All fossil pieces from the rediscovered Bernsteinsammlung (amber collection) of the Phyletisches Museum were provided with unique specimen identifiers. The identifiers are in the “Inv.-Nr. Pa.” series, which corresponds to the older accessions of the museum (von Knorre and Beutel 2018). The amber pieces were stored in three drawers in the attic of the museum and have been untouched for around 50 years. Presumably, due to the strong temperature and humidity fluctuations, the overall condition was quite poor with the fossils having brittle, deteriorated surfaces (Bisulca et al. 2012). Some of the pieces were found to be coated with an unknown type of varnish, which emitted a distinct smell during grinding. Other specimens were unconventionally glued to cardstock. We infer that the glue was some sort of epoxy, as this was commercially available in the late 1940s (Chen et al. 2019), and that the pieces were likely glued by E. Uhlmann sometime after acquisition (see section 3.1.1 below). We observed that the epoxy had “eaten” its way into the amber over time, resulting in dissolved surfaces. We removed the epoxy completely through careful grinding.

To facilitate the identification of inclusions, all amber specimens were manually ground and polished. Grinding was done with waterproof single silicon carbide abrasive paper (Robert Bosch GmbH, Robert-Bosch-Platz 1, 70839 Gerlingen, Germany) soaked in water (Sadowski et al. 2021). Initial sandpaper grainsize was chosen for each specimen individually, depending on condition and on the distance of the inclusions to the surface. In general, we used the following grainsize steps: 180, 400, 500, 1000, 1200, and 2000. Whenever possible, we made an effort to generate flat surface windows for viewing. Grinding was performed with sandpaper over a glass plate in one hand and the amber piece in the other hand. Amber pieces and glass plate were cleaned with water between every grain size step.

We experimented with different polishes such as chalk and Peek Polish (Peek Polish International, 51 Waterloo Road London NW2 7TX, United Kingdom). As the latter contains residual petroleum (Peek Premium Polish Paste Safety Data Sheet), it was used very cautiously but produced good results. Finally, we found that excellent results could be achieved by using the toothpaste Colgate® Sensation White Aktivkohle Zahnpasta (Colgate-Palmolive, 300 Park Avenue New York, NY, United States). A dab of toothpaste placed on microfibre cloth with a small amount of water was used to polish the amber pieces on all sides, until most of the remaining small scratches were no longer visible. To finish a single piece took up to three hours, as constant control under a desktop-mounted magnifying glass was mandatory to prevent grinding off inclusions.

To isolate specific inclusions, some amber specimens were cut into smaller pieces using a Dremel® 3000 (Robert Bosch GmbH, Dremel, 1800 W. Central Rd., Mt. Prospect, Illinois, U.S.) with a thin saw blade (0.1 mm) attachment. As the Dremel’s minimum speed is 10000/min, which was too fast for freehand amber cutting, we used a defective Proxxon FBS 240/E (Proxxon Inc., 130 US Hwy 321 SW, Hickory, NC 28602 USA) to create guiding cuts. The lowest revolution rate of the Proxxon is given as 5000/min, but the one that we used had a markedly lower speed. With the guiding cuts done, the faster Dremel® could be used safely. This process required very straight cuts to prevent the blade from bending and getting caught in the amber, which would occasionally cause breaking or splintering. Alternatively, we also used a hardwood saw (Heckenrose 3 fein, Augusta-Heckenrose, Werkzeugfabriken GmbH & Co KG, Rudolf-Diesel-Straße 36, 71154 Nufringen, Germany) of 0.3 mm, in this case with the tool fixed in place and the amber pulled over the blade. After polishing and cutting, the specimens were carefully dried using a microfiber cloth.

The final curatorial step was to store the fossils in individually shaped moulds of PE-Foam (SV-Schaumstoffe GmbH, Junkerstraße 10, 82178 Puchheim, Germany) in insect drawers. A slit was cut into the foam above each object, into which the matching label was inserted. A small note was placed under each amber piece (Museumspapier altweiß mit Alkalipuffer, Klug Conservation, Zollstraße 2, 087509 Immenstadt, Germany) with the corresponding inventory number.

As most of the specimens in the three PMJ Pa drawers lacked reliable collection information, we undertook a series of tests to determine whether the fossils had the properties of amber or copal. Specifically, we evaluated melting behavior, autofluorescence, hardness, solubility, and density. All tests were compared between the PMJ Pa material and known samples, which were either from Ethiopian, Baltic, or Burmese ambers. As the Ethiopian pieces were a loan from MAIG (Museum of Amber Inclusions, University of Gdańsk, Gdańsk, Poland), we did not sample these destructively.

To confirm that our specimens were products of plant resin rather than artefacts in plastic material, we heated a needle and attempted to insert it into the test samples to test melting behavior. We checked fluorescence using a handheld LED UV flashlight and a 6 pi LED special UV lamp with a 120° angle of radiation. To test hardness, we scratched the test pieces either against our fingernails or vice versa. For the solubility test, we used a Dremel disc saw with a diamond blade to cut small pieces from the test samples, which we then placed in 99.5% Acetone; after a few minutes, we removed the samples and pressed them between our fingers. For the density test, we filled two dishes with fresh water, and added salt to saturation to one of them, after which we placed known and unknown samples in both liquids. Finally, we sent a representative sample of 10 specimens that were either labeled as “copal” or “amber” to the International Amber Association (IAA; Gdańsk, Poland) for UV fluorescence and Fourier-transformed infrared spectroscopy (FT-IR) (Table 1).

Specimens were examined at the Phyletisches Museum primarily with a Zeiss Stemi SV 11 stereomicroscope and a Zeiss Axioskop compound microscope. For the stereomicroscope a maximum magnification of 40× was used. For the light microscope we used the magnifications 50×, 100× and 200×.

To remove minute scratches, the amber pieces selected for photography were polished in three successive steps with ST5000, ST7000 wet abrasive paper (Starcke, Melle, Germany), and Peek polish (Tri-Peek International, Saffron Walden, United Kingdom) or Colgate® Sensation White Aktivkohle Zahnpasta (Colgate-Palmolive, 300 Park Avenue New York, NY, United States).

For overview photographs, stacks of partially focused images were taken of the amber pieces with a Canon EOS R5 equipped with a Canon EF 100 mm f/2.8L Macro IS USM lens (Canon, Krefeld, Germany), which was mounted on a Kaiser copy stand. For focus bracketing, the internal camera software was used. The scene was illuminated with a Euromex LE.5211-230 cold light source for stereomicroscopy (Euromex, Papenkamp, Netherlands) equipped with three gooseneck lamps to adjust light conditions and prevent reflections. Underneath the camera a blurred glass plate was positioned over a black sprayed Kapa® box. The amber pieces were placed on the glass plate in a petri dish filled with distilled water as suggested by Sadowski et al. (2021). The amber was held in place with UHU Patafix (UHU, Bühl, Germany). Some pieces were photographed without water to obtain a better spatial impression.

For shots of details from single embedded specimens, a Canon Eos 7D Mark II (Canon, Krefeld, Germany) equipped with a Mitutoyo M Plan Apo 10 microscopic lens (Mitutoyo, Kawasaki, Japan) was used. To perform stack shots, the camera was mounted on a StackShot macro rail (Cognisys, Traverse City, USA). Two flashlights (Yongnuo Photographic Equipment, Shenzhen, China) illuminated the scene. The amber pieces were placed on a cover slip. Plasticine was used to level the surface. A drop of glycerine was placed on the surface of the amber piece, and the glycerine was then covered by an additional cover slip. Additional detail images were taken at the Museum für Naturkunde Berlin (MfN), where manual stacks were taken using a Zeiss Axioscope 5 with a Zeiss Achromat S 1,0 FWD 63 (Carl Zeiss AG, Oberkochen, Germany), mounted with a Canon EOS 80D (Canon, Krefeld, Germany) via a T2-T2 1,6× SLR tube.

All photographs were developed with Adobe Lightroom classic (v.11.5) (Adobe, San Jose, USA). The images (option: standard) were denoised with Topaz DeNoize AI (Topaz Labs, Dallas, USA). Zerene stacker 1.04 (Zerene Systems LLC, Richland, USA) was used to fuse the images (option: align & stack all (PMax).

Nine specimens (Table 2) were scanned using synchrotron radiation (SR-µ-CT) at the Imaging Beamline P05 (IBL) (Haibel et al. 2010; Greving et al. 2014; Wilde et al. 2016) operated by the Helmholtz-Zentrum Hereon at the storage ring PETRA III (Deutsches Elektronen Synchrotron—DESY, Hamburg, Germany). A photon energy of 18 keV and a sample to detector distances of 30–50 mm were used. Projections were recorded using a 50 MP CMOS camera system with an effective pixel size of 0.46 µm. 4001 projections were recorded for each tomographic scan at equal intervals between 0 and π, with an exposure time of 350 ms. When specimens were too large to fit into the field of view in the z-axis, we scanned overlapping sections and subsequently stitched them together. Tomographic reconstruction was done by applying a transport of intensity phase retrieval and using the filtered back projection algorithm (FBP) implemented in a custom reconstruction pipeline (Moosmann et al. 2014) using MATLAB (Math-Works) and the Astra Toolbox (Palenstijn et al. 2011; van Aarle et al. 2015, 2016). For further processing, raw projections were binned two times resulting in an effective pixel size of the reconstructed volume of 0.913 µm. For segmentation and visualization, the 32-bit .tif image sequences were converted to 8-bit files and downsampled twofold with Fiji (Schindelin et al. 2012), resulting in an effective pixel (voxel) size of 1.826 µm.

Two samples, PMJ Pa 5809 (‡Amphientomum knorrei Weingardt, Bock & Boudinot, sp. nov.) and PMJ Pa 5884 (Dorylus nigricans molestus) were dated using ¹⁴C analysis, as recommended by Delclòs Martínez et al. (2020). Amber pieces with a mass of 5 mg (PMJ Pa 5809), 6 mg (PMJ Pa 5884) and 5 mg (PMJ Pa 5889) were cut off and sent to Beta Analytic (4985 S.W. 74^th Court, Miami, FL, USA 33155). Dating analyses resulted in estimated age ranges with certain likelihoods. The age range with the highest likelihood was chosen (Table 3). Based on our identifications of the copal inclusions and the history of the collections, East Africa was selected as the geographic reference for dating. We chose the psocodean as taxonomic work on the group is challenged, with presently only a handful of specialists working on them (e.g., Mockford 2018). We also chose the putative Baltic Dorylus due to its potential evolutionary and biogeographic implications.

µ-CT-image stacks were segmented and 3D-reconstructed using Amira 6.0.1 (Thermo Fisher Scientific) and Dragonfly 2022.1 (Object Research Systems,). Image (tif) stacks and isosurfaces were exported with the Amira macro “Multi-Export” (Engelkes et al. 2018). The isosurfaces were reduced and smoothed with following parameters: ItTotal: 5; smooth: iteration: 4, lambda: 0.6; reduction: 0.7. Tiffs were volume rendered in VGStudio Max 2.0 (Volume Graphics) using the option Phong reflection model. The isosurfaces were further smoothed (modifier: smooth and option shade smooth) with Blender 3.2.0 (Blender Foundation). The 3D-Models were uploaded to the 3D repository Sketchfab (URL: https://sketchfab.com) using the free blender plugin: Sketchfab for Blender 1.5.0 (URL: https://github.com/sketchfab/blender-plugin/releases/tag/1.5.0).

Image plates were compiled using Adobe Photoshop (v. 24.1.0) (Adobe, San Jose, USA). Lettering was added with Adobe Illustrator (v. 27.2).

Specimens evaluated in the present study, and also for the use of comparison to each other, were from the following collections:

BEBC Brendon E. Boudinot research collection, Frankfurt am Main, Germany

MAIG Museum of Amber Inclusions, University of Gdansk, Poland.

PMJ Phyletisches Museum Jena, Germany.

MNHB Museum für Naturkunde Berlin, Germany.

MWC Michael Weingardt research collection, Jena, Germany.

SMNS Staatliches Museum für Naturkunde Stuttgart, Germany.

USNM U.S. National Museum of Natural History, Washington, D. C., U.S.A.

The raw scan data will be made available at MorphoSource upon acceptance.

3.1.1. Fossil provenance

Collector. Unknown.

Date. Unknown. It is likely that the 161 fossil pieces in the collection were acquired by the Phyletisches Museum in several batches between 1920 and 1930 (see below).

Circumstantial evidence. By the handwriting of the label for specimen PMJ Pa 5827, the date of collection is estimated between 1920 and 1930 (personal communication with Uwe Dathe, Historische Sammlungen ThULB, 15.07.2022 and Alexander Gehler, Geowissenschaftliches Zentrum der Georg-August-Universität, 14.07.2022). The label for specimen PMJ Pa 5827 has two different scripts on each side, which cannot be clearly assigned to any handwriting samples in the museum’s archives. Furthermore, the former curator of the museum, Dietrich von Knorre, stated that Eduard Uhlmann (1888–1974) was in charge of the small “Bernsteinsammlung” (amber collection); Uhlmann was a scientific assistant, later conservator at the Phyletisches Museum Jena while it was under the directorship of Ludwig Plate (director from 1909–1935), and later became associate professor (ao. [außerodentlicher] Professor 1950) and director of the Phyletisches Museum from 1952 till 1954.

In accordance with the statements above, there are handwritten labels on nine specimens (Pa 5828, 5829, 5836, 5871, 5873, 5874, 5875, 5882, and 5885) which can be assigned to E. Uhlmann and his wife Frida Uhlmann (born Preiss, 1894–1981). This assessment is unambiguous as the Uhlmanns re-sorted numerous drawers in the entomology collection, which were labelled by her (Krogmann et al. 2007). Moreover, von Knorre stated (25.07.2022) that Uhlmann (Fig. 1) bequeathed the small amber collection to him after retirement, which von Knorre later added to the museum’s collection. We screened the archives of the Friedrich Schiller University (21.07.2022) and the Phyletisches Museum (21–22.07.2022) but did not discover invoices or personal communication of Uhlmann and Plate. In sum, how exactly the small amber collection found its way into the museum cannot be confidently resolved at present.

Listed localities. Of the 161 numbered pieces, 76 lacked locality information and one additional number was associated with human-made beads without further information (Pa 5911; not included in the total count). Of the remaining 85 pieces, 21 have labels indicating a Baltic origin and three are marked as copal from East Africa, such as Pa 5829, which has a handwritten label stating “Kopalinsect Ost-Afrika Diluvium”. One of the putative Baltic pieces, Pa 5827, has a handwritten label “Fundstück von Kleinkuhren Samland” (Filino), while 15 are marked with “Samland?” and the derivation of 51 of the other presumptive amber specimens is indicated as Samland Bernsteinwerke Königsberg on the labels, with an authentic invoice from the “Preuß. Bergwerks- und Hütten-Akt.-Ges. Zweigniederlassung Bern-steinwerke Königsberg Pr. for 1 M” for Pa 5863. For a discussion of the Samland and Kleinkuhren localities, see also section 4.2.

Qualitative tests. Despite the labels, the locality or localities of origin are not as clear as the first impression suggested. As it is not possible to discriminate copal and amber reliably based on the visual appearance alone (e.g., Federman 1990), we conducted a series of qualitative tests (see section 2.3), most of them at the Phyletisches Museum. Overall, the qualitative tests were contradictory and inconclusive. All pieces labelled as amber floated in saltwater and all labelled as copal sunk, apparently confirming the original labels. However, not a single piece in the collection emitted blue light under UV, while amber specimens of known provenance from the BEBC (burmite, succinite) and MAIG (Ethiopian amber) did. By burning the different samples from the PMJ Pa collection, every piece sizzled and produced black smoke, as expected for amber. Only one sample (Pa 5809) showed clear white smoke, but burning a larger piece yielded black smoke. Thus, the expected difference—that amber burns and copal melts—was not observed. When only small pieces were left, both copal and amber melted. Most conspicuous was a sweet odor produced by burning a larger piece of copal (from Pa 5809), which clearly distinguished it from true amber. This scent was not detected with smaller pieces, however.

Quantitative tests. For quantitative testing, additional small pieces were cut off and sent to the International Amber Association (IAA, 1 Warzywnicza Street, 80-838 Gdańsk, Poland) and checked with the UV/VIS and the FT-IR method (Table 1). The results from these analyses contradicted those from the qualitative tests and showed that all light-yellow pieces were copal, including those with Dorylus. In contrast, the darker, orange-colored pieces were true succinite (Baltic amber), including the one with †Yantaromyrmex. Finally, the ¹⁴C dating yielded an age of only ~145 years old for the piece containing the Dorylus. Therefore, this and other pieces with similar biotic inclusions are identifiable as Defaunation resin (sensu Solórzano-Kraemer et al. 2020), while the estimated age for the Amphientomum-bearing specimen is ~390 years old.

Biotic evidence. Ultimately, our final interpretations of the geological source of the putative amber pieces were based on the combined weight of evidence from the IAA results and the insect inclusions themselves. While it was exciting to consider the possibility that the Dorylus, Lepisiota, Pheidole, and Crematogaster ants were first records from Baltic amber, our µ-CT scan of the Dorylus revealed that it is identifiable as an extant subspecies. Moreover, this subspecies, Dor. nigricans molestus (Figs 2A–C, 3, Appendix 1: Fig. A1), was previously recorded from reliably identified Tanzanian copal by DuBois (1998) (see section 4.2.2 below). Another critical element of biotic evidence was the large piece PMJ Pa 5827, which contained another Do. n. molestus as well as a distinctive platypodine beetle (Curculionidae; “ambrosia beetles”, “pinhole borers”). Based on our µ-CT rendering of the beetle, Bjarte Jordal (01 Nov 2022) identified it as either Doliopygus serratus or Dol. cf. serratus (Figs 4A–C, Appendix 1: Fig. A2), which in either case represents bark beetle populations that are extreme generalists, and presently distributed throughout Southern, Central, and West Africa (Beaver and Löyttyniemi 1985). Several other amber pieces from the PMJ Pa contained sweat bees (Meliponini) that closely resembled specimens from known African copal pieces at the SMNS. All these aforementioned fossil pieces were light yellow, while other specimens—without these distinctive taxa—were dark and of a pinkish red color. Among these darker specimens was an ant definitively identifiable as †Yantaromyrmex geinitzi, the type species of a Baltic-amber-endemic genus. In section 3.2, we outline the complete list of taxa that we identified in the PMJ Pa collection, as well as the results of investigations on the origin of material compared to the point of starting our investigations.

In Table 4 we provide a list of all specimens sorted taxonomically, including reference to the museum accession numbers.

The initial situation of the material, with a rather chaotic sorting, shows that almost half of the material was without evidence of origin. Around 40% were labelled as Baltic amber, while 3 pieces were labeled as copal (Suppl. material 1). Overall, almost 10% were labelled, but with a question mark, which is not reliable information in any case.

After identification, provenance research and chemical analysis the origin of most of the material could be solved confidently. One of the biggest surprises was that 7 pieces are of Kauri origin. Now, more than half of the material is of clear Baltic origin. The biggest switch was from pieces without evidence and the 3 pieces labelled as copal, as with 44% a sizable number in the collection is indeed East-African copal (Suppl. material 2).

3.3.1. Order Psocodea: Synopsis of higher taxa in the PMJ Pa

Mostly families represented by material in the PMJ Pa are reviewed below. The identification of PMJ Pa material was based on the keys of Smithers (1990), with Taylor (2013) for the amphientomid. Specific details about the fossil deposits and their ages in this and other synopsis header sections are drawn from Paleobio (2022); we have included these for the geological information and to ease future divergence-dating phylogenetic analysis. Taxonomic information was drawn in part from the Psocodea Species File Version 5.0 (Johnson et al. 2023) database.

3.3.1.1. Family Amphientomidae Enderlein, 1903. [Note 1]

Amphientominae Enderlein, 1903 amber species:

I. Genus Amphientomum Pictet, 1854. [Note 2].

A. Oise amber [France, Le Quesnoy; Eocene, Ypresian, 56.0–47.8 Mya].

1. †Am. parisiense Nel, Prokop, De Ploeg & Millet, 2005.

B. Baltic ambers [Eocene, 37.8–33.9 Mya].

2. †Am. (Amphientomum) leptolepis Enderlein, 1905. [Note 3].

3. †Am. (Amphientomum) paradoxum Pictet, 1854. [Type species!]

4. †Am. (Palaeoseopsis) colpolepis Enderlein, 1905.

C. African resin [ca. 390 ± 30 years].

5. ‡Am. knorrei Weingardt, Bock & Boudinot, sp. nov. [Note 4].

II. Genus Lithoseopsis Mockford, 1993.

D. Mexican amber [Miocene, 23.0–16.0].

1. †Li. elongata (Mockford, 1969).

III. Genus †Proamphientomum Vishnyakova, 1975.

E. Taimyr amber [Russia; Cretaceous, 85.8–83.5 Mya].

1. †Pr. cretaceum Vishnyakova, 1975.

Amphientomidae amber species incertae sedis:

IV. Genus †Arcantipsocus Azar, Nel & Néraudeau, 2009.

F. Charentese amber [France; Cretaceous, 105.3–99.6 Mya].

1. †Ara. courvillei Azar, Nel & Néraudeau, 2009. [Note 5].

Note 1. As part of our ongoing psocodean revisionary investigations, we provide an extended discussion of Amphientomum (see below).

Note 2. A list of all extinct and extant Amphientomum species is provided in Table 5. With the addition of the new species described herein, there are 20 species attributed to this genus (Johnson et al. 2023), of which four are extinct; it is unknown if the species described herein is still extant.

Note 3. The species †Am. leptolepis might be a variant of †Am. paradoxum, as these two taxa are distinguished by only a few characters (Enderlein 1911). Specifically, †Am. leptolepis is differentiated from †Am. paradoxum by the following: (1) fore wing with very long and slender scales; (2) sides of the scales parallel; and (3) the count of ctenidiobothria on the hind basitarsus is 36 (vs. 29–34 in †Am. paradoxum) (Enderlein 1911). It should be noted that the description of †Am. leptolepis was based on two specimens (Enderlein 1911, p. 295) and no additional information on this species was published since then, to the best of our knowledge. Overall, we consider the scale shape as not fully reliable, yet we retain the species status of these two taxa pending more detailed study.

Note 4. We describe this new species from African resin in the PMJ Pa collection. See section 3.1.1.1.4 below for our treatment of this taxon.

Note 5. Mockford et al. (2013) synonymized †Arcantipsocidae Azar, Nel & Néraudeau, 2009 with Amphientomidae, arguing that a dark, thickened pterostigma is a homoplastic feature across the order Psocodea, thus cannot be relied upon singly for placement in Psocomorpha. They left †Arcantipsocus unplaced within the family but further recognized features that the genus shares with modern Amphientomidae, i.e., the hindwing venation and the shape of maxillary palps, head, and forewings.

3.3.1.2. Family Liposcelididae Broadhead, 1950

Liposcelidinae Broadhead, 1950 amber and copal species:

I. Genus Liposcelis Motschulsky, 1852.

A. Baltic ambers [Eocene, 37.8–33.9 Mya].

1. †Li. atavus Enderlein, 1911.

B. Mexican amber [Miocene, 23.0–16.0].

2. †Li. sp. [Note 1]. [f].

C. Zanzibar copal [Pleistocene?].

3. ‡Li. resinata (Hagen, 1865).

Embidopsocinae Broadhead, 1950 amber species:

II. Genus Belaphopsocus Badonnel, 1955.

D. Dominican amber [Miocene, 20.4–13.8 Mya].

1. †Bs. dominicus Grimaldi & Engel, 2006.

III. Genus Belaphotroctes Roesler, 1943.

B. Mexican amber [Miocene, 23.0–16.0].

1. †Bt. ghesquierei Badonnel, 1949.

E. Zhangpu amber [Miocene, 16.0–13.8].

1. †Bt. grimaldii Engel & Wang, 2022.

IV. Genus Embidopsocus Hagen, 1866.

F. Oise amber [France, Le Quesnoy; Eocene, Ypresian, 56.0–47.8 Mya].

1. †Em. eocenicus Nel, de Ploëg & Azar, 2004.

A. Baltic ambers [Eocene, 37.8–33.9 Mya].

2. †Em. pankowskiorum Engel, 2016.

G. Bitterfeld amber [Eocene, 38.0–33.9 Mya].

3. †Em. saxonicus Günther, 1989.

Liposcelididae amber species incertae sedis:

II. Genus †Cretoscelis Grimaldi & Engel, 2006.

H. Kachin amber [Myanmar; Cretaceous, 99.6–93.5 Mya].

1. †Csc. burmitica Grimaldi & Engel, 2006.

Note 1. Mockford (1969) recognized a species of Liposcelis from Mexican amber that he left undescribed as insufficient structural detail, i.e., cuticular microsculpture and chaetotaxy, was observable.

3.3.1.3. Family Archipsocidae Pearman, 1936

Archipsocinae Pearman, 1936 amber species:

I. Genus Archipsocopsis Badonnel, 1948.

A. Mexican amber [Miocene, 23.0–16.0].

1. †Ari. antigua (Mockford, 1969).

II. Genus Archipsocus Hagen, 1882.

B. Baltic ambers [Eocene, 37.8–33.9 Mya].

1. †Aru. puber Hagen, 1882.

3.3.2. Taxon description (Psocodea)

Family Amphientomidae Enderlein, 1903

Subfamily Amphientominae Enderlein, 1903

The present key is modified after Badonnel (1967) using information from Enderlein (1911), Badonnel (1955, 1979), Smithers (1964, 1999), Li (1999, 2002), and Nel et al. (2005). Note that the species Amphientomum aelleni Badonnel, 1959 is excluded from the key as only the nymphal stage is described. This species, however, can be identified using an illustration of the unique coloration pattern of the head (Badonnel 1959, p. 763, fig. 1).

3.3.1.3. Further considerations of Amphientomum

Several hypotheses on relationships between species of Amphientomum have been proposed. According to Badonnel (1979), Am. flexuosum and Am. loebli are sibling species (= espèce voisine), but differ in details of coloration, the reduction of the basal section of Rs in the forewing of Am. loebli, and by the presence of a sclerite of the subgenital plate in this species, which is lacking in Am. flexuosum (Badonnel 1979). Moreover, both species differ in size and the number of femoral spines (Badonnel 1979), but whether these differences are stable and statistically significant is presently unclear. Both species occur in West Africa and western Central Africa (Angola, Ivory Coast, Nigeria). Additionally, Smithers (1999) proposed that Am. annulitibia is likely closely related to Am. flexuosum, but noted that it differs in the shape of the anterior margin of the phallosome and in the coloration of the head.

There are few studies on the ecology and faunistics of the East African Psocodea, including for instance the family Amphientomidae (Broadhead and Richards 1982; Georgiev 2022a). One species whose description is based on a single nymph from a Congolese cave might indicate weak troglophilic tendencies (Badonnel 1959). As Badonnel (1959) already discussed, a detailed study on the morphology of nymphal stages of Amphientomidae is needed before the specimen can be reliably placed in a genus. Several genera and the subfamily Tineomorphinae can nevertheless be excluded from the list of potential taxa based on the presence of profemoral spines and the shape of the lacinia (Badonnel 1959), while the genus Amphientomum seems to be the most likely taxon based on the few described characters that the nymph shares with species of Amphientomum. Several species of the genus are known to be attracted by light traps (Badonnel 1979; Smithers 1999).

A character of special interest is the coxal interlocking device described here for the new species (Fig. 15A, B). It consists of a hemispherical outgrowth on the right metacoxa, which fits into a corresponding cavity on the opposite side. A similar device was described by Pearman (1935) for the amphientomine genera Syllysis and Nephax, but never before for a species of Amphientomum. It is conceivable that the interlocking device is part of a jumping mechanism, but there are no observational data for amphientomids that would support this hypothesis. Interestingly, another interlocking device has been described for species of Lepidopsocidae, in this case on the mesocoxae but with a similar arrangement of a hemispherical protuberance on the right and a cavity on the left side, and involvement in jumping was also suggested in this case (Menon 1938; Ramesh et al. 2020). The occurrence of these devices on different segments represents a remarkable case of parallel evolution in the Psocodea. The evolutionary background still requires clarification.

A candidate homolog for the coxal interlocking devise is the Pearman’s organ, a unique structure of Psocodea, which consists of a mirror and rasp on the inner side of the metacoxae in most psocids (Mockford 2018). It can be missing (e.g., Pachytroctidae, Sphaeropsocidae, and likely in Liposcelididae, where the metacoxae are distinctly separated; Yoshizawa and Lienhard 2010; Mockford 2018) or certain substructures can be reduced (e.g., Amphientomidae with only the mirror part; Weidner 1972; but see discussion above and Mockford 1993 for a different view on the presence of the rasp in the family). We were unable to observe any rasping structures on the metacoxae in ‡Am. knorrei sp. nov. It is possible that the structure was in fact present but is indistinct due to artifacts and the quality of preservation of our specimen. As the mirror was also not visible in the SR-µ-CT scan and photographs or using light microscopy (maximal resolution 200 ×), it appears plausible that this interlocking device represents a modification of the Pearman’s organ, especially due to the functional requirement of intercoxal contact. Pearman (1935) proposed the same hypothesis on the origin of this device in other Amphientomidae. Smithers (1999) described the “tympanum” of the metacoxae of Am. annulitibia Smithers, 1999, with a thick and strongly raised edge but without an illustration. Nevertheless, it is conceivable that a similar structure is present in this species (described as inconspicuous by Smithers 1999), and that the device is more strongly developed in ‡Am. knorrei sp. nov. In any case, the device described here differs from the mesocoxal interlocking mechanism observed in Lepidopsocidae, which probably represents a non-homologous neoformation.

The lack of a robust phylogeny of Amphientomidae impedes the systematic placement of new species. As many genera are only defined by diagnostic characters and not apomorphies (e.g., Taylor 2013), the possibility of non-monophyly of these groups cannot be excluded and paraphyly may even be widespread in the family. It is therefore important for future research to establish a robust phylogeny of amphientomids, based on molecular data and morphological apomorphies. Badonnel (1967) explicitly mentioned that the use of different characters would lead to different systematic placements of his newly described species. This underlines the lack of morphological support for groupings in the family. Badonnel also commented on the minimal variability of the genitalia in Amphientomum, as only the basal common shaft of the parameres and the spermapore sclerite are suitable for diagnostic purposes. Mockford (2018) supported this observation, stating that external genitalia of both sexes are uniform throughout the entire amphientomids, with a y-shaped phallosome and three ovipositor valves, with a bilobed external valve (V3) fused over a part of its length with the dorsal valve (V2). The shape of the forewings can also be used as a diagnostic feature (Badonnel 1967), but sexually dimorphic wing shapes in a single species have to be considered (Badonnel 1967). The shape of the lacinia is an important diagnostic feature (Badonnel 1967) and is displayed together with the wing venation in Figs 13–15 for all species of Amphientomum with available data. As a phylogenetic analysis and a detailed study of type material would be beyond the scope of this study, we refrain from providing an updated diagnosis of Amphientomum or the subgenus Amphientomum (Palaeoseopsis). However, our study is a first step towards a phylogenetic investigation of the genus.

3.3.1.4. Additional observations and remarks for Psocodea in the PMJ Pa collection

A single apterous liposcelidid trapped in the large piece of resin PMJ Pa 5827 together with many other syninclusions is slightly deformed and almost transparent (Fig. 16C). It is a nymph, as only 8 antennomeres are visible, and the specimen is only ca. 0.9 mm long. As no external metafemoral tubercle is visible, it is likely that the single specimen belongs to the subfamily Embidopsocinae. To assign it to a genus of this subfamily and to make sure that the tubercle on the metafemur is indeed absent, further trimming and cutting of the amber piece would be necessary for higher magnification. However, as many syninclusions could be affected by this we refrained from further processing. As the specimen is miniscule, we could not observe other taxonomically relevant characters (Lienhard 1991), for instance the number of tarsomeres. Even though we do not assign the specimen to a genus, we could verify that the last maxillary palpomere is conical, and not enlarged as in Belapha Enderlein, 1917, Belaphopsocus Badonnel, 1955, Belaphotroctes Roesler, 1943 or Troctulus Badonnel, 1955 (Lienhard 1991). This makes a placement in these genera unlikely.

Additionally, one single macropterous specimen of the family Archipsocidae is present in the resin piece PMJ Pa 5825 (see Figs 16A, B, 17A, B, 18A, B, Appendix 1: Fig. A4). The presence of large wings indicates that this individual is a female (Mockford 1993). As the genital region is deformed and blocked by a large air bubble, it is unclear whether ovipositor valvulae are present or not. The visible, apically flattened and very wide subgenital plate resembles the homologous structure of Archipsocopsis (Mockford 1993, fig. 248; Georgiev 2022b, fig. 30). The results of the FT-IR analysis (IAA) suggest that the resin containing the archipsocid belongs to a kauri-type (see SI: Report No 41995_17022023), a tree species today only occurring in New Zealand (Steward and Beveridge 2010). This suggests that the specimen might be part of the fauna of this country and the first possible record of the family in this region (Lienhard 2016). That the specimen belongs to the genus Archipsocopsis is tentatively suggested by the shape of the subgenital plate. As further confirmation is required, we treat it here as Archipsocidae gen. et sp. indet. We briefly characterize this specimen as follows: Female. Head large and globular, with comparatively small compound eyes. Antennae relatively short and with relatively thick flagellum. Three ocelli present but indistinct. Macropterous. Fore wings and hind wings completely developed. Tarsi with two articles. Ovipositor seemingly entirely missing. Subgenital plate wide and flattened posteriorly. Body length: 1.46 mm. Forewing length: 1.71 mm. Hindwing length: 1.45 mm.

3.3.2. Order Hymenoptera: Subfamilial and tribal synopsis of fossil Formicidae in the PMJ Pa collection

In this section, we review the fossil record of each subfamily of ants that occurs in the Phyletisches Museum amber collection, with the exception of Formicinae, for which our review is restricted to the tribes Plagiolepidini and Camponotini. See section 3.3.1 for details about deposit and age source information.

3.3.2.1. Subfamily Dorylinae Leach, 1815

Amber fossils: Identifiable to species:

A. Baltic ambers [Eocene, 37.8–33.9 Mya].

I. Genus †Procerapachys Wheeler, 1915. [Note 1].

1. †Pr. annosus Wheeler, 1915. [w, m]. [Type species of genus].

2. †Pr. favosus Wheeler, 1915. [w].

3. †Pr. sulcatus Dlussky, 2009. [w].

B. Dominican amber [Miocene, 20.4–13.8 Mya].

II. Genus Acanthostichus Mayr, 1887.

5. †A. hispaniolicus de Andrade, 1998b. [w].

III. Genus Cylindromyrmex Mayr, 1870.

6. †Cy. antillanus de Andrade, 1998a. [q].

7. †Cy. electrinus de Andrade, 1998a. [q].

8. †Cy. inopinatus de Andrade, 1998a. [q].

IV. Genus Neivamyrmex Borgmeier, 1940.

9. †N. ectopus Wilson, 1985. [w]. [see Wilson 1985b].Copal fossils: Identifiable to species:

C. East African copal [Holocene, 145 years based on our ¹⁴C analysis; < 36 Kya more generally, based on Solórzano-Kraemer et al. 2020].

V. Genus Dorylus Fabricius, 1793.

10. D. nigricans molestus (Gerstäcker, 1859). [w]. [See section 4.2.1; Fig. 3].

D. Colombian copal [Pleistocene(?); DuBois 1998].

VI. Genus Neivamyrmex Borgmeier, 1940.

11. N. iridescens Borgmeier, 1950. [w].

Note 1. After the morphological and phylogenomic revision of the Dorylinae by Borowiec (2016, 2019), the generic identities of the species attributed to †Procerapachys are uncertain. Ongoing work by multiple research groups will resolve at least some of these questions. At least one specimen in the BEBC is meaningfully identifiable as Lioponera (unpubl. data).

3.3.2.2. Subfamily Dolichoderinae Forel, 1878

3.3.2.2.1. Genus †Yantaromyrmex Dlussky & Dubovikoff, 2013 [Note 1]

Amber fossils: Identifiable to species:

A. Baltic ambers [Eocene, 37.8–33.9 Mya].

1. †Y. constrictus (Mayr, 1868). [w].

2. †Y. geinitzi (Mayr, 1868). [w, q, m]. [Note 2].

3. †Y. intermedius Dlussky & Dubovikoff, 2013. [w].

4. †Y. mayrianum Dlussky & Dubovikoff, 2013. [w].

5. †Y. samlandicus (Wheeler, 1915). [w].

Note 1. †Yantaromyrmex is an extinct genus of ants that is endemic to ambers from Baltic sources. The genus is of unknown phylogenetic affiliation with other Dolichoderinae, although it has been hypothesized to be close to the so-called “DNAPPTOFI” clade of the Leptomyrmecini (sensu, Ward et al. 2010; Boudinot et al. 2016), more specifically, as an ancestor of the clade containing Anonychomyrma and Iridomyrmex (Dlussky and Dubovikoff 2013). Addressing this hypothesis is outside the scope of the present study, so we retain the placement of this fossil in Leptomyrmecini and note explicitly that this is an uncertain and untested relationship.

Note 2. The PMJ Pa specimen identified as †Y. geinitzi (PMJ Pa 5856) was identified using the key of Dlussky and Dubovikoff (2013).

3.3.2.3. Subfamily Formicinae Latreille, 1802

3.3.2.3.1. Tribe Plagiolepidini Forel, 1886

Amber and copal fossils: Identifiable to species:

I. Genus Lepisiota Santschi, 1926.

A. African copal [Holocene, < 36 Kya (Solórzano-Kraemer et al. 2020)].

1. †Le. cf. canescens [w]. [Note 1].

II. Genus Plagiolepis Mayr, 1861.

B. Baltic amber [Eocene, 37.8–33.9 Mya].

2. †Pl. klinsmanni Mayr, 1868. [w].

• Wheeler, 1915: 101 (m); Dlussky, 2010a: 65 (“ergatoid” q).

3. †Pl. kuenowi Mayr, 1868. [w].

= †Pl. balticus: Dlussky, 2010a: 69.

• Dlussky, 2010a: 70 (q).

4. †Pl. singularis Mayr, 1868. [q].

5. †Pl. solitaria Mayr, 1868. [m].

6. †Pl. squamifera Mayr, 1868. [w].

7. †Pl. wheeleri Dlussky, 2010. [w].

C. Rovno amber [Ukraine; Eocene, 38.0–33.9 Mya].

8. †Pl. minutissima Dlussky & Perkovsky, 2002. [m].

D. Saxonian (Bitterfeld) amber [Germany; Eocene (Priabonian), 38.0–33.9 Mya].

9. †Pl. paradoxa Dlussky, 2010. [m].

E. Sicilian amber [Italy; Oligocene, 11.6–5.3 Mya].

10. †Pl. labilis Emery, 1891. [w].

III. Genus Acropyga Roger, 1862.

F. Dominican amber [Miocene, 20.4–13.8 Mya].

11. †Acropyga glaesaria LaPolla, 2005. [q, m].

Species inquirendae:

II. Genus Plagiolepis.

G. Uncertain [Cenozoic?].

(1.) †Pl. succini André, 1895. [Note 2].

Note 1. This specimen (PMJ Pa 5807) may be the first fossil or sub-fossil Lepisiota recorded to date. The specimen is dark, has the bituberculate form of the propodeum associated with L. canescens, and lacks spines on the petiolar node. Lepisiota is also the third genus (of nine) from the Plagiolepidini to be recovered from fossil-bearing sediments, the other two being Acropyga and Plagiolepis. Ants of this tribe were formerly lumped with the Prenolepis genus group of the Lasiini (see Bolton 2003; Boudinot et al. 2022a).

Note 2. As reported by Dlussky (2010a, b), there is considerable uncertainty about the generic identity and geological origin of the fossil(s) described as †Pl. succini by André (1895). In André’s own words (paraphrased, p. 83), “despite its great size for [Plagiolepis], this species resembles [Anoplolepis] custodiens [F. Smith]”, with the following diagnostic note (also paraphrased): “but it is distinguished by its size and the thicker and less elongate antennae”. Dlussky (2010a, b) was unable to find the specimen(s) originally described by André and speculated that the taxon may be from East African copal, i.e., misidentified as Baltic amber. This species should, therefore, be continued to be considered of uncertain status until the original or additional material becomes available.

3.3.2.3.2. Tribe Camponotini Forel, 1878

Overview. The Camponotini is a comparatively diverse tribe of Formicinae with > 1950 valid species attributed to eight extant genera (Bolton 2023), which have received revisionary attention subsequent to the major phylogenomic study of Blaimer et al. (2015) (see Ward et al. 2016; Ward and Boudinot 2021). Because this clade represents a major radiation of ants (e.g., Rafiqi et al. 2020), fossils are particularly important for understanding and modeling the tempo and mode of formicine evolution. The diversity of fossils attributed to this tribe, moreover, necessitates this overview. Prior to the present study, Camponotini contained two monotypic and valid fossil genera (†Chimaeromyrma brachycephala Dlussky, 1988, †Pseudocamponotus elkoanus Carpenter, 1930), one monotypic fossil genus that is a synonym of Camponotus Mayr, 1861 (†Palaeosminthurus juliae Pierce & Gibron, 1962), one fossil species attributed to Polyrhachis (†Po. annosus Wappler et al., 2009), and 29 valid and 5 invalid fossil species placed in Camponotus itself (Bolton 2023).

Based on our assessment of all 38 of these fossil species plus the three species attributed to the Camponotus-like form taxon †Camponotites Steinbach, 1967, we propose the following revision to the fossil record of Camponotini below. In brief, we transfer one species out of the subfamily Formicinae to Liometopum (I), we recognize one as-yet unidentified copal specimen of Camponotus (II) and one fossil species of Polyrhachis (III), we leave †Chimaeromyrma and †Pseudocamponotus incertae sedis in Camponotini (IV, VII), we transfer one Baltic fossil species from Camponotus to †Eocamponotus gen. nov. (V), we revive †Palaeosminthurus and consider it unidentifiable while transferring it out of Camponotini as incertae sedis in Formicinae (VI), and finally, we transfer 29 fossil species from Camponotus to the form genus †Camponotites, which we treat as a catch-all that is incertae sedis in Camponotini (VIII). We also provide detailed annotations for our synopsis of fossil Camponotini (see the “Notes”).

Finally, we point out that all future studies on fossils that may possibly be associated with Camponotus or Camponotini should critically evaluate the morphological evidence for placement in any of the extant genera particularly in reference to Ward et al. (2016) for workers and Ward and Boudinot (2021) for workers and alates (the wing venation characters apply equally to males and queens). A recent work which ignored these studies is Takahashi and Aiba (2023), which misidentified multiple specimens as Camponotus. If a species name must be given to a fossil that cannot be placed in any of the extant genera of Camponotini based on synapomorphies, we strongly encourage authors to place these fossils in the form genus †Camponotites so that uncertainty is explicitly recognized and to prevent the propagation of errors in macroevolutionary analysis.

I. Transferred to Liometopum (Dolichoderinae):

Compression fossil:

A. Shanwang formation [China, Linqu County; Miocene (Burdigalian), 20.4–16.0 Mya].

a. Liometopum Mayr, 1861.

= †Shanwangella Zhang, 1989. Syn. nov.

(1.) †L. palaeopterum (Zhang, 1989). Comb. nov. [q]. [Note 1].

= †S. palaeoptera Zhang, 1989.

• Combination in Camponotus: Hong and Wu 2000: 19.

II. Genus Camponotus Mayr, 1861, subgenus indet. [Note 2].

Copal fossil: Identifiable to species:

B. East African copal [Holocene, < 36 Kya (Solórzano-Kraemer et al. 2020)].

1. Ca. sp. THIS STUDY. [w]. [Note 3].

III. Genus Polyrhachis Smith, F., 1857.

Compression fossil:

C. Varvara formation [Miocene, 7.2–5.3 Mya].

1. †Po. annosus Wappler et al., 2009. [Note 4].

IV. Genus †Chimaeromyrma Dlussky, 1988, incertae sedis in tribe.

Amber fossil:

D. Sakhalin amber [Eocene, 47.8–41.3 Mya].

1. †Ch. brachycephala Dlussky, 1988. [Note 5].

V. Genus †Eocamponotus Boudinot, gen. nov. [Note 6].

Type species: †Eo. mengei (Mayr, 1868) by original designation.

Amber fossils: Identifiable to species:

E. Baltic ambers [Eocene, 37.8–33.9 Mya].

1. †Eo. mengei (Mayr, 1868). [w]. Comb. nov. [Note 7].

= †Eo. igneus (Mayr, 1868). Comb nov.

• Synonymized by Wheeler (1915): 138.

VI. Genus †Palaeosminthurus Pierce & Gibron, 1962 stat. rev., incertae sedis in Formicinae, unidentifiable hence invalid stat. nov.:

Phosphatized fossil:

F. Barstow formation, Calico member [USA, California; Miocene (Hemingfordian), 20.4–16.0 Mya].

(–.) †Pa. juliae Pierce & Gibron, 1962. [m]. Comb. rev.; unidentifiable, hence invalid, stat. nov. [Note 8].

= Formerly unresolved junior homonym of Camponotus juliae Emery, 1903.

• Transferred to Formicidae: Najt, 1987: 152.

• Status as species: Bolton, 1995b: 311.

• Transferred to Camponotus: Snelling, R.R. pers. comm. to Bolton, B. 2004, in Bolton (2023).

VII. Genus †Pseudocamponotus Carpenter, 1930, incertae sedis in tribe, unidentifiable hence invalid stat. nov.

Compression fossil:

G. Elko formation [Eocene, 37.2–28.4 Mya].

(–.) †Ps. elkoanus Carpenter, 1930. [q]. Unidentifiable, hence invalid, stat. nov. [Note 5].

VIII. Genus †Camponotites Steinbach, 1967, incertae sedis in Formicinae. [Note 9].

Amber fossil: Unidentifiable at neontological genus level:

H. Fushun amber [China, Liaoning, Jijuntun formation; Eocene (Lutetian), 47.8–41.3 Mya].

(1.) †Ctt. tokunagai (Naora, 1933). [q?]. Comb. nov. [Note 10].

Compression fossils: Unidentifiable at neontological genus level: [Note 11].

I. Green River formation [USA, Colorado; Eocene, 50.3–46.2 Mya].

(2.) †Ctt. vetus (Scudder, 1877). [q?]. Comb. nov. [Note 12].

J. Bouldnor formation, Bembridge Marls member [Great Britain; Eocene (Priabonian), 38.0–33.9 Mya].

(3.) †Ctt. cockerelli (Donisthorpe, 1920). [m]. Comb. nov.

= †Leucotaphus cockerelli Donisthorpe, 1920.

• Combination in Camponotus: Dlussky and Perfilieva 2014: 417.

K. Florissant formation [USA, Colorado; Eocene, 37.9–33.9 Mya].

(4.) †Ctt. fuscipennis (Carpenter, 1930). [q]. Comb. nov.

(5.) †Ctt. microcephalus (Carpenter, 1930). [q]. [Note 13]. Comb. nov.

(6.) †Ctt. petrifactus (Carpenter, 1930). [w]. Comb. nov.

L. Brunstatt, horizon d2 [France; Early Oligocene, 33.9–28.4 Mya].

(7.) †Ctt. compactus (Förster, 1891). [q]. Comb. nov.

(8.) †Ctt. vehemens (Förster, 1891). [m]. Comb. nov.

• Théobald, 1937: 218 (w, q, m).

= Senior synonym of †Ca. miserabilis Förster, 1891: Théobald, 1937: 218.

M. Creek near Bechlejovice [Czechia; Oligocene (Rupelian), 33.9–28.1].

(9.) †Ctt. novotnyi (Samšińák, 1967). [q]. Comb. nov.

N. Rott formation [Germany, Orsberg; Oligocene: 28.4–23.0 Mya].

(10.) †Ctt. lignitus (Germar, 1837). [q]. Comb. nov.

= †Formica lignitum Germar, 1837.

• Combination in Camponotus: Mayr, 1867: 51.

O. Niveau du gypse d’Aix Formation [France; Oligocene (Chattian), 28.1–23.0 Mya].

(11.) †Ctt. longiventris (Théobald, 1937. [q, m]. Comb. nov.

(12.) †Ctt. theobaldi (Özdikmen, 2010). [m]. Comb. nov.

• Replacement name for †Ca. saussurei Théobald, 1937.

(13.) †Ctt. penninervis (Théobald, 1937). [m]. Comb. nov.

A. Shanwang formation (see above).

(14.) †Ctt. ambon (Zhang, 1989). [q?]. Comb. nov.

(15.) †Ctt. ampullosus (Zhang, 1989). [q?]. Comb. nov.

(16.) †Ctt. curviansatus (Zhang, 1989). [q]. Comb. nov.

(17.) †Ctt. gracilis (Zhang, 1989). [m?]. Comb. nov.

(18.) †Ctt. longus (Zhang, 1989). [q]. Comb. nov.

(19.) †Ctt. microthoracus (Zhang, 1989). [q]. Comb. nov.

(20.) †Ctt. plenus (Zhang, 1989). [q]. Comb. nov.

(21.) †Ctt. shanwangensis (Hong, 1984). [q]. Comb. nov.

(22.) †Ctt. pictus (Zhang et al., 1994) [q]. [Note 14]. Comb. nov. • Previously junior primary homonym of Ca. ligniperda pictus Forel, 1886.

(23.) †Ctt. xiejiaheensis (Hong, 1984). [m]. Comb. nov.

• TYPE SPECIES of †Rabidia Hong, 1984.

• Combination in Oecophylla: Zhang, 1989: 297.

• Combination in †Camponotites Dlussky et al., 2008: 616.

P. Radoboj [Croatia; Miocene (Sarmatian), 12.7–11.6 Mya].

(24.) †Ctt. heracleus (Heer, 1849). [m]. Comb. nov.

= †Formica heraclea Heer, 1849.

• Combination in Camponotus: Mayr, 1867: 52.

• Also described as new by Heer, 1850: 116.

(25.) †Ctt. induratus (Heer, 1849). [m]. Comb. nov.

= †Formica indurate Heer, 1849.

• Dlussky and Putyatina 2014: 249 (m, q).

• Combination in Camponotus: Mayr, 1867: 52.

• Also described as new by Heer, 1850: 116.

(26.) †Ctt. oeningensis (Heer, 1849). [q]. Comb. nov.

= †Formica obesa oeningensis Heer, 1849.

• Combination in Camponotus and raised to species: Cockerell, 1915: 486.

Q. Joursac [France; Miocene, 11.6–7.2].

(27.) †Ctt. obesus (Piton, 1935). [q?]. [Note 15]. Comb. nov.

R. Montagne d’Andance Saint-Bauzile, Privas [France, Ardèche; Miocene (Turolian), 8.7–5.3 Mya].

(28.) †Ctt. crozei (Riou, 1999). [q]. Comb. nov.

S. Brunn-Vösendorf [Austria; Miocene (Messinian), 7.2–5.3 Mya]

(–.) †Ctt. ullrichi (Bachmayer, 1960). [wing]. Comb. nov.; unidentifiable, hence invalid, stat. nov. [Note 16].

T. Willerhausen clay pit [Germany; Pliocene (Piacenzian), 3.6–2.6 Mya].

(29.) †Ctt. silvestris Steinbach, 1967. [q].

• †Camponotites Steinbach, 1967 TYPE SPECIES.

• Redescribed: Dlussky et al. 2011: 452.

(30.) †Ctt. steinbachi Dlussky et al., 2011. [q].

Note 1. The species †Ca. palaeopterus (Zhang, 1989) was originally attributed to its own genus, †Shanwangella Zhang, 1989 before being placed in Camponotus by Hong and Wu (2000). The fossil cannot be attributed to Formicinae at all, however, due to the presence of cross vein 2rs-m, which never occurs in Formicinae; this crossvein encloses the second submarginal cell of the wing. Because the specimen has 2rs-m, lacks postpetiolation of abdominal segment III, and does not have cinctation of abdominal segment IV (= presence of transverse sulci), we transfer the species to Dolichoderinae. Therein, the specimen is recognizable as an alate of Liometopum given its large size (discal and submarginal cells ~ 1 mm long), its massive gaster, and details of the venation. †Liometopum palaeopterum comb. nov. was discovered in Shanwang in the Shandong Province, reasonably within the current distribution of the extant species L. sinense Wheeler, 1921 (Del Toro et al. 2009). Moreover, it is plausible that the fossil represents an ancestral population given that L. sinense is the only species currently known from China, at least to present knowledge. Unfortunately, gynes and males are unknown for the extant species. The unusual head and antennae of †L. palaeopterum comb. nov. are here interpreted as preservational artefacts.

Note 2. Camponotus and the tribe Camponotini more broadly is one of the most challenging taxonomic puzzles in the Formicidae, and not merely due to the massive size of these taxa (1084 valid species and 411 valid subspecies are currently attributed to Camponotus at the date of writing, Bolton 2023). Although some genera in the tribe are reasonably identifiable based on external morphology (e.g., Ward et al. 2016), others, such as the fundamental distinction between Colobopsis—which is sister to all other Camponotini—and the hyperdiverse Camponotus is challenging even with extant material in hand and under the microscope (Ward and Boudinot 2021). For these reasons, we substantially revise the fossil system of Camponotus in order to meet the twin aims of: (1) cleaning up the useless species names attributed to Camponotus, and (2) discouraging uncritical use of these fossils for macroevolutionary analysis (e.g., Klimeš et al. 2022). Toward these aims, we have: (a) provided a new genus name for †Camponotus mengei, †Eocamponotus gen. nov., as this fossil cannot be confidently placed in Camponotus yet; (b) transfer red 29 fossil taxa from Camponotus to the form genus †Camponotites, which we treat as incertae sedis in Formicinae; and (c) transfer red one species out of the Formicinae altogether.

Note 3. The minor worker specimen (PMJ Pa 5829) is difficult to identify, given the limited resources available. We attempted to identify the specimen using Emery’s (1925) key to Old World Camponotus subgenera, Brian Taylor’s Ants of Africa website (Taylor 2022), and AntWeb (2022). Ultimately, we were unable to obtain a satisfying identification. In mesosomal form and postcephalic setation, the specimen resembles Camponotus (Myrmacraphe) furvus Santschi, 1911 but it differs in head shape and by having shorter palps. In Taylor’s key, the specimen runs to C. acvapimensis Mayr, 1862, yet it differs in mesosomal form. We therefore conservatively consider the specimen unidentified.

Note 4. †Polyrhachis annosa neatly meets the expectations for Polyrhachis as it clearly has lateral petiolar spines, which is synapomorphic condition of the genus. Based on the limited preservation, we do not have confidence that this species will be placeable either in the stem or crown of the genus based on morphology. However, given that the Varvara formation is young, being dated at between 7–5 Myo, we do not think that it is defensible to place this fossil in a separate genus. We therefore leave this species in Polyrhachis with the hope that future phylogenetic work will resolve the polarity of petiolar spines within the genus.

Note 5. The monotypic genera †Chimaeromyrma and †Pseudocamponotus have been treated as incertae sedis in Camponotini, for which we see no specific morphological evidence to question these otherwise harmless placements. We choose to retain †Chimaeromyrma brachycephala Dlussky, 1988 as a valid genus and species as it is possible that the identification of this amber fossil may be refined through the application of µ-CT at some point in the future. As for †Pseudocamponotus elkoanus Carpenter, 1930, we doubt that this fossil will ever be identifiable given the lack of wing venation and very limited preservation of the single known specimen. Given that there is insufficient preservation to confidently place the species to tribe, we consider †Ps. elkoanus to be unidentifiable, hence invalid stat. nov. We do not synonymize †Pseudocamponotus with †Camponotites, however, as the former would take priority and we prefer the latter name as the form genus, given that it has the proper suffix (-tites) to indicate paleontological uncertainty.

Note 6. We erect the genus †Eocamponotus gen. nov. for †Camponotus mengei and its junior synonym †Ca. igneus as, although these fossils are sufficiently preserved for species diagnosis, formal combined-evidence analysis failed to support a relationship with any particular genus of the Camponotini (Boudinot et al. 2022a). In so doing, we aim to preclude the usage of this fossil in macroevolutionary analysis as a calibration point for the genus Camponotus. At most, this fossil species can be used as a calibration for the Camponotini; whether as a stem or crown calibration point, however, is much less certain, and a conservative approach would be for the stem of the tribe.

Note 7. We briefly note that †Ca. mengei was described alongside †Ca. igneus by Mayr (1868), and that the latter was accepted in a number of articles until Wheeler (1915) concluded that the latter is a subjective synonym of the former based on an examination of 103 Camponotus specimens and Mayr’s types of †Ca. mengei. Wheeler reported that the type specimens of †Ca. igneus were in the collection of Franz Anton Menge, which is presumably in Gdańsk, Poland. The valid species was originally considered to be a Ca. (Tanaemyrmex) but was recently suggested by Radchenko and Perkovsky (2021) to be Ca. (Camponotus) due particularly to the form of its clypeus, and the shapes of the head and mesosoma, without further specification. Reevaluation of the morphological affinities of this species is necessary.

Note 8. †Ctt. juliae (Pierce & Gibron, 1962) comb. nov. is represented by a single male that was phosphatized in a calcareous nodule in the Calico member of the Miocene-aged Barstow formation in the Mojave Desert of California. The taxonomic history of this fossil is unusual. In its original description in Pierce and Gibron (1962), the fossil was classified as a new species of a new genus representing a new family of symphypleonan Collembola: †Palaeosminthurus juliae (†Palaeosminthuridae). These names went unnoticed for more than two decades, until the collembolist Dr Judith Najt (see Deharveng et al. 2017) observed that the preserved head, scape, thorax, and leg remnants of the fossil belong to a hymenopteran, which she identified as Camponotus (see Najt 1987). Subsequently, Roy R. Snelling examined the fossil, presumably at the Los Angeles County Museum, and concluded that the taxon is a junior synonym of Camponotus festinatus (Buckley, 1866) (Snelling 2006), an identification that was communicated to Barry Bolton in 2004 but went unpublished by the time of Roy Snelling’s death in 2008. Bolton provisionally accepted this hypothesis in his taxonomic catalog (Bolton 2023). Here, after critical consideration of the available morphological evidence, we exclude the species from Camponotus, and revive the genus †Palaeosminthurus stat. rev., which we consider to be incertae sedis in Formicinae and unidentifiable hence invalid stat. nov. Specifically, we attempted to run the specimen through the male-based key to all Nearctic genera of Smith (1943) and that of Boudinot for all New World formicine genera (see section 3.7.H of Boudinot 2020); there is simply too little structural detail preserved to render a meaningful identification of this fossil. Unless a method like laminar µ-CT may be applied successfully, we anticipate that this fossil will remain unidentifiable at the genus and tribal levels among the Formicinae.

Note 9. Here, we recognize the form taxon †Camponotites to which we transfer 29 species. †Camponotites should be categorically precluded from usage as calibration points for macroevolutionary analysis of the Formicinae or Formicidae. A balance for retaining and actively using this form genus is that the fossils in this taxon may be useful for paleogeographic study, so invalidation of this name may result in loss of paleostratigraphic information. Future work involving direct re-examination of these fossils is necessary to determine some of the taxa now placed in †Camponotites may perhaps be placed with more confidence among the genera of Camponotini—such as †Ctt. novotnyi (Samšińák, 1967) comb. nov. as this fossil is quite well preserved and clearly displays the major synapomorphy of the Camponotini, namely that the antennal sockets are separated posteriorly from the posterior clypeal margin. For this example, however, the remainder of the specimen is too poorly preserved to allow the fossil to be meaningfully associated with any extant genus of the Camponotini.

Note 10. From the illustration provided in the original description, it is not possible to confidently identify the specimen as a member of Camponotus. We retain this species as incertae sedis in the genus to encourage future work on the fossil, if possible.

Note 11. All of these fossils could be considered unidentifiable to species, hence invalid, but are here treated as incertae sedis in Camponotus to highlight their existence. Critically, because of the lack of morphological information, it is possible that a number of these taxa belong to other genera of Camponotini (see Ward et al. 2016 and Ward and Boudinot 2021). Reexamination of the original material is necessary in all cases.

Note 12. While it may be tempting to use †C. vetus as a calibration for Camponotus or the Camponotini, it cannot be confidently attributed to any living genus, subgenus, or species group due to insufficient morphological information.

Note 13. The generic placement of †Ca. microcephalus is dubious and should be confirmed through direct examination of type and additional material.

Note 14. Although †Camponotus pictus Zhang et al., 1994 is a junior primary homonym of Ca. ligniperda pictus Forel, 1886, we transfer this species to the form taxon †Camponotites, rendering a nomen novum unnecessary.

Note 15. The compression-fossil taxon †Ca. obesus is represented by fragments of the mesosoma, legs, and metasoma, all of which are preserved in a dorso-anterolateral oblique view. These remains are suggestive of Camponotus but otherwise cannot be identified meaningfully. Identification, in this case, is primarily driven by the rough similarity and absence of other ~14 mm long ants with an apparently rounded mesosoma in modern Western Europe.

Note 16. Bachmayer (1960) described †Ctt. ullrichi based on a single forewing from a Miocene-aged deposit in Austria. This ~10.3 mm long wing cannot be attributed to Camponotus or the Camponotini because free M diverges from Rs+M proxi‹mad 2r-rs by more than twice the length of that cross vein. In Camponotini, free M diverges at or distad 2r-rs (see Char. 499 on p. 293 of Boudinot et al. 2022b; also, Ward and Boudinot 2021). Because only Myrmelachistini in Formicinae have the split of Rs+M well proximad 2r-rs, and as these ants are miniscule, we would prefer to consider the taxon incertae sedis in Formicidae. However, as this would necessitate the recognition of another “trashbin’’ form taxon, we elect to place the fossil in †Camponotites and to consider it unidentifiable, hence subjectively invalid stat. nov.

Between the taxonomic and color qualities of the resin pieces, coupled with the IAA results (Table 1), we were able to sort the fossils into Baltic amber and African copal categories to our satisfaction. Initially, the conflict among the qualitative tests and the results from the IAA confounded us. After considering these results, the true significance of the biotic inclusions, however, became clear, resolving the conundrum of fossil source. From our perspective, it was deeply surprising to find specimens interpreted as copal from the IAA tests that were directly labelled as from “Samland Kleinkuhren” in the PMJ Pa collection, especially as there are no reports, to our knowledge, that copal has been found in the Samland Peninsula (See also 4.2.2.). While we were initially skeptical that labels and objects might have been mixed in the small PMJ Pa collection, this seems to be the most likely scenario. We note that von Knorre kept the collection as it was given to him, thus the collection has not been seen and processed until now. The evidence from the PMJ Pa amber collection clearly shows a shift in source, especially from “without” to a confidently assigned one and simply false labeling could be resolved, and thus presents in a small scale what it might look like in another, larger collection.

At the bottom line, the scenario we encountered with this collection illustrates two critical points: (1) the importance of the correct labeling of any specimen, further underscoring the value of accurate corresponding information for contemporary and future research (King 1975; Corado 2005; Donovan and Riley 2013); and (2) the necessity of skepticism for fossils, even when having label data, as it was only the combination of biotic and chemical data that allowed us to draw confident conclusions about “amber” provenance in the present study. While the chemical or biotic evidence alone may have been sufficient, our uncertainty was not resolved until we had both lines of evidence, which clearly showed the inadequacy of the qualitative tests that have been supposed to differentiate between amber and copal.