Formica species (Hymenoptera, Formicidae, Formicinae) in late Eocene Rovno amber

A new species, Formica ribbeckei Radchenko & Perkovsky, sp. nov., is described based on four workers from late Eocene Rovno amber (Ukraine). It most resembles F. flori Mayr, 1868 but differs from the latter mainly by the 5-segmented maxillary palps with the preapical segment subequal in length to the apical one, and by the shorter first funicular segment. Fossil F. luteola Presl, 1822, F. trigona Presl, 1822, F. macrognatha Presl, 1822 and F. quadrata Holl, 1829 are considered incertae sedis in Formicidae. Thus, ten valid Formica Linnaeus, 1758 species (including F. ribbeckei) are known now from late Eocene European ambers. The diversity of Formica in the early and middle Eocene deposits of Eurasia and North America is considered. It is assumed that the genus Formica most likely arose in the early Eocene.


Introduction
The ant fauna of late Eocene European ambers (Priabonian, 33.9-37.8 Ma) is the best studied among all fossil myrmecofaunas worldwide. More than 17,500 ant specimens After careful investigation, somewhat unexpectedly we found that this Formica specimen differs from the common F. flori by many important features, allowing the description of a new species, F. ribbeckei Radchenko & Perkovsky, sp. nov. The discovery of a new Formica species in three localities -Gulyanka (Zhitomir region), Klesov and Vladimirets District (Rovno Region) together with P. klinsmanni, which was previously known from Klesov, supports a similar, late Eocene age, of all these Rovno amber deposits, since reliably dated late Eocene flora is known from Gulyanka (Pimenova 1937;Sokoloff et al. 2018).

Material and methods
We examined all available Formica inclusions from Rovno amber (about 100), four of which belong to F. ribbeckei Radchenko & Perkovsky, sp. nov. One piece of amber contains the holotype specimen (worker) of the described species together with a worker of P. klinsmanni (Fig. 2), and three other each contain one worker of F. ribbeckei Radchenko & Perkovsky, sp. nov. The holotype and two paratype specimens are deposited in the Schmalhausen Institute of Zoology of NAS of Ukraine, Kiev (SIZK), the paratype specimen No. KH-F-057 is in the collection of M. Khomych (Ukraine).
The photographs were taken with Leica Z16 APO microscope equipped with Leica DFC 450 camera processed by LAS Core software.
Not all features of the examined specimen were properly visible and measurable, hence we measured only well visible details (accurate to 0.01 mm), particularly: HL -maximum length of the head in dorsal view, measured in a straight line from the anteriormost point of clypeus to the mid-point of occipital margin; maxHW -maximum width of the head in dorsal view just behind (above) the eyes; minHW -minimal width of the head in dorsal view at the level of clypeus; SL -maximum length of the scape measured in a straight line from its apex to the articulation with condylar bulb; OL -maximum diameter of the eye; GL -length of the gena, measured from the anterior margin of the eye to the articulation with the mandible; ML -diagonal length of the mesosoma in lateral view from the anterior margin of the neck shield to the posterior margin of the metapleuron; MH -height of the mesosoma, measured from the upper level of the mesonotum perpendicularly to the lower margin of mesopleuron; PL -maximum length of the petiole, measured from the posterodorsal margin of the petiole to the articulation with the propodeum; PW -maximum width of the petiole in dorsal view; PH -maximum height of the petiole in profile, measured from the uppermost point of the petiolar scale perpendicularly to the lowest point of the petiole; HTL -maximum length of the hind tibia; FSL1…2 -length of the funicular segments 1 and 2; FSW1-width of the funicular segment 1; MP4…5 -length of segments of the maxillary palps from 4 and 5; TL -approximate total length is calculated as the sum of HL + ML + PL + length of the gaster. , HWI= maxHW/minHW, SI1=SL/HL, SI2=SL/max-HW, OI=OL/HL, GI= GL/OL, PI1=PL/HL, PI2=PL/PH, MI=ML/MH, FSI1=FSL1/  FSL2, FSI2=FSL1/FSW1, MPI=MP5/MP4. Diagnosis. Worker: maxillary palps 5-segmented, long, reaching occipital foramen, its apical and preapical segments subequal in length; gastral tergites with sparse decumbent pubescence, distance between setae greater than their length; first funicular segment ≤ 1.2 times as long as second one, and ≤ 2.25 times as long as broad.

Taxonomy
Description. Body length 4.7-6.5 mm. Head slightly longer than wide, distinctly narrowing anteriorly (maxHW/minHW 1.31-1.36), with broadly rounded occipital corners and weakly convex occipital margin. Frontal area well defined, sculptured similarly to frons. Eyes large (OI 0.3-0.35), situated distinctly behind (above) midlength of sides of head, genae distinctly longer than maximum diameter of eyes (GI 1.10-1.15). Ocelli distinct, forming equilateral triangle. Frontal groove weakly developed, short and very shallow. Clypeus with distinct longitudinal medial carina, its anterior margin slightly angulated. Antennal scape longer than head, somewhat widened apically, surpassing occipital margin for ca. 1/4 of its length. First funicular segment 1.1-1.2 times as long as second one and ca. twice as long as wide, second and third segments of same length and 1.7 times as long as wide. Maxillary palps 5-segmented but long, reaching occipital foramen, their third segment the longest, apical segment subequal to preapical one; labial palps 4-segmented. Mandibles with 6-8 sharp teeth, apical tooth distinctly larger, but less than twice as long as preapical one.
Mesosoma quite slender, 2.34-2.43 times as long as high, mesonotum slightly raised above pronotum. Dorsal surface of propodeum convex, posterior one almost straight and inclined posteriorly, both surfaces converging at rounded angle. Propodeal spiracles elongate-oval, ca. twice as long as wide. Petiolar scale rather high, its anterior surface convex, posterior one flat and almost straight, sides very slightly convex; its upper margin convex, without notch (seen in anterior or posterior views), in profile quite narrow, with crest. Meso-and metatibiae each with a single, long simple spur.
Body surface with very fine microsculpture, appears quite shiny, especially on genae. Mesosoma, petiole, legs and antennae without standing setae. Several coarse semierect setae present on vertex, frons and clypeus; gastral tergites and sternites with similar setae. Meso-and metatibiae usually without row of coarse bristles, only with 2-4 bristles on distal third of flexor surface. Mesosoma and petiole with very fine, short and dense whitish decumbent pubescence, setae longer than distance between them; head dorsum with sparse pilosity, distance between setae greater than their length. Gastral tergites with sparse decumbent pilosity, distance between setae greater or at most equal to their length. Eyes bare.
Measurements (in mm) and indices are given in the Tables 1, 2. Queens and males. Unknown. Etymology. The species is dedicated to Mr Hans-Werner Ribbecke (Mecklenburg-Vorpommern, Germany), who donated us the holotype specimen.   Wheeler, 1915, F. kutscheri Dlussky 2008, F. palaeopolonica Dlussky, 2008and F. zherikhini Dlussky, 2008 by the absence of erect or suberect setae on the tibiae and antennal scape. Among other species with bare tibiae and scape, F. ribbeckei Radchenko & Perkovsky, sp. nov. differs from F. radchenkoi Dlussky, 2008 by the absence of notch on the anterior clypeal margin. It resembles F. phaethusa Wheeler, 1915 by the 5-segmented maxillary palps, but its palps are much longer and reach the occipital foramen; additionally, it differs by the longer antennal scape and by the angulated anterior clypeal margin. New species differs from F. strangulata Wheeler, 1915 by the clypeal shape, the first funicular segment shorter, and the promesonotal and propodeal shape. Both F. flori and F. gustawi Dlussky, 2002  We examined approximately 100 Formica specimens of five species from Rovno amber in varying degrees of preservation: F. flori (38), F. gustawi (15), F. palaeopolonica (1), F. radchenkoi (2, including the holotype) and F. ribbeckei Radchenko & Perkovsky,sp. nov. (5). These represent half of the known amber Formica species.
The most abundant species in all ambers is F. flori, followed by F. gustawi. The latter clearly differs from F. flori by the dense decumbent pubescence on the gastral ter-gites. Thirty three specimens of F. gustawi have been recorded from all ambers (Dlussky 2002b(Dlussky , 2008Dlussky and Rasnitsyn 2009;Perkovsky 2016), but its separation with F. flori was not agreed upon by all researchers until 2002 when F. gustawi was described. Therefore, when previously examined Formica material will be re-evaluated, the abundance of F. gustawi may increase. The situation with F. ribbeckei Radchenko & Perkovsky, sp. nov. may be similar, as this species is very similar to F. flori, and we cannot rule out that F. ribbeckei Radchenko & Perkovsky, sp. nov. can be found in other late Eocene European ambers after re-investigation of corresponding material.
Nine fossil species originally assigned to genus Formica were first described from Baltic amber in the first half of 19 th century (Presl 1822;Holl 1829;Giebel 1856). However, the type specimens of these species have been lost, and based on their inadequate descriptions, it is impossible to attribute these species to Formica or any other ant genera. As a result, Mayr (1868) suggested that F. cordata Holl, 1829 belongs to Pheidole Westwood, 1839, and F. lucida Giebel, 1856 is not an ant, but a parasitoid Braconidae wasp, and Handlirsch (1907) proposed to consider most of them incertae sedis in Formicidae.

F. luteola
The body length of F. luteola is ca. 3.2 mm (too small for Formica), the body is light yellow ("Flavescenti-alba") and the head is rounded and somewhat cordate ("Caput rotundatum, fere cordatum"). These features are not known to be present in either extant or extinct Formica species. F. trigona is also too small for Formica (ca. 2.5 mm) with a triangular head ("Caput trigonium") and long antennae ("antennae longitudine capitis et thoracis"). These features raise doubts that this species should even be included to the family Formicidae. F. macrognatha is also quite small (ca. 3.7 mm), with short legs ("Pedes breves"), and its head is large and triangular ("Caput… magnum, trigonum"), precluding its placement in Formica.
Our data also confirm the exclusion of the following species from Formica: F. nigra and F. parvula are too small, with body lengths of 2.1 and 1.5 mm, respectively. F. gibbosa is too large: ca. 10.5 mm, its head is ovoid ("ovoideum") and the mesosoma is extremely elevated ("Thorax ovoideus, eximie elevates"). And, based on the original description, F. quadrata Holl, 1829 cannot be unambiguously assigned to any ant genus, and we agree with its exclusion from Formica.

Discussion
Accordingly, only ten Formica species are reliably known from late Eocene European ambers. Dlussky (2008) suggested that most of these species are quite specialized and only F. flori and F. gustawi can be considered the putative ancestors of extant Formica species. They belong to the F. fusca Linnaeus, 1758 species-group. Despite that F. ribbeckei Radchenko & Perkovsky, sp. nov. resembles F. flori, it has a reduced number of maxillary palps, and therefore we do not consider it to be one of the ancestors of the extant species of the F. fusca group.
The Formica fusca-group is considered ancestral to other extant Formica species (Dlussky 1967a;Radchenko 2016). Its representatives are characterized by a non-specialized head structure, mesosoma, petiole, maxillary palps, funicular segments, as well as a weakly developed polymorphism and a low level of social organization. Their gynes are large, which is associated with the independent founding of new colonies.
Formica radchenkoi most likely belongs to the more derived subgenus Raptiformica Forel, 1913, which includes slave-making species. Similarly, F. parexsecta Dlussky & Putyatina, 2014, described from the early Miocene deposits of Radoboj (Croatia), resembles extant species from the morphologically specialized subgenus Coptoformica Müller, 1923. Representatives of this subgenus have an emarginated occipital margin, short maxillary palps and often reduced body pubescence. Queens are quite small compared to those of the other Formica subgenera (Dlussky 1967a;Radchenko 2016). The emarginated occipital margin of the head allows an increase in the mass and strength of the mandibular muscles, which is associated with their nesting habits: workers cut grass to build nesting mounds. The small body size of gynes is associated with their biology: they are unable to independently establish a new colony and are temporary social parasites of Serviformica Forel, 1913 species (Dlussky 1967a).
Until recently, the oldest undoubted Formica species were known from the late Eocene. These are 10 species known from European amber: 8 species from Baltic amber and 7 from Scandinavian, Bitterfeld and Rovno ambers taken together. Cockerell (1920) described F. heteroptera from the Bagshot Beds in Bournemouth (Dorset, England) (Lutetian) based on a forewing, but the described features of this wing indicate that this specimen does not belong to Formica (Perfilieva, personal communication), from which it is formally excluded here. Théobald (1937) described three new Formica species and recorded impressions of F. flori (gyne) from the Upper Rhine Graben of Kleinkems (Baden-Württemberg, Germany) (latest Eocene); however, because the short antennal scape does not reach the occipital margin, this specimen differs not only from F. flori but also from all extinct and extant Formica species. The affinities of F. tripartita Théobald, 1937, F. alsatica Théobald, 1937and F. cf. septula Théobald, 1937 with Formica from late Eocene ambers remains uncertain. Interestingly, Formica ants have not been found in the Bembridge Marls, Isle of Wight, England (late Eocene) (Dlussky and Perfilieva 2014). This could be due to unsuitable environmental conditions at Bembridge for Formica and other epigeic ants (Radchenko and Perkovsky 2021).
Two new Formica species, F. biamoensis Dlussky et al., 2015 (worker) and F. paleosibirica Dlussky et al., 2015 (male), have recently been described from the latest Eocene deposit of the Russian Far East (Bol'shaya Svetlovodnaya) (Dlussky et al. 2015). F. biamoensis resembles F. flori and F. gustawi in general appearance, but the affinity of the second species to other known Formica is uncertain. Four species have also been recorded from the United States: F. eoptera Cockerell, 1923 (described based on a forewing) from Texas, and three species have been described based on winged gynes from Florissant, Colorado (latest Eocene) by Carpenter (1930). Although Carpenter (1930) doubted the attribution of F. eoptera to Formica and suggested considering it incertae sedis in Formicidae, Bolton (2021) concurred with Cockerell (1923). Based on the original description and illustration, there is no evidence to contradict the assignment of this species to Formica, and thus, we agree with Bolton.
The oldest species, assigned to Formica, F. arcana Scudder, 1877, was described by Scudder (1877) based on a forewing from the early Eocene of Quesnel (Okanagan Highlands, Canada), but Archibald et al. (2018) doubted that this species belonged to Formica or even Hymenoptera. However, the original description of the wing venation does not contradict the main features of Formica wing venation (Perfilieva, personal communication). Another undescribed Formica species was recently recorded from the early Eocene Palana Formation (India) (Patel et al. 2019). However, the interpretation of this specimen seems to be incorrect, and based on a photograph of the imprint (loc. cit., Fig. 10) we believe that this specimem does not exactly belong to Formica and, most likely it is not an ant at all.
Recently, Dlussky (2008) and Dlussky and Perfilieva (2014) indicated that there are some Formica compression fossils from the latest Ypresian Messel (47.8 Ma) and Lutetian Eckfeld (44.3 Ma) Lagerstätten (Germany). LaPolla and Greenwalt (2015) described males of F. annosa from the Kishenehn Formation (Lutetian, ca. 46 Ma) in northwestern Montana, USA. Despite that many important diagnostic details are usually unrecognizable from imprints (in contrast to amber fossils), the described males resemble extant males of the F. fusca-group in general appearance and visible features. This observation may serve as confirmation of its ancestral position in the genus.
Extant Formica species are one of the dominant groups of ants in the temperate humid zones of the Northern Hemisphere, where they mainly inhabit forests and meadows. Some species live in semi-arid regions (steppes, Mediterranean landscapes), but in deserts and semi-deserts, they occur only in intrazonal wet places. The distribution of many species extends beyond the Arctic Circle, and species can be found at elevations of 3600 m in the Alps, 3950 m in the Pamir, 4300 m in the North American Cordillera and even up to 4800 m in the Himalayas. Some species are also distributed in the mountains of tropical regions (Mexico, Taiwan, Myanmar), and F. fusca has been introduced and naturalized in Cuba (Dlussky 1967a;Gregg 1969;Radchenko 2016).
Ecologically, Formica is a keystone genus (together with Lasius Fabricius, 1804 and Myrmica Latreille, 1804) in the temperate zone of the Holarctic. Many species (e.g. members of the subgenera Formica s. str., Coptoformica, some Serviformica) dominate in ant communities (Dlussky 1967a;Zakharov et al 2013;Zakharov 2018). All Formica species are epigeic ants; they build nests in the ground, often with soil mounds, or make "ant-hills" from small pieces of branches, conifer needles or grass. They actively forage not only on the ground but also in the grass and on trees and bushes where they collect honeydew of aphids and other homopterans and prey on a variety of invertebrates.
It can be assumed that the extinct Formica species had a similar lifestyle. Homopteran insects were common by the late Eocene and are present in European ambers, and numerous syninclusions of aphids and ants have been recorded (Perkovsky 2006(Perkovsky , 2007(Perkovsky , 2008(Perkovsky , 2009(Perkovsky , 2010(Perkovsky , 2011aPerkovsky and Wegierek 2018). Most likely, workers of one of the most numerous amber ants, F. flori (as well as other Formica species), were trapped in resin during foraging on tree trunks when collecting honeydew.
The mixed character of the Baltic amber ant fauna, including both extant temperate and tropical genera, was already mentioned by Wheeler (1910Wheeler ( , 1915. The same can be said about all other late Eocene European ambers (e.g. see Dlussky and Rasnitsyn 2009). The composition of amber fauna can be explained by the climate of the late Eocene, which was equable and has no current analogues (Radchenko and Perkovsky 2021 and references therein).
The genus Formica most likely arose in the early Eocene, diversified in the late Eocene, and its species became dominant in ant communities. Then they continuously evolved to the present and have now become the dominant group of ants in the temperate zone of the Northern Hemisphere.