segments opaque, completely preserved, segments VI–X darkened; no prominent posterolateral projections; three last abdominal segments not elongated compared to proximal segments; preserved segments of left cercus light brown, first segment seems darker (Fig. 2B,C).

Genitalia The shape of genitalia is typical for Rhithrogena. Penial lobes stout and straight, clearly divergent apically; one lobe width distant from each other at the base; penial lobes not broadened at the base; tip of each lobe truncates obliquely, with inner margin shorter, not curved outwards; penial vesicles well-developed; apical, inner apical or outer teeth not visible (Fig. 4A–C). Titillators poorly visible, with unclear shape, seems to have a wide base; left titillator whitish, seems membranous, folded longitudinally and pointed apically; right titillator also looks membranous and visualised as two poorly visible outlines (see Fig. 4B,C). Gonobasal plate uniformly dirty brown; gonostyli and penial lobes of the same colour; posterior margin of gonobasal plate arcuately incurved in the middle; two slightly asymmetrical and obtusely rounded protuberances laterally near gonostylomere I; gonostyli with four gonostylomeres; basal gonostylomere I slightly elongated, asymmetrical; gonostylomere II longest, without hump on inner margin, 4 × longer than gonostylomere III; gonostylomere IV shortest, as long as 3/4 of gonostylomere III (Fig. 4A).

Remarks

Georges Demoulin provided a very short description of this species, mainly axed on legs proportions and shape of genitalia. However, the structure of the thorax was not mentioned and analysed in the original description⁵. Our investigation of thoracic sutures combined with genital characters, especially the shape of penial lobes, confirms with high probability Demoulin’s attribution of this specimen to the genus Rhithrogena. On the other hand, the previously described female subimago of the extinct genus Burshtynogena Godunko & Sontag, 2004 from Eocene Baltic amber most probably also belongs to Rhithrogeninae, closely resembling the fore- and hind wing vernation of the fossil specimen of R. sepulta¹². However, the posterior part of the mesonotum in the holotype of B. fereci Godunko & Sontag, 2004 is invisible, thus the arrangement of lateroparapsidal and medioparapsidal sutures is impossible to indicate. Based on this consideration, the systematic position of Burshtynogena relative to Cinygmini and Rhithrogenini is still unresolved¹². Nevertheless, the absence of the prosternal transversal ridge in R. sepulta and B. fereci and the anteriorly narrowing furcasternal impression define the position of both taxa within Rhithrogeninae as mentioned above (see fig. 1, 2 in Godunko & Sontag¹²; Kluge^29,35; Fig. 3C). Additionally, in contrast to Burshtynogena, R. sepulta can be differentiated by the hind wing proportions, with a width/length ratio 0.56–0.57 (in contrast to B. fereci with a ratio about 0.45; see Godunko & Sontag¹²), and by nearly oval-shaped furcasternal protuberances typical for Rhithrogena (in contrast to almost rectangular protuberances in B. fereci).

Rhithrogena sepulta differs from all recent and fossil representatives of Rhithrogena by:

1. (i)

   The shape of penial lobes not broadened at the base, as it is the case in R. bogoescui Sowa & Degrange, 1987 (Rhithrogena alpestris group), but similar to the species attributed to R. diaphana group (Alba-Tercedor, 1998). The tip of each penial lobes of R. sepulta is obliquely truncate (inner part of lobe is shorter), not curved outwards as in R. zernyi Bauernfeind, 1991 (R. diaphana group). Ventral penial membrane well-preserved, most likely reaching at least 1/2 of penial shaft (2/3 in R. zernyi). In contrast to the many species of Rhithrogena, the apical, inner apical and outer teeth are absent, at least not visible in ventral view^46,59. In ventral view, in the R. alpestris and R. diaphana group only the inner apical tooth is visible; in the R. germanica group apical teeth are not visible and the inner apical tooth is missing; in R. hybrida group, the outer apical tooth is not visible and the inner apical tooth is small but present, while in the R. loyolaea group it is well visible and strongly developed; in the R. semicolorata group the outer apical tooth is distinctly visible and lastly, in the R. sowai group the apical tooth is usually not visible (for more details see Bauernfeind & Soldán⁴⁶).

2. (ii)

   Unlike in almost all Rhithrogeninae (including Cinygma), there are a few simple crossveins in the forewing pterostigma of R. sepulta (Fig. 2B,C). The same character state was also described for Burshtynogena¹².

3. (iii)

   The leg proportions of R. sepulta are well described in detail by Demoulin⁵ and were re-examined in the present contribution (see e.g. Table 1). However, similarly, as noted by Kluge^8,35, leg proportions are inappropriate to define subgenera due to the variability of this character. This was also mentioned by Tshernova & Belov⁷⁹ in the diagnosis of Cinygmula.

4. (iv)

   R. sepulta differs from most of the species in the subfamily Rhithrogeninae in its relatively small body size (Table 1). It resembles the representatives of the R. diaphana group defined by Bauernfeind & Soldán⁴⁶, a complex of species with relatively narrow head, small and slender body. However, R. sepulta can be separated from this group by the shape of the genitalia (see above).

Other characters that would allow the fossil species to be assigned to one of the known subgenera or groups of species, thus allowing to establish a systematic position among Rhithrogena s.l., are not available for this study.

Subfamily Heptageniinae Needham, 1901.

Genus Heptagenia Walsh, 1863.

Heptagenia Walsh, 1863; Proc. Entomol. Soc. Philadelphia 2: 197.

Type species Palingenia flavescens Walsh, 1862; Proc. Acad. Nat. Sci. Philadelphia, 13–14: 373; designation by Eaton⁸⁰ in Entomol. Mon. Mag. 5: 90.

Included species The genus Heptagenia comprises about 30 extant species in the Holarctic realm (except of North Africa, Central America as far south as Guatemala) and the Oriental region⁴⁶. The Eocene fauna of the genus Heptagenia comprises five formal species described by Demoulin⁵, which were re-examined in the current study. Two specimens (male and female imagines), which G. Demoulin indicated as Rhithrogena without assignment to any species, are also attributed to the genus Heptagenia here. Although a detailed study of the systematic position of H. atypica is expected in the future, in the current study we place this species within Heptagenia. At the same time, we transfer H. senex, originally described in the genus Heptagenia⁵, to the genus Kageronia (see below).

Heptagenia (s.l.) atypica Demoulin, 1968

(Figures 5, 6 and 7; Table 2).

Heptagenia atypica Demoulin, 1968; Dtsch. Entomol. Z., 15: 258, fig. 24a, b [original description]

Heptagenia atypica Demoulin, 1968: Keilbach⁷¹: Dtsch. Entomol. Z., 29: 208 [list of fossil mayfly taxa]; Hubbard⁷²: Fossilium Catalogus I: Animalia, 129: 50 [list of fossil mayfly taxa]; Spahr⁷³: Stuttgarter Beitr. Naturk. Ser. B, 182: 8 [list of fossil mayfly taxa]; Wichard & Weitschat⁷⁴: Entomol. Mitt. Löbbecke-Mus. + Aquazoo Beiheft, 4: 25 [list of fossil mayfly taxa]; Weitschat & Wichard⁷⁵: Atlas of Plants and Animals in Baltic Amber: 92 [list of fossil mayfly taxa]; Godunko & Neumann⁴⁴: Ann. Zool., 56(1): 177 [list of fossil mayfly taxa]; Wichard et al.⁷⁶): Aquatic Insects in Baltic Amber: 291 [list of fossil mayfly taxa]; Varela-Hernández et al.⁷⁷: Can. J. Earth Sci., 59: 483 [tabular list of fossil mayfly taxa].

“Heptagenia“ atypica Demoulin, 1968: Kluge⁸: Paleontol. J., 2: 112 [list of fossil taxa]

atypica Demoulin 1968 [Heptagenia]: Kluge²⁹: The Phylogenetic System of Ephemeroptera: 208 [list of fossil Heptageniidae].

Heptagenia atypica: Yang et al.¹⁹: Cretac. Res., 147: 7 [tabular record of fossil Heptageniidae].

Type material Holotype male imago, Eocene, Baltic amber, originally labelled: “Holotype Fam. Ephemeridae. Demoulin 1968, fig. 24 13”; “G. Demoulin det., 1966 Heptagenia? atypica sp. n. ♂ imago MB.I.2240”; housed in MNB, Simon collection (Fig. 5A).

Type locality and horizon Eocene, Eocene, Priabonian, 37.7–34 Ma. The pieces of amber with embedded holotype of H. atypica originates from the unknown Eocene deposits of Baltic amber in former Königsberg [Kaliningrad] Region. As with other amber inclusions from the Simon collection, this specimen is embedded in Gum dammar and mounted using the “Königsberg method” (Fig. 5B; see also Material and methods).

Diagnosis Male imago Heptagenia atypica is classified within Heptageniinae due to following character combination: (a) thorax of Heptageniinae type: prosternum with distinct transverse ridge and narrow longitudinal keel; mesonotal sutures transverse; lateroparapsidal sutures not laterally curved, connected to medioparapsidal sutures; furcasternal protuberances narrowing anteriorly; (b) cubital venation with two pairs of long intercalaries; (c) tarsomere I of all legs distinctly separated from tibia.

It is distinguished from other extinct and extant Heptagenia by the following character combination: (i) body length 5.24 mm; forewing length 4.92/5.16 mm [all as preserved]; (ii) compound eyes broadly separated medially at the distance 3.5 × wider than frontal ocellus maximum width; (iii) forewings relatively short, not narrow; (iv) crossvenation of both wings well developed; no crowed veins; no spots and joints between and around veins; (v) hind wings 0.31–0.33 length of forewings; (vi) first tarsomere of foreleg 0.6 × length of tarsomere II; (vii) hind tarsomere I slightly shorter than tarsomere II; (viii) penial lobes short, small, not laterally expanded; dorsal fold external spine flattened, rounded, weakly projecting above penial lobe; no lateral spines at penes base; (ix) titillators small, acutely pointed, moderately hooked; (x) gonobasal plate moderately concave anteriorly, without projections.

Redescription of the holotype [male imago] (Figs. 5, 6 and 7). The specimen is well preserved in a piece of translucent amber which is mounted on a microscopic slide, and visible in dorsoventral aspect. Thorax, wings and abdomen partially covered with resin containing cracks, gas bubbles, and various small inclusions. Right foreleg and left middle leg not preserved; abdomen and genitalia complete; left hind leg and cerci absent.

Colouration The specimen is yellowish to intensively brown coloured; head and thorax darker than most of abdominal segments, except of segments VIII–X and genitalia dark brown coloured; abdominal segments I–VII partly translucent; legs of same colour as thorax (Fig. 5A–D).

Measurements Body length 5.24 mm; length of forewings 4.92/5.16 mm; length of hind wings [right] 1.60 mm; hind/forewing length ratio 0.31–0.33. See Table 2 for other measurements.

Head Compound eyes round and large, medially broadly separated by a gap 3.5 × wider than frontal ocellus maximum width; border between upper and lower portions of compound eyes indiscernible. Ocelli well preserved, lateral ocelli larger than frontal ocellus. Facial keel short, obtuse. Antennae uniformly dark brown (Fig. 5F).

Thorax Prosternum with distinct transverse ridge; longitudinal keel narrow, weakly expressed; MNs transverse; LPs well expressed, not curved laterally, connected to MPs; median impression between FSp narrowing anteriorly (Figs. 5E, 6A–C).

Wings opaque, translucent. Crossvenation well-developed; forewing crossveins between Sc and R not crowded; no dark clouds and spots around crossveins; no blackish joints of crossveins (Figs. 5D, 6D,E).

Forewings relatively short, not narrow; basitornal margin as long as 0.43 of tornoapical margin length; tornus placed ca. 0.32 from the wing base. Longitudinal venation pale, yellowish-brown; veins slightly darker basally; crossvenation slightly paler than longitudinal veins, well-developed, without clusters between Sc and R field. Pterostigma with 10–13 simple veins; C and Sc yellowish-brown; RS fork close to wing base, at less than 1/4 of its length; iRS long, connected with RSp by 9–10 crossveins; iRS free, approximated to RSp; MA fork slightly asymmetrical, forked after ca. 0.62 of its length; iMA shortened connected to MA₁ and MA₂ by 3 crossveins from each side; MP asymmetrical, forked after 0.22 of its length, MP₂ arise from MP₁; iMP long, connected to MP₁ and MP₂ by 3–5 crossveins from each side; CuA closely approximated to CuP basally; CuA and CuP smoothly curved toward wing margin; two crossveins between CuA and CuP basally; two pairs of intercalaries in cubital field well distinguishable; no additional small intercalaries in cubital field; AA well visible, connected to CuP by 1–2 crossveins; additionally three longitudinal veins in anal field. Basally connected small, marginal intercalaries present between RS and CuA (Table 2; Fig. 6E).

Hind wings opaque, 0.31–0.33 × of forewings length. Costal process near wing base, apex acute, angle ca. 120°, outer margin smoothly curved. Venation brown to light brown; no crossveins visible, except between C–MA. Crossveins: 9 (C–Sc), 10 (Sc–RA), 5 (RA–RSa). MA–RSp fork at 0.13 of vein length. iRS long, basally connected to RSa and RSp. MA forked at 0.59. Cubital veins forked; furcation indistinct (Table 2; Fig. 6D).

Legs yellowish-brown to intensively brown; forelegs paler than middle and hind legs; right foreleg femur-tibia-tarsus [tarsal segments] length ratio: 1 : 1.17 : 0.94 [0.15 : 0.25 : 0.17 : 0.20 : 0.17]; right middle leg length ratio: 1 : 0.87 : 0.57 [0.09 : 0.12 : 0.12 : 0.12 : 0.12]; right hind leg length ratio: 1 : 0.85 : 0.54 [0.10 : 0.11 : 0.12 : 0.10 : 0.11]. Pretarsal claws of each leg with outer claw hooked and inner claw blunt (Table 2; Fig. 7A–D).

Abdominal segments completely preserved, segments I–VII translucent, VIII–X opaque; no posterolateral projections on abdominal segments. Abdominal terga uniformly light brown to brown; segments VIII–X dark brown to blackish brown; nerve ganglia not distinguishable. Cerci absent (Fig. 5C,D).

Genitalia Gonostyli with four gonostylomeres; gonostylomere I slightly wider than long; gonostylomere II ca. 4 × length of gonostylomere III, and ca. 5.9 × length of gonostylomere IV; gonobases short, narrow, not prominently visible. Gonobasal plate moderately concave anteriorly, without projections. Penes short, small, not expended laterally; penial sclerites and folds poorly distinguishable; external spine of dorsal fold flattened, rounded, weakly protruding above lobe. Penes not laterally expanded, no lateral spines at penes base; no dorsal spines. Titillators small, acutely pointed, moderately hooked (Fig. 7E–F).

Remarks

As the species name H. atypica itself suggests, Demoulin clearly recognised that its adult characters differ from the diagnostic features of Heptagenia s.str. The establishment of this species was accompanied by a very brief description and commentary (Demoulin⁵: 258–259, fig. 24a, b), in which he noted that, despite the good preservation of the male imaginal genitalia, the gonobases and penes “do not strictly conform to the Heptagenia type; they are rather reminiscent of Epeorus”.

Small discrepancies were detected between the information presented in the original description and the results of our re-examination. These differences concern body and wing length, as well as the proportions of leg segments. The tarsal formulas of the fore- and hind legs calculated in the original description differ only slightly from the morphometric data obtained in our study (Table 2; Demoulin⁵). Tarsomere I is only slightly shorter than the fifth tarsomere on the foreleg, and markedly shorter than tarsomere II, a condition typical of Heptagenia s.l.

Furthermore, the placement of H. atypica within Heptagenia s.l. is supported by the presence of a distinct transverse ridge on the prosternum, in combination with lateroparapsidal sutures that do not curve laterally and reach the mesoparapsidal sutures, as well as a furcasternal impression narrowing anteriorly (Figs. 5E, 6A–C). In addition to these characters that are diagnostic for Heptagenia s.l., the species exhibits the simple veins in the forewing pterostigma that are typical of Heptageniinae.

On the other hand, the compound eyes in Heptagenia atypica are distinctly separated by a gap exceeding the diameter of the frontal ocellus. This character clearly differentiates the fossil species from all representatives of Heptagenia s.str. Among extant Palaearctic taxa, widely spaced male compound eyes are also observed in H. orbiticola Kluge, 1986 (in Tshernova et al.⁵⁶); however, this species belongs to the subgenus Kageronia. Representatives of Kageronia are readily distinguished from other taxa of Heptagenia s.l., including H. atypica, by the absence of a distinct longitudinal keel on the prosternum, as emphasised by Webb & McCafferty⁶².

Another set of characters separating representatives of different subgenera within Heptagenia s.l. relates to the relative proportions of tarsomeres I and II of the fore- and hind legs in adult stages. The taxonomic relevance of this character was first highlighted by Kimmins⁸¹, who noted that the relative lengths of the first and second tarsomeres of the hind leg in imagines differ markedly among genera and subgenera of the Heptageniidae. In that study, several species then assigned to Heptagenia are now placed within Ecdyonurinae, for example Electrogena lateralis (Curtis, 1834). For the extant species Kageronia fuscogrisea, Kimmins⁸¹ stated that the first tarsomere of the hind leg is longer than the second one, whereas in Heptagenia (including Dacnogenia) the first hind tarsomere is shorter than, or subequal to, tarsomere II. Subsequently, this character has been repeatedly employed in adult-based diagnoses and comparative morphological analyses. It was used by Flowers⁶³ (p. 96) to separate “true Heptagenia” from the complex of genera traditionally assigned to the Ecdyonurinae, by Saaristo & Savolainen⁸² (p. 188, figs. 1–8) in their analysis of diagnostic adult characters of Heptagenia sulphurea (Müller, 1776) and H. dalecarlica Bengtsson, 1912 based on Scandinavian material, by Tomka & Zurwerra⁵⁵ (p. 119) in a key to the genera of the Heptageniidae, and by Kluge⁸ (p. 112) in the description of a fossil representative of Kageronia from Baltic amber. The same character was further discussed by Jacob et al.⁶⁴ (p. 93), Blackburn et al.⁸³ (p. 260), and Bauernfeind & Soldán⁴⁶ (pp. 308, 313) in their treatments of adult diagnostic characters within Heptageniidae. Heptagenia atypica exhibits a shortened first segment of the hind tarsus, as observed in most representatives of the Heptageniinae, and in this respect clearly differs from Kageronia and Cinygma (Table 2; Fig. 7C,D).

The relative length of foretarsomeres I and II in adult Heptageniinae, particularly in males, has also been repeatedly cited as a diagnostic character for generic delimitation. In his classical series of contributions, Eaton⁸⁴ incorporated the foretarsal formula into adult diagnoses of numerous genera within Heptageniidae and other mayfly families. In several cases, he stated for Heptageniinae that the relative length ratios of first to second foretarsal segments would be constant in males and females, respectively.

Based on these data, Traver⁸⁵ (p. 107) used this character in her key to male imagines for North American Heptageniidae. For Cinygma, Arthroplea, Iron and Epeorus, Traver⁸⁵ noted that the “first joint of fore tarsus [is] equal or almost equal to second joint, or slightly exceeding it in length”, and separated these genera from Heptagenia, Stenonema, Anepeorus and Rhithrogena, in which the first segment of the foretarsus was described as “considerably shorter” than the second one. When further separating genera of the Heptageniinae, Traver⁸⁵ characterised Heptagenia by the presence of a “first joint of fore tarsus very short (1/6 to 1/4 of second)”, whereas in Stenonema the first foretarsal segment was indicated as ranging from 1/3 to 2/3 the length of the second. It should be noted that nearly a century ago the systematics of the Stenonema generic complex had not yet been sufficiently developed, and many of the species currently included in this complex had not yet been described. Consequently, in Traver⁸⁵ the stated condition of the foretarsal segment proportions most probably refers to a composite character state encompassing what is now recognised as three distinct genera (or subgenera in other classifications) within this complex, namely Stenonema, Stenacron, and Maccaffertium. This interpretation is supported by recent studies conducted by Beacon⁶⁵ (pp. 300, 301) on different species of Stenacron, which demonstrated that the length of foretarsomere I represents 35–55% of the length of foretarsomere II, i.e. on average approximately 40–45%, as concluded by the author.

A comparable approach was adopted in the key of Flowers & Hilsenhoff⁸⁶, as well as in the comprehensive monograph on European mayflies by Bauernfeind and Soldán (2012). In the latter work, adult Heptagenia s.str. and Dacnogenia are characterised by a first foretarsal segment corresponding to approximately 1/5 to 1/4 of the length of segment II, whereas for Kageronia fuscogrisea it is stated that in male imagines the first tarsal segment is long, while in H. orbiticola it is short.

In contrast, in their global key to heptageniid genera, Tomka & Zurwerra⁵⁵ proposed a different diagnostic criterion for some Heptageniinae, expressing the length of the first foretarsal segment in relation to the length of the tibia in both males and females. Finally, for males of the small genus Cinygma, Lehmkuhl⁸⁷, followed by Webb & McCafferty⁶², indicated the length of the first foretarsal segment as 2/3–7/8 of the length of the second segment, corresponding to a ratio of 0.67–0.875.

In Heptagenia atypica, the preserved right foreleg shows a tarsal segment I/II ratio of 0.6 (Table 2), a value closer to that observed in Cinygma, Kageronia and the Stenonema generic complex than in Heptagenia s.str. and Dacnogenia.

The most obvious differences between H. atypica and recent species of Dacnogenia, Kageronia, and Heptagenia s.str. are observed in the structure of the male genitalia. In contrast to the subgenera listed above, which are characterised by large, distally elongated and folded penial lobes lacking distinct sclerites, the penes of H. atypica are short and small, without lateral expansion. It should also be emphasised that in this extinct species the penial sclerites and folds are poorly distinguishable. Ventrally, only the external spine of the dorsal fold can be reliably recognised; however, in H. atypica it is flattened, rounded and only weakly projecting above the lobe, unlike in extant species of the subgenera listed above (Fig. 7D).

In the shape and structure of the penial lobes, H. atypica is very close to the Nearctic species of the Stenonema generic complex, which comprises Stenonema, Maccaffertium, Macdunnoa, and Stenacron, treated at either generic or subgeneric rank (for a detailed discussion, see above). However, males of H. atypica differ from those of Stenonema, in which the penial lobes are large and L-shaped, by having small, apically rounded lobes (Fig. 7E,F; see Webb & McCafferty⁶²).

Additionally, in H. atypica the crossveins are not clustered between Sc and RA in the middle part of the forewing (Fig. 6E). It separates the extinct taxon from species of Stenonema, in which such crossveins clusters are present and often surrounded by dark, brownish to blackish clouds (the same pattern is observed in many other members of the Stenonema generic complex; see Webb & McCafferty⁶² and Beacon⁶⁵). The absence of the clouds around the crossveins in H. atypica are most probably not related to the condition of fossilisation and correspond to the natural wing plate colouration. The colour pattern with presence specific clouds around of the crossveins is described in some other Heptageniidae from fossil resins. For example, clouds around the crossveins are clearly visible in the Miocene species Maccaffertium annae Macadam & Ross, 2016 from Mexican amber, belonging to the Stenonema generic complex, as well as in several species of Borinquena Traver, 1938 (Leptophlebiidae Banks, 1900) from Dominican amber^22,88,89).

Heptagenia atypica differs considerably from Maccaffertium, as its males possess the same type of penial morphology as Stenonema, that is, subquadrate penial lobes with a pair of spines on the dorsal surface^62,65. In addition, Maccaffertium exhibits a cluster of crossveins between Sc and the R field of the forewings, which is absent in H. atypica.

Heptagenia atypica is readily distinguished from recent North American species of Stenacron by the absence of robust spines on the lateral margins at the base of the penes, as well as by the absence of a black spot linking several (typically two or three) crossveins between RA and RSa₁ of the forewings⁶².

Heptagenia atypica is very close to the small genus Macdunnoa, which currently includes three Nearctic species^{29,90,91,92,93,94,95,96}. The shared characters in the males include: (i) relatively small body size and small fore- and hind wings; these values range from 5.24 mm/4.92–5.16 mm/1.6 mm in H. atypica, to 6.0–8.0 mm/7.0–8.0 mm/2.2–2.4 mm in Macdunnoa; hind wings are as long as 0.3 of the forewing length in both taxa; (ii) widely separated compound eyes, with the distance between them being 3.5 × the maximum width of the frontal ocellus in H. atypica, and up to 5 × in Macdunnoa; (iii) a gonobasal plate that is concave and lacks marked projections in both taxa; (iv) penial lobes fused along most of their length, relatively small, not expanded laterally, and nearly rounded lateroapically; (v) titillators acute apically and swollen at least basally (Fig. 7E,F; Burks⁹²: p. 182, 184, fig. 363; Lehmkuhl⁹⁰: p. 860–862, fig. 3; Flowers⁹³: p. 25, 27, 28, figs. 13, 14; Berner & Pescador⁹⁴: p. 190–191; Webb & McCafferty⁶²: p.8, 42, fig. 162). Finally, the proportions of the segments of the fore- and hind legs, as well as the tarsal formulas of the two taxa, are similar (compare Table 2 and Flowers⁹³: p.25).

On the other hand, the external spine of dorsal fold in H. atypica is clearly delineated, projecting above the anterior margin of the penial lobes slightly more than in species of Macdunnoa. The titillators of H. atypica are basally broadened and clearly taper apically, whereas in recent species of Macdunnoa they remain massive [or swollen in the terminology of Flowers⁹³, almost along their entire length.

We do not exclude the possibility that H. atypica, which is morphologically close to Macdunnoa in the male imago, represents either the sister group to this Nearctic genus or is nested within it. On the other hand, as noted above, reliable diagnoses for taxa of the Stenonema generic complex have only been proposed for larvae, while adult morphological characters may exhibit a certain degree of variability. Considering all considerations expressed above and aiming to maintain stability in the systematics of Heptageniinae, in the current study we refer to H. atypica as a taxon of Heptagenia s.l., emphasising the need of further investigation.

From a biogeographical perspective, the presence of a mayfly taxon in Baltic amber that is systematically close to Nearctic taxa is not surprising. Heptagenia s.l. dominates throughout the Holarctic and is also distributed in the Oriental Region. In addition to the small Palaearctic subgenera Kageronia and Dacnogenia, the subgenus Heptagenia s.str. comprises more than 30 species, with centres of diversity in North America and the eastern Palaearctic (see Kluge N.J., web page: https://insecta.bio.spbu.ru/z/Eph-phyl/_Heptagenia-f8g7.htm). Species belonging to the Stenonema generic complex are recorded exclusively from the Nearctic Region, and as fossils only from Miocene Mexican amber²². Finally, the genus Rhithrogena, which is also considered here, is Holarctic²⁹.

The assignment of H. atypica to the Stenonema generic complex is not the first case of fossil taxa from the Eocene of Europe belonging to extant Nearctic genera and families. Species of the genera Analetris Edmunds, 1972 (Acanthametropodidae Edmunds, 1963) and Siphloplecton Clemens, 1915 (Metretopodidae Traver, 1935), which today occur only in the running waters of the USA and Canada, have previously been described from Baltic amber^39,95. The number of mayflies with Holarctic distribution, including taxa recorded from the Baltic amber is even greater. In addition to the genera mentioned above in the family Heptageniidae (including Arthroplea), these also include Siphlonurus Eaton, 1868 (Siphlonuridae Ulmer, 1920), Metretopus Eaton, 1901 (Metretopodidae Traver, 1935), Paraleptophlebia Lestage, 1917 (Leptophlebiidae), Ephemerella Walsh, 1863, and Timpanoga Needham, 1927 (both Ephemerellidae Klapálek, 1909)^5,6,15,29.

Heptagenia (? Dacnogenia) bachofeni Demoulin, 1968

(Figures 8, 9; Table 3).

Eintagsfliege oder Ephemeride: Bachofen-Echt⁹⁶: Der Bernstein und seine Einschlüsse: 77, fig. 66 [first record and image]

Heptagenia bachofeni Demoulin, 1968; Dtsch. Entomol. Z., 15: 256, figs 22a–c (holotype), 23a–e (paratype) [original description]

“Heptagenia” bachofeni Demoulin, 1968: Kluge⁸: Paleontol. J., 2: 112 [list of fossil taxa]

Heptagenia bachofeni Demoulin, 1968: Hubbard⁷²: Fossilium Catalogus I: Animalia, 129: 49 [list of fossil mayfly taxa]; Spahr⁷³: Stuttgarter Beitr. Naturk. Ser. B, 182: 7 [list of fossil mayfly taxa]; Wichard & Weitschat⁷⁴: Entomol. Mitt. Löbbecke-Mus. + Aquazoo Beiheft, 4: 25 [list of fossil mayfly taxa]; Weitschat & Wichard⁷⁵: Atlas of Plants and Animals in Baltic Amber: 92 [list of fossil mayfly taxa]; Wichard et al.⁷⁶: Aquatic Insects in Baltic Amber: 291 [list of fossil mayfly taxa]; Varela-Hernández et al.⁷⁷: Can. J. Earth Sci., 59: 483 [tabular list of fossil mayfly taxa]

Heptagenia?bachofeni Demoulin, 1968: Demoulin⁶: Bull. Inst. R. Sci. Nat. Belg., 46(2): 5 [description of an additional specimen]

Heptagenia bachofeni? Demoulin, 1968: Keilbach⁷¹: Dtsch. Entomol. Z., 29: 208 [list of fossil mayfly taxa]

bachofeni Demoulin 1968 [Rhithrogena]: Kluge²⁹: The Phylogenetic System of Ephemeroptera: 208 [list of fossil Heptageniidae]

Heptagenia bachofeni: Yang et al.¹⁹: Cretac. Res., 147: 7 [tabular record of fossil Heptageniidae]

Type material Holotype: male imago, Eocene, Baltic amber, originally labelled: “G. Demoulin det., 1967 Heptagenia bachofeni sp. n. N20 ♂ imago Abt. 22a2c Holotype N 20”; “Bayer. Staatssammlung Paläontologie u. Geologie Nr. 1958 VIII N20 Heptagenia bachofeni (männl. Imago) Oligozän Samland, Kaliningrad Holotyp Demoulin (1968), Abb. 22a-c München”; inventory number 1958 VIII N20; housed in BSPG, A. Bachofen-Echt collection (Fig. 9D).

Paratype [designated as Allotype in original description]: female imago, Eocene, Baltic amber, originally labelled: “G. Demoulin det.. 1967 Heptagenia? bachofeni sp. n. N2 ♀ imago Abb. 23a-e”; “Bayer. Staatssammlung Paläontologie u. Geologie Nr. 1958 VIII N2 (Heptogenia [sic!] bechofeni [sic!] sp.n., weibl., Imago, Abb. 23a-e, G. Demoulin det. 1967; Orig. zu Demoulin (1968): Dt. Entomolog. Zeitschr., NF Bd. 15.-) München”; “Bayer. Staatssammlung Paläontologie u. Geologie Nr. 1958 VIII N2 Oligozän Samland, Kaliningrad coll. Bachofen-Echt München” (Fig. 10A).

Type locality and horizon Eocene, Eocene, Priabonian, 37.7–34 Ma. The specimens belong to the BSPG collection, established by Adolf Bachofen-Echt in the first half of the twentieth century, and are embedded in amber pieces originating from Eocene Baltic amber deposits in Samland (Königsberg) (Fig. 10A).

Diagnosis Male imago The attribution of H. bachofeni to Heptageniinae and particularly to Heptagenia (? Dacnogenia) is supported by the combination of following characters: (a) thoracic morphology, i.e.: prosternum with distinct transverse ridge and weakly expressed longitudinal keel; MNs transverse; LPs not curved laterally, distally connected to MPs; median impression of mesothoracic furcasternum narrowing anteriorly; (b) forewing cubital venation with two pairs of long intercalary veins; (c) tarsomere I of preserved leg well separated from tibia.

Additionally: (i) body length 6.42 mm; forewing length 7.44/7.46 mm [all as preserved]; (ii) compound eyes large, dorsally flattened, medially separated by a distance narrower than width of frontal ocellus; (iii) crossvenation well developed, not clustered between Sc and R field; crossveins without dark clouds or blackish joints; (iv) hind wings relatively long, 0.36 × forewing length, with well-developed costal process; (v) penial lobes fused almost fussed, moderately expanded laterally, separated apically by shallow, narrow V-shaped cleft; external spine rounded; (vi) titillators small, acutely pointed; (vii) posterior margin of gonobasal plate broadly convex between gonostylomeres I (Figs. 8 and 9A–C, E–G).

Redescription of the holotype [male imago] (Figs. 8 and 9). The specimen is incomplete, with head, thorax, wings, abdomen, right hind leg, and cerci preserved; other legs missing. The body is embedded in resin in lateral aspect, with the head rotated perpendicular to the longitudinal body axis. The compound eyes are flattened. Both pairs of wings overlap, rendering the wing venation pattern hardly distinguishable. Cracks and small clouds of Verlumung obscure the forewing field between the R and MA veins on the left side; anal field of the left hind wing is partly folded. The abdomen is markedly bent at segments VI–VIII, whereas the preceding segments I–V are deformed and flattened. The genitalia are well preserved and complete. Measurements of the cerci are provided in Table 3.

Colouration Body mostly uniformly coloured, brown to dark brown. Lateral sides of thorax slightly paler, yellowish brown. Both wings translucent, hyaline; pterostigma translucent. Abdominal segments brown to dark brown (Fig. 8A,B).

Measurements Body length 6.42 mm; length of forewings 7.44/7.46 mm; length of hind wings 2.65/2.70 mm; hind/forewing length ratio 0.36; length of cerci 12.60/13.68 mm, as preserved. See Table 3 for additional measurements.

Head Facial keel small, obtuse. Antennae intensively brown, relatively short. Ocelli well visible, relatively large. Compound eyes ovoid, large, dorsally flattened, medially separated by gap narrower than width of frontal ocellus; border between upper and lower portions of compound eyes indiscernible (Fig. 8C,F).

Thorax Prosternum mostly covered by compound eyes; with distinct transverse ridge; presence of longitudinal keel unclear due to obstructed view; MNs transverse at crossings with MLs and MPs; LPs not curved distally, connected to MPs; median impression of mesothoracic FSp narrowing anteriorly (Fig. 8D–F).

Wings opaque, translucent. Crossvenation well-developed, without crowded crossveins between Sc and R field of forewings; no darker clouds around crossveins; no dark spots joining or nearly joining of crossveins (Fig. 9A–C).

Forewings narrow; basitornal margin short, 0.53 × length of tornoapical margin; tornus situated ca. 0.32 of wing length. Longitudinal venation brown, distally slightly paler; crossvenation light brown, slightly paler than longitudinal veins; several crossveins forked. Pterostigma with 14–15 crossveins, 1–2 of them forked; C and Sc darkest, brown to blackish brown distally. RS fork close to wing base, at less than 1/4.4 of length; iRS long, connected to RSp by 9–10 veins, arising from RSa₁. MA slightly asymmetrical, forked after 0.54 of length; MA₁ and MA₂ connected to iMA by several crossveins; iMA long. MP asymmetrical, forked after 0.27 of length; iMP long. CuP smoothly curved toward wing base; two weakly visible basal crossveins between CuA and CuP basally. Two pairs of cubital intercalaries poorly distinguishable; no free intercalaries in cubital field. At least two longitudinal veins of anal field weakly visible; crossvenation indistinct. Basally connected marginal intercalaries scattered at least between MA₁ and CuA; other intercalaries not visible (Table 3; Fig. 9A–C).

Hind wings 0.36 × forewing length. Costal process apically bluntly pointed, well projected. Venation light brown to brown; cross-venation well-developed. MA–RSp fork at 0.15 from wing base; iRS long, basally connected to RSa and RSp. MA and MP present; MA fork at 0.55 from vein base. Cubital venation indistinguishable.

Leg [right hind leg] brown, with femur-tibia-tarsus [tarsal segments] length ratio: 1 : 0.87 : 0.27 [0.03 : 0.06 : 0.06 : 0.05 : 0.07]; pretarsal claws dissimilar (one pointed, other blunt) (Table 3).

Abdominal segments completely preserved; segments III–V partly translucent, twisted, bent at segments V–VII level; posterolateral projections on lateral sides of terga not prominent; nerve ganglia not visible; cerci light brown (Figs. 8B, 9E,F).

Genitalia Gonostyli with four segments; gonostylomere II 3.5 × length of gonostylomere III and ca. 5.2 × length of gonostylomere IV; basal segment ca. 2.0–2.5 × as long as wide. Gonobases short, subequal in length and width. Posterior margin of gonobasal plate with median projection moderately expanded. Penial lobes fused along most of their length, separated by shallow, narrow V-shaped cleft; lobes moderately expanded laterally; external spine rounded, well protruding above lobe; apices of lobes rounded. Titillators small, acutely pointed (Fig. 9E–G).

Redescription of the paratype [female imago] (Fig. 10). The specimen is incomplete and preserved in dorsoventral aspect, embedded in a piece of pale, translucent amber. The wings are present and slightly curved; the anal field of the right forewing is partly twisted, and both forewings are affected basally and centrally by resin cracks on the right side and by blackish shadows on the left. The left hind wing is also twisted. Both fore and middle pairs of legs are preserved, whereas the hind legs are absent. The abdomen is moderately curved and complete, with the segments neither twisted nor flattened. The cerci are completely preserved (Table 4; Fig. 10A,B).

Colouration Body pale to dark brown; blackish maculae scattered over entire body; left forewing with blackish clouds. Legs brown to dark brown; left fore- and middle legs distally unusually dark brown to black. Unpigmented areas of fore- and hind wings translucent, hyaline; pterostigma translucent (Fig. 10A,B).

Measurements Body length 6.42 mm; forewings length 7.24/7.28 mm; right hind wing length 2.50 mm; hind/forewing length ratio 0.34; length of cerci 14.20/14.70 mm, as preserved. See Table 4 for additional measurements.

Head not extended; facial keel small, obtuse. Antennae brown, slightly longer than head; scape, pedicel and flagellum same colour with head. Ocelli relatively large; frontal ocellus smaller than lateral ocelli. Compound eyes moderately enlarged, widely separated medially; maximum distance between eyes more than 7 × width of frontal ocellus (Fig. 10E).

Thorax of structure typical of some Heptageniinae: prosternum with distinct transverse ridge and longitudinal keel; MNs distinctly transverse; LPs not distally curved, connected to MPs; median impression of mesothoracic FSp narrowing anteriorly (Fig. 10F).

Wings opaque and translucent in parts that not covered by blackish clouds. Crossvenation well-developed, relatively numerous, evenly distributed within vein fields; no crowded crossveins between Sc and R field of forewings; no dark clouds around crossveins; no dark spots joining or nearly joining of crossveins (Fig. 10B).

Forewings elongated, not narrow, with length of basitornal margin 0.55 × of tornoapical margin length; tornus situated ca. 0.3 × of wing length from wing base. Longitudinal and crossvenation pale, yellow to light brown. Crossvenation with simple veins only. Pterostigma translucent, with 10–11 crossveins. RS forked app. after 0.22 of length; iRS long, approximated to RSp, and connected to it by 9 veins. MA slightly asymmetrical, forked app. after 0.5 of length; MA₁ and MA₂ connected to iMA by at least 4–5 crossveins; iMA long. MP asymmetrical, forked app. after 0.24 of length; iMP long. Cubital venation scarcely distinguishable, with two pairs of intercalaries connected to CuA and CuP, as well as between each other by several crossveins. At least four longitudinal veins of anal field weakly visible; crossvenation indistinct. All small intercalaries along basitornal and tornoapical margins connected to longitudinal veins; no small basally free intercalaries (Table 4; Fig. 10B).

Hind wings 0.34 × forewing length. Venation poorly distinguishable. Several longitudinal veins in cubital and anal fields (Fig. 10B).

Legs yellowish-brown to blackish brown; right foreleg femur-tibia-tarsus [tarsal segments] length ratio: 1 : 1.10 : 0.83 [0.12 : 0.23 : 0.21 : 0.12 : 0.15]; left foreleg length ratio: 1 : 1.05 : 0.67 [0.09 : 0.18 : 0.13 : 0.15 : 0.12]; right middle leg length ratio: 1 : 0.94 : 0.45 [0.08 : 0.10 : 0.09 : 0.11 : 0.07]; left middle leg length ratio: 1 : 0.95 : 0.47 : [0.08 : 0.11 : 0.09 : 0.11 : 0.08]. Pretarsal claws of preserved leg with outer claw hooked and inner claw blunt (Table 4; Fig. 10C).

Abdominal segments well preserved, partly translucent, covered by unspecific blackish maculae; no posterolateral projections on abdominal segments. Subgenital plate relatively narrow broad, ca. 3.0 × as long as wide, convex, widely obtuse apically. Subanal plate moderately elongated, broadly triangular, with shallow apical cleft (Fig. 10D). Cerci well preserved, almost complete.

Remarks

Demoulin⁵ described H. bachofeni based on two adult specimens, designating a male imago as holotype, which is characterised by comparatively small body and wing dimensions. A female specimen of similar size was designated by that author as allotype and is treated in the present study as a paratype of H. bachofeni. Both specimens were described and illustrated only briefly (see Demoulin⁵: p. 256–258, figs. 22, 23). In discussing these findings, Demoulin briefly noted a certain resemblance of the males H. bachofeni and H. senex, at the same time emphasising differences in overall size and in the structure of the penial lobes. With regard to the conspecificity of the male and female assigned to H. bachofeni, Demoulin identified no diagnostic evidence beyond their small body and wing size in support of this attribution.

Indeed, the two specimens of the type series of H. bachofeni show no substantial differences in thoracic morphology or in the venation patterns of either wing pair. Considering their similar size characteristics, which we were able to confirm in the present study (see Table 4), we support their putative assignment to Heptagenia (Dacnogenia) (see below). Thus, we retain the conspecifity of male and female as H. bachofeni.

Subsequently, Demoulin⁶ (p. 5, 6, fig. 5a, b) briefly described a female subimago and tentatively also assigned it to H. bachofeni, but explicitly pointing out the uncertainty of this attribution. In addition to the similarity in body size reported by Demoulin, the observed wing venation pattern is also consistent with that of the type material of H. bachofeni. This female subimaginal specimen was deposited in the Muséum national d’Histoire naturelle, Paris. However, the specimen could not be located in the collection and therefore could not be examined in greater detail for the present contribution. Consequently, we provisionally retain the systematic assignment proposed by Demoulin⁶.

Heptagenia bachofeni undoubtedly belongs to the subfamily Heptageniinae, as indicated by the general pattern of wing venation and thoracic morphology. Both the male holotype and the female paratype possess a prosternum of characteristic shape, with a distinct transverse ridge and a concave surface posterior to this ridge. In addition, the median impression of the mesothoracic furcasternum is strongly narrowed anteriorly (Figs. 8F and 10F; see Kluge^{29,33,35,50,51}). The thoracic characters exclude the assignment of H. bachofeni to the subfamilies Rhithrogeninae or Ecdyonurinae. Conversely, the shape of the male penial lobes corresponds well to Heptagenia (including Dacnogenia and Kageronia) (Fig. 9G).

The systematic attribution of the male specimen of H. bachofeni is complicated by the loss of most legs and by the anterior portion of the prosternum being obscured by the rotated head and compound eyes, which prevents a reliable judgement on the presence or absence of a longitudinal keel. However, such a keel is present in the female of H. bachofeni. On this basis, an assignment of H. bachofeni to Kageronia can be excluded, as adults of that genus lack a distinct longitudinal keel on prosternum. The assignment of the male of H. bachofeni to Kageronia can be also excluded, as the preserved hind tarsomere 1 is markedly shorter than tarsomere 2 (Table 3). Additionally, the attribution of the male to Kageronia is contradicted by the absence of small lateral spines at the base of the penes (Fig. 9G; see also Webb & McCafferty⁶²). The same character, together with the absence of 2–3 crossveins in the R field of the forewings connected by a transverse dark dash, also clearly distinguishes H. bachofeni from extant representatives of Stenacron. Other taxa of the Stenonema generic complex differ from H. bachofeni by the following combinations of characters: (i) L-shaped penial lobes and the presence of a distinct cluster of 2–3 crossveins between the Sc and R field of the forewings (as in Stenonema); (ii) the presence of 3–6 densely crowded crossveins in the interspace between Sc and RSa₁ (as in Maccaffertium); and (iii) subquadrate penial lobes combined with widely separated compound eyes (as in Macdunnoa).

The male holotype of Heptagenia bachofeni is most similar to the extant species Heptagenia (Dacnogenia) coerulans Rostock, 1878, sharing several characters of the male genitalia, namely: (i) penial lobes that are not tall, moderately expanded laterally, and bent at an almost right angle; (ii) penial lobes with a relatively shallow V-shaped apical emargination, nearly contiguous medially; and (iii) well-developed dorsomedial spines of the penes, with their lateral tips slightly hooked. At the same time, the shape of the posterior margin of the gonobasal plate in H. bachofeni is convex, with a moderately expanded median projection, in contrast to the concave posterior margin in H. coerulans, which bears small rounded lateral protuberances near the base of the gonostyli (Fig. 9G; see also Kluge⁵¹; Webb & McCafferty⁶²; Bauernfeind & Soldán⁴⁶).

Nevertheless, when assessing the systematic position of H. bachofeni, its attribution to Dacnogenia should remain questionable, as the legs (particularly the forelegs) are not preserved and the natural coloration of the body and wings is also not retained in the fossils.

Among both fossil and extant taxa, H. bachofeni is readily distinguished by its comparatively small body and wing size, in combination with the morphology of the penial lobes and the outline of the posterior margin of the gonobasal plate.

Heptagenia (? Heptagenia) baltica (Demoulin, 1968) comb. nov.

(Figures 11 and 12; Table 5).

Cinygma? baltica Demoulin, 1968; Dtsch. Entomol. Z., 15: 254, fig. 20a, b (holotype) [short description].

Cinygma baltica: Larsson⁹⁷: Entomonograph, 1: 80 [list of fossil taxa]; Hubbard⁷²: Fossilium Catalogus I: Animalia, 129: 48 [list of fossil mayfly taxa]; Spahr⁷³: Stuttgarter Beitr. Naturk. Ser. B, 182: 7 [list of fossil mayfly taxa]; Carpenter⁹⁸: Treatise on invertebrate palaeontology, 3: 25 [list of fossil mayfly taxa]; Wichard & Weitschat⁷⁴: Entomol. Mitt. Löbbecke-Mus. + Aquazoo Beiheft, 4: 25 [list of fossil mayfly taxa]; Weitschat & Wichard⁹⁹: Wasserinsekten im Bernstein, 4: 92 [list of fossil mayfly taxa]; Weitschat & Wichard⁷⁵: Atlas of Plants and Animals in Baltic Amber: 92 [list of fossil mayfly taxa]; Kluge²⁹: The Phylogenetic System of Ephemeroptera: 208; Wichard et al.⁷⁶: Aquatic Insects in Baltic Amber: 291 [list of fossil mayfly taxa]

“Cinygma” baltica Demoulin, 1968: Kluge⁸: Paleontol. J., 2: 112 [list of fossil taxa]

Cynigma [sic!] baltica: Keilbach⁷¹: Dtsch. Entomol. Z., 29: 208 [list of fossil mayfly taxa].

Type material Holotype: male subimago, Eocene, Baltic amber, originally labelled: “Holotype Fam. Ephemeridae Demoulin 1968, fig. 20 12”; “G. Demoulin det. 1966 Cinygma? baltica sp. n. ♂ subimago MB.I.2245”; inventory number MB.I.2245; housed in MNB, Simon collection.

Type locality and horizon Eocene, Eocene, Priabonian, 37.7–34 Ma. The male subimago is preserved in a piece of Baltic amber originating from Eocene deposits of the Königsberg region. As with many other amber specimens comprising the Simon Collection housed at the MNB, the specimen is mounted using the “Königsberg method” (Fig. 11A).

Diagnosis Male subimago The assignment of H. baltica to Heptageniinae is supported by: (a) thoracic morphology, i.e.: MNs nearly transverse; LPs not laterally curved, connected with MPs; median impression of mesothoracic furcasternum narrowing anteriorly; prosternum with distinct transverse ridge and weakly expressed longitudinal keel; (b) forewing cubital venation with two pairs of long intercalary veins; (c) tarsomere I of all legs well separated from tibia;

Additionally (i) body length 7.80 mm; length of forewing 8.40 mm [all as preserved]; (ii) upper portions of compound eyes large, moderately separated medially by a distance slightly exceeding maximum width of frontal ocellus; (iii) crossvenation not clustered between Sc and R field; crossveins without dark clouds or blackish joints; (iv) foreleg tarsomere I approximately 0.74 × length of tarsomere II; (v) penial lobes closely positioned, not laterally expanded; dorsal fold moderately protruding above lobes; (vi) posterior margin of styliger plate deeply and broadly incised, forming U-shaped notch (Figs. 11B–E and 12A,C,E,F).

Description of the holotype [male subimago] (Figs. 11 and 12). The specimen is complete and well preserved both dorsally and ventrally. The right foreleg is partially preserved, with entire femur and part of tibia intact, whereas the left foreleg is almost complete, retaining the first two tarsomeres, with the third tarsomere truncated. Only left middle and hind legs are completely preserved. The left wings are fully extended; the right forewing is vertically torn at approximately one quarter of its estimated original length, while the right hind wing is complete. One cercus is preserved, but detached from the body.

The specimen is embedded in a rectangular piece of yellow to ochre resin, darkened along its margins, and mounted on a microscope slide. It is additionally surrounded by an irregularly shaped, carved impression with a black outline (Fig. 11A). The body, from head to approximately ninth abdominal segment, is dorsal and ventral unevenly covered by a thin layer of Verlumung. Left forewing and both hind wings are preserved in an outstretched position, with venation almost entirely visible; only the distal portion of the forewing bears a dark brown smear, although the venation remains discernible.

Colouration General coloration dark brown to black. Body coated with golden and silvery crust, most evident along outlines, also present on wings, abdomen, and genitalia. Compound eyes brown, almost entirely covered by Verlumung. Facial keel brown to dark brown, with scattered darker markings. Mesonotum brown, covered with Verlumung; sutures discernible. Wing membranes translucent; black detrital particles at apex and along costal brace of forewing, extending partly towards costal margin and anal field. Hind wing translucent or covered with bright golden layer. Legs with dark and golden coating, or partly translucent with golden hue. Abdomen greyish due to Verlumung, with silvery and golden sheen.

Measurements Body length 7.80 mm; length of left forewing 8.40 mm; length of hind wing 2.44/2.52 mm; hind/forewing length ratio 0.30; length of left cercus 10.16 mm as preserved [detached]. See Table 5 for other measurements.

Head General coloration brown to greyish brown. Anterior margin of head blunt; facial keel short, obtuse. Compound eyes relatively large, sub-spherical, not contiguous medially; separated by a distinct gap slightly wider than frontal ocellus. Left compound eye greyish brown; right compound eye covered by Verlumung. Lower portions of compound eyes not discernible. Ocelli dark brown, relatively large, well visible despite surrounding Verlumung. Antennae opaque; scape and pedicel present, with Verlumung more pronounced on these segments (Fig. 11B,C).

Thoracic morphology characteristic of Heptagenia s.l.: mesonotum with distinct transverse MNs; LPs straight, not laterally curved, connecting with MPs; prosternum with distinct transverse ridge, and weakly protruding longitudinal keel; medial impression of mesothoracic furcasternum tapering anteriorly (Fig. 11D,E).

Wings opaque to translucent. Venation typical of Heptageniidae, with two pairs of intercalaries in cubital field. Forewing densely reticulated with crossveins; no clusters of crossveins between Sc and R field; no dark clouds around crossveins; no dark spots joining or nearly joining of crossveins. Hind wing with well-developed crossvenation and fewer basally connected veins along posterior margin. Microtrichia present (Fig. 12A,B).

Forewings longitudinal veins light brown to bright yellowish gold; crossveins concolourous with longitudinal veins. Costa and Sc thick, light brown to translucent gold; RS forked near wing base, at ca. 0.25 of its length; iRS long, connected to RSp by eight visible crossveins, proximally equidistant from RSa₁ and RSp. MA fork slightly asymmetrical, dividing at approximately 0.65 of its length; iMA long, MA₁ and MA₂ connected to iMA by two crossveins. MP asymmetrical, forked after ca. 0.25 of its length; MP₁ and MP₂ basally connected by a single crossvein; iMP long, connected to MP1 by 6 crossveins and to MP₂ by 4 crossveins. CuA closely approximated to base of CuP; CuP not curved towards wing base; two pairs of long intercalary veins between CuA and CuP; two cua–cup crossveins between CuA and CuP basally; three cup–a1 crossveins between CuP and A₁ near base. Several small, basally connected intercalary veins present along basitornal margin, absent between RA and iRS, and not visible in anal sector (Fig. 12A).

Hind wings opaque, elongate, with straight costa. Venation light brown to brown; crossvenation well-developed; 6 crossveins between C and Sc, 9 to 10 between Sc and RA, 2 between RA and RSa; RSp apically connected to RSa by 4 crossveins and to MA₁ by 3 cross-veins; iRS short, connected to RSa and RSp on each side by 2 cross-veins; MA forked, longer than MP; cubital furcation indistinct; CuA relatively long. Marginal basally connected intercalaries along posterior margin of wing. Costal process weakly developed, slightly curved, situated near wing base (Fig. 12B).

Legs brown to light brown with golden sheen; femora, tibiae, and tarsi concolourous; tarsomere I of left foreleg as long as 0.74 of tarsomere II length; left middle leg femur-tibia-tarsus [tarsal segments] length ratio: 1 : 0.92 : 0.40 [0.05 : 0.10 : 0.10 : 0.05 : 0.10]; left hind leg length ratio: 1 : 0.10 : 0.55 [0.08 : 0.08 : 0.08 : 0.06 : 0.15]. Pretarsal claws dissimilar (Fig. 12D,E).

Abdominal segments completely preserved, opaque; no posterolateral projections; preserved cercus brown, elongate, detached from body.

Genitalia. Gonostylomeres II–IV relatively straight; gonostylomere II ca. 3.6 × length of segment III; apical gonostylomere IV nearly equal to gonostylomere III; gonostylomere I ca. 2 × as long as wide. Gonobases short, nearly quadrate. Posterior margin of gonobasal plate deeply and broadly incised medially by U-shaped notch; two protuberances near base of gonostyli. Penial lobes partially obscured by amber cracks and inclusions, apices only visible dorsally; lobes closely positioned, not laterally expanded, moderately divergent; dorsal fold slightly protruding apically. Dorsolateral spines and titillators not visible (Fig. 12C,F).

Remarks

Demoulin⁵ (p. 254–255, fig. 20a, b) provided a brief description of “Cinygma” baltica, accompanied by illustrations of the specimen in dorsal view, the tarsus of the left hind leg, a fragment of the cercus, and the genitalia in ventral view. However, several aspects of the description and the interpretation of the figures produced by G. Demoulin are inaccurate. For example, both the description and the illustration of the penial lobes in ventral aspect depict them as tubular (originally described as “baguette”-shaped), although the ventral surface of the fossil specimen is affected by inclusions and resin flows. As a result, almost no details are discernible in the area surrounding the gonobasal plate, gonostyli and penes. Nevertheless, the gonobasal plate and the gonostyli are otherwise correctly illustrated by the author⁵.

Comparing to the original description and illustration of H. baltica, Kluge⁸ investigated a male imago from Baltic amber and identified it as the extant species H. (K.) fuscogrisea (Retzius, 1783) (specimen deposited in the collection of the Swedish Museum of Natural History [SMNH], inventory no. Ar 8984), differing from H. baltica in having narrow penial lobes.

The apical portions of the penial lobes of H. baltica are visible in dorsal view. Their outline corresponds well to the shape of the penial lobes observed in some representatives of the family Heptageniidae, and in particular in the subfamily Heptageniinae. It should be taken into account that in male subimagines the morphology of the penial lobes may differ markedly from that of the imago. The presence of a well-defined and moderately protruding apical part of the dorsal fold in the penes of the fossil subimago may indicate that, at the imago stage, the genitalia possess a shape typical of Heptagenia (including Kageronia and Dacnogenia), with lobes that are distally broadened and folded, without distinct sclerites⁴⁶.

The use of leg proportions, and especially the ratios of tarsomeres, are problematic in the case of H. baltica, as only three basal segments of the left foreleg are preserved. Demoulin⁵ used the proportions of the foreleg (i.e. the lengths of the first and second tarsomeres) as a character to distinguish Cinygma from Rhithrogena. However, this character, when considered in combination with the structure of the posterior margin of the gonobasal plate and the presence of a longitudinal keel on the prosternum, is in fact a key feature for separating H. baltica from other representatives of Heptageniinae (see below). In addition, Demoulin⁵ illustrated the hind tarsus with the first tarsomere longer than the second. In reality, the first tarsomere is equal in length to the second (Table 5), as in many other taxa of the Heptageniidae.

Thus, in contrast to the original generic attribution of H. baltica to Cinygma proposed by Demoulin⁵, which the author himself regarded as tentative, we assign this fossil species to Heptagenia within the subfamily Heptageniinae, based on the following considerations:

1. (i)

   prosternum with a distinct transverse ridge and a markedly concave surface posterior to this ridge (in contrast to Rhithrogeninae and Ecdyonurinae, excluding Cinygma);

2. (ii)

   longitudinal keel present, but weakly expressed and poorly visible owing to partial coverage by resin cracks and Verlumung (in contrast to all Kageronia);

3. (iii)

   median impression of the mesothoracic furcasternum convergent anteriorly (in contrast to Ecdyonurinae);

4. (iv)

   lateroparapsidal suture of the mesonotum not curved laterally from the medioparapsidal suture (in contrast to Rhithrogeninae);

5. (v)

   pterostigmatic area with simple crossveins only, not divided into two clusters by a longitudinal vein between C and Sc (in contrast to Cinygma, in which the pterostigmatic crossveins are separated by a longitudinal vein);

6. (vi)

   forewings lacking clusters of crossveins in the Sc–R field and lacking coloured clouds around cross-veins, as well as lacking dark spots joining or nearly joining the crossveins (in contrast to Stenonema, Stenacron, and some Maccaffertium);

7. (vii)

   tarsomere I of the left foreleg approximately 0.74 × the length of tarsomere II (close to K. fuscogrisea and Cinygma, in which the length of the first tarsomere is about half or more than half of the foretarsomere II, and in contrast to other Heptageniinae, in which tarsomere I comprises only about 1/5 to 1/4 of the length of tarsomere II);

8. (viii)

   penial lobes with an expanded and moderately protruding dorsal fold (in contrast to the Stenonema generic complex).

We leave the subgeneric attribution of H. baltica unresolved. This is due to the limited visibility of most part of the penial lobes, which precludes a detailed description and comparison of the shape and structure of the folds. In addition, the character relating to the structure of the foreleg tarsus should be used with considerable caution, as only a single left foretarsus is preserved, which is partially damaged and derived from a subimaginal specimen.

Finally, key to distinguishing H. baltica from all other species of Heptagenia, both extant and fossil, is a combination of characters, including the presence of a prosternal longitudinal keel, the relative length of the first tarsomere of the foreleg, and the outline of the posterior margin of the gonobasal plate.

Heptagenia (s.l.) demoulini Godunko, Benhadji & Staniczek sp. nov.

(Figures 13, 14, 15 and 16; Table 6).

Rhithrogena sp.: Demoulin, 1968; Dtsch. Entomol. Z.,15: 262, fig. 28a, b [short description]; Keilbach⁷¹: Dtsch. Entomol. Z., 29: 208 [list of fossil mayfly taxa]; Hubbard⁷²: Fossilium Catalogus I: Animalia, 129: 50 [list of fossil mayfly taxa]; Spahr⁷³: Stuttgarter Beitr. Naturk. Ser. B, 182: 8 [list of fossil mayfly taxa]; Carpenter⁹⁸: Treatise on invertebrate paleontology, 3: 25 [list of fossil mayfly taxa]; Wichard & Weitschat⁷⁴: Entomol. Mitt. Löbbecke-Mus. + Aquazoo Beiheft, 4: 25, fig. 5 [list of fossil mayfly taxa]; Weitschat & Wichard⁷⁵: Atlas of Plants and Animals in Baltic Amber: 92, fig. 43 [list of fossil mayfly taxa]; Wichard et al.⁷⁶: Aquatic Insects in Baltic Amber: 291 [list of fossil mayfly taxa]

Type material Holotype: male imago, Eocene, Baltic amber, originally labelled: “Fam. Ephemeridae 10”; “G. Demoulin det., 1966 Rhithrogena sp. ♂ imago Demoulin, 1968, fig. 28.”; inventory number MB.I. 2235; housed in MNB, Simon collection.

Type locality and horizon Eocene, Priabonian, 37.7–34 Ma. The specimen, marked by Demoulin⁵ as Rhithrogena sp., originates from the Eocene deposits of Baltic amber in the Königsberg Region. The piece of amber is additionally embedded in translucent resin and mounted on slide [the size of the microscopic glass] (Fig. 13A–C).

Derivation of name This species is named in honour of Georges Demoulin, who was the first to examine the type specimen, in recognition of his many years of dedicated work on both extant and extinct mayfly taxa.

Diagnosis Male imago The specimen originally designated by Demoulin⁵ as Rhithrogena sp. is attributable to Heptagenia within Heptageniinae based on the following combination of characters: (a) thoracic morphology, i.e. prosternum with distinct transverse ridge and weakly expressed longitudinal keel; mesonotal suture transverse; lateroparapsidal suture not distally curved, connected to medioparapsidal suture; median impression of mesothoracic furcasternum narrowing anteriorly; (b) cubital field of forewing with two pairs of long intercalary veins; (c) tarsomere I of all preserved legs well separated from tibia.

The new species established here, based on the specimen recorded by Demoulin, can be separated from all representatives of Heptagenia s.l. (including Dacnogenia, Kageronia and Stenonema; see Kluge²⁹) by the following characters: (i) body length 9.40 mm; length of forewing 8.44/8.52 mm [all as preserved]; (ii) compound eyes medially separated by a gap narrower than width of central ocellus; (iii) forewings relatively narrow, with short basitornal margin; (iv) forewings without clustered crossveins between Sc and R field; crossveins without dark clouds or blackish joints; (v) crossvenation in anal field of forewing moderately developed, absent in anal field of hind wing; basally connected small intercalary veins weakly developed; (vi) hind wings 0.35 × forewing length; (vii) foreleg tarsomere I ca. 0.25 × length of tarsomere II; (viii) penial lobes relatively small, laterally not expanded; no visible spines on base of penes laterally; dorsomedial spine well-developed, expanded apically; (ix) relatively large sclerite similar to titillator, as long as preserved penial lobes; (x) gonobasal plate flat anteriorly, without projections (Figs. 13B–E, 14, 15 and 16).

Description of the holotype [male imago] (Figs. 13, 14, 15 and 16). The specimen is almost complete, relatively well-preserved, visible in lateral aspect within translucent amber, preserved on microscopic slide. Base of compound eyes, dorsal and ventral thorax, and pleura covered by few scattered clouds or trails of Verlumung. The right lateral and ventral surfaces of the body are covered with diffuse clouds of Verlumung. All wings preserved, folded on each other, venation more or less visible but hardly recognizable in cubital and anal fields of fore- and hind wings; middle of forewings with dark shadows between costal and iMP vein (Figs. 13B,C and 15A,B). Distal portions of forewings covered by resin cracks. All legs preserved. The abdominal tip is almost fully preserved, with several bubbles present around the genitalia (Figs. 15E,F and 16A,B). The body decomposition had probably already started before the specimen was coated with tree resin, as abdominal segments VI–X are deteriorated. The cerci are partially damaged.

Colouration General preserved colour golden pomegranate to shiny bronze. Compound eyes with lighter golden pomegranate colour than thorax. Facial keel translucent and of same pomegranate colour. Mesonotum dark golden pomegranate, with diffuse clouds of Verlumung; wing surface pleated at mid forewings length. Legs more or less opaque, brown to golden brown. Abdomen with well-preserved opaque segments I–V; segments VI–X lighter, damaged and translucent (Fig. 13B,C).

Measurements Body length 9.40 mm; length of forewing 8.44/8.52 mm; length of hind wing 2.92/2.96 mm; hind/forewing length ratio 0.35; length of left cercus 8.80 mm as preserved. See Table 6 for other measurements.

Head Facial keel obtuse. Antennae dark golden brown or pomegranate. Ocelli, well visible, relatively large. Compound eyes ovoid and large, medially separated by a gap smaller than width of central ocellus; border between upper and lower portions of compound eyes weakly distinguishable, covered partly with Verlumung (Fig. 13D,E).

Thorax Prosternum with distinct transverse ridge and weakly expressed longitudinal keel; MNs transverse; LPs not curved distally, connected to MPs; median impression of mesothoracic FSp narrowing anteriorly (Figs. 13D,E and 14A–D).

Wings partly opaque. Forewings with two pairs of intercalaries in cubital field as in Heptageniidae; crossvenation well-developed and very dense; no clusters of crossveins between Sc and R field; crossveins without dark clouds, and without blackish joints. Hind wings with regular crossvenation (Fig. 15A,B).

Forewings relatively narrow; basitornal margin short, as long as 0.56 × the length of tornoapical margin; tornus positioned ca. 0.36 × of wing length from wing base. Longitudinal venation light brown; veins coloration is altered after mid-wing length due to folding; crossvenation light brown, extremely developed with numerous cells. Pterostigma with 11 simple veins; C and Sc brown; RS fork close to wing base, at less than 1/4 of its length; iRS long, connected with RSp by 11 crossveins, proximally close to RA; iRS arise from RSa₁; MA fork slightly asymmetrical, forked after ca. 0.57 of its length; iMA long, MA₁ and MA₂ connected to iMA by 6–7 crossveins; MP asymmetrical, forked after 0.25 of its length, MP₂ arise from MP₁; iMP long, connected to MP₁ and MP₂ by 3 crossveins from each side; CuA closely approximated to CuP basally; CuP smoothly curved toward wing base; one visible crossvein basally between CuA and CuP; two pairs of cubital intercalaries poorly distinguishable, not accompanied by basally free intercalaries; two longitudinal veins of anal field poorly visible, crossvenation not distinguishable. Intercalaries in RS connected to longitudinal veins by 3–11 crossveins; basally connected marginal intercalaries present between MA₁ and MP₂ (Table 6; Fig. 15A).

Hind wings opaque; shape indistinct, as anal margin of visible part of left wing folded over posterior margin, obscuring both anal and cubital fields, as well as most of MA field. Costal process situated near wing base, acutely pointed apically; apex angular (ca. 120°); outer margin smoothly curved. Venation brown to light brown; 9 crossveins between C and Sc, 10 between Sc and RA, and 5 between RA and RSa; MA–RSp fork situated after 0.13 of vein base; iRS relatively long, basally connected to RSa and RSp; MA forked at 0.59 of its length; cubital veins forked, though furcation indistinct (Table 6; Fig. 15B).

Legs yellow to blackish brown; forelegs paler than middle and hind legs; right foreleg femur-tibia-tarsus [tarsal segments] length ratio: 1 : 1.10 : 1.48 [0.12 : 0.48 : 0.44 : 0.34 : 0.10]; left foreleg length ratio: 1 : 1.15 : 1.68 : [0.13 : 0.53 : 0.51 : 0.38 : 0.13]; right middle leg length ratio: 1 : 1.02 : 0.60 [0.09 : 0.12 : 0.12 : 0.12 : 0.15]; left middle leg length ratio: 1 : 1.03 : 0.59 [0.09 : 0.12 : 0.12 : 0.12 : 0.14]; right hind leg length ratio: 1 : 0.92 : 0.51 [0.08 : 0.10 : 0.10 : 0.10 : 0.13]; left hind leg length ratio: 1 : 0.89 : 0.58 [0.10 : 0.11 : 0.11 : 0.11 : 0.15]; pretarsal claws dissimilar (one slender and pointed, other spatulate and blunt) (Table 6; Fig. 15C,D,G).

Abdominal segments partly preserved; segments I–V opaque, segments VI-X translucent and damaged; posterolateral projections on lateral sides of terga not prominent. Preserved abdominal terga I–V with same coloration as sterna; nerve ganglia not visible. Only left cercus preserved, light brown, partly damaged (Fig. 13B,C).

Genitalia Gonostyli with four gonostylomeres; length of gonostylomere II is 7.5 × longer than segment III; gonobases not well visible, relatively narrow. Gonobasal plate poorly preserved, appears flat anteriorly, without any projections. Only the right penial lobe is preserved, visible in ventrolateral aspect; dorsomedial spine well-developed, distally expanded by protruded, slender apical part of fold. Penial lobes not laterally expanded; no visible spines on base of penes laterally. Apically pointed sclerite similar to titillator near penial base, relatively large, nearly as long as preserved penial lobes (Figs. 15E,F and 16A,B).

Remarks

The specimen MB.I. 2235 from the MNB collection was identified by Demoulin⁵ as a male imago belonging to the genus Rhithrogena. A very brief description was provided, including the tarsal formulae of the fore- and hind legs, accompanied by two drawings: one showing the general body habitus and the other depicting a fragment of the right cercus. Although Demoulin noted that the genitalia were of the type characteristic of the genus (orig. Rhithrogena), he refrained from a detailed examination of their structure and from describing a new species, stating that such an analysis was not feasible. Nevertheless, based on the body size of this specimen, Demoulin⁵ stated that it could not be confused with R. sepulta, but did not address its relationship with other fossil species of Heptageniidae that he had described.

The body and forewing measurements given by Demoulin⁵ (p. 262) differ from those obtained in our re-examination of the specimen MB.I. 2235. In the original publication, the body length was reported as 8.1 mm, compared with 9.4 mm in our study; the forewings were reported as 7.9 mm, compared with 8.44–8.52 mm; the hind wings measured 3.0 mm to 2.92–2.96 mm respectively, and the cerci were reported as 8.6 mm, compared with 8.8 mm in our measurements.

We consider this specimen to represent the genus Heptagenia s.l., primarily based on the prosternum and the shape of a single preserved penial lobe. Additionally, the generic placement of H. demoulini sp. nov. within Heptagenia s.l. is supported by a hind leg tarsomere I that is shorter than segment II, and by the hind tarsus being shorter than 2/3 length of the hind tibia (Table 6; both characters listed by Bauernfeind & Soldán⁴⁶). In combination with the shortened tarsomere I of the foretarsi, both characters provisionally exclude Kageronia and Cinygma from a closer relationship with H. demoulini sp. nov.

We refrain from assigning a more precise systematic position of the male imago within Heptagenia due to the poor preservation of the genitalia. It is, however, worth to note that several diagnostic features of H. demoulini sp. nov. distinguish it from other representatives of Heptagenia s.l. and justify its description as a new species:

1. (i)

   The distance between the compound eyes in H. demoulini sp. nov. is smaller than the width of the central ocellus, as in other representatives of Heptagenia s.l., except for H. orbiticola Kluge, 1986, in which the compound eyes are widely separated (Bauernfeind & Soldán⁴⁶: pp. 313, 314).

2. (ii)

   Similar to other representatives of this genus, the extinct species exhibits only simple crossveins in the pterostigma. Other characteristics of the fore- and hind wing venation are typical of Heptageniinae.

3. (iii)

   Contrary to the tarsal formula reported by Demoulin⁵ for H. demoulini sp. nov., our measurements point to a slightly longer tarsomere I than tarsomere V on the right foreleg, whereas on the left foreleg tarsomeres I and V are equal in length (Table 6). Such tarsomere proportions are characteristic of most species of Heptagenia s.l.

4. (iv)

   At the same time, the foreleg tarsomere I, both in Demoulin’s original description and according to our measurements, is much shorter than tarsomere II, which clearly excludes the possibility of assigning H. demoulini sp. nov. to Kageronia. Additionally, a placement of this extinct species within Kageronia can be ruled out based on the presence of a longitudinal keel on the prosternum and the absence of small spines laterally on the shaft of the single preserved penial lobe (both characters noted by Webb & McCafferty⁶²).

5. (v)

   The shape of the penial lobes in H. demoulini sp. nov. is typical of the genus Heptagenia and is characterised by a well-developed dorsomedial spine, whose distal portion is expanded by the protruding, slender apical part of the fold. Additionally, the penial lobes are not laterally expanded. Although only a single penial lobe is preserved, its orientation and angle, together with the preserved putative titillator, indicate that in the original undamaged condition, both lobes would have been widely separated and widely spread laterally. Based on these features, the genitalia of the male imago of H. demoulini sp. nov. are structurally similar to those of Recent species of Heptagenia s.str., and in particular to Heptagenia (Heptagenia) dalecarlica Bengtsson, 1912. However, the fossil species differs in the length of the presumed titillators, which are longer than in all other Heptagenia species. We also cannot confidently exclude the possibility of its placement within Dactogenia, which is characterised by a well-developed dorsomedial spine on the penial lobes^35,46,50,64.

Despite the notable similarity in the shape and structure of the penial lobes between H. demoulini sp. nov. and species of Heptagenia s.str., we refrain from assigning the extinct species to a more precise systematic position. This approach is warranted because the genitalia of the holotype are partially damaged, and the morphology of the titillators appears atypical for the subgenus.

Heptagenia (s.l.) gleissi (Demoulin, 1968)

(Figures 17 and 18; Table 7).

“Mayfly encased in amber“ in Gleiss¹⁰⁰: Die Eintagsfliegen: 40, fig. 23 [unnamed photo of mayfly specimen in right lateral view]

Heptagenis gleissi in Demoulin 1968: Dtsch. Entomol. Z., 15: 259, fig. 25a–f.

Heptagenia gleissi Demoulin, 1968: Keilbach⁷¹: Dtsch. Entomol. Z., 29: 208 [list of fossil mayfly taxa]; Hubbard⁷²: Fossilium Catalogus I: Animalia, 129: 50 [list of fossil mayfly taxa]; Spahr⁷³: Stuttgarter Beitr. Naturk. Ser. B, 182: 8 [list of fossil mayfly taxa]; Wichard & Weitschat⁷⁴: Entomol. Mitt. Löbbecke-Museum + Aquazoo Beiheft, 4: 25 [list of fossil mayfly taxa]; Weitschat & Wichard⁷⁵: Atlas of Plants and Animals in Baltic Amber: 92 [list of fossil mayfly taxa]; Godunko & Neumann⁴⁴: Ann. Zool., 56(1): 177 [list of fossil mayfly taxa]; Wichard et al.⁷⁶: Aquatic Insects in Baltic Amber: 291 [list of fossil mayfly taxa]; Varela-Hernández et al.⁷⁷: Can. J. Earth Sci., 59: 483 [tabular list of fossil mayfly taxa]

“Heptagenia “ gleissi Demoulin, 1968: Kluge⁸: Paleontol. J., 2: 112 [list of fossil taxa].

gleissi Demoulin 1968 [Heptagenia]: Kluge²⁹: The Phylogenetic System of Ephemeroptera: 208 [list of fossil Heptageniidae]

Heptagenia gleissi: Yang et al.¹⁹: Cretac. Res., 147: 7 [tabular record of fossil Heptageniidae].

Type material Holotype: male imago, Eocene, Baltic amber, originally labelled: “Holotype Fam. Ephemeridae Demoulin 1968, Fig. 25 14”; “Heptagenia gleissi sp. n. ♂ imago MB.I.2239”; housed in MNB, Simon collection.

Type locality and horizon Eocene, Priabonian, 37.7–34 Ma. The specimen from unknown deposits of the former Königsberg [Kaliningrad] Region; embedded in gum dammar and mounted on a microscopic slide using the “Königsberg method” (Fig. 17A).

Diagnosis Male imago Heptagenia gleissi (Heptageniinae) is clearly distinguished from all known fossil representatives of the subfamily by the following combination of characters: as in other representatives, (a) the thorax bears a prosternum provided with a distinct transverse ridge; the remaining characters of the mesonotum and mesosternum are as in H. atypica; (b) cubital venation with two pairs of long intercalaries; (c) tarsomere I of all legs distinctly separated from the tibia. In addition: (i) body length 11.40 mm; forewing length 1010/10.40 mm [all as preserved]; (ii) compound eyes broadly separated medially by a distance 5.6 × greater than the maximum width of the frontal ocellus; (iii) forewings relatively short and not narrow; (iv) crossvenation of both wings well-developed, without crowded veins, spots, or joints between or around veins; (v) hind wings 0.34–0.36 times the length of the forewings; (vi) tarsomere I of the foreleg 0.34 × the length of tarsomere II; (vii) hind tarsomere I markedly shorter than tarsomere II; (viii) penile lobes short and small, not laterally extended, rounded distally; lobes widely separated by a broad, V-shaped gap; no lateral spines at base of penes; (ix) titillators very small and acutely pointed; (x) gonobasal plate broadly incised medially by a U-shaped notch, with two weak projections on each side of the notch.

Redescription of the holotype (male imago) (Figs. 17 and 18). The specimen is well preserved and embedded in a piece of translucent amber mounted on a microscopic slide, where it is visible in lateral aspect. Thorax, wings, and abdomen are partly obscured by resin containing cracks, small gas bubbles, various minute inclusions and Verlumung on head and part of thorax. Tarsomeres IV and V of both foretarsi are damaged or absent; the left hind leg is missing, most of the right middle leg is absent, and the cerci are partly damaged.

Colouration General colouration yellowish to brown; head and thorax darker brown than abdomen; abdominal segments I–VII lighter and partly translucent, segments VIII–X and genitalia dark brown; legs brown to dark brown; wings hyaline with brownish veins, without distinct colour pattern or clouds around crossveins (Fig. 17A).

Measurements Body length 11.40 mm; length of cerci 8.60 mm [incomplete]; length of forewing 10.10–10.40 mm; length of hind wing 3.52–3.60 mm; hind/forewing length ratio 0.34–0.36. See Table 7 for other measurements.

Head Compound eyes large, widely rounded, distinctly separated medially; distance between inner margins of compound eyes ca. 5.6 × exceeding maximum width of frontal ocellus; upper and lower portions of compound eyes not markedly differentiated. Ocelli well developed; lateral ocelli larger than frontal ocellus. Facial keel relatively small, short and weakly developed. Antennae slender, uniformly brown (Fig. 17B–E).

Thorax typical of Heptageniinae; prosternal transversal ridge poorly visible, visualised laterally by narrow ridge below forecoxae; presence or absence of a longitudinal keel cannot be evaluated; mesonotal sutures transverse; lateroparapsidal sutures not curved laterally and connected to medioparapsidal sutures; furcasternal protuberances poorly visible laterally, moderately converge distally (Figs. 17B–E).

Wings folded, partly overlapping; opaque to translucent; partly covered by blackish spots, cracks, minute inclusions, gas bubbles. Venation well visible, yellow to dark brown. Crossvenation well-developed. No crowded veins between Sc and R of forewings. No dark clouds or spots around forewing crossveins; crossvein joints not blackened (Fig. 18A,B).

Forewings relatively elongate, not narrow. Length of basitornal margin ca. half as long as length of tornoapical margin; tornus at ca. 0.28 × wing length from wing base. Venation well preserved, light coloured; crossvenation well-developed. Pterostigma with 14–15 simple crossveins. RS fork near wing base, after ca. 0.18 of its length. iRS long, connected with RSp by 9–10 crossveins; basally connected to RSa₁, moderately approximated to RSp. MA fork slightly asymmetrical, after ca. 0.60 × of its length; iMA shortened, connected to MA₁ and MA₂ by 3–5 crossveins on each side. MP asymmetrical, forked after ca. 0.20 × of its length; MP₂ arising from MP₁; iMP long, basally approximated to MP₂. Cubital field typical for Heptageniidae, with four long intercalary veins; at least one basal crossvein connecting CuA and CuP; CuA nearly straight; CuP smoothly curved toward wing margin. AA connected to CuP by 2–3 crossveins; additionally three poorly recognisable longitudinal veins in anal field. Small marginal intercalaries basally connected, present between RSa₁ and CuA (Table 7; Fig. 18A).

Hind wings opaque; length 0.34–0.36 × of forewing length. Costal process bluntly pointed, well prominent, outer margin smoothly curved. Venation pale light brown; crossvenation poorly visible, simple, sparse; at least 7 crossveins between C–Sc, 10–12 between Sc–RA, up to 7 between RA–RSa. MA–RSp fork at ca. 0.12 of vein length. iRS long, basally connected to RSa and RSp; MA forked at ca. 0.54. CuA visible; other cubital veins unclear, furcation indistinct (Table 7; Fig. 18B).

Legs yellow to dark brown, slightly paler than thorax and abdomen; forelegs slightly paler than other legs; right foreleg segment ratio not calculated (see Remarks); left foreleg femur-tibia-tarsus [tarsal segments] length ratio: 1 : 1.30 : 1.14* [0.17 : 0.49 : 0.53* : – : –]; left middle leg length ratio: 1 : 0.97 : 0.52 [0.09 : 0.11 : 0.11 : 0.09 : 0.12]; right hind leg length ratio: 1 : 0.59 : 0.43 [0.05 : 0.09 : 0.09 : 0.08 : 0.11]; measurements of partially preserved legs in Table 7; pretarsal claws dissimilar as in other species described here (Fig. 17F).

Abdominal segments I–VII light brown, partly translucent, VIII–X slightly darker; posterolateral projections absent. Terga smooth, without conspicuous ornamentation. Cerci long, multisegmented, incompletely preserved.

Genitalia Gonostyli slender, with four segments; gonostylomere I ca. 5.6 × longer than gonostylomere III; gonostylomeres II and III subequal in length; basal segment nearly trapezoidal, wider than long. Gonobases nearly oblong-shaped, not prominent. Gonobasal plate broadly incised medially by U-shaped notch; two weak projections on each side of U-shaped notch. Penial lobes short, small, not laterally extended; lobes rounded distally, widely separated by broad V-shaped gap; lateral sides without spines basally. Penial sclerites and folds indistinct; external spine of dorsal fold flattened, rounded, not protruding beyond lobe; ventral penial surface with small, poorly distinct spines; dorsal surface of lobes with prominent tips of penial sclerites, clearly visible in lateral view as pointed projections. Titillators small, sharply pointed, poorly visible due to Verlumung (Fig. 18C,D).

Remarks

Demoulin⁵ provided a detailed description of the male imago of H. gleissi, naming the new species in honour of Dipl.-Biol. Horst Gleiss (*1930–†2020), who had earlier used a specimen of the future holotype, without assigning it a name, to illustrate a fossil mayfly preserved in Baltic amber¹⁰⁰. Here, we expand the original description by adding detailed characteristics of the thorax, refined measurements of the leg segments, and a comprehensive analysis of the male genitalia.

In the present study, we deliberately refrain from analysing the length ratios of tarsomeres I and II of the right foreleg of H. gleissi. Examination of the holotype indicates that the proportions of this leg are altered. Demoulin⁵ already noted this peculiarity and suggested that the modified proportions may result from leg regeneration. By contrast, the left foreleg shows the first two tarsomeres to be well preserved, with tarsomere I being 2.9 × shorter than tarsomere II, which is close to the values reported for males of Stenacron⁶⁵, within the Stenonema generic complex (Table 7; Fig. 17F).

The illustrations provided by Demoulin⁵ are likewise highly accurate. They include detailed representations of the wing venation, the proportions of the tarsomeres of both forelegs, and the well preserved right hind leg. Moreover, in the fig. 25, Demoulin⁵ depicted the male genitalia in ventral (fig. 25b) and lateral (fig. 25c) view, faithfully rendering the structure of the gonobasal plate and forceps, the widely spaced penial lobes separated by a relatively deep median incision, and the pointed apical portions of the penial sclerites, visible on the dorsal surface of these lobes.

Beyond the morphological description of H. gleissi, Demoulin⁵ briefly compared this species with H. senex, emphasising similarities in the structure of the gonostyli, while simultaneously noting differences in the shape of the penial lobes.

The described set of characters confirms the initial attribution of H. gleissi to Heptageniinae. This conclusion is supported by the cubital venation of the forewings, with two pairs of intercalary veins, and by the first tarsomeres of the middle and hind legs being both well separated from the tibiae, both of which are common features of all Heptageniidae. In addition, the presence of a transverse prosternal ridge, in combination with a transverse mesonotal suture, lateroparapsidal sutures that are not curved laterally, and distally converging furcasternal protuberances of the mesosternum, further corroborates this placement (Fig. 17B–E).

Although the presence of a prosternal longitudinal keel can neither be confirmed nor refuted in the male imaginal holotype, H. gleissi can be confidently excluded from the subgenus Kageronia. This exclusion is based on the relative proportions of the hind leg tarsomeres, as tarsomere I of the right hind leg is markedly shorter than tarsomere II, whereas the opposite condition is observed in Kageronia and Cinygma. In addition, the structure and shape of the penial lobes, lacking large, distally elongated and folded lobes, clearly distinguish H. gleissi from species of Kageronia, as well as from all extant and fossil representatives of Heptagenia s.str. and Dacnogenia (Fig. 18C,D).

Thus, following the concept of the subfamily Heptageniinae adopted in this study, the systematic position of H. gleissi at present can be defined within Heptagenia (s.l.). In several respects, this fossil species resembles H. atypica and may be closely related to extant representatives of the Stenonema generic complex. This interpretation is supported by the position of the male compound eyes, which are markedly separated by a distance exceeding five times the width of the base of the frontal ocellus. Additional similarities include the proportions of tarsomeres I and II (0.34 in H. gleissi, compared with values ranging from 0.35 to 0.55, for example, in Stenacron)⁶⁵. However, owing to the variability of this ratio observed in the forelegs of the holotype of the fossil species (see above; Table 7), this character should be treated with caution.

Finally, similar to the fossil H. atypica and extant representatives of the Stenonema generic complex, the penial lobes of H. gleissi lack clearly defined sclerites and folds that markedly protrude beyond the lobe, as observed in Heptagenia (including Kageronia and Dacnogenia) (Fig. 18C,D).

Bogoescu & Tabacaru⁶⁷ were among the first authors to illustrate and describe in detail the structure of the penes in several European species of Heptageniidae. Their drawings of the penial lobes of two species, Heptagenia sp. and K. fuscogrisea, shown in dorsal, lateral and apical views, provide a clear understanding of the relative positions of the sclerites, which are transformed into broad, folded structures. In these extant Heptageniinae, the dorsal fold of each lobe forms a conspicuous external and internal spine laterally, as subsequently described by Bauernfeind & Soldán⁴⁶. These spines represent the apical parts of the ‘inner sclerite’ and ‘outer sclerite’ sensu Bogoescu & Tabacaru⁶⁷, designated in their fig. 2C and D as si [inneres Sklerit] and se [äußeres Sklerit], respectively. A particular characteristic of Heptagenia s.str., Dacnogenia and Kageronia is the dorsoapical portion of the fold, which is markedly protruded above the penial lobes and was marked as sv [Ventralsklerit] by Bogoescu & Tabacaru⁶⁷.

Whereas the male genitalia of H. atypica are not visible in lateral view owing to the specific placement of the specimen in the amber piece, the right penial lobe of H. gleissi is clearly exposed laterally, allowing an analysis of the distal portions of the folded sclerites (Fig. 18D). Demoulin⁵ drew attention to this feature by illustrating the genitalia of the holotype male imago of H. gleissi in lateral view (his figure 25c), indicating two delicate spines. In fact, three delicate spines are visible laterally on the right side of the penial lobe of H. gleissi (Fig. 18D; spines marked by red arrows). The outermost and most robust spine, situated on the ventral side of the lobe, corresponds to the tip of the dorsoapical part of the fold, i.e. to the apex of the Ventralsklerit (ventral sclerite) sensu Bogoescu & Tabacaru⁶⁷. The second spine, likewise relatively robust, belongs either to the tip of the dorsoapical fold of the left lobe or to the Inneres Sklerit (inner sclerite) of the right lobe as defined by Bogoescu & Tabacaru⁶⁷ (p. 277, fig. 2C). In many extant species of Heptageniinae, this inner spine is not visible in lateral view of the penial lobes. The third spine, which is isolated and acutely pointed, represents the distal end of the äußeres Sklerit (outer sclerite). None of these spines is visible on the ventral side of the genitalia of H. gleissi (Fig. 18C).

At the same time, H. gleissi occupies an isolated position based on the genital morphology. The widely separated penial lobes, divided by a distinct V-shaped gap, in combination with the configuration of the outer margin of the gonobasal plate, which is broadly incised medially by a U-shaped notch bearing two projections, clearly distinguishes this fossil species from other extinct Heptageniinae, as well as from species of the Stenonema generic complex.

Finally, H. gleissi falls within the size range observed for the majority of Heptageniinae with respect to both body length and wing dimensions (Table 7). Its body size and forewing length broadly overlap with those reported for extant genera such as Heptagenia, Kageronia, Stenonema, and Maccaffertium, while the hind wing dimensions are likewise consistent with proportions typical of the subfamily.

In any case, a comparative study of extant representatives of the Stenonema generic complex and extinct taxa is necessary, particularly given that fossil representatives appear to have been rather diverse in Cenozoic resins (Fig. 19; see also discussion about systematic position of H. atypica above).

Heptagenia (s.l.) ligata Demoulin, 1968

(Figure 20; Table 8).

Heptagenia ligata Demoulin, 1968; Dtsch. Entomol. Z., 15: 260, fig. 26 (holotype) [original description].

Heptagenia ligata Demoulin, 1968: Keilbach⁷¹: Dtsch. Entomol. Z., 29: 208 [list of fossil mayfly taxa];

Hubbard⁷²: Fossilium Catalogus I: Animalia, 129: 49 [list of fossil mayfly taxa]; Spahr⁷³: Stuttgarter Beitr. Naturk. Ser. B, 182: 7 [list of fossil mayfly taxa]; Wichard & Weitschat⁷⁴: Entomol. Mitt. Löbbecke-Mus. + Aquazoo Beiheft, 4: 25 [list of fossil mayfly taxa]; Weitschat & Wichard⁷⁵: Atlas of Plants and Animals in Baltic Amber: 92 [list of fossil mayfly taxa]; Wichard et al.⁷⁶: Aquatic Insects in Baltic Amber: 291 [list of fossil mayfly taxa]; Varela-Hernández et al.⁷⁷: Can. J. Earth Sci., 59: 483 [tabular list of fossil mayfly taxa]; Yang et al.¹⁹: Cretac. Res., 147: 7 [tabular record of fossil Heptageniidae].

ligata Demoulin, 1968 [Heptagenia]: Kluge²⁹: The Phylogenetic System of Ephemeroptera: 208 [list of fossil Heptageniidae; as other Heptagenia placed to Radulapalpata incertae sedis].

Type material Holotype female imago, Eocene, Baltic amber. The specimen is embedded in a thin amber stone with several cracks, carved with its inventory number 4539. Housed in GZG Geowissenschaftliches Zentrum der Universität Göttingen, Germany, W. Stantien & M. Becker collection.

Type locality and horizon Eocene, Priabonian, 37.7–34 Ma. The piece of amber with embedded holotype of H. ligata originates from the unknown Eocene deposits of Baltic amber in former Königsberg [Kaliningrad] Region (Fig. 20).

Diagnosis Female imago The placement of H. ligata within Heptageniinae, as well as suggestions for its current assignment to Heptagenia (s.l.), are supported by the following characters: (a) thoracic structure, i.e.: prosternum apparently bearing a transverse prosternal ridge, inferred from a sclerotized fold between the bases of the forelegs; right furcasternal protuberance of the mesosternum apparently bent inward, likely resulting in a median impression narrowing anteriorly; (b) cubital venation of forewing with two pairs of long intercalary veins; (c) foretarsus of middle and hind legs distinctly articulated to tibia, not fused with it;

additionally: (i) body length 6.90 mm; forewing length up to 7.42 mm (as preserved); (ii) compound eyes moderately large, medially separated by a distance equal to the width of the frontal ocellus; (iii) crossvenation poorly to moderately developed, without darker spots or nodal joints; a distinct group of two closely spaced crossveins present between C and MP₁; (iv) hind wings relatively short, 0.29 × forewing length (proportion calculated only for completely preserved wings); (v) subgenital plate narrow and nearly trapezoidal; subanal plate of comparable outline (Fig. 20A,B).

Redescription of the holotype [female imago] (Fig. 20A,B). The specimen is almost complete, but exhibits pronounced distortion. Head, thorax, and left forewing are preserved in dorsal view; however, the left forewing is basally inverted and twisted, such that its ventral surface is exposed dorsally. The remainder of the body, including the abdomen, legs, and the other wings, is bent almost perpendicularly beneath the anterior part of the body and is therefore only visible in ventral view. The specimen is embedded in amber of an orange hue.

All wings are present but variably altered. The left forewing is basally twisted, resulting in distortion of the basal regions of both the anal and costal fields. The left hind wing closely adjoins approximately three quarters of the ventral thoracic surface. The right forewing is folded medially towards the ventral side and is apically incomplete, possibly due to sectioning. The right hind wing follows the general torsion of the body and appears to be slightly folded apically. The left foreleg and both middle legs, as well as the right hind leg, are absent, whereas the remaining legs are fully preserved. The abdomen is complete but distinctly curved. The cerci are incomplete (Table 8; Fig. 20A,B).

Colouration Body dark brown; left forewing with blackish apical edge and an air bubble trapped centrally between longitudinal veins MA₂ (just after the bifurcation of MA) and MP₁. Legs brown to brown to dark brown distally and on the edges. Wings translucent (Fig. 20A,B).

Measurements Body length 6.90 mm; length of right forewing 4.96 mm, length of left forewing 7.42 mm.; length of right hind wing 2.12 mm, length of left hind wing is ca. 2.10 mm; hind/forewing length ratio 0.29; length of cerci 2.74 mm, as preserved. See Table 8 for additional measurements.

Head small; facial keel narrow, distinct/apically acute (well visible in ventral view). Antennae as long as head width across eyes; scape, pedicel and flagellum slightly lighter than head colour. Ocelli relatively small; frontal ocellus smaller than lateral ocelli. Compound eyes moderately large, closely approximate medially; maximum distance between eyes nearly equal to width of frontal ocellus (Fig. 20A,B).

Thorax and its sutures poorly visible, obscured by folded fore- and hind wings, especially ventrally (Fig. 20A,B). Prosternum between bases of forelegs with folded structure; largest fold possibly a sclerotised fold related to transverse prosternal ridge. Mesosternum poorly distinguishable; tip of right furcasternal protuberance curved inward, possibly indicating that median impression between both protuberances narrows anteriorly (Fig. 20A,B).

Wings translucent. Crossvenation poorly to moderately developed, more or less evenly distributed within vein fields. Sc and R fields of forewings well anastomosed by crossvenation, distally approximated in left forewing due to deformation; venation distally slightly denser beyond mid-forewing between C and RSp. Demoulin⁵ stated “A l’aile I, le ptérostigma est faiblement anastomosé; à mi-longueur, entre le bord costal et MP₁, les nervules transverses sont rapprochées et forment une espèce de ligature transverse”. A group of two crowded crossveins present between C and MP₁ slightly before mid-length of forewing; no spots and joints between and around crossveins.

Forewings elongated, narrow. Longitudinal and crossvenation light brown to brown. Venation locally denser at mid-wing between C and RSp. Pterostigma translucent, intersected by a stone fissure running predominantly along Sc but still discernible; ca. 13 crossveins present, two forked. RS forked before 1/4 wing length; iRS long, distinctly separated from RSp and connected to it by eight crossveins. MA slightly asymmetrical, forked at approximately 0.6 × wing length; MA₁ and MA₂ connected to iMA by two to three crossveins; iMA long, basally free. MP asymmetrical, forked at approximately 0.10–0.20 × wing length; iMP long, basally free. Cubital venation with two pairs of intercalaries; crossvenation in cubital field well developed. At least four longitudinal veins visible in anal field. All small intercalaries along basitornal and tornoapical margins connected to longitudinal veins; no small basally free intercalaries present (Table 8; Fig. 20A,B).

Hind wings 0.29 × forewing length. Venation poorly discernible/visible (Fig. 20A,B).

Legs translucent amber-orange to brown to dark brown. Length of right foreleg 5.30 mm; right foreleg femur-tibia-tarsus [tarsal segments] length ratio: 1 : 1.11 : 0.68 [0.12 : 0.16 : 0.13 : 0.15 : 0.15]. Length of right middle leg ca. 5.02 mm; femur-tibia-tarsus [tarsal segments] length ratio: 1 : 1.03 : 0.36 [– : – :?0.11 : 0.12 : 0.12]. Length of left hind leg 4.04 mm; femur-tibia-tarsus [tarsal segments] length ratio: 1 : 1.03 : 0.56 [0.09 : 0.13 : 0.10 : 0.13 : 0.12]. Pretarsal claws with outer claw hooked, inner claw blunt (Table 8; Fig. 20A,B).

Abdominal segments preserved but twisted and distorted; predominantly opaque due to post-mortem or entrapment-related torsion, partly translucent near left foreleg; no prominent posterolateral projections. Posterolateral projections absent. Subgenital and subanal plates as described by Demoulin⁵; subgenital plate narrow, trapezoidal or close to subrectangular; subanal plate of similar shape (Fig. 20A,B).

Remarks

Some aspects of the forewing venation approximate those of Maccaffertium. However, with comparable probability, the female imago may also be closely attributable to Stenonema. Demoulin (1968) hypothesized that this female specimen could represent the female of one of the Heptageniinae species described in the same paper from males only (i.e., H. atypica, H. gleissi, H. senex), but considered this unlikely because the characteristically compact, ligature-like forewing venation would imply an unlikely degree of sexual dimorphism. At present, there are no clear diagnostic differences between females of these two genera, with the exception of the characteristic abdominal colour pattern in adult extant species of Stenonema. Unfortunately, this pattern is not preserved in the fossil specimen, and therefore the precise generic placement remains unresolved.

Webb & McCafferty⁶² reported the presence of a group of crowded crossveins in the first 2–3 interspaces below the bullae (e.g. between Sc and RSa₁, occasionally iRS or RSp) in adult females of extant Stenonema, whereas in Maccaffertium females the distribution of crowded crossveins was recorded in the first 3–6 interspaces, extending from Sc to MP₁. Based on this character, H. ligata appears to be closer to Maccaffertium.

On the other hand, Bednarik & McCafferty⁵⁴ documented pronounced variability in both number and position of crowded crossveins in the forewings of North American Stenonema. Some species exhibit distinct crowding of crossveins (e.g. S. femoratum (Say, 1823) and S. pudicum (Hagen, 1861)), whereas others entirely lack such crowding (e.g. S. terminatum (Walsh, 1862)). This variability substantially reduces the diagnostic value of this feature at generic level.

Due to the absence of a preserved abdominal colour pattern on the terga, the female cannot be confidently associated with Macdunnoa. However, the absence of crowded crossveins between C and RSa₁ in the forewing of Macdunnoa allows to exclude H. ligata from that genus^62,90.

The genus Stenacron can be readily excluded based on the absence of the characteristic black spot connecting the crossveins between RA and RSa₁, as well as the lack of blackish clouds surrounding the crossveins between C and RSa₁⁶⁵.

At the same time, H. ligata differs slightly from extant species of Maccaffertium (to which it appears most similar), as well as from Stenonema, in the more proximal position of paired crowded crossveins, located between Sc and MP_1, slightly before mid-length of the forewing.

The shape of the subgenital and subanal plates in H. ligata is comparable to that observed in some extant species of Maccaffertium and Stenonema (see e.g. Bednarik & McCafferty⁵⁴; Kondratieff et al.¹⁰¹). Finally, the relatively small body size and forewing length of H. ligata correlate more closely with those of extant Maccaffertium females, in which imagines and subimagines of both sexes typically exhibit a body length ranging from 5 to 15 mm and a forewing length of 5–17 mm. In contrast, females of Stenonema are generally larger, with mean values of approximately 10–14 mm in body length and 12–16 mm in forewing length.

Georges Demoulin’s contribution to the study of Eocene Mayflies: a short overview

Already in his first contribution to mayflies in Eocene Baltic amber, Georges Demoulin¹⁰² demonstrated exceptional skill as a meticulous observer and an accomplished systematist. The systematic position and diagnostic characters of Cronicus anomalus (Pictet, 1856) (in Pictet-Baraban & Hagen³¹) were discussed, and the taxon was interpreted as a genus of Coloburiscinae, at that time classified within Isonychiidae (see Supplementary Table S1).

In his second contribution, Demoulin³ outlined in the introductory notes a clear methodological approach, which was reiterated in each of his subsequent papers on extinct taxa preserved in fossil resins, namely the redescription of both newly described and previously known fossil species within an up-to-date systematic framework. Within this initial phase of research on fossil mayflies, Demoulin relied on the assistance of several colleagues and international institutions, including Mr Jules de Walsche (*1917–†1973) from Brussels, who made Electrogenia dewalschei Demoulin, 1956 available for study from his private collection. In this publication, Demoulin³ not only described a new genus and species, but also proposed a brief key to the genera of Arthropleinae, based on adult and larval characters.

In a later paper, Demoulin⁴ acknowledged the contribution and assistance of Dr S.G. Larsson from the Copenhagen Zoological Museum [ZMUC] (Zoologisk Museum København), through which the presence of several insect orders, including Ephemeroptera, was documented. Six specimens were made available to Demoulin, who described four species based on this material (Supplementary Table S1).

Demoulin⁵ established the largest part of the mayfly taxa from Baltic amber, accompanied by a detailed analysis of their morphology and systematic position. Initially, he systematised all available data of the inventory of fossils discovered by F.J. Pictet and H. Hagen³¹, as published in the important major monograph edited by G.C. Berendt³⁷. Their original descriptions were insufficient, owing to the understandable state of invertebrate science at the time. In addition to this historical material represented by specimens housed at the Museum für Naturkunde Berlin, Demoulin examined further forty specimens deposited in both institutional and private collections. This work was made possible through the assistance of Prof. Dr R. Dehm (Bayerische Staatssammlung für Paläontologie und Historische Geologie, Munich), Prof. Dr W. Hennig (Staatliches Museum für Naturkunde Stuttgart, at that time temporarily located in Ludwigsburg), and Dr H. Jaeger (Institut für Paläontologie und Museum der Humboldt-Universität, Berlin). Thus, Demoulin⁵ published his study of six specimens originating from the A. Bachofen-Echt collection (Munich), twelve specimens from the Stantien & Becker collection (formerly Stuttgart, now Göttingen), a further twelve specimens loaned from the G.C. Berendt collection, and fourteen specimens from the W. Simon collection, both housed in Berlin.

Georges Demoulin illustrated and classified specimens from different collections within a single study, arranged according to systematic order. However, he warned that the illustrations represent projections of the specimens onto a single plane, potentially leading to distortions in the apparent proportions of structures not oriented perpendicular to the viewing axis. As reported above, these limitations were taken into account in the analysis of the morphology of fossil species. In addition to the description of new taxa, Demoulin⁵ compiled a key to six families recorded from Baltic amber and reported a total of 16 genera and 26 species (Supplementary Table S1).

Shortly thereafter, Demoulin⁶ documented and illustrated additional specimens embedded in Eocene Baltic amber and provided information on the synonymy of several previously known fossil taxa (Supplementary Table S1). The material analysed in this study originated from the collections of the National Museum of Natural History, Paris (four specimens), and ZMUC (five specimens), made available on loan with the assistance of Ms S. Kelner-Pillault and Dr Sv.G. Larsson. In total, six species belonging to four families were analysed⁷ (Supplementary Table S1).

Finally, in his last contribution on mayflies from Eocene Baltic amber, additional male characters of Cronicus anomalus were described and analysed based on a single subimaginal specimen housed in the Brussels Museum¹⁰³ (Supplementary Table S1).

Notes on the type specimens of Electrogenia dewalschei Demoulin, 1956, Succinogenia larssoni Demoulin, 1965, and Heptagenia senex Demoulin, 1968

In the present contribution, these three type species are treated separately from the other taxa, for several reasons discussed here:

Electrogenia dewalschei belongs to the monospecific genus Electrogenia, established by Demoulin³ based on a single male imago preserved in Baltic amber. In 2024, we submitted several requests to the IRSNB to obtain the holotype of E. dewalschei for re-examination. However, the holotype could not be located, and its fate following the studies reported by Demoulin³ remains unclear. Consequently, any remarks on the systematic position of Electrogenia can, at present, be made solely based on the original description and accompanying illustrations.

1. a)

   The original description of E. dewalschei is relatively detailed and is accompanied by illustrations showing a general ventral view of the body, the head in dorsal view, the apices of the foretarsi and claws, the right middle and hind tarsi, the male genitalia, and a ventral fragment of the right cercus. The author provided the tarsal formula of all preserved legs, as well as information on wing venation, colouration, and the genitalia. At the same time, the thoracic characters are described only very briefly, as typical of Heptageniidae, without further specification, and none of the figures accompanying the description depict thoracic sutures or sclerites.

2. b)

   Demoulin³ noted that the proportions of Electrogenia tarsomeres resemble those of Heptagenia. He also pointed out a reduced density of crossvenation in the forewings, comparing Electrogenia with species of the extant genera Compsoneuria Eaton, 1881 and Compsoneuriella Ulmer, 1939. However, the number of crossveins in Electrogenia on each of the preserved forewings consistently exceeds 100 (24–25 in the costal and subcostal fields), whereas for Compsoneuria Sartori¹⁰⁴ reported a total of 45–65 crossveins, with fewer than 12 in the C and Sc fields. For Compsoneuriella, Sartori¹⁰⁴ indicated numerous crossveins (> 15) in the costal and subcostal fields. Webb & McCafferty⁶² likewise documented a reduction in the number of crossveins in Compsoneuria, and Kluge²⁹ further noted the presence of four to five nearly regular cross lines formed by crossveins, a feature characteristic of all Compsoneuriini Wang & McCafferty, 2004. Subsequently, Sartori¹⁰⁴ clarified that in Compsoneuriella these cross lines are mostly absent or, if present, localised in the apical part of the radial field. In Electrogenia, however, as evident from figure 1a in Demoulin³, the crossveins are not arranged into distinct rows.

3. c)

   According to Demoulin³, the most important diagnostic character distinguishing Electrogenia from other representatives of Heptageniidae is the presence of an unbranched MA in the hind wing. The author emphasised that within this family, such a condition is exceptional and otherwise known only in closely related Arthroplea Bengtsson, 1909 (Arthropleidae Balthasar, 1937) within Heptagenioidea Ulmer, 1920. Indeed, this character was listed by Kluge²⁹ (pp. 161, 164, fig. 53D) as one of the three adult apomorphies of Arthroplea. However, as demonstrated by the extant genus Compsoneuria (Heptageniidae), both cross and longitudinal venation of the hind wings may be strongly reduced and simplified, with MA and MP entirely absent as referred by Sartori¹⁰⁴ (pp. 5, 6, fig. 2). Moreover, a MA fork is also absent in the hind wing of Cinygma dimicki McDunnough, 1934, whereas it is present in two other species of the genus Cinygma belonging to Heptageniidae³⁵. Therefore, in the case of the single fossil specimen of E. dewalschei, it remains questionable to draw conclusions about its familial affiliation based solely on this character. By contrast, Demoulin³ assigned considerable weight to this character, proposing that Arthroplea and Electrogenia be placed in a separate subfamily, Arthropleinae, within Heptageniidae. The combined phylogenetic study of Ogden et al.¹⁰⁵ indeed revealed that Arthroplea nests within Heptageniidae.

4. d)

   Another apomorphic character listed by Kluge²⁹ for adults of Arthroplea is the absence of a connection between RSa₂’’ and RSa₁, as a result of which the former has an intercalary vein that originates more distally than iRSa (Kluge²⁹: pp. 161, 164, fig. 53C; Bauernfeind & Soldán⁴⁶: p. 242). Nevertheless, in Electrogenia, judging from the illustration in Demoulin³, the connection between RSa₂’’ and RSa₁ appears to be typical of many Ephemeroptera, including Heptageniidae.

5. e)

   Finally, a character considered as diagnostic for Arthroplea is the presence of imaginal and subimaginal gonostyli composed of a total of five gonostylomeres. This condition has been reported by multiple authors and was listed by Kluge²⁹ as an imaginal apomorphy. However, Demoulin³ (p. 96, fig. 1b) illustrated the gonostyli of E. dewalschei as consisting of only four gonostylomeres, a condition typical of Heptageniidae and many other mayflies.

The characters of Arthroplea discussed above are not unique apomorphies and also occur in several other lineages of Ephemeroptera²⁹. Bauernfeind & Soldán⁴⁶ identified a distinctive genital structure they evaluated as autrapomorphy of Arthroplea. It is characterised by a pair of median titillators and a pair of long, pointed spines situated basolaterally on the penial base [a pair of ventral titillators, i.e. immovable spines, located on the ventral side proximad of the median titillators]. This condition was previously listed and illustrated (see Burks⁹²: pp. 152, 198, 201, figs. 321, 393; Studemann et al.²⁴: pp. 152, 154–156, figs. 24–32; Bauernfeind & Soldán⁴⁶: pp. 242, 243; Kluge²⁹: p. 164). In E. dewalschei, the details of the penes are poorly discernible, as already noted by Demoulin³, who described them as “constitué apparemment de deux lobes presque entièrement fusionnés, à extrémités distales libres” [apparently consisting of two lobes almost entirely fused, with free distal ends].

Finally, for adult males of Arthroplea, Bauernfeind & Soldán⁴⁶ noted a foretarsus measuring 1.0–1.3 times the length of the foretibia, and a hind tarsus in which segment II is distinctly shorter than segment I. Although the ratio of foretarsus length to foretibia length in Electrogenia was neither described nor illustrated by Demoulin³, the original illustration of the hind leg clearly shows the first tarsomere to be much shorter than the second. Moreover, the first tarsomeres of the right middle and right hind legs were depicted as fused, i.e. not articulated, with the tibia³.

Based on the original description and illustrations, there is no compelling evidence to assume a closer relationship of Electrogenia with Arthroplea. Indeed, the holotype of E. dewalschei lacks several key characters typical of adult males of extant Arthroplea. Conversely, several characters that could be interpreted as supporting an affinity of Electrogenia with Arthroplea, or more broadly with Heptagenioidea, are not exclusive to these groups and also occur in other lineages, including Siphlonuroidea.

The presence of two pairs of long intercalary veins in the cubital field of the forewing, apart from Arthroplea and other Heptageniidae, is also observed in some representatives of Ametropodidae Bengtsson, 1913 and Metretopodidae Traver, 1935 (e.g. Siphloplecton Clemens, 1915). The middle and hind tarsi of Electrogenia illustrated by Demoulin³, with a shortened first tarsomere fused (i.e. unarticulated) to the tibia, are likewise characteristic of certain Siphlonuroidea, particularly Metretopodidae. An unbranched MA in the hind wing is also typical of Ametropodidae.

Finally, the male genitalia as depicted by Demoulin³ show notable similarities to those of some Metretopodidae, especially extinct representatives of the Siphloplecton picteti and S. jaegeri groups as defined by Staniczek & Godunko³⁸ (figs. 6D, 7B, 8E, 9B, 10H, 11B, 12F, 13B) and Godunko et al.³⁹ (figs. 4B, 5F, 7). These similarities include the shape of the distal margin of the gonobasal plate, the proportions and number of gonostylomeres, and particularly the structure of the penial lobes, which are elongated and rhomboidal, and fused for most of their length. At the same time, all known species of the S. jaegeri group are clearly distinguished by the presence of only three long intercalary veins in the cubital field of the forewing.

Thus, the systematic position of Electrogenia should at present be regarded as unresolved, pending on the rediscovery and re-examination of the holotype of the type species. In particular, a detailed description of the thoracic morphology is required, with special emphasis on the shape and position of the furcasternal protuberances of the mesosternum, as well as a reassessment of the articulation between the first tarsomere of the middle and hind legs and the tibia.

Succinogenia larssoni is the type species of the monospecific genus Succinogenia Demoulin, 1965, which was established based on a nymph preserved in Baltic amber. At present, S. larssoni remains the only species of Ephemeroptera described exclusively from larval material originating from this type of fossil resin. This undoubtedly represents a unique record, given that mayflies (imagines and subimagines) constitute no more than approximately 4% of all amber inclusions of aquatic insects in Baltic amber¹⁰⁶. Only a few heptageniid nymphs have so far been reported from this amber, apparently referable to the subfamilies Heptageniinae and Ecdyonurinae^74,75,76,107. Apart from the Eocene S. larssoni, only a single additional mayfly species has been described from fossil resins, namely the monospecific genus Kachinophlebia zhouchangfai Chen & Zheng, 2022, based on a solitary specimen from mid-Cretaceous Kachin amber¹⁰⁸.

While examining mayfly material described by Demoulin, we contacted Lars Vilhelmsen (Deputy Museum Director of Education, ZMUC) with a request to access the relevant specimens. Owing to his kind assistance, we were able to study the mayfly types cited in Demoulin⁴, including the holotype nymph of S. larssoni (Fig. 21A,B; ZMUC collection, inventory number 900397).

Examination of the holotype revealed that the head of the nymph is now severely damaged compared with its original condition, as documented in the original illustration (fig. 5 in Demoulin⁴). This circumstance substantially hampers the assessment of the systematic position of the larva, as diagnostic concepts proposed for extant genera and tribes of Heptageniidae rely primarily on larval characters^29,33,35. For this reason, we postpone a detailed reinvestigation of the holotype of S. larssoni until non-destructive imaging tools, preferably micro-CT, can be applied, ideally in combination with comparative material from other Eocene heptageniid larvae mentioned above (see also Cechowicz et al.¹⁰⁷).

Heptagenia senex was described by Demoulin⁵ based on a male imago from the collection of W. Simon (MNB), where it was stored together with other material prepared using the “Königsberg method” (Fig. 22A–C; holotype, male imago, MB.I.2234). This well-preserved specimen was described and illustrated in detail by Demoulin, who provided convincing evidence for assigning H. senex to the genus Heptagenia. In particular, this conclusion is supported by the characteristic structure of the penes, with markedly folded penial lobes lacking distinct sclerites (see Demoulin⁵: pp. 255, 256, fig. 21b). Nevertheless, the assignment of H. senex to one of the three subgenera currently recognised within Heptagenia remains problematic and represents a more complex taxonomic issue.

The prosternum of the male holotype lacks a distinct longitudinal keel but bears a well-developed transverse ridge, a condition characteristic of Kageronia⁶². However, because the forelegs of the holotype are damaged and most tarsal segments are missing (see Table 9), it is not possible to assess the proportions of the individual segments or their precise length ratios relative to the foretibia. The same limitation applies to the tarsi of the poorly preserved right hind leg (Table 9). The only completely preserved left foretarsomere measures at least 0.21 of the foretibia length, whereas in males of both extant and extinct species of Heptagenia s.str. and Dacnogenia this ratio typically ranges from 0.10 to 0.15. Thus, the foretarsomere in H. senex is clearly not shortened, a feature that may further support the tentative assignment of this species to Kageronia. Based on these preliminary results, and subject to the reservations outlined above, we tentatively place Heptagenia senex in the subgenus Kageronia, i.e. Heptagenia (? Kageronia) senex.

Approximately four decades ago, Kluge⁸ described and illustrated a male imago from Baltic amber that he considered to be very similar to the extant species Heptagenia (Kageronia) fuscogrisea (specimen Ar 8984, housed in the collection of SMNH [Naturhistoriska riksmuseet], Stockholm). This assessment was based on the structure of thorax, wing venation, the proportions of the tarsomeres of the fore- and hind legs, and, in particular, the structure of the male genitalia. This led Kluge⁸ to regard the fossil male as indistinguishable from of the extant H. (K.) fuscogrisea. Moreover, the general shape of the penes in H. (K.) fuscogrisea, the fossil specimen identified by Kluge as “H. (K.) fuscogrisea”, and H. senex is remarkably similar, further complicating the interpretation of taxonomic relationships among these materials. Moreover, all examined material exhibits similarity in body and wing size, as well as in wing venation.

To facilitate a comprehensive comparative analysis of this material, we assembled an extensive series of extant specimens of H. (K.) fuscogrisea from various localities across the European part of the range of this trans-Palaearctic species. Earlier, we contacted the late Christina Franzén-Bengtson (*1942–†2023) of the Department of Palaeozoology, SMNH, with a request to examine the male specimen described and associated with H. (K.) fuscogrisea by Kluge⁸, and we were kindly granted access to this material.

Having assembled all the aforementioned material, we plan to undertake a comprehensive study of fossil evidence potentially indicating the presence of Kageronia in Baltic amber. In addition to classical morphological approaches, this investigation will incorporate non-destructive imaging techniques, such as micro-CT and SEM, particularly in the analysis of genital structures. A redescription of the holotype of H. senex will form part of this separate contribution.

Summary of the systematic position of the Heptageniid taxa described by Georges Demoulin

Studies of mayflies from Baltic amber conducted by Georges Demoulin^{3,4,5,6,102,103} represent a major contribution to our understanding of the spatial and temporal diversification of Cenozoic biota. Remarkably, Demoulin’s original species descriptions, illustrations and interpretations of the systematic position of Ephemeroptera are characterised by a high degree of accuracy and an exceptionally rigorous scientific standard. Our re-examination of the original material has demonstrated that the family- and genus-level assignments proposed by G. Demoulin are correct in most cases. This conclusion applies not only to representatives of Heptageniidae, but also to other fossil mayfly taxa from Baltic amber that we have revised over the past two decades^30,39,45,48.

The Eocene representatives of Heptageniidae described by Demoulin belong to subfamilies and genera characterised by high extant species diversity and a complex, in many respects controversial systematic framework. Divergent approaches to the classification of Heptageniidae, differing concepts of the rank and circumscription of supraspecific taxa, and pronounced morphological variability within genera (particularly evident in adult stages), substantially complicate the interpretation of fossil material, which is already inherently challenging.

The systematic interpretation of fossil representatives of Heptageniidae is further complicated by the limited set of diagnostic characters available in the imago. Demoulin’s conclusions were based primarily on features of wing venation, measurements and proportions of the body, particularly those of the leg segments, and general aspects of male genital morphology, which fully conformed to the taxonomic standards of his time. However, subsequent studies have demonstrated that many of these characters are themselves highly variable and are often insufficient for the reliable delimitation of genera or subgenera, particularly in cases where the larval stages remain unknown⁴⁶.

In his discussions of the diagnoses of subfamilies, tribes and genera of Heptageniidae, N.J. Kluge^{29,33,35,50,51} relied predominantly on larval characters, emphasising the considerable variability in adults. A similar conclusion is reflected in the diagnosis and generic keys of Heptageniidae provided by Webb & McCafferty⁶².

Considering evolutionarily conservative and systematically stable characters of adult specimens, Kluge^29,34 and Kluge et al.⁴⁷ paid particular attention to thoracic morphology, and for the first time proposed its use as a diagnostic tool in the study of fossil mayflies from Baltic amber, based on the male imago of Heptagenia⁸. The same approach is applied in the present study to the species and genera of Heptageniidae described by Demoulin. Reassessment of these taxa using thoracic characters confirms the presence of the subfamilies Heptageniinae and Rhithrogeninae in Eocene Baltic amber on the basis of the material described by Demoulin, as well as Ecdyonurinae, as recorded in our later studies^10,11,107.

In particular, Rhithrogena sepulta is reliably assigned to Rhithrogeninae based on the characteristics of pro- and mesosternum, and the shape and configuration of the mesothoracic sutures. This confirms Demoulin’s original attribution, but based on more morphological characters. The genus Heptagenia reported by Demoulin from Baltic amber is generally well diagnosed owing to the characteristic male genitalia. Our examination of the type material, focused on a detailed analysis of thoracic morphology and a reassessment of leg segment proportions, confirmed the generic affiliation of most specimens and allowed their subgeneric placement within the genus Heptagenia to be discussed with a higher degree of confidence (e.g. H. bachofeni, H. baltica and H. senex).

At the same time, a number of species represented by type specimens in unsatisfactory condition are deliberately treated here as Heptagenia in a broad sense, which within the framework of the present study corresponds to the subfamily Heptageniinae (H. atypica, H. gleissi and H. demoulini sp. nov.). This approach is adopted in order to maintain nomenclatural and systematic stability of the genus Heptagenia and the subfamily Heptageniinae as a whole. Future, more detailed investigations of fossil material provisionally associated with the Stenonema generic complex should clarify the systematic status of problematic species that are temporarily left in limbo within Heptagenia s.l. Should the assumption of a rich and diverse Stenonema generic complex fauna in Baltic amber be confirmed, it would become possible to associate H. atypica and H. gleissi more confidently with the tribe Stenonematini, whereas H. demoulini sp. nov. could be referred to Heptagenini, were it not for the unsatisfactory preservation of key diagnostic characters in the holotype of male imago.

In discussing the systematic position of H. ligata, we deliberately restrict its placement to Heptagenia s.l., owing to the poor visibility of morphological characters in the dark and weakly translucent amber piece, as well as to the limited set of diagnostic characters available in female imagines, which does not allow a reliable assignment of this extinct species to a lower-ranking taxon.

The poor preservation of the larval holotype of S. larssoni constitutes a serious obstacle to a precise assessment of its systematic position. A potential solution lies in the application of non-destructive imaging techniques, particularly micro-CT, to investigate any preserved diagnostic structures. A similar approach appears advisable when addressing questions related to the systematic position of H. senex and its relationships with fossil and extant representatives of the genus Kageronia, and probably also in the case of fossil representatives of the Stenonema generic complex.

Finally, in the case of E. dewalschei, our conclusions are necessarily very preliminary and are based exclusively on the original description and illustrations provided by Demoulin³, as the type specimen formerly housed in the IRSNB has not yet been relocated. Several adult characters may point towards an affinity of this species rather with Metretopodidae than with Heptageniidae. In addition to purely morphological evidence and its interpretation, this hypothesis is further supported by the remarkable diversity of Metretopodidae in Baltic amber, represented in particular by the genus Siphloplecton, in which males of some species groups exhibit genital structures resembling those of E. dewalschei. It is noteworthy that Demoulin^5,6 examined representatives of Siphloplecton from Baltic amber; however, following his original description of E. dewalschei³, he never returned to this species.

Thus, the revised material originally described by Georges Demoulin provides an opportunity to reassess the broader context of Heptageniidae diversity in the Eocene of Europe. It is convincingly demonstrated that the family is represented in fossil resins by a diverse assemblage of genera and species encompassing all three subfamilies. Heptageniinae are the most comprehensively documented, based on the studies of Demoulin^5,6, Kluge⁸, and the present contribution; Rhithrogeninae are additionally known from the description of the genus Burshtynogena Godunko & Sontag, 2004; and Ecdyonurinae are represented by a recently discovered male imago of Ecdyonurus (Helvetoraeticus) and by the genus Nestormeus Godunko, 2004^10,11,12,107.

Biogeographic considerations

Following the mass extinction at the Cretaceous–Palaeogene boundary (c. 65 Ma) and, probably, during the early Palaeocene, temperatures across parts of the Eurasian continent were broadly comparable to those recorded in North America¹⁰⁹. The Paleocene–Eocene Thermal Maximum (PETM), at approximately 55 Ma, was not only associated with profound transformation of the terrestrial biosphere, but also with a rise in deep-ocean temperatures of about 5–6 °C. This coincided with the onset of large-scale collision between the Indian lithospheric plate and Eurasia. Concurrently with this tectonic convergence, which persisted throughout the Eocene (c. 56–34 Ma), rifting and subsequent opening of the North Atlantic Ocean commenced, with its most active phase occurring around 60–50 Ma.

These tectonic processes resulted in major reorganisation of oceanic and atmospheric circulation patterns, promoting enhanced hydrological cycling and the establishment of more humid climatic conditions in several regions of the Northern Hemisphere. Such environmental changes coincided temporally with a series of significant biotic events, including large-scale redistribution and dispersal of biota, likely driven by the combined effects of climatic shifts and changing palaeogeographic configurations¹¹⁰. In parallel, progressive fragmentation of the Laurasian landmass led to the final separation of North America and Europe by approximately 55–50 Ma, exerting an additional influence on dispersal routes and the emerging biogeographical structure of continental biotas.

As a consequence, numerous invertebrate taxa, including aquatic insects such as mayflies, which had previously been widely distributed across Laurasia, became isolated within what are now the Nearctic and Palaearctic regions. Some of these taxa have retained a Holarctic distribution pattern to the present day and simultaneously represent characteristic elements of European Eocene amber faunas, for example the genera Rhithrogena, Heptagenia, Ameletus Eaton, 1885, Metretopus, Siphlonurus, Ephemerella, Paraleptophlebia and members of Arthropleinae Balthasar, 1937, while others are essentially cosmopolitan (e.g. Baetis Leach, 1815, Labiobaetis Novikova & Kluge, 1987, Pseudocloeon Klapálek, 1905, Cloeon Leach, 1815, Caenis Stephens, 1836)^2,4,5,44,48.

In contrast, a second group of taxa, although present in the European Eocene and described as amber inclusions, subsequently disappeared from the Palaearctic but persist as elements of the recent Nearctic fauna. This group includes Analetris, Siphloplecton, representatives of Baetiscidae Edmunds & Traver, 1954, as well as numerous yet undescribed taxa belonging to the Stenonema generic complex, briefly discussed in the present study^{38,39,95,111,112}. Finally, a third group comprises taxa documented from European Eocene ambers whose extant distributions are now restricted to the western Palaearctic, such as the genera Calliarcys Eaton, 1881 and Ecdyonurus (Helvetoraeticus)^107,113.

The causes underlying the disappearance of many extant Nearctic taxa from Europe remain unclear, as does the timing of this exodus. According to our unpublished data, the taxonomic structure of dominant groups in other European Eocene fossil resins slightly differs from that observed in Baltic amber. In Oise amber (Paris Basin, Le Quesnoy; Early Eocene [Ypresian, Sparnacian], ca. 55–53 Ma), the assemblage is dominated by Baetidae (cf. Baetopus Keffermüller, 1960) and Rhithrogeninae, and both groups having extant representatives distributed in the Holarctic and Oriental regions. In contrast, Rovno amber (Ukraine/Belarus; Late Eocene [Priabonian], ca. 37.8–33.9 Ma) is characterised by the predominance of the genus Paraleptophlebia (Leptophlebiidae), accompanied by a smaller proportion of Rhithrogenini and other Heptageniidae incertae sedis. Finally, Bitterfeld amber, which has been associated by some authors with Upper Oligocene (Chattian) deposits dated to ca. 28.1–23.0 Ma¹¹⁴, yields the lowest number of known mayfly inclusions; all of them are represented exclusively by adult specimens of Paraleptophlebia.

While Ephemeroptera are represented in Europe by both compression fossils (e.g. Masselot & Nel¹³) and inclusions in fossil resins, records of mayflies from North American deposits of the same time interval are comparatively scarce. In North America, Ephemeroptera are predominantly known from compression fossils, consisting mainly of larval remains and, more rarely, adult specimens. These fossils are typically assigned to extant genera and families and exhibit primarily Nearctic, and in some cases Oriental biogeographic affinities. The most substantial records of Ephemeroptera from Eocene deposits of North America originate from the Ruby River Basin, the Green River and Florissant formations, and the Republic locality^{21,115,116,117,118,119}. In contrast, records from the eastern and south-eastern Palaearctic are even more limited and are largely restricted to widely distributed families, including Siphlonuridae, Ephemeridae, Heptageniidae, and Leptophlebiidae^9,120.

More generally, fossil records of mayflies from the Paleogene to the Quaternary across the Northern Hemisphere, as well as the extent to which their systematic placement can be assessed with confidence, remain extremely limited. This scarcity constrains robust inferences regarding the timing of origin and extinction of individual lineages and hampers regional comparisons of evolutionary turnover. At the species level, however, the influence of Miocene orogenetic events on diversification and the establishment of intrageneric diversity is well supported. This has been demonstrated for several Palaearctic mountain systems using Epeorus (Caucasiron) as a model taxon, highlighting the role of Neogene landscape evolution in shaping present-day diversity patterns^20,121.

In the Recent fauna, Heptageniidae comprise more than 600 species in 36 genera worldwide, of which 14 genera occur in the Palaearctic. Approximately 30% of species, representing 16 genera, are distributed in the Nearctic (see Kluge N.J., web page: https://insecta.bio.spbu.ru/z/Eph-phyl/_Radulapalpata.htm). The family is moderately represented in the Afrotropic, with about 20 described species in two genera, and in the Oriental Region, with approximately 35 species in 11 genera. In contrast, its presence in the Neotropics is highly restricted, being limited to only three species in three genera². The Holarctic and Neotropical genus Rhithrogena Eaton, 1881, comprising more than 150 species, is the most diverse within Heptageniidae and the third most species-rich genus within Ephemeroptera.

The present-day distribution of Heptageniidae is reflected in the biogeographic patterns of its constituent tribes^1,29,62. The widely distributed tribe Heptageniini includes representatives occurring throughout the Holarctic and Oriental regions (e.g. Heptagenia). In contrast, the tribes Ecdyonurini and Leucrocutini (subfamily Ecdyonurinae) exhibit more restricted distributions: Ecdyonurus Eaton, 1871 is prevalent in the western Palaearctic, whereas Nixe Flowers, 1980 and Leucrocuta Flowers, 1980 are confined to the eastern Palaearctic and the Nearctic. The tribe Rhithrogenini is broadly distributed across the Holarctic, exemplified by the genus Rhithrogena.

Several tribes and genera within Heptageniidae show even narrower distributional ranges, being confined to particular biogeographic regions. The so-called “Stenonema complex”²⁹ is characteristic of the Nearctic and includes Stenacron Jensen, 1974 (closely related to Kageroniini Wang & McCafferty, 2004), together with Maccaffertium Bednarik, 1979, Stenonema Traver, 1933, and Macdunnoa Lehmkuhl, 1979, all of which belong to the tribe Stenonematini Wang & McCafferty, 2004.

The predominantly Northern Hemisphere diversity of Heptageniidae, together with the restriction of all currently known fossil records of the family to this region, provides strong support for a Laurasian origin, a view that represents a broad consensus among researchers^1,2,33. Moreover, available evidence indicates that well before the final fragmentation of Laurasia, at least since the Early Cretaceous, Heptageniidae were already a widespread and established component of freshwater ecosystems across this ancient landmass¹⁰⁷.

This interpretation is further supported by the occurrence of Rhithrogeninae at both high latitudes, like in the late Campanian Wapiti Formation of Canada, and at lower latitudes, in Cenomanian Burmese amber deposits. These occurrences are documented by fossil larvae¹⁶ as well as by adult males and females¹⁹. In the case of the subfamily Ecdyonurinae, Edmunds¹²², followed by Kluge³³ and Barber-James et al.², hypothesised that its centre of origin and early diversification was located in Laurasia. This hypothesis is partly supported by the discovery of a male preserved in mid-Cretaceous Burmese amber²⁰.

At present, no further Cretaceous fossil records of Ecdyonurinae have been reported, and subsequent occurrences of this subfamily are restricted to Eocene European amber deposits^10,11,107. Similarly, the earliest unequivocal fossil records of Heptageniinae are currently known only from Baltic amber (present contribution), although an earlier and more extensive Laurasian distribution prior to the Eocene cannot be excluded. For example, if the fossil larva described by Lewis & Wehr²¹ from an Eocene deposit near Republic, Washington, north-western USA (Klondike Mountain Formation), and tentatively identified by those authors as “?Stenonema or?Heptagenia”, indeed belongs to either Stenonematini or Heptagenini, then these tribes must have been components of freshwater ecosystems for at least the interval between ca. 49–48 and 38–33.9 Ma, within a broad geographic range extending from present-day Europe to western North America.

Although the final fragmentation of Laurasia was largely completed by the early Eocene, approximately 60–55 Ma, biotic exchange between the northern continents persisted throughout much of the Paleogene. During the early to middle Eocene (ca. 56–40 Ma), Europe existed as a highly fragmented archipelago bordered by shallow epicontinental seas, a configuration that promoted regional endemism while still permitting episodic intercontinental dispersal during periods of lowered sea level^123,124. Climatic conditions during this interval were dominated by a greenhouse regime, culminating in the Early Eocene Climatic Optimum (EECO; ca. 53–50 Ma), when globally elevated temperatures supported extensive boreotropical forests across the Northern Hemisphere^110,125. These forests provided ecologically continuous corridors favourable for the dispersal of thermophilic and freshwater-associated insects.

Within this palaeogeographic, climatic and palaeoclimatic framework, Heptageniidae, inferred to have originated in Laurasia during the Middle Jurassic (ca. 165 Ma) and to have diversified into major subfamilies well before the mid-Cretaceous (ca. 95–90 Ma), represented a component of these early Paleogene exchange systems^2,126.

Intercontinental dispersal during the Paleocene–Eocene (ca. 66–40 Ma) was facilitated by a dynamic system of northern land bridges that operated intermittently and asynchronously. Transatlantic exchange was possible via the North Atlantic Land Bridge, incorporating the De Geer Route during the latest Cretaceous to early Paleocene (ca. 70–60 Ma) and the later Thulean Route during short-lived intervals of the early Eocene (ca. 57–56 Ma). These routes linked Fennoscandia, Greenland and northeastern North America under warm climatic conditions^123,127. In parallel, Beringia provided a transcontinental connection between Eurasia and western North America during multiple Paleogene intervals, particularly following the retreat of the Western Interior Seaway after the Paleocene and during climatically permissive phases of the early to middle Eocene^21,127.

Comparable Paleogene Holarctic distributions followed by regional extinction have been documented in several other insect groups, including aquatic Plecoptera (Nemouridae, Leuctridae)¹²⁸, Diptera (Bombyliidae, Keroplatidae)^124,129, and Coleoptera (Staphylinidae)¹³⁰. Progressive climatic cooling from the middle Eocene onward, culminating in the Eocene–Oligocene transition at ca. 34 Ma, led to the collapse of boreotropical ecosystems and widespread extinction of subtropical faunal elements in Europe¹¹⁰. Consequently, the recent Holarctic distribution patterns observed within Heptageniidae are best interpreted as relicts of a broader Paleogene or earlier range, shaped in its final phase by early Eocene faunal interchange across northern land bridges and subsequent regionally asynchronous extinction events. Such a model is applicable to the aquatic insect groups discussed above, as well as to mayfly taxa, particularly Heptageniidae.

Palaeoecology

Heptageniidae in Baltic amber: ecology versus taphonomy

Extant representatives of the mayfly family Heptageniidae exhibit a pronounced ecological affinity for fast-flowing, well-oxygenated lotic environments. This rheophilic specialisation is expressed through a suite of morphological adaptations, including a dorsoventrally flattened body, enlarged abdominal gills, and, particularly in many genera of Rhithrogeninae, gill lamellae modified into plate-like structures functionally analogous to suction discs, which facilitate firm attachment to the substrate via modified setae on their gills^46,131. These adaptations enable the body to align parallel to laminar flow and confer resistance to dislodgement under turbulent hydraulic conditions (Brittain¹³²; for a recent synthesis see Ditsche et al.¹³³).

The occurrence of larval ecomorphotypes [ectomorphs; life forms] in Eocene resins, closely resembling those of extant Heptageniidae^4,107, provides compelling evidence for the presence of suitable freshwater habitats during this period. These habitats likely comprised a heterogeneous network of running waters, ranging from shallow, high-gradient, stony headwater streams predominantly inhabited by Rhithrogeninae and Ecdyonurinae to medium-sized and large lowland or submontane rivers supporting Heptageniinae, Arthropleidae, and other meta- to hyporhithral mayfly taxa, such as Analetris, certain species of Siphloplecton, and Baetiscidae (see also Cechowicz et al.¹⁰⁷). The extant genera of Heptageniidae identified in Baltic amber are exclusively rheophilous, with numerous living species associated with running waters. Accordingly, based on the ecomorphological and taxonomic structure of the fossil assemblages, and drawing on modern analogues, it can be inferred with a high degree of confidence that these represent diverse and abundant rheophilous communities, rather than indicating that Eocene Heptageniidae possessed a different ecological specialization than today.

Eocene fluvial systems would therefore have provided the hydrodynamic and physicochemical conditions necessary to sustain diverse heptageniid populations. In recent ecosystems, Vuataz et al.¹³⁴ demonstrated that the strong rheophility of Rhithrogena species, together with their strict preference for highly oxygenated lotic habitats, have played a key role in their diversification within Alpine streams. Comparable ecological constraints may reasonably be inferred for Eocene taxa. Accordingly, the existence of a dense and dynamic hydrographic network, combined with episodic habitat fragmentation driven by orogenic processes and marine transgressions, likely acted as a major evolutionary driver for aquatic insect lineages such as Heptageniidae, as well as Baetidae and Leptophlebiidae, contributing to the rapid diversification of mayflies from the Eocene to the present^20,121.

In contrast, the taphonomic context of fossil deposits yielding Heptageniidae across different geological periods and regions is frequently interpreted as predominantly lacustrine (for example Masselot & Nel¹³; Demers-Potvin & Larsson¹³⁵) or as reflecting low-energy depositional settings such as river backwaters (Sinitshenkova¹⁴). This apparent discrepancy raises important questions concerning taphonomic bias and the palaeoecological fidelity of such assemblages¹⁶. In particular, it remains unclear whether these palaeolakes represented cold, well-oxygenated environments capable of sustaining autochthonous heptageniid populations, or whether the fossil record instead reflects allochthonous accumulation following post-mortem transport from adjacent lotic habitats.

When reconstructing the palaeoecological composition of fossil mayfly assemblages within a given deposit, it is therefore essential to integrate not only the broader geological and sedimentological context but also key biological characteristics of Ephemeroptera. These include larval habitat specificity, seasonal and diel downstream drift dynamics, the generally limited dispersal capacity of adults, and the compensatory upstream flight behaviour of ovipositing females. Collectively, these factors are critical for disentangling primary ecological signals from taphonomic overprinting in fossil mayfly communities.

Because mayflies preserved in Eocene resins from Europe encompass a broad taxonomic spectrum represented predominantly by adult life stages, palaeoecological reconstruction is further facilitated through analogy with extant phylogenetically related taxa for which larval stages and ecological requirements are well documented. Given the evolutionarily constrained dispersal capacity of short-lived adult mayflies, characterised by limited sustained flight ability, it is highly probable that both fossil and extant taxa exhibited comparable behavioural patterns, with mating swarms forming near running waters that provided suitable habitats for their larvae.

Within this framework, an important feature of the qualitative composition of Eocene resin inclusions, reflecting evolutionarily conserved behavioural traits of winged stages in Ephemeroptera, is the conspicuously high proportion of subimagines. It may be inferred that shortly after emergence from the larval stage, subimagines of both sexes dispersed from the water surface into adjacent riparian vegetation, including shrub layers and tree canopies, where they remained for a transitional period ranging from several hours to several days prior to moulting into the imago stage.

It is during this ecologically and behaviourally constrained interval that subimagines were most susceptible to entrapment in viscous tree resins. This taphonomic pathway strongly suggests that resin-producing plants were present either directly along stream margins or in close proximity to aquatic habitats. Consequently, the dominance of subimaginal inclusions in Eocene ambers provides indirect yet robust evidence for the spatial coupling of forested riparian zones and freshwater bodies, thereby reinforcing paleoenvironmental reconstructions that invoke well-developed lotic ecosystems embedded within densely vegetated landscapes.

Implications for Heptageniidae regarding the origin and fluvial transport of amber inclusions

A postulated uplift of the Fennoscandian Shield, approximately 40 Ma ago, has traditionally been interpreted as giving rise to a large fluvial system, commonly referred to as the Eridanus River or Baltic River System, which transported sediments and amber from northern Europe towards the North Sea Basin^136,137. The size of this system has been estimated to be comparable to that of large modern deltaic complexes¹³⁷. According to this classical model, the river flowed from the highlands of the Russian–Scandinavian mainland (now Lapland) through the present-day Gulf of Bothnia and the Baltic Sea into the North Sea, discharging via a large Eridanus delta¹³⁶.

However, the existence of a continuous, large-scale Eridanus River remains debated. Recent studies have emphasised that direct geological evidence for such a transcontinental river system is lacking^138,139. In particular, no continuous sedimentary structures unequivocally attributable to a single, long-lived river have been identified. Instead, these authors propose that amber transport may have occurred through a network of smaller, regionally restricted river systems, combined with coastal reworking and marine redistribution processes¹³⁸.

Within the traditional framework, the river system has been considered responsible for transporting amber-bearing sediments and associated biota from Fennoscandia to the North-West European Basin. Fennoscandia has further been interpreted as hosting extensive resin-producing forests, previously thought to be dominated by Pinaceae^96,140,141. However, this interpretation has been revised in light of more recent palaeobotanical and geochemical evidence. In particular, Wolfe et al.¹⁴² and subsequent studies^{138,139,143,144} suggest that the primary source of Baltic amber was more likely related to members of the Sciadopityaceae, rather than Pinaceae. This reinterpretation has important implications for reconstructing the structure and ecology of Eocene amber forests.

The palaeoenvironment of the supposed Eridanus basin has been described as consisting of fast-flowing, well-oxygenated lotic systems draining upland regions. Nevertheless, in the absence of preserved upstream channel systems, this reconstruction remains hypothetical. Alternative models suggest that amber and associated organic material accumulated in a range of depositional environments, including coastal plains, estuaries, lagoons and shallow marine basins^139,145. Such environments would have facilitated both the preservation and secondary redistribution of amber, including floating remains of terrestrial and aquatic organisms. In this context, the occurrence of aquatic insect inclusions, including representatives of the Heptageniidae, does not necessarily imply the existence of a single extensive river system, but may instead reflect a mosaic of smaller freshwater habitats within a heterogeneous landscape.

The buoyancy of amber would have further enhanced its dispersal potential, allowing transport not only by rivers but also by coastal currents and storm-driven processes¹³⁶. This multiphase transport model provides a plausible explanation for the wide geographic distribution of Baltic amber, including its occurrence in the North Sea Basin and adjacent regions, without requiring a single large river system.

Eocene amber inclusions, including aquatic insects such as Heptageniidae, have been interpreted as reflecting the prevailing bioclimate of the time. The Eocene climate is generally considered to have been warm temperate to subtropical, with mean annual temperatures ranging from approximately 15 to 25°C^146,147. However, alternative interpretations emphasise that the European Eocene climate was spatially heterogeneous and temporally dynamic, influenced by global events such as the Palaeocene–Eocene Thermal Maximum^110,147 and subsequent Middle Eocene climatic fluctuations.

Moreover, recent work^139,143,144 suggests that amber forests may have developed under more humid and potentially paratropical conditions than previously assumed, with diverse plant communities rather than simple pine-dominated assemblages. Consequently, direct comparisons with modern temperate montane ecosystems should be treated with caution. This also affects the interpretation of associated aquatic taxa, as their ecological requirements may have been broader or different under Eocene conditions.

The occurrence of rheophilic taxa such as Heptageniidae has often been interpreted as evidence of well-developed river systems. However, this interpretation is not unequivocal. Such taxa may also have inhabited a variety of flowing-water environments, including smaller streams and marginal aquatic habitats within a complex landscape. In addition, ecological and climatic uniformitarianism may not strictly apply, as this family, based on modern analogues, occurs across a relatively wide range of conditions, including different types of running waters and climatic zones on all continents except Antarctica².

The Eocene was characterised by pronounced climatic variability, including the Palaeocene–Eocene Thermal Maximum (~ 55.8 Ma), subsequent cooling phases, and the Middle Eocene Climatic Optimum^109,147. These fluctuations likely influenced both vegetation and hydrological systems, potentially leading to shifts in ecosystem structure over time. Therefore, the assumption of long-term ecological stability in the amber-producing environments requires careful re-evaluation, particularly when interpreting the ecological significance of aquatic insect taxocenes.

Baltic amber has recently been dated to approximately 37.7–34 Ma, corresponding to the Priabonian stage of the Late Eocene⁷⁸. Although this interval broadly coincides with still relatively warm climatic conditions, the ecological and depositional context of amber formation and accumulation was likely more complex than previously assumed.

Finally, the disappearance of the Eridanus system has traditionally been linked to the advance of continental ice sheets during the Pleistocene. However, if the existence of a large, unified Eridanus River is uncertain, its inferred disappearance must likewise be treated cautiously. Instead, the evolution of northern European drainage systems may reflect a progressive reorganisation of multiple smaller basins in response to tectonic and climatic changes.

Summarising, while the Eridanus River concept provides a convenient framework for interpreting amber transport, increasing evidence supports a more complex scenario involving fragmented drainage networks, diverse resin-producing forests, and significant coastal and marine reworking processes. Within this framework, the presence of Heptageniidae is best interpreted as evidence for the existence of running-water habitats, rather than as definitive support for a single, large-scale river system, as both models provide a broad range of ecological conditions conducive to the development of diverse Heptageniidae taxocenes.

The revision of Baltic amber Heptageniidae originally described by Georges Demoulin highlights both the enduring value of his pioneering work and the importance of reassessing fossil taxa in the light of modern systematic concepts. Although Demoulin worked with a limited set of characters available at the time, his interpretations of generic and familial affinities were, in most cases, remarkably accurate and remain largely valid today.

The present study demonstrates that the incorporation of additional adult characters-particularly those derived from thoracic morphology-provides a more robust framework for evaluating the systematic position of fossil mayflies. Such characters substantially improve the resolution of taxonomic interpretations based on adult material and facilitate clearer comparisons between fossil and extant lineages within Heptageniidae.

At the same time, our results emphasise the inherent limitations of the fossil record, especially in cases of incomplete preservation, sexual dimorphism, or reliance on single specimens. For several taxa treated herein, cautious retention within broadly defined systematic categories is therefore preferable to premature reassignment, as it ensures nomenclatural stability while explicitly acknowledging existing uncertainties.

Based on the re-examination of eleven fossil specimens, we propose several nomenclatural changes, including new combinations and revised generic placements. These include the transfer of Cinygma baltica Demoulin, 1968 to Heptagenia (? Heptagenia) baltica (Demoulin, 1968) comb. nov., and the reassignment of two specimens previously identified as Rhithrogena sp. to Heptagenia, with the male imago described herein as Heptagenia (s.l.) demoulini sp. nov.

Characters indicating a possible placement of the two known adult specimens of Heptagenia bachofeni Demoulin, 1968 in the subgenus Dacnogenia are discussed. With regard to the holotype of Heptagenia ligata Demoulin, 1968, and in view of the limited number of diagnostic characters available for reliably assessing the systematic position of many heptageniid genera based on female imagines, we deliberately restrict its placement to Heptagenia s.l.

We further advance a preliminary suggestion that Heptagenia senex Demoulin, 1968 may be associated with Kageronia; however, a more detailed analysis of the relationships between extant and fossil representatives of this subgenus requires the examination of additional material and is therefore deferred to a separate contribution. The systematic position of Heptagenia (s.l.) atypica Demoulin, 1968 and H. (s.l.) gleissi Demoulin, 1968 is discussed in the context of their possible relationships with the extant Nearctic Stenonema generic complex.

Additionally, we analyse morphological evidence that may indicate the assignment of Electrogenia dewalschei Demoulin, 1956 to the family Metretopodidae; however, any definitive conclusion must await the rediscovery and re-examination of the holotype. Succinogenia larssoni Demoulin, 1965 remains a taxon of uncertain systematic position within Heptageniidae owing to the extensive damage to the larval holotype.

The study further underscores the importance of integrating non-destructive imaging techniques, such as micro-CT and SEM, into future investigations of amber inclusions. These approaches hold particular promise for resolving long-standing questions concerning the systematic position of problematic taxa known from poorly preserved or unique specimens.

Beyond their taxonomic implications, refined interpretations of Baltic amber Heptageniidae contribute to a deeper understanding of the evolutionary history and paleobiogeography of Ephemeroptera. The presence of taxa in Eocene Europe showing affinities to extant Nearctic lineages reflects complex historical connections within the Holarctic freshwater fauna and highlights the value of amber fossils for reconstructing patterns of faunal exchange and diversification through deep time.