Research Article

The rise and fall of the Eocene fly genus Kelneria (Diptera, Keroplatidae)

Alicja Pełczyńska^‡§, Evgeny E. Perkovsky^§|, Brendon E. Boudinot^¶, Adrian Richter^¶, Wiesław Krzemiński^#, Tomasz Kołodziej^¤, Pawel Korecki^¤«, Katarzyna M. Sowa^¤, Agnieszka Soszyńska^‡

‡ Department of Invertebrate Zoology and Hydrobiology, Faculty of Biology and Environmental Protection, University of Lodz, Łódź, Poland

§ Natural History Museum of Denmark, University of Copenhagen, Copenhagen, Denmark

| I.I. Schmalhausen Institute of Zoology, National Academy of Sciences of Ukraine, Kyiv, Ukraine

¶ Senckenberg Forschungsinstitut und Naturmuseum Frankfurt, Frankfurt am Main, Germany

# Institute of Systematics and Evolution of Animals, Polish Academy of Sciences, Kraków, Poland

¤ National Synchrotron Radiation Centre SOLARIS, Jagiellonian University, Krakow, Poland

« Institute of Physics, Jagiellonian University, Krakow, Poland

Corresponding author: Alicja Pełczyńska ( alicja.pelczynska@edu.uni.lodz.pl )

Corresponding author: Pawel Korecki ( pawel.korecki@uj.edu.pl )

Academic editor: Mónica M. Solórzano Kraemer

This is an open access article distributed under the terms of the Creative Commons Attribution License (CC BY 4.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.

Citation: Pełczyńska A, Perkovsky EE, Boudinot BE, Richter A, Krzemiński W, Kołodziej T, Korecki P, Sowa KM, Soszyńska A (2026) The rise and fall of the Eocene fly genus Kelneria (Diptera, Keroplatidae). Arthropod Systematics & Phylogeny 84: 677-703. https://doi.org/10.3897/asp.84.e197461

ZooBank: urn:lsid:zoobank.org:pub:CFB7FB9D-8247-472A-BECE-032F42368E70

Abstract

The extinct Eocene fly genus Kelneria Matile, 1979 is characterized by extreme modification of male genitalia within the family Keroplatidae. However, their morphology has remained poorly understood, as fossil material cannot be examined using traditional entomological techniques. To overcome this limitation, we applied micro-computed tomography (µCT) and synchrotron radiation micro-computed tomography (SRµCT) to reconstruct the genitalia through segmentation of scan data, enabling reassessment of the spatial organization and homology of their structures.

Our analyses indicate that Kelneria was both morphologically distinctive and well represented in Eocene amber assemblages. They also support the synonymization of the Oise amber genus Eokelneria Hebert, Ngô-Muller & Nel, 2025 with Kelneria Matile, 1979, demonstrating that the characters used to establish Eokelneria fall within the range of variation observed in Kelneria. We further demonstrate that the species-level diversity of Kelneria has been underestimated and describe three new species: Kelneria rovnensis Pełczyńska & Perkovsky sp. nov., K. erroris Pełczyńska, Krzemiński & Soszyńska sp. nov. and K. szymoni Pełczyńska, Krzemiński & Soszyńska sp. nov. Additionally, identification of representatives in Rovno and Oise ambers extends the geographic and temporal range of the genus to the early Eocene (c. 53 Ma).

Taken together, these findings show that Kelneria was an abundant, diverse, and widespread genus in the Eocene but has no modern representatives. The reasons for this extinction remain unknown; however, contrasting distribution of Eocene and extant representatives of the tribe Robsonomyiini suggests that this disappearance may reflect a broader faunal turnover in Europe after the Eocene.

Keywords

Baltic amber, Eocene–Oligocene transition, Eokelneria, fossil insects, micro-CT, new synonymy, Rovno amber, synchrotron radiation micro-CT (SRµCT)

1. Introduction

Insect genitalia evolve faster than any other body part, resulting in their extreme structural diversity (Hosken and Stockley 2004; Sinclair et al. 2013). The order Diptera provides a striking illustration of this phenomenon, as even within taxa with otherwise highly similar general morphology, genital structures may differ profoundly, forming the primary basis for species delimitation (Marshall 2012; Smith and Mayfield 2015). In some nematoceran lineages, genital modifications are so extensive that establishing structural homologies across taxa becomes challenging (Sinclair et al. 2013). An extreme example is the extinct keroplatid genus Kelneria Matile, 1979, whose highly differentiated terminalia diverge markedly from those of all other known members of the family (Matile 1990).

Kelneria belongs to the subfamily Macrocerinae, in which males typically bear well-developed gonostyli terminating in two tooth-like projections, as seen in Macrocera lutea Meigen, 1804. In striking contrast, the gonostyli of Kelneria are drastically reduced, approaching complete loss. The epandrium is likewise highly modified, bearing a pair of unique lateral processes and the aedeagus with associated membranous structures is remarkably strongly sclerotized.

Given that genital morphology is closely associated with reproductive performance and fitness across animals (Dale Broder et al. 2020), such extensive modifications likely had important functional implications. Their detailed investigation is therefore of interest not only for the purpose of species identification, but also for clarifying the structural organization and variation of these highly modified genitalia.

In extant Diptera, such analyses rely on genital dissection and KOH maceration, procedures that separate and clear sclerotized structures, thereby removing obscuring tissues and allowing individual elements to be examined, manipulated, and compared under light microscopy in different orientations. This enables accurate assessment of their articulation and spatial relationships (e.g. Porto et al. 2015; Cumming and Wood 2017).

However, all known Kelneria species are preserved exclusively as Eocene amber inclusions, precluding the application of these techniques. As a result, investigations of their genitalia have so far been restricted to light-microscopic examination of intact structural complexes, in which individual elements remain overlapped and obscured. Consequently, observations are limited to externally visible features, hindering comprehensive interpretation of their morphology. To overcome these constraints, we applied micro-CT imaging to digitally dissect and reconstruct the male terminalia, enabling examination of their spatial organization and reassessment of structural homologies.

However, it is not only the extraordinary genital morphology, unknown among extant Keroplatidae, that makes this genus noteworthy. Kelneria is the most abundant keroplatid genus in Baltic amber, but despite this apparent prevalence, only four species have been described to date: K. abundare (Meunier, 1904), K. ciliata (Meunier, 1904), K. filiformis (Meunier, 1904), and K. setosa Matile, 1979. This discrepancy between remarkable abundance and the limited number of described species suggested that the species-level diversity of the genus has been underestimated and prompted its reinvestigation.

Application of μCT and SRμCT enabled detailed reconstruction of the male terminalia and facilitated recognition of morphological differences among the examined specimens. Consequently, in the present study we describe three new Kelneria species, although additional morphotypes were recognized within the examined material. Furthermore, our observations document its occurrence not only in Baltic but also in Rovno and Oise ambers, extending both the geographic and temporal range of the genus.

2. Materials & Methods

2.1. Materials

For the purpose of this study, 682 unsorted keroplatid inclusions from Baltic amber, dated to the early Priabonian (c. 36–35 Ma; Ross et al., 2026) and Rovno amber (c. 36–35 Ma; Chemyreva et al., 2024; Eskov et al., 2026) were examined. Of these, 339 specimens were identified as Kelneria, accounting for 49.7% of all examined Keroplatidae and the vast majority of Macrocerinae. Specifically, Kelneria comprised 152 of 316 keroplatid inclusions from Rovno amber (48.1%) and 187 of 366 inclusions from Baltic amber (51.1%). Additionally, type specimens (all derived from Baltic amber) from the type series of K. abundare (GZG.BST.3081, former Königsberg catalogue no. Z1582; GZG.BST.3077 = Z3063; GZG.BST.3059 = Z620), K. ciliata (GZG.BST.3048 = Z2450), K. filiformis (GZG.BST.3072 = Z2953), and K. setosa (GZG.BST.3079 = Z2680; GZG.BST.3068 = Z1247) were examined.

2.2. Focus-stacking photography

Photographic documentation was obtained using a Canon EOS 5D Mark IV digital camera equipped with a Canon MP-E 65 mm f/2.8 1–5× macro lens. Image stacks acquired at successive focal planes were combined into extended-depth-of-field images using Helicon Focus 8. Line drawings were prepared in CORELDRAW 2018 based on the photographs, and measurements were taken using IMAGEJ (Schneider et al. 2012).

2.3. µ-CT scanning and µ-CT reconstruction

Scanning was performed using a customized version of the TOMOSCOPE XS PLUS 200 (Werth Messtechnik, Gießen, Germany). The specimen was scanned using 60 kV voltage, 120 µA current, an exposure time of 1500 ms and 2200 projections while averaging three images of every projection position. The resulting volume had a resolution of 2.46 µm^3 and after contrast adjustment in WINWERTH (Werth Messtechnik, Gießen, Germany) was exported as a 16-bit .rek file.

Data segmentation was performed in DRAGONFLY 3D WORLD version 2024.1 (Comet Technologies Canada Inc.) using a non-commercial license. To segment with a minimum of manual labelling, removing background graininess was the central problem. The fossil reconstruction procedure developed and performed here is as follows: (1) A segmentation threshold was set to highlight the darkest areas, which corresponded to the cavity of the fossil and black noise pixels; (2) a one-click segmentation was made by using “add to new”; (3) standard cleaning was done by using the 3D paintbrush to delete blocks of non-subject pixels and by using the “remove islands” function; (4) the region of interest (ROI) was inverted; (5) the “close” morphological function was used to collapse bright points together; (6) the ROI was re-inverted; (7) remaining noise was deleted using remove islands; (8) the “dilation” function was used until the dilated ROI exceeded the boundary of the amber cavity; (9) the “smooth” function was used with a kernel of 5 a few times as this produced more smoothing than 3; (10) the “erode” function was used a few times until the ROI was slightly inside of the boundary of the body cavity; (11) the ROI was inverted and used to mask following the standard export protocol (e.g., Boudinot et al. 2024).

Two alternative approaches were taken. (A) At step 10 (erosion), the ROI was over-eroded and the cavity was masked with the ROI using the maximum value (10,000). The ROI was then inverted and used to mask the background at the default value (1). (B) At step 8, instead of dilating, the ROI could be smoothed directly and maximum value cavity filling was run, with subsequent background masking. Combinations of the main workflow and alternative workflows A and B were run to explore the data. The best instance for the genitalia was found when following steps 1–7 of the main workflow was run, then manual segmentation was performed to label pixels known to be part of the genitalia but which had been removed by the less sensitive methods.

Rendering was performed in DRAGONFLY on the best model, with lighting and shadows enabled and the dataset set to maximum solidity and cubic interpolation.

2.4. Synchrotron X-ray µ-CT (SRµCT)

The X-ray µCT scans were performed at the PolyX beamline (Sowa et al. 2023) of SOLARIS National Synchrotron Radiation Centre, Kraków, Poland (Szlachetko et al. 2023). The tomographic data were measured using polychromatic X-ray beam from bending magnet (1.3T) attenuated with 0.5 mm aluminium absorber and 250 µm Be and 150 µm CVD diamond windows, resulting in the beam of 15 keV central energy. The sample was placed on a rotational stepper motor approx. 14.5 m from the source. The scans were acquired in continuous mode (2001 equiangular X-ray projections) with a single frame exposure time of 540 ms. Projections were acquired with a white-beam X-ray Microscope (Peter Optique) equipped with 10x magnification objective and LuAG:Ce 10 µm-thick scintillator. Sample-to-scintillator distance was set to 40 mm. A PCO edge 5.5 sCMOS camera with 6.5 µm pixel size was used for image acquisition. Raw images were corrected with flat and dark frames and post-processed with stripe suppression and automatic alignment procedures. Phase retrieval was performed with the Paganin method (Paganin et al. 2002) and δ/β=100. Tomographic reconstruction was performed using ASTRA TOOLBOX (van Aarle et al. 2016) with an effective voxel size 0.72 µm. 3D reconstruction was further segmented and visualized in SLICER 3D 5.10.0 (Fedorov et al. 2012).

2.5. Fourier transform infrared spectroscopy (FTIR)

Fourier transform infrared spectroscopy (FTIR) was used to obtain IR spectra of the investigated amber specimens. The analyses were performed with a Nicolet iS5 FTIR spectrometer equipped with a diamond crystal attenuated total reflectance (ATR) attachment at the Institute of Systematics and Evolution of Animals, Polish Academy of Sciences in Kraków, Poland (ISEA PAS). The obtained spectra have been archived in the institutional database, in accordance with the guidelines proposed by Zakrzewska et al. (2020).

2.6. Morphological terminology

Nomenclature used in this study primarily follows Cumming and Wood (2017), as outlined in the Manual of Afrotropical Diptera (Volume 1). Note that because the term “paramere” has been applied to virtually all genitalic structures, we also refer to it in the figures as the lateropenite, as this musculated sclerite is unique to and synapomorphic of Holometabola (Boudinot 2018). Wing vein nomenclature (Fig. 1) follows the terminology proposed by Pełczyńska et al. (2026), consistent with that adopted in our previous studies (Pełczyńska et al. 2024, 2025).

Figure 1.

Original schematic illustration of the hypothetical ground plan of the Kelneria wing, showing individual sections of the radial and medial sectors. Wing vein nomenclature follows Pełczyńska et al. (2026). Remarks: the Mb vein in Kelneria is present only in some species, where it appears as a weakly defined trace on the wing. — Abbreviations: C = costal vein; h = humeral crossvein; Sc = subcostal vein; sc–r = subcostal–radial crossvein; Rb = radiobasal vein; Rs = radial sector; R₁ = anterior branch of radius; R_2+3+4+5 = stem of the R_2+3+4+5 fork; R₂₊₃ = second branch of radius; R₄₊₅ = third branch of radius; frm = radio-medial fusion; Mb = mediobasal vein; bM₁₊₂ = basal part of M₁₊₂; dM₁₊₂ = distal part of M₁₊₂, stem of the M₁₊₂ fork; M₁ = first branch of media; M₂ = second branch of media; M₃₊₄ = fourth branch of media; m–cu = medio-cubital crossvein; Cu = cubital vein; A₁ = first anal vein.

3. Results

3.1. Systematic Palaeontology

Order Diptera Linnaeus, 1758

Infraorder Bibionomorpha Hennig, 1948

Superfamily Sciaroidea Billberg, 1820

Family Keroplatidae Rondani, 1856

Subfamily Macrocerinae Rondani, 1856

Tribe Robsonomyiini Matile, 1990

Genus † Kelneria Matile, 1979

2025 †Eokelneria Hebert, Ngô-Muller & Nel: 53, figs 1, 2 [syn. nov.]

Type species.

†Kelneria setosa Matile, 1979 by original designation.

Diagnosis.

Relatively small Macrocerinae with head bearing three closely grouped ocelli forming a triangle; antennae medium-sized, in males exceeding the length of the thorax but shorter than the wing; apical flagellomere ending in an apiculus; anepisternum bearing a few anterodorsal setae, rest of thoracic pleura bare; wings relatively short and broad; vein R1 short, not reaching the mid-length of the wing; basal cell small, not exceeding one-fourth of the wing length; radiomedial fusion and stem of the medial fork short; tibial spurs longer than the tibial diameter; pretarsus with well-developed empodium exceeding the claws; pulvilli small; male genitalia with strongly reduced gonostyli, appearing as small lobes at the outer angles of the gonocoxites; epandrium notched, bearing a pair of lateral processes.

Included species.

†Kelneria abundare (Meunier, 1904); †Kelneria ciliata (Meunier, 1904); †Kelneria filiformis (Meunier, 1904); †Kelneria setosa Matile, 1979; †Kelneria rovnensis sp. nov.; †Kelneria erroris sp. nov.; †Kelneria szymoni sp. nov.; Kelneria oisensis (Hebert, Ngô-Muller & Nel, 2025)

Remarks.

The two fossil genera, Kelneria and Eokelneria, both occurred in the Eocene of Europe and share numerous morphological similarities. The general habitus of Eokelneria closely resembles that of Kelneria. Both taxa are similar in antennae (length and presence of an apiculus on the apical flagellomere), length of tibial spurs (exceeding the width of the apical tip of the tibia), overall wing venation (Fig. 2).

Figure 2.

Comparison of wing venation in Kelneria setosa and Kelneria oisensis comb. nov. Blue dots indicate: 1 apical margin of the basal cell; 2 termination of frm; 3 fork of M₁₊₂; 4 fork of R_2+3+4+5. The shape of the basal cell is highlighted in yellow. The wing of Kelneria oisensis comb. nov. is redrawn based on the photographic documentation provided in the original description by Hebert et al. (2025). The missing portion of the wing is reconstructed and indicated by dashed lines.

The principal character used by Hebert et al. (2025) to distinguish Eokelneria from Kelneria is the presence of a row of tibial macrochaetae, described as “large setae inserted into very distinct sockets”, without specifying on which pair of legs these macrochaetae are present. However, Matile (1990) noted that in some specimens of Kelneria a row of dorsal setae, distinctly thicker and darker than the surrounding setation, extends over the apical half of the hind tibia. He further suggested that tibial macrochaetae in Macrocerinae may have evolved through a process of gradual straightening and elongation of microchaetae, as observed in some species of Kelneria from Baltic amber. A thicker and more prominent row of setae is also present on the hind tibiae of K. rovnensis sp. nov. In our interpretation, enlarged tibial setae and macrochaetae represent different degrees along a continuous spectrum of morphological variation rather than distinct character states. Consequently, we consider the distinction between them to be largely terminological. In our view, the proposed separation of Eokelneria does not reflect clear generic-level morphological differences, but rather differences in the interpretation of tibial chaetotaxy and in the distinction between macrochaetae and microchaetae. The observed variation in tibial setation is therefore considered to fall within the range of intrageneric variation in Kelneria. It should further be noted that E. oisensis, the only known representative of the genus, was described from a single specimen interpreted as a female. The generic diagnosis of Kelneria relies primarily on characters of the male terminalia. The absence of male genital characters supporting generic separation, together with the overall morphological congruence with Kelneria, does not justify the recognition of a separate genus. Consequently, Eokelneria Hebert, Ngô-Muller & Nel, 2025 is here proposed as a new junior synonym (syn. nov.) of Kelneria Matile, 1979.

Kelneria rovnensis Pełczyńska & Perkovsky, sp. nov.

https://zoobank.org/2C2D2ACF-2095-4DBF-8FC1-1AB2F7 5BB688

Figures 3, 4, 5, 6, 7

Material.

Holotype: MALE, preserved in an 11 × 10 × 4 mm piece of Rovno amber (SIZK K-32069) housed in the Schmalhausen Institute of Zoology, National Academy of Sciences of Ukraine, Kyiv, Ukraine (SIZK); (Figs 3A, 3B, 19[IR spectrum]). Syninclusions: SIZK K-32066–K-32068 (Chelonariidae, Formicidae, Nematocera, stellate hairs).

Type locality and age.

Late Eocene (Priabonian, c. 36–35 Ma), Rovno amber (Volhynian Uplift), Pugach quarry, Klesov, Rovno Oblast, Ukraine.

Diagnosis.

Antennae approx. 0.8× wing length, with distinctly separated flagellomeres densely covered with relatively long microtrichia; pedicel subequal in length to scape; palpus very large, with apical palpomere distinctly elongated, longer than subsequent one and longer than first flagellomere; Sc ending in C approx. at level of rb cell tip; m–cu joining cubital vein before level of Rs base; R₂₊₃ approx. 0.4 length of R_2+3+4+5 fork stem; epandrium deeply and broadly notched, bearing long, slender, spinulose lateral processes extending slightly beyond apices of gonopods. Female unknown.

Description.

Body (Fig. 3A): approx. 2.5 mm long; wing 1.9 mm long; antennae 1.5 mm long. — Head (Fig. 4A): wider than long; eyes large, well separated, occupying most of lateral surface of head capsule; distinct cerebral sclerite present, posteriorly rounded, with dorsal surface bearing dense, thick setae; ocelli present, lateral ocelli situated on dorsal surface of cerebral sclerite rather than at its margin; position of median ocellus obscured by antennae. — Antennae (Figs 3A, 4A): with 2+14 segments; 0.8× wing length in male, scapus annular in shape, approx. as wide as long; pedicel slightly broader but approx. equal in length to scape, globular in shape, about 1.8× wider than first flagellomere; flagellum with 12 distinctly separated segments, covered densely with relatively long microtrichia; first flagellomere approximately 3.2× as long as broad; subsequent flagellomeres progressively decreasing in length, except the terminal one, which is longer than preceding segment, terminating in a thin apiculus. — Mouthparts (Fig. 4A): palpus large, 1+4 segmented; small palpiger visible; all four maxillary palpomeres longer than broad; second and third palpomeres subequal in length; apical palpomere distinctly elongated, 2.6× longer than the preceding one and 1.5× longer than the first flagellomere; labella very large, well developed, approx. as long as the apical palpomere. — Wing (Fig. 4B, C): broad, 2.2× longer than wide, membrane hyaline without microtrichia and any visible markings; C with microtrichia throughout length; microtrichia visible on dorsal surface of radial and medial veins; C terminates at tip of wing, after end of R₄₊₅, on approx. two third of distance between end of R₄₊₅ and M₁; Sc ending in C approx. at the level of tip of rb cell; sc-r present on approx. half of the Rb length; R₁ ending in C approx. half-length of wing, just after level where M₁₊₂ forks, approx. at the level of Cu termination; R₂₊₃ approx. 0.4 the length of R_2+3+4+5 fork stem; frm ending just after level at which A₁ reaches wing margin; a faint trace of Mb is present, dividing the basal cell into two; M₁₊₂ fork stem approx. 4.8× longer than frm, ending just before level of Cu termination; M₁ approx. 3.2× longer than M₁₊₂ fork stem; m₂ cell opening 1.3× wider than opening of cell M₁; M₃₊₄ base weakened; m₃₊₄ cell opening 1.3× wider than opening of m₂ cell; basal part of M₁₊₂ distinct; m–cu distinct, joining cubital vein before level of Rs base; Cu reaching wing margin; A₁ ending on wing margin; A₂ absent. — Thorax (Fig. 4A): higher than long; scutum weakly convex, densely covered with long, thick setae; scutellum with a row of very long trichia along its margin; anepisternum narrowed dorsally, higher than long, with few setae in the dorsoanterior part; katepisternum higher than long, bare; anepimeron reaching the ventral margin of the pleura, bare; laterotergite and mediotergite bare. — Legs (Fig. 5B, C): fore coxa the longest, with long setae covering entire anterior surface; mid coxa with a several setae anteroapically and with at least two setae anteroexternally; hind coxa the shortest, with a several setae posteroexternally; femora densely and irregularly covered with thin setae; tibiae covered with short, thin, irregularly arranged setae on whole surface, additionally on posterior surface of hind tibia visible row of thicker and more robust setae; fore tibiae with a sensory pit and a single spur, more than 1.5× longer than the apical width of the tibia; mid and hind tibiae without sensory pits, bearing two long spurs, subequal in length, (spurs 1.8 and 2.9× longer apical width of tibia, respectively); claws short; empodium big, longer than claws. — Abdomen (Fig. 3A): densely covered with long setae, all eight segments visible; segment I short, segment II the longest, segment III and IV subequal in length; subsequent segments progressively decreasing in length. — Male genitalia (Figs 5A, 5D–G, 6, 7): epandrium deeply and broadly notched, bearing slender, spinulose lateral processes terminating just before apices of gonopods, several apical setae present; gonocoxites short, broad, fused, oriented parallel to each other; gonostyli strongly reduced, appearing as two small lobes at outer angles of gonocoxites; cerci small; not exceeding length of epandrial processes, with visible marginal setation; anal cone prominent; paired lateral parameres present on either side of aedeagus; phallosome broadened; dorsoventrally flattened, strongly sclerotized.

Figure 3.

Holotype of Kelneria rovnensis Pełczyńska & Perkovsky sp. nov. (specimen SIZK K-32069). A Habitus; B amber piece containing specimen.

Figure 4.

Holotype of Kelneria rovnensis Pełczyńska & Perkovsky sp. nov. (SIZK K-32069). A Close-up of the head and thorax, lateral view; B wing; C interpretative drawing of wing venation. — Abbreviations: flg I = flagellomere I; scp = scape; ped = pedicel; plp I–IV = palpomeres I–IV; la = labellum; oc = ocellus; sct = scutum; anepst = anepisternum; anepm = anepimeron; kepst = katepisternum; ltg = laterotergite; h = humeral crossvein; Sc = subcostal vein; sc–r = subcostal–radial crossvein; Rs = radial sector; R₁ = anterior branch of radius; R₂₊₃ = second branch of radius; R₄₊₅ = third branch of radius; frm = radio-medial fusion; M₁ = first branch of media; M₂ = second branch of media; M₃₊₄ = fourth branch of media; m–cu = medio-cubital crossvein; Cu = cubital vein; A₁ = first anal vein.

Figure 5.

Holotype of Kelneria rovnensis Pełczyńska & Perkovsky sp. nov. (specimen K-32069). A Male genitalia, lateral view; B fore leg; C mid and hind legs. Volumetric renders of the genitalia (SRµCT); D ventral view; E lateral view; F dorsal view; G right quarter profile. — Abbreviations: t VIII = tergite VIII; st VIII = sternite VIII; epand = epandrium; cerc = cercus; gc = gonocoxite; gs = gonostylus; tb II–III = tibiae II–III; ta I–V = tarsomeres I–V. Remarks: spurs are indicated by yellow arrows; the empodium is indicated by a blue arrow; a row of more robust setae is indicated by white arrows.

Figure 6.

Comparison of homologous structures of the male terminalia in Macrocera lutea Meigen, 1804, and Kelneria rovnensis sp. nov. The drawing of M. lutea is redrawn from Matile (1990). Remarks: The dashed line at the boundary between the gonocoxite and gonostylus indicates the poor visibility of the separating margin, most likely resulting from a partial fusion of these two structures. The arrow marks the inferred line of separation.

Figure 7.

Volumetric renders of the genitalia and terminal abdominal segments of Kelneria rovnensis Pełczyńska & Perkovsky sp. nov. (specimen K-32069; µCT). A Left lateral view; B left quarter profile; C right lateral view; D right quarter profile; E ventral view; F dorsal view.

Etymology.

The species epithet rovnensis refers to Rovno region of Ukraine, from which the holotype originates.

Remarks.

The specimen was found in the clear piece of amber with weight 4 g and size 38 × 10 × 25mm; distinction between gonostyli and gonocoxites is unclear; shape of the anal angle of the wing is hypothesized and indicated by a dashed line, as the anal field of the wing of the holotype is folded.

Kelneria erroris Pełczyńska, Krzemiński & Soszyńska, sp. nov.

https://zoobank.org/CFB7FB9D-8247-472A-BECE-032F42 368E70

Figures 8, 9, 10

Material.

Holotype: MALE, preserved in a 15 × 10 × 3 mm piece of Baltic amber (GZG.BST.03049) housed in the Geowissenschaftliches Zentrum, Universität Göttingen, Göttingen, Germany (GMUG) (Figs 8A, 8B, 19[IR spectrum]).

Figure 8.

Holotype of Kelneria erroris Pełczyńska, Krzemiński & Soszyńska sp. nov. (specimen GZG.BST.03049). A Habitus; B amber piece containing specimen.

Figure 9.

Holotype of Kelneria erroris Pełczyńska, Krzemiński & Soszyńska sp. nov. (specimen GZG.BST.03049). A Close-up of the head and thorax, lateral view; B wing; C interpretative drawing of wing venation. — Abbreviations: flg I = flagellomere I; scp = scape; ped = pedicel; plp III–IV = palpomeres III–IV; la = labellum; oc = ocellus; sct = scutum; anepst = anepisternum; anepm = anepimeron; kepst = katepisternum; ltg = laterotergite; med = mediotergite; h = humeral crossvein; Rb = radiobasal vein; Sc = subcostal vein; Rs = radial sector; R₁ = anterior branch of radius; R₂₊₃ = second branch of radius; R₃₊₄ = third branch of radius; frm = radio-medial fusion; M₁ = first branch of media; M₂ = second branch of media; M₃₊₄ = fourth branch of media; m–cu = medio-cubital crossvein; Cu = cubital vein; A₁ = first anal vein.

Figure 10.

Holotype of Kelneria erroris Pełczyńska, Krzemiński & Soszyńska sp. nov. (specimen GZG.BST.03049). A Male genitalia, lateral view; B volumetric render of the genitalia in lateral view (SRµCT); C interpretative drawing of genitalia, lateral view; D volumetric render of the genitalia in dorsal view (SRµCT); E fore leg; F mid leg; G hind leg; H apex of tarsus, mid leg. — Abbreviations: t VIII = tergite VIII; st VIII = sternite VIII; epand = epandrium; aed = aedeagus; gc = gonocoxite; gs = gonostylus; tb I–III = tibiae I–III; ta IV = tarsomere IV; claw = pretarsal claw. Remarks: spurs are indicated by yellow arrows; the empodium is indicated by a blue arrow; the colour coding of the genital structures is consistent with that used in Figure 1.

Diagnosis.

Antennae approx. 0.7× wing length, with distinctly separated flagellomeres densely covered with very short microtrichia; palpus small, with apical palpomere subequal in length to the preceding one and shorter than first flagellomere; Sc ending in C distinctly before level rb cell tip; m–cu joining cubital after level of Rs base; R₂₊₃ short, approx. 0.2 length of R_2+3+4+5 fork stem; epandrium with deep, triangular notch, bearing large, triangular lateral processes with very short, dense apical setation, terminating approx. at the level of gonopods apices. Female unknown.

Type locality and age.

Late Eocene (Priabonian, c. 36–35 Ma), Baltic amber (Sambian Peninsula), Kaliningrad Oblast, Russia.

Description.

Body (Fig. 8A): approx. 2.2 mm long; wing 1.9 mm long; antennae 1.3 mm long. — Head (Fig. 9A): wider than long; eyes large, well separated, occupying most of the lateral part of head capsule, distinct cerebral sclerite present, rounded posteriorly, projected posteriorly, dorsal surface bearing several setae approximately half the length of the setae on the scutum; three ocelli present, forming triangle, median ocellus not reduced, lateral ocelli situated on the dorsal surface of the cerebral sclerite, not on its margin. — Antennae (Fig. 9A): with 2+14 segments; 0.7× wing length in male, scapus annular in shape, wider than long; pedicel bigger than scape, as long as broad, globular in shape, about 1.7× wider than first flagellomere; flagellum with 12 distinctly separated segments, densely covered with long microtrichia; first flagellomere approximately 2.7× as long as broad; subsequent flagellomeres progressively decreasing in length, except for the terminal one, which is longer than the preceding one, terminating in a small apiculus. — Mouthparts (Fig. 9A): palpi small; three apical palpomeres visible; all longer than broad; apical and subapical palpomeres subequal in length; approx. 0.7× length of first flagellomere. — Wing (Fig. 9B, C): broad, 2.3× longer than wide, membrane hyaline without microtrichia and any visible markings; C with microtrichia throughout length; microtrichia visible on dorsal surface of all veins except transverse ones; C terminates at tip of wing, after end of R₄₊₅, on approx. two third of distance between end of R₄₊₅ and M1; Sc very short, ending in C distinctly before the level of tip of rb cell; R₁ ending in C before half-length of wing, approx. at level of M₁₊₂ fork, just before the level of Cu termination; R₂₊₃ approx. 0.2 the length of R_2+3+4+5 fork stem; frm ending before level at which A₁ reaches wing margin; a faint trace of Mb is present, dividing the basal cell into two; M₁₊₂ fork stem approx. 4.5× longer than frm, ending just before level of Cu termination; M₁ approx. 3.1× longer than M₁₊₂ fork stem; m₂ cell opening 1.3× wider than opening of m₁ cell; M₃₊₄ base weakened and widely interrupted; m₃₊₄ cell opening 0.9× opening of m₂ cell; basal part of M₁₊₂ distinct; m–cu distinct, joining cubital vein after level of Rs base; Cu reaching wing margin; A₁ termination not visible, as wing is folded; A₂ absent. — Thorax (Fig. 9A): higher than long; scutum weakly convex, densely covered with long setae arranged in two dorsocentral and two acrostichal rows; scutellum with a row of very long trichia along its margin; anepisternum narrowed dorsally, higher than long, with at least three setae on the dorsoanterior part; katepisternum higher than long, bare; anepimeron reaching the ventral margin of the pleura, bare; laterotergite and mediotergite bare. — Legs (Figs 8A, 10B–D): fore coxa the longest, sparsely setulose on anterior surface; hind coxa the shortest, with a several setae posteroexternally; femora irregularly covered with thin setae; tibiae covered with short, thin, irregularly arranged setae on whole surface; fore tibiae with a sensory pit and a single spur, 1.6× longer than the apical width of the tibia; mid and hind tibiae without sensory pits, bearing two long spurs, subequal in length, (approx. 2.4× longer apical width of tibia); claws short; empodium big, longer than claws. — Abdomen (Fig. 8A): covered with long, thin setae, all eight segments visible; segment I short, segment II the longest, subsequent segments progressively decreasing in length. — Male genitalia (Figs 10A–D): epandrium with deep, triangular notch; bearing large, triangular lateral processes with very short, dense apical setation, terminating approx. at level of gonopods apices; gonocoxites short, broad, triangular, partially fused, oriented parallel to each other; gonostyli strongly reduced, appearing as two small lobes at outer angles of gonocoxites; cerci not visible; phallosome bilobed, strongly sclerotized.

Etymology.

The species epithet erroris refers to a taxonomic error whereby the holotype of this species was previously included in the type series of K. abundare.

Remarks.

Crossvein sc–r is not illustrated in the wing drawing, as its likely positioned close to the wing base and cannot be discerned in the holotype; shape of the anal angle of the wing and apical tip of vein A₁ is hypothesized and indicated by a dashed line, as anal field of the wing of the holotype is folded; specimen designated herein as the holotype originates from the type series of K. abundare as defined by Meunier (1904); this misinterpretation was subsequently retained in the generic revision by Matile (1979) and is corrected in the present study.

Kelneria szymoni Pełczyńska, Krzemiński & Soszyńska, sp. nov.

https://zoobank.org/FF4198A6-B02F-49BB-B3F0-F20968 D3E535

Figures 11, 12, 13, 14

Material.

Holotype: MALE, preserved in a 29 × 19 × 5 mm piece of Baltic amber (specimen MP/5375a). Paratype: FEMALE, preserved in the same piece of amber (specimen MP/5375b) deposited in the Institute of Systematics and Evolution of Animals, Polish Academy of Sciences, Kraków, Poland (ISEA PAS) (Figs 19A, 19C, 19[IR spectrum]).

Diagnosis.

Antennae approx. 0.7× wing length in male, 0.5× wing length in male, with distinctly separated flagellomeres densely covered with very short microtrichia; palpus small, with apical palpomere longer than the preceding one and shorter than first flagellomere; Sc ending in C distinctly before level rb cell tip; m–cu joining cubital before level of Rs base; R₂₊₃, approx. 0.4 length of R_2+3+4+5 fork stem; epandrium with deep, triangular notch, bearing long and narrow triangular processes, terminating approx. at the level of gonopods apices.

Type locality and age.

Late Eocene (Priabonian, c. 36–35 Ma), Baltic amber, exact locality unknown.

Description.

Body (Fig. 11A): male approx. 2.6 mm long; wing 2.2 mm long; antennae approx. 1.6 mm long; female approx. 3.3 mm long; wing 2.6 mm long; antennae approx. 1.4 mm long. — Head (Fig. 13A): wider than long; eyes large, well separated, occupying most of lateral part of head capsule; distinct cerebral sclerite present, rounded posteriorly, projected posteriorly, dorsal surface bearing several setae approximately half length of setae on scutum; three ocelli present, closely grouped, forming triangle, median ocellus not reduced. — Antennae (Figs 11A, 13A): robust, with 2+14 segments; 0.7× wing length in male, 0.5× wing length in female; scapus short, annular in shape, wider than long; pedicel bigger than scape, as long as broad, globular in shape, approx. 2.1× wider than first flagellomere; flagellum with 12 distinctly separated segments, covered densely with very short microtrichia; first flagellomere approx. 3.4× long as broad, subsequent flagellomeres progressively decreasing in length, except terminal one, which is longer than subsequent one, ending with small, weakly developed apiculus. — Mouthparts (Fig. 13A): palpi small; two apical palpomeres visible; longer than broad; apical palpomere distinctly longer than subapical one; approx. 0.6× length of first flagellomere. — Wing of male holotype (Fig. 12A, C): broad, 2.1× longer than wide; membrane hyaline without microtrichia and any visible markings; C with microtrichia throughout its length; microtrichia visible on dorsal surface of all veins, except for transverse veins and Rs; C terminates at the tip of the wing, beyond the end of R₄₊₅, at approximately three quarters of the distance between the end of R₄₊₅ and M₁; h close to the wing base; Sc short, ending in C distinctly before the level of the tip of the rᵦ cell; sc–r present, shortly after the level of h; R₁ ending in C shortly before half-length of the wing, distinctly before the level at which R_2+3+4+5 forks, approximately at the level of Cu termination; R_2+3+4+5 forks distal to the level at which M₁₊₂ forks, beyond half the width of the m₃₊₄ cell; R₂₊₃ approximately 0.4× the length of the R_2+3+4+5 fork stem; frm ending slightly distal to the level at which A₁ is expected to reach the wing margin; a faint trace of Mᵦ present, dividing the basal cell into two; M₁₊₂ fork stem approximately 4.4× longer than frm, ending shortly after the level of Cu termination; M₁ approximately 2.9× longer than the M₁₊₂ fork stem; m₂ cell 1.3× wider than m₁ cell; M₃₊₄ base weakened and partially atrophied; m₃₊₄ cell 1.1× wider than m₂ cell; basal part of M₁₊₂ and m–cu distinct; Cu reaching the wing margin; A₁ ending on the wing margin; A₂ absent; anal angle folded, therefore its shape not discernible. — Wing of female paratype (Fig. 12B, D): broad, 2× longer than wide; membrane hyaline, without microtrichia or visible markings; C with microtrichia throughout its length; microtrichia present on the dorsal surface of all veins except the transverse veins and Rs; C terminates at the wing tip, beyond the end of R₄₊₅, at approximately three quarters of the distance between the end of R₄₊₅ and M₁; h close to the wing base; Sc short, ending in C distinctly before the level of the tip of the rᵦ cell; sc–r present, shortly after the level of h; R₁ ending in C at approximately half the wing length, distinctly before the level at which R_2+3+4+5 forks, shortly after the level of Cu termination; R_2+3+4+5 forks distal to the level at which M₁₊₂ forks, at approximately half the width of the m₃₊₄ cell; R₂₊₃ approximately 0.4× the length of the R_2+3+4+5 fork stem; frm ending distal to the level at which A₁ reaches the wing margin; a faint trace of Mᵦ present, dividing the basal cell into two; M₁₊₂ fork stem approximately 3.9× longer than frm, ending at approximately the level of Cu termination; M₁ approximately 2.8× longer than the M₁₊₂ fork stem; m₂ cell 1.3× wider than m₁ cell; base of M₃₊₄ weakened and partially atrophied; m₃₊₄ cell 1.1× wider than m₁ cell; basal part of M₁₊₂ and m–cu distinct; Cu reaching the wing margin; A₁ ending on the wing margin; A₂ absent; anal angle rounded. — Thorax (Fig. 13A): higher than long; scutum weakly convex, covered with dense, long setation arranged in two dorsocentral and two acrostichal rows; scutellum ovoid in lateral view, with a row of very long trichia along its margin; anepisternum triangular in shape, approximately as long as high, bare; katepisternum rectangular, higher than long, bare; anepimeron small, not reaching ventral margin of pleura, bare; laterotergite and mediotergite bare. — Legs (Figs 11A, 13B–D): fore coxa the longest, covered densely with long setae on entire anterior surface; mid coxa with a few setae anteroapically; hind coxa the shortest, with a row of setae anteroexternally; femora irregularly covered with long, thin setae; tibiae covered with long, irregularly arranged setae; fore tibiae with a sensory pit and a single spur, longer than apical width of tibia (approx. 1.1× longer apical width of tibia); mid and hind tibiae without sensory pits, bearing two long spurs, subequal in length (spurs 2.1 and 2.2× longer apical width of tibia, respectively); claws short; empodium large, longer than claws. — Abdomen (Fig. 11A): covered with long, thin setae, all eight segments visible; segment I short, segment II the longest, subsequent segments progressively decreasing in length. — Male genitalia (Fig. 14A–C, E, F, H, I): epandrium deeply notched, bearing long, narrow, triangular processes reaching apical tips of gonocoxopodites; gonocoxites large, triangular, partially fused, positioned parallel to each other; gonostyles very small, reduced to two lobes at external angle of gonocoxites; cerci ending approx. at the level of epandrial processes tips, densely covered with setae; phallosome strongly sclerotized. — Female genitalia (Fig. 14D, G): tergite X well developed; cerci biarticulate, first segment short, second elongate (nearly 3× longer than subsequent one); sternite VIII divided into two.

Figure 11.

Holotype (male, specimen MP/5375a) and paratype (female, specimen MP/5375b) of Kelneria szymoni Pełczyńska, Krzemiński & Soszyńska sp. nov. A Habituses; B amber piece containing specimens.

Figure 12.

Holotype (male, specimen MP/5375a) and paratype (female specimen, MP/5375b) of Kelneria szymoni Pełczyńska, Krzemiński & Soszyńska sp. nov. A: Male wing; B female wing, C interpretative drawing of male wing venation; D interpretative drawing of female wing venation. — Abbreviations: h = humeral crossvein; Sc = subcostal vein; sc–r = subcostal–radial crossvein; Rs = radial sector; R₁ = anterior branch of radius; R₂₊₃ = second branch of radius; R₄₊₅ = third branch of radius; frm = radio-medial fusion; M₁ = first branch of media; M₂ = second branch of media; M₃₊₄ = fourth branch of media; m–cu = medio-cubital crossvein; Cu = cubital vein; A₁ = first anal vein.

Figure 13.

Holotype (male, specimen MP/5375a) of Kelneria szymoni Pełczyńska, Krzemiński & Soszyńska sp. nov. A Close-up of the head and thorax, lateral view; B fore leg; C hind leg; D tip of the tarsus, hind leg; female. — Remarks: spurs indicated by yellow arrows; empodium indicated by blue arrow. — Abbreviations: flg I = flagellomere I; scp = scape; ped = pedicel; plp = palpus; la = labellum; oc = ocellus; sct = scutum; sctl = scutellum; anepst = anepisternum; anepm = anepimeron; kepst = katepisternum; ltg = laterotergite; med = mediotergite; t VIII = tergite VIII; t X = tergite X; st VIII = sternite VIII; epand = epandrium; aed = aedeagus; gc = gonocoxite; gs = gonostylus; cerc I–II = cerci I–II; tb I–II = tibiae I–II; ta IV = tarsomere IV; claw = pretarsal claw.

Figure 14.

Holotype (male, specimen MP/5375a) and paratype (male, specimen MP/5375b) of Kelneria szymoni Pełczyńska, Krzemiński & Soszyńska sp. nov. A Male genitalia, lateral view; B volumetric render of male genitalia in lateral view (SRµCT); C drawing of male genitalia, lateral view; D female genitalia; E volumetric render of male genitalia in dorsal view (SRµCT); F drawing of male genitalia, dorsal view; G drawing of female genitalia, lateral view; H volumetric render of male genitalia in ventral view (SRµCT); I drawing of male genitalia, ventral view. — Abbreviations: t VIII = tergite VIII; t X = tergite X; st VIII = sternite VIII; epand = epandrium; aed = aedeagus; gc = gonocoxite; gs = gonostylus; cerc I–II = cerci I–II. Remarks: the colour coding of the genital structures is consistent with that used in Figure 6.

Etymology.

The species epithet szymoni is derived from the name Szymon and is honoring Szymon Kaczmarek (University of Lodz, Poland), who provided the holotype specimen.

Remarks.

The two specimens are interpreted as conspecific because the position of the male relative to the female strongly suggests that they were caught in resin either while in copula (and subsequently separated by a flow of resin), during the initiation of copulation, or immediately after copulation. Sexual dimorphism is evident in overall body proportions, with the female being larger and more robust, the antennae of the female are markedly shorter (approx. 0.5× wing length), than those of the male (approx. 0.7× wing length), slight differences are also observable in wing venation R1 terminates slightly before the mid-length of the wing in the male (ending approximately at the level of Cu termination), whereas in the female it reaches approximately the mid-length of the wing (ending distinctly distal to the level of Cu termination), as shown in Fig. 12C, D.

3.2. Morphology

Micro-CT reconstruction of the male terminalia of K. rovnensis sp. nov. (Fig. 15) provides the first three-dimensional insight into the genital morphology of the genus and largely confirms the interpretation proposed by Matile (1979, 1990), while additionally revealing structures that were not discernible under light microscopy. In particular, paired lateropenites are clearly visible in the reconstructed model.

Figure 15.

Volumetric renders of the genitalia and terminal abdominal segments of K. rovnensis sp. nov. (specimen K-32069) (µCT): A Posteroventral oblique view; B posterior view.

3.3. Identification key to the males of Kelneria Matile 1979

1	Epandrial notch broad and rounded, with lateral processes arising along the lateral margins of the epandrium; processes slender, spinulose in lateral view (Fig. 16A, B)	2
1’	Epandrial notch deep and triangular; lateral processes appearing triangular in lateral view (Fig. 16C, D)	4
2	Antennal flagellomeres indistinctly separated, flagellum filiform, setation of flagellomeres very short (Fig. 17A)	*K. filiformis*
2’	Antennal flagellomeres distinctly separated, flagellomeres covered with long setation (Fig. 17B)	3
3	Maxillary palpi large, apical palpomere longer than first flagellomere (Fig. 18A)	K. rovnensis sp. nov.
3’	Maxillary palpi small, apical palpomere shorter than first flagellomere (Fig. 18B)	*K. ciliata*
4	Epandrial processes narrow, distinctly incised along the ventral margin of the epandrium, reaching gonopodal apices (Fig. 15C)	K. szymoni sp. nov.
4’	Epandrial processes broad, ventral margin not distinctly incised (Fig. 15D, E)	5
5	Gonocoxite about as long as wide (Fig. 15D)	*K. abundare*
5’	Gonocoxite distinctly longer than wide (Fig. 15E)	K. erroris sp. nov.

3.4. FTIR spectra

The FTIR spectra of the examined specimens display a doublet of peaks in the 1260–1160 cm^–1 range, with the strongest absorption peak at approximately 1150 cm^–1 (Fig. 19). This feature is known as the “Baltic shoulder”, a diagnostic spectral signature associated with the presence of succinic acid (Drąg et al. 2022; Wolfe et al. 2016).

Figure 16.

Male genitalia of Kelneria species: A Dorsal view; B–E lateral views. — Remarks: the colour coding of the genital structures is consistent with that used in Figure 6.

Figure 17.

Comparison of Kelneria species, heads in lateral view with close-ups of the flagellomeres.

Figure 18.

Comparison of Kelneria species. A, B Heads in lateral view, photographs, with lengths of the first flagellomere and apical palpomere indicated; C, D corresponding drawings. — Abbreviations: flg I, flagellomere I; palp, palpomere.

Figure 19.

Fourier transform infrared spectroscopy (attenuated total reflectance) spectra obtained from investigated amber specimens; the Baltic shoulder wavelength of occurrence marked.

4. Discussion

4.1. Abundance of Kelneria and disappearance of the Robsonomyiini

Kelneria is remarkably abundant in Baltic and Rovno ambers; in contrast, the genus is known from a single specimen in Oise amber. However, the implications of this striking disparity in relative abundance among those deposits are difficult to assess. Baltic and Rovno ambers originated from amber forests distributed across broad geographic regions, effectively capturing multiple microhabitats and reflecting large-scale palaeoenvironments (Perkovsky et al. 2007; Sadowski et al. 2017). In contrast, Oise amber derives from a comparatively more localized deposit of different botanical origin, being associated with angiosperm resin production (Nel et al. 2004; Brasero et al. 2009). As a result, Oise amber likely samples a narrower and ecologically distinct forest habitat. Therefore, it remains unclear whether the apparent rarity of Kelneria in Oise amber reflects its genuine scarcity in Europe at that time, which can potentially be related to the warmer climate of the early Eocene (Thompson et al. 2025), represents an earlier stage of the genus diversification, or just its low representation within the specific habitat captured by this deposit. More generally, even the high frequency of Kelneria in Baltic and Rovno ambers does not necessarily reflect its true abundance in the original forest ecosystems. Resin is a selective trap; consequently, amber does not faithfully record the overall composition of arthropod communities but rather preferentially preserves organisms inhabiting the trunk and the immediate surroundings of the resin-producing tree (Solórzano-Kraemer et al. 2018). Furthermore, taphonomic bias in the case of fossil resin results from numerous factors, including behavioral and ecological ones, like the tendency to hide in bark crevices or the presence of particular mating behaviours. Daily and seasonal activity (Vilhelmsen et al. 2024) may have also played an important role, since the viscosity of the resin and the rate of its flow are influenced by temperature and humidity in the air. Moreover, the characteristic scent of resin may have served as a repellent for some dipterans, while simultaneously serving as an attractant for others (Krzemińska et al. 1993). The number of these factors means that this genus did not necessarily have to be dominant but simply fell victim to sticky resin more frequently. Nevertheless, given its overwhelming representation, accounting for nearly half of all keroplatid inclusions, it appears unlikely that this pattern can be explained solely by preservation bias.

Our observations of Kelneria in Baltic, Rovno and Oise ambers indicate that the genus dates back at least to the earliest Eocene (c. 55–53 Ma) and was widespread and common in Eocene Europe (Brasero et al. 2009). This corresponds with the high diversity of the tribe Robsonomyiini on the continent during that epoch. Today, however, the tribe, represented by five extant genera, Calusamyia Coher, 2011, Langkawiana Ševčík, 2009, Micrepimera Matile, 1990, Robsonomyia Matile & Vockeroth, 1980, and Srilankana Matile, 1990 is absent from the modern European fauna. Its distribution is now restricted primarily to Asia, with additional representatives occurring in Madagascar and North America (Fig. 20).

Figure 20.

Distribution of extant and fossil representatives of the tribe Robsonomyiini (yellow = extant; red = fossil). Numbers indicate different Eocene amber deposits: 1 Oise amber; 2 Baltic amber; 3 Rovno amber. Map created using the SimpleMappr online generator (simplemappr.net) and modified in CorelDRAW 2018 (coreldraw.com/en/product/coreldraw).

Comparable disjunct distributional patterns have been documented in other insect groups known from Eocene ambers. For example, in Coleoptera, the genus Dysanabatium was abundant in Eocene Europe but is today restricted to Southeast Asia (Bogri et al. 2018). In Hymenoptera, the family Scolebythidae, also recorded from Baltic amber, is currently distributed across Africa, Australia, the Neotropics and Southeast Asia, but is absent from Europe (Perkovsky and Rasnitsyn 2013). Among the ants (Formicidae), there are numerous examples including the entire genera that no longer occur in Europe (Oecophylla, Prenolepis, Gesomyrmex, Carebara, Gnamptogenys, Nylanderia, Tetraponera, among others; e.g., Dlussky & Rasnitsyn, 2009). Similarly, within Mecoptera, the genus Panorpodes is today restricted to Eastern Asia and North America, yet four species have been described from Baltic amber (Soszyńska-Maj and Krzemiński 2015). This corresponds well with the observation that the closest modern analogues of the Baltic amber forest are, generally, most likely warm-temperate forests of East Asia and North America (Sadowski et al. 2017).

Notably, more species of Robsonomyiini are known from Baltic amber alone than from all extant zoogeographic regions combined (Table 1). It seems that already in the Eocene the genus strongly preferred microthermal conditions: a potential second specimen from Oise amber has still not been found (A. Nel, pers. com., 2026). Extant species are known by so few specimens, that in any five kilograms of succinite with inclusions we could find more Robsonomyiini specimens than in all the world's collections of extant dipterans together (our unpublished data).

Table 1.

Download as

CSV

XLSX

Species distribution of the extant and fossil representatives of the tribe Robsonomyiini. Abbreviations: Af, Afrotropical; Pa, Palearctic; Ne, Nearctic; Or, Oriental.

Genus	Extant distribution				Fossil record
Genus	Af	Pa	Ne	Or	Baltic amber	Rovno amber	Oise amber
Kelneria	—	—	—	—	6	1	1
Calusamyia	—	—	1	—	—	—	—
Langkawiana	—	—	—	1	—	—	—
Micrepimera	1	—	—	2	2	—	—
Robsonomyia	—	1	1	—	2	—	—
Srilankana	—	—	—	1	—	—	—
Total:	1	1	2	4	10	1	1

This pattern suggests that the tribe may have reached its peak diversity in the European amber forests. Consequently, the post-Eocene climatic transition, which led to the replacement of those warm, mixed evergreen forests by more temperate and increasingly seasonal vegetation in Europe, may have contributed to their disappearance from the region (Utescher et al. 2021; Lyubarsky et al. 2023; Wu et al. 2024).

However, unlike Robsonomyia and Micrepimera, which persisted outside the continent, Kelneria is entirely absent from the extant fauna. This may reflect a more geographically restricted distribution of the genus, potentially lacking access to southeastern climatic refugia during the late Eocene cooling (Hutchinson et al. 2021).

4.2. Survival hypothesis

Data from the Paleobiology Database (PBDB 2025) show that a total of 545 dipteran genera are known from the Eocene of Europe, of which as many as 241 are regarded as extinct (approx. 44%). Because many fossil insect species were described before modern concepts to separate genera were established, it is possible that many may require reassignment to possibly extinct genera. If this expectation were borne out, the real scale of their extinction may be even higher. Yet, the possibility that Kelneria survived into the present day was previously hypothesised by Matile (1981). Although this may appear unlikely, it cannot be excluded, as demonstrated by the keroplatid genus Palaeoplatyura Meunier, 1899, first described from Baltic amber and later found in the extant Holarctic fauna. Such cases, however, did not concern only single genera, but even entire dipteran families, as exemplified by the Tanyderidae (Lukashevich and Krzemiński 2009). At the same time molecular data from North America obtained from BOLD Systems reveals a higher number of BINs than described species, suggesting that the extant diversity of the tribe remains only partially documented. Thus, Kelneria itself, or more plausibly its close relatives, may yet be discovered in modern fauna.

4.3. Species comparison

Although the general habitus, including body proportions, and the wing venation pattern are relatively uniform in Kelneria, pronounced interspecific variation is observed in the male genitalia, which remain the primary basis for species-level identification within the genus. Although the gonostyli are strongly reduced in all species and appear as small lobes at the outer angles of the gonocoxites, the remaining genital structures differ markedly. The biggest variation concerns the shape of the epandrium, particularly the shape of the epandrial notch. This notch may be broad, shallow and rounded, as in K. rovnensis sp. nov. (Fig. 15A), K. ciliata, K. filiformis and K. setosa, or deep and distinctly triangular, as in K. abundare, K. erroris sp. nov., and K. szymoni sp. nov. Consequently, the shape of the epandrial processes also varies. These may be slender and spinulose, short as in K. rovnensis sp. nov., or more elongate as in K. setosa, or triangular, broad as in K. abundare and K. erroris sp. nov., or narrow and distinctly incised as in K. szymoni sp. nov.

In addition to genital morphology, important diagnostic information is provided by the mouthparts and antennae. Regarding morphology of the mouthparts, among the currently recognised species, Kelneria rovnensis sp. nov., described from Rovno amber, is the most distinctive. This species differs from other representatives of the genus in possessing large and well-developed palpi (Fig. 5A). However, given the limited understanding of the adult biology of Macrocerinae, the ecological significance of this modification remains speculative. Antennae in the genus Kelneria differ in the thickness of the flagellum, the distinctness of separation between individual flagellomeres, and their setation. All species described herein possess distinctly separated flagellomeres with relatively long setation and thus differ from the previously described species K. filiformis, in which the flagellum is filiform and the setation very short.

5. Concluding remarks

The evolutionary history of Kelneria presents a striking combination of morphological distinctiveness, numerical abundance, and ultimate disappearance. During the Eocene, the genus was not only geographically widespread, occurring in Oise, Baltic, and Rovno ambers, but also numerically dominant within Keroplatidae. Such abundance in Baltic and Rovno amber inclusions, even when accounting for taphonomic bias, suggests that Kelneria constituted a substantial component of forest-associated dipteran communities. Its disappearance, therefore, cannot be interpreted as the loss of a rare or ecologically insignificant taxon. Kelneria appears to represent a lineage that flourished in the equable microthermal climate of warm-temperate, humid amber forests but possibly failed to persist under middle-latitude climate with much colder winters following the Eocene–Oligocene Transition (EOT) and the restructuring of European forest ecosystems (Hutchinson et al. 2021; Utescher et al. 2021; Wu et al. 2024). Further, the case of Kelneria highlights the value of integrating modern imaging techniques with classical taxonomy. The genus exhibits one of the most extreme modifications of male terminalia within Keroplatidae, and without micro-CT reconstruction its structural organisation would have remained only partially resolved. By enabling three-dimensional assessment of complex genital structures, micro-CT has refined our interpretation of structural homologies.

Taken together, these findings demonstrate that Kelneria was a distinctive, abundant, diverse, and geographically widespread Eocene genus with no known extant representatives. The reasons for its disappearance remain unknown. However, the distributional contrast between Eocene and extant representatives of the tribe Robsonomyiini suggests that this loss may have been part of a broader faunal turnover in Europe after the Eocene.

6. Declarations

Authors’ contributions. A.P. took the lead in writing the manuscript and was responsible for material preparation, photography, and graphic illustrations. A.P. and W.K. were responsible for taxonomic decisions. B.B. and A.R. performed the laboratory micro-CT scanning, prepared the 3D reconstructions and conducted the morphological analyses. T.K. performed the synchrotron radiation micro-computed tomography (SRµCT) measurements. K.M.S. and P.K. designed and implemented the synchrotron imaging methodology and performed the 3D reconstructions. E.P. provided the examined specimen and assisted in the interpretation of the results. A.S. and W.K. contributed to the research concept and design, provided access to material, supervised the project, and secured funding. All authors critically revised the manuscript and approved the final version.

Funding. This research was funded by the National Science Center, Poland (grant no. 2020/37/B/NZ8/03042). We also thank HMWK (Hessian Ministry of Science and Arts) through the IWB-EFRE program, project number: 20009100, and SOSA (Senckenberg Ocean Species Alliance) for financing the Werth micro-CT scanner under the project title “3D-Forschung mittels hochauflösender µCT für den digitalen Zwilling von Objekten”. Research at the National Synchrotron Radiation Centre SOLARIS is supported by the Ministry of Science and Higher Education, Poland, under contract no. 1/SOL/2021/2. Evgeny Perkovsky was supported by Scholars at Risk Ukraine (SARU) program for 2026 funded by Orient’s Fond.

Conflict of interests. The authors declare that they have no conflict of interests

Use of AI. The authors acknowledge the use of ChatGPT (OpenAI, GPT-5.3; accessed April 2026) for stylistic and language improvement of the manuscript.

7. Acknowledgments

The authors thank Katarzyna Kopeć (Institute of Systematics and Evolution of Animals, Polish Academy of Sciences) for performing the spectroscopy analyses.

8. References

Aarle W van, Palenstijn WJ, Cant J, Janssens E, Bleichrodt F, Dabravolski A, Beenhouwer, JD, Batenburg, KJ Sijbers J (2016) Fast and flexible X-ray tomography using the ASTRA toolbox. Optics Express 24: 25129–25147, https://doi.org/10.1364/OE.24.025129

Bogri A, Solodovnikov A, Żyła D (2018) Baltic amber impact on historical biogeography and palaeoclimate research: oriental rove beetle Dysanabatium found in the Eocene of Europe (Coleoptera, Staphylinidae, Paederinae). Papers in Palaeontology 4: 433–452. https://doi.org/10.1002/spp2.1113

Boudinot BE (2018) A general theory of genital homologies for the Hexapoda (Pancrustacea) derived from skeletomuscular correspondences, with emphasis on the Endopterygota. Arthropod Structure & Development 47: 563–613. https://doi.org/10.1016/j.asd.2018.11.001

Boudinot BE, van de Kamp T, Peters P, Knöllinger K (2024) Male genitalia, hierarchical homology, and the anatomy of the bullet ant (Paraponera clavata; Hymenoptera, Formicidae). Journal of Morphology 285: e21757. https://doi.org/10.1002/jmor.21757

Brasero N, Nel A, Michez D (2009) Insects from the Early Eocene amber of Oise (France): diversity and palaeontological significance. Denisia 26: 12.

Broder ED, Ghalambor CK, Handelsman CA, Ruell EW, Reznick DN, Angeloni LM (2020) Rapid evolution and plasticity of genitalia. Journal of Evolutionary Biology 33: 1361–1370. https://doi.org/10.1111/jeb.13700

Chemyreva VG, Legalov AA, Perkovsky EE (2024) A new genus of Ambositrinae (Hymenoptera, Diapriidae) from Rovno amber and remarks on the Eocene distribution of the subfamily. Ecologica Montenegrina 79: 104–112. https://doi.org/10.37828/em.2024.79.9

Cumming JM, Wood DM (2017) Adult morphology and terminology. In: Kirk-Spriggs AH, Sinclair BJ (Eds) Manual of Afrotropical Diptera. Volume 1. Suricata 4. SANBI Publications, Pretoria, 89–133.

Dlussky GM, Rasnitsyn AP (2009) Ants (Insecta: Vespida: Formicidae) in the upper Eocene amber of central and eastern Europe. Paleontological Journal 43: 1024–1042. https://doi.org/10.1134/S0031030109090056

Drąg K, Mroczkowska-Szerszeń M, Magdalena D-S, Żeliszewska G (2022) Identification of treated Baltic amber by FTIR and FT-Raman – A feasibility study. Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 279: 121404. https://doi.org/10.1016/j.saa.2022.121404

Dale Broder E, Ghalambor CK, Handelsman CA, Ruell EW, Reznick DN, Angeloni LM (2020) Rapid evolution and plasticity of genitalia. Journal of Evolutionary Biology 33: 1361–1370. https://doi.org/10.1111/jeb.13700

Eskov KY, Vorontsov DD, Marusik YM, Perkovsky EE (2026) First pholcommatine cobweb spider from Eocene Rovno amber, with a delimitation of the subfamily Pholcommatinae (Araneae: Theridiidae). Palaeoentomology 9: 45–55. https://doi.org/10.11646/palaeoentomology.9.1.6

Fedorov A, Beichel R, Kalpathy-Cramer J, Finet J, Fillion-Robin J-C, Pujol S, Bauer C, Jennings D, Fennessy F, Sonka M, Buatti J, Aylward S, Miller JV, Pieper S, Kikinis R (2012) 3D Slicer as an image computing platform for the Quantitative Imaging Network. Magnetic Resonance Imaging 30: 1323–1341. https://doi.org/10.1016/j.mri.2012.05.001

Hebert F, Muller VN, Nel A (2025) First Keroplatidae Robsonomyiini (Diptera: Sciaroidea) from the lowermost Eocene Oise amber (France). Zootaxa 5739: 53–59. https://doi.org/10.11646/zootaxa.5739.1.9

Hosken DJ, Stockley P (2004) Sexual selection and genital evolution. Trends in Ecology & Evolution 19: 87–93. https://doi.org/10.1016/j.tree.2003.11.012

Hutchinson DK, Coxall HK, Lunt DJ, Steinthorsdottir M, de Boer AM, Baatsen M, von der Heydt A, Huber M, Kennedy-Asser AT, Kunzmann L, Ladant J-B, Lear CH, Moraweck K, Pearson PN, Piga E, Pound MJ, Salzmann U, Scher HD, Sijp WP, Śliwińska KK, Wilson PA, Zhang Z (2021) The Eocene–Oligocene transition: a review of marine and terrestrial proxy data, models and model–data comparisons. Climate of the Past 17: 269–315. https://doi.org/10.5194/cp-17-269-2021

Krzemińska E, Krzemiński W, Haenni J-P, Dufour R (1993) W bursztynowej pułapce. Muzeum Przyrodnicze Instytutu Systematyki i Ewolucji Zwierząt PAN, Kraków, 141 pp.

Lukashevich ED, Krzemiński W (2009) New Jurassic Tanyderidae (Diptera) from Asia with first find of larvae. Zoosymposia 3: 155–172. https://doi.org/10.11646/zoosymposia.3.1.13

Lyubarsky GY, Perkovsky EE, Vasilenko DV (2023) Unexpected diversity of Xenoscelinae in Priabonian European amber: The third xenosceline species from Rovno amber. Life 13: 636. https://doi.org/10.3390/life13030636

Marshall SA (2012) Flies: The natural history and diversity of Diptera. Firefly Books, Richmond Hill, 616 pp.

Matile L (1979) Un nouveau genre de Keroplatidae de l’ambre oligocène de la Baltique (Diptera: Mycetophilidae). Revue Française d’Entomologie 1: 36–41.

Matile L (1981) Description d’un Keroplatidae du Crétacé moyen et données morphologiques et taxinomiques sur les Mycetophiloidea (Diptera). Annales de la Société entomologique de France (N.S.) 17: 99–123. https://doi.org/10.1080/21686351.1981.12278265

Matile L (1990) Recherches sur la systématique et l’évolution des Keroplatidae (Diptera, Mycetophiloidea). Mémoires du Muséum national d’Histoire naturelle, Série A, Zoologie 148: 1–682.

Matile L, Vockeroth JR (1980) Description d’un genre nouveau de Keroplatidae de l’ouest Nord-Américain (Diptera: Mycetophiloidea). The Canadian Entomologist 112: 545–548. https://doi.org/10.4039/Ent112545-6

Nel A, De Ploëg G, Millet J, Menier J-J, Waller A (2004) The French ambers: a general conspectus and the lowermost Eocene amber deposit of Le Quesnoy in the Paris Basin. Geologica Acta 2: 3–8. https://doi.org/10.1344/105.000001628

Paganin D, Mayo SC, Gureyev TE, Miller PR, Wilkins SW (2002) Simultaneous phase and amplitude extraction from a single defocused image of a homogeneous object. Journal of Microscopy 206: 33–40. https://doi.org/10.1046/j.1365-2818.2002.01010.x

Pełczyńska A, Krzemiński W, Blagoderov V, Vilhelmsen L, Soszyńska A (2024) Eocene amber provides the first fossil record and bridges distributional gap in the rare genus Robsonomyia (Diptera: Keroplatidae). Scientific Reports 14: 9252. https://doi.org/10.1038/s41598-024-59448-y

Pełczyńska A, Blagoderov V, Krzemiński W, Soszyńska A (2025) Long story of long antennae: Mesozoic origins of the subfamily Macrocerinae (Diptera: Keroplatidae). Acta Geologica Sinica (English Edition) 99: 1225–1248. https://doi.org/10.1111/1755-6724.15345

Pełczyńska A, Krzemiński W, Soszyńska A (2026) Overlooked diversity of the predatory fungus gnat Macrocera in Baltic amber. Acta Palaeontologica Polonica 71: 219–245. https://hdl.handle.net/10520/EJC84578

Perkovsky EE, Rasnitsyn AP (2013) First records of Scolebythidae and Chrysididae (Hymenoptera, Chrysidoidea) in Rovno amber. Vestnik Zoologii 47: 103–108. https://doi.org/10.2478/vzoo-2013-0010

Perkovsky EE, Rasnitsyn AP, Vlaskin AP, Taraschuk MV (2007) A comparative analysis of the Baltic and Rovno amber arthropod faunas: representative samples. African Invertebrates 48: 229–245. https://doi.org/10.10520/EJC84578

Porto D, Melo GAR, Almeida EAB (2015) Clearing and dissecting insects for internal skeletal morphological research with particular reference to bees. Revista Brasileira de Entomologia 60: 109–113. https://doi.org/10.1016/j.rbe.2015.11.007

Ross AJ, Bojarski B, Szwedo J (2026) A critical review of the age of Baltic amber from the Samland Peninsula, Russia. Earth and Environmental Science Transactions of the Royal Society of Edinburgh 1–14. https://doi.org/10.1017/S1755691025100960

Sadowski E-M, Schmidt AR, Seyfullah LJ, Kunzmann L (2017) Conifers of the “Baltic amber forest” and their palaeoecological significance. Stapfia 106: 1–73.

Schneider CA, Rasband WS, Eliceiri KW (2012) NIH Image to ImageJ: 25 years of image analysis. Nature Methods 9: 671–675. https://doi.org/10.1038/nmeth.2089

Sinclair BJ, Cumming JM, Brooks SE (2013) Male terminalia of Diptera (Insecta): a review of evolutionary trends, homology and phylogenetic implications. Insect Systematics & Evolution 44: 373–415. https://doi.org/10.1163/1876312X-04401001

Smith TJ, Mayfield MM (2015) Diptera species and functional diversity across tropical Australian countryside landscapes. Biological Conservation 191: 436–443. https://doi.org/10.1016/j.biocon.2015.07.035

Solórzano-Kraemer MM, Delclòs X, Clapham ME, Arillo A, Peris D, Jäger P, Stebner F, Peñalver E (2018) Arthropods in modern resins reveal if amber accurately recorded forest arthropod communities. Proceedings of the National Academy of Sciences of the United States of America 115: 6739–6744. https://doi.org/10.1073/pnas.1802138115

Soszyńska-Maj A, Krzemiński W (2015) New representative of the family Panorpodidae (Insecta, Mecoptera) from Eocene Baltic amber with a key to fossil species of genus Panorpodes. Palaeontologia Electronica 2015: 1–17. https://doi.org/10.26879/546

Sowa KM, Wróbel P, Kołodziej T, Błachucki W, Kosiorowski F, Zając M, Korecki P (2023) PolyX beamline at SOLARIS—Concept and first white beam commissioning results. Nuclear Instruments and Methods in Physics Research Section B: Beam Interactions with Materials and Atoms 538: 131–137. https://doi.org/10.1016/j.nimb.2023.02.033

Szlachetko J, Szade J, Beyer E, Błachucki W, Ciochoń P, Dumas P, Freindl K, Gazdowicz G, Glatt S, Guła K, Hormes J, Indyka P, Klonecka A, Kołodziej J, Kołodziej T, Korecki J, Korecki P, Kosiorowski F, Kosowska K, Kowalski G, Kozak M, Kozioł P, Kwiatek W, Liberda D, Lichtenberg H, Madej E, Mandziak A, Marendziak A, Matlak K, Maximenko A, Nita P, Olszowska N, Panaś R, Partyka-Jankowska E, Piszak M, Prange A, Rawski M, Roman M, Rosmus M, Sikora M, Sławek J, Sobol T, Sowa K, Spiridis N, Stępień J, Szczepanik M, Ślęzak M, Ślęzak T, Tyliszczak T, Ważny G, Wiechecki J, Wilgocka-Ślęzak D, Wolanin B, Wróbel P, Wróbel T, Zając M, Wawrzyniak A, Stankiewicz M (2023) SOLARIS National Synchrotron Radiation Centre in Krakow, Poland. The European Physical Journal Plus 138: 10. https://doi.org/10.1140/epjp/s13360-022-03592-9

The Paleobiology Database (n.d.) World Wide Web address: https://paleobiodb.org [accessed on 05.05.2026]

Thompson N, Salzmann U, Hutchinson DK, Strother SL, Pound MJ, Utescher T, Brugger J, Hickler T, Hocking EP, Lunt DJ (2025) Global vegetation zonation and terrestrial climate of the warm Early Eocene. Earth-Science Reviews 261: 105036. https://doi.org/10.1016/j.earscirev.2024.105036

Utescher T, Erdei B, François L, Henrot A-J, Mosbrugger V, Popova S (2021) Oligocene vegetation of Europe and western Asia—Diversity change and continental patterns reflected by plant functional types. Geological Journal 56: 628–649. https://doi.org/10.1002/gj.3830

Vilhelmsen L, Perkovsky EE, Jenkins Shaw J (2024) Rogue sawflies: rare late Eocene amber fossils provide new calibration points for dating the evolution of Tenthredinoidea (Insecta: Hymenoptera). Journal of Systematic Palaeontology 22: 2348774. https://doi.org/10.1080/14772019.2024.2348774

Wolfe AP, McKellar RC, Tappert R, Sodhi RNS, Muehlenbachs K (2016) Bitterfeld amber is not Baltic amber: Three geochemical tests and further constraints on the botanical affinities of succinite. Review of Palaeobotany and Palynology 225: 21–32. https://doi.org/10.1016/j.revpalbo.2015.11.002

Wu M, Kunzmann L, Li S-F, Teodoridis V, Zhou Z, Su T (2024) Vegetation changes across the Eocene–Oligocene transition: global signals vs. regional development. Science China Earth Sciences 67. https://doi.org/10.1007/s11430-023-1335-8

Zakrzewska M, Singh H, Wagner-Wysiecka E, Giłka W (2020) Minute and diverse in fossil sticky stuff: Tanytarsini (Diptera: Chironomidae) from early Eocene Indian Cambay amber. Zoological Journal of the Linnean Society 189: 1398–1425. https://doi.org/10.1093/zoolinnean/zlz159