Two Eocene species of peacock mites (Acari: Tetranychoidea: Tuckerellidae)

Tuckerellidae are suctorial plant feeders with a wide host range, exploiting both aboveground and underground plant parts. A new species, Tuckerella weiterschani n. sp., is described from Eocene (ca. 40 mya) Baltic amber, and Tuckerella fossilibus Khaustov, Sergeyenko & Perkovsky, 2014 from Eocene Rovno (Ukrainian) amber is redescribed, each based on single holotype specimens. The two mites are morphologically distinct, with most striking differences concerning posterior, H-row of idiosomal setae. Tuckerella fossilibus has unusually large, obovate setae h1, while the other, setae in H-row are robust and fagellate with three different lengths. The new species, T. weiterschani, has small, narrowly lanceolate setae h1, while flagellate h2–h8 are fine, filiform and similar in length. Setae h1–h8 in the holotype of T. weiterschani are held erect and directed forward over the mite’s body, in a position reminiscent of the apparently defensive posture displayed by recent Tuckerella species. Tarsal setations of the two fossils and comparison with those of the recent Tuckerellidae advances discussion on the evolutionary significance of setal variation in acariform mites.


Introduction
Plant-feeding trombidiform mites -gall and rust mites, tetranychoid, some tydeoid, eupodoid and heterostigmatic mites -are presently ubiquitous on higher plants. Phytophagy was repeatedly acquired by different lineages of Trombidiformes, and some of these acquisitions might have happened as early as in the Mesozoic (Lindquist 1998). There is, however, very little direct evidence of this evolution, for the fossil record of phytophagous mites is extremely scant. The oldest Tetrapodili, four species of Triasacaroidea, are described from the Triassic (ca 230 mya) amber of South Dolomites in Italy (Schmidt et al. 2012;Sidorchuk et al. 2014). A middle/late Eocene (ca 35 mya) Eriophyidae specimen from Maslin fossil beds in South Australia is described as a species in a Recent genus, Aculops Keifer, 1966 (Southcott andLange 1971). Two possible spider mite species are described from Eocene (56 to 33 mya) Baltic amber by C.L. Koch in Koch and Berendt (1854): Panonychus gibbus (C.L. Koch, 1854) and Schizotetranychus? brevipes (C.L. Koch, 1854), but their morphology is only partly known (Oudemans 1937). One species of a peacock mite (Tetranychoidea: Tuckerellidae), Tuckerella fossilibus Khaustov, Sergeyenko, Perkovsky, 2014, is known from Eocene Ukrainian (Rovno) amber. It was described from a single poorly preserved specimen. A close preparation was not undertaken due to the risks involved, and although this resulted in its description being incomplete, there was not doubt as to the species identity. Recently, a new peacock mite fossil was found in Eocene Baltic amber by Thomas Weiterschan, who has kindly lent the specimen for our study. The mite, fossilized in a defensive posture characteristic for extant tuckerellids, appeared to represent a distinct species, new to science.
By this contribution, we aim to describe the new species, Tuckerella weiterschani n. sp., and augment the description of T. fossilibus after closer examination of its holotype.

Material
The Baltic amber piece with a single inclusion, representing a new species of Tuckerella, originally in the collection of Thomas Weiterschan (TW), collection identification number TW 1288. The piece is transferred to the collection of the Geological-Paleontological Museum and University of Hamburg (GPIH), and assigned a new collection number GPIH 4598. The piece is translucent, light yellow in colour. The fragment measuring 1x0.7x0.7 mm, containing the mite, was extracted and polished on six sides. For a review, and further data on the geology, of Baltic amber, see Weitschat and Wichard (2002). The Holotype of Tuckerella fossilibus is a Rovno amber piece, found in Klesov (Fedorovka) locality in Ukraine. The geology and fauna of this locality is reviewed by Perkovsky et al. (2010). The specimen is deposited in I.I. Schmalhausen Institute of Zoology, National Academy of Science, Kiev (SIZK), collection number KF-2237. A fragment, sized 2x1x0.5 mm, of the piece was mounted in the Buehler EpoThin epoxy resin between two 1 cm round coverslips. The fragment has a fracture that runs through the anterior part of the mite specimen. The preparation is labeled and glued with a hot-melt adhesive to a 1.5 mm-thick transparent plastic frame with a round aperture, allowing for observation from dorsal and ventral sides. The frame was also labeled.

Amber preparation
Amber pieces were cut and polished following protocols of Sidorchuk (2013), with tools described by Sidorchuk and Vorontsov (2016).

Imaging
All observations were done with light microscopy. Image stacks were obtained with Zeiss AxioImager A2 compound microscope with dry lenses and differential interference contrast (DIC) illumination with AxioCam 506 color digital camera. Bright field and polarized light micrographs were obtained with Nikon Eclipse 800 microscopes with 60x water immersion lens, Nikon DS-Fi1 and Nikon D7000 digital cameras. All images were corrected for light, color, noise and sharpness using Adobe Lightroom, then processed with Helicon Focus algorithm A, always with some detail manually transferred from the individual focal planes to the resulting images. These images are further referred to as "layered". Original image stacks are available through Figshare,doi:10.6084/m9.figshare.c.3743399. Drawings are made in Adobe Photoshop with stacks of the light micrographs as a background or in InkScape with a live capture from the Amscope MU900 camera mounted on Nikon Eclipse 800 microscope as a background (Sidorchuk and Vorontsov 2014).

Terminology
The terminology is that developed by F. Grandjean for acariform mites (see Travé andVachon 1975 for review andbibliography, andLindquist 1985 for application to Tetranychoidea); shapes of setae are described using botanical glossary by Beentje (2010). Idiosomal setae designations follow Beard et al. (2013), i.e. prodorsal setae are named sc 1 , sc 2 , v 1 , v 2 , setal row D contains 6 setae and row E contains 3 setae. There are three versions of the homologization of tuckerellid tarsal setae to those of the other acariform mites: (1) Quiros and Baker 1984, (2) Lindquist 1985, reproduced in Servin and Otero 1989 Bochkov et al. 2009. We accept here Lindquist's version, in which the primilateral setal pair (pl) is considered present in the larva of Tetranychoidea (including Tuckerellidae), the tectal setal pair (tc) and seta l 1 " are added in postlarval stages, dorsal seta d and antelateral setal pair (a) are absent. Leg setal counts are written from trochanter to tarsus as the number of tactile/eupathidial setae followed by the number of solenidia in parentheses. Leg setation formulae represent the cumulative data from both legs. For example, when on the right tibia III the setae l' and v' are visible, while on the left these are d and v', observed setal count is d+l'+v'=3 setae.
Comparative data on the leg setation was obtained from the following species and works: T. pavoniformis Ewing, 1922and T. ornata Tucker, 1926-from Baker and Pritchard 1953T. litoralis Collyer, 1969, T. japonica Ehara, 1975, T. anommata Smith Meyer and Ueckermann, 1997, T. channabasavannai Mallik and Kumar 1992, T. revelata Beard and Walter, 2005from the original descriptions of these species. Hypotheses on the setal homologies are ours, but for the latter species Jennifer Beard has generously supplied additional sketches with her setal homologies, confirming our hypotheses. In addition, Evert E. Lindquist (EEL) has kindly examined and sketched for us tarsi I and tarsus II of a male T. flabellifera Miller, 1964 and tarsus I of female T. ornata, kept at the Canadian National Collection of Insects, Arachnids and Nematodes, Ottawa); for these, setal homologies are by EEL.

Measurements
The measurements were taken from edge to edge of a sclerite or seta in its widest part, choosing where possible the structures oriented perpendicularly to the optical axis of the microscope. As the orientations were non-ideal, all measurements are to be treated as the minimal estimates, and are rounded to closest 5 if over 50. The unit is micrometer (μm).
Gnathosoma (Figs 1, 2, 4) -Only right palp clearly visible. Palpgenu dorsally with seta d, reaching just beyond tip of tibia. Palptibia with large, hooked tibial claw and three setae: d, slightly longer than palpgenu; short l", reaching to tip of tibial claw; medium-sized l', about as long as palptibia+tibial claw. Palptarsus completely covered by tibial claw in dorsal view, and elongate, roughly necktie-shaped as visible in lateral view, 14 long, phaneres not visible. Subcapitulum long and narrow, with alveolus of seta m visible basally and one pair of small peg-like adoral setae visible apically (Fig. 2, arrow). Peritremes prominent and emergent (Figs 1H,4A). Stylophore and cheliceral stylets not visible.
Holotype and type repository -Originally from private collection of Thomas Weiterschan, coll. # TW 1288, housed in the Geological-Paleontological Museum and University Hamburg (GPIH), collection number GPIH 4598. The holotype is the only inclusion in a rectangular light yellow amber piece measuring 1x0.7x0.7 mm, polished on six sides, stored in plastic O-ring-capped vial with water and preservative Thymol (Fig. 1A).
Type locality, occurrence and geological age -Type locality is Baltic Sea Coast, amber in Blue Earth fossil beds, dated late Eocene. Only known from the type specimen and locality.
Etymology -The new species is named in honor of Thomas Weiterschan, who discovered the specimen, promoted its study and assured its holotype deposition in a public repository.
Remark -The holotype and only known specimen of T. weiterschani has emergent peritremata, leg I tarsal setae l 1 " and leg I genual and tibial setae (l 1 ) present. These features suggest that the specimen is at least tritonymphal (Servin andOtero 1989, Beard andWalter 2005). In the presence of the tapered setae v 1 and lateral idiosomal setae and in the absence of clear view on the genital area, however, we cannot distinguish between an adult and a tritonymphal stage. Emergent peritremes in immature tuckerellid mites, not mentioned in the literature to our best knowledge, do occur at least in four Recent species starting from the deutonymphal stage (J. Beard, pers. comm.), and are thus inconclusive. The Diagnosis below is hence preliminary, and only uses those characters known to be relatively stable through the last two stages of the tuckerellid ontogeny.
Differential diagnosis -By the presence of seven pairs of subequal filiform setae h 2 -h 8 the new species is similar to five Recent species: T. coleogynis Jorgensen, 1967, T. elegans Rossi de Simons, 1972, T. eloisae Servin and Otero, 1989, T. hypoterra McDaniel and Morihara in McDaniel et al. 1975and T. spechtae Womersley, 1957. The new species differs from T. spechtae and T. hypoterra by the relative positions of setal pairs f 1 and f 2 , f 1 being inserted clearly posterior to f 2 in T. weiterschani, as opposed to slightly anterior in these two species; in addition, T. spechtae has setae h 6 shorter than the other flagellate setae. From T. eloisae the new species differs by its much smaller size (total body length -including gnathosoma -310 in T. weiterschani vs. 480 in tritonymph, 473 -566 in adults of T. eloisae). T. elegans, only slightly larger than T. weiterschani (adult idiosomal length 342 -370 in T. elegans vs. 275 in T. weiterschani), and T. coleogynis differs from the new species by its setae v 1 having a smooth edge and uneven reticulation vs. these setae widely lancetiform, with tapered tips in T. weiterschani (the validity of this character is only certain if T. weiterchani holotype is an adult female), and having setae h 2 -h 8 as long as body, vs. at least 1.5 times body length in T. weiterschani. T. coleogynis, whose size is unknown, is the most similar to T. weiterschani, differing in the shape of dorsolateral idiosomal setae (only significant if T. weiterchani holotype is an adult female) and distribution of setulae on the setae h 2 -h 8 : present only in basal part in T. coleogynis, evenly distributed throughout the length in T. weiterschani. Idiosomal dorsum (Figs 5, 6, 7B) -Cuticular ornamentation, where visible anterodorsally, irregularly reticulate (Fig. 5A). Number of dorsal idiosomal setae as in T. weiterschani n. sp. Setae v 2 , sc 1 and sc 2 fan-like (Fig. 5C, D); v 1 considerably different from other prodorsal setae, with smooth margins and weakly tapered tip (Figs. 5D, 7). Two pairs of eyes located anterolaterally to bases of setae sc 1 ; posterior eye distinctly larger than anterior (Fig. 7B, oc). All setae of C, D, E and F-rows broad, flat, ovate to obovate in shape: c 1 -c 3 , d 1 -d 3 and e 1 , e 2 of medium size, subequal (16 -18 x 24 -28); f 1 and f 2 slightly smaller than e 1 and e 2 ; f 2 situated on same transverse level as f 1 . Lateral setae c 5 -c 7 , d 4 -d6 and e3 distinctly larger than central, more elongate, with rounded tips, largest d 6 28 x 60. Setae h 1 large, flat, obovate (27 x 55), setae h 2 -h 8 flagellate: h 2 , h 4 , h 5 , h 7 long, barbed, tapered (180, 270, 180, 220 respectively),  h 6 short, thick, heavily barbed, blunt (40); h 3 , h 8 intermediate in length and thickness, barbed, tapered (80). Only anterior margins of dorsal shields C and D clearly visible (Fig. 6).
Remarks - Khaustov et al. (2014) in original description of T. fossilibus were not able to identify its life stage. Now, after more detailed description we can suggest that it is probably a deutonymph. Our opinion is based on the setation of leg I, which shows: (1) two tarsal solenidia, (2) absence of the tarsal seta l 1 ', (3) three obovate setae on each genu and tibia. According to observations by Quiros-Gonzalez and Baker (1984), Lindquist (1985), Beard and Walter (2005) and Beard et al. 2013, although there is some interspecific variation, the second tarsal solenidion is deutonymphal, tarsal seta l 1 ' is tritonymphal, and three obovate setae on first genu and tibia are only present in deutonymphs: known protonymphs are found to have 1-2 obovate setae per segment, while tritonymphs and adults have five.
Addition to the differential diagnosis -Certain caution is necessary when using this comparison, for the immature specimen is being compared to adults. Based on the available descriptions (Quiros-Gonzalez and Baker 1984, Lindquist 1985, Beard and Walter 2005, Beard et al. 2013, ontogenetic changes of the opisthonotal setation may affect setae v 1 (in T. revelata they are obovate, with rough edge, in larva, with outline smoothing out in the subsequent stages), leaf-shaped C-H-row setae (relatively narrower and more tapered in the immatures) and flagellate H-row setae (becoming relatively shorter during ontogeny). T. fossilibus is so far unique tuckerellid species known to have obovate setae h 1 similar in size to the lateral setae of C-, D-and E rows. By the shape of proterosomal setae v 1 and the relative shortness of setae h 1 and h 6 , deutonymphal T. fossilibus is similar to adults of T. pavoniformis (Ewing, 1922), T. indica Prasad, 1973 andT. channabasavannai Malik andKumar, 1992. It differs from them by much larger setae h 1 and by the differentiated shapes of the other H-row setae (T. pavoniformis has setae h 2 -h 5 , h 7 and h 8 all similar, whip-like (Baker and Pritchard 1953), as do T. indica and T. channabasavannai). By the shape of setae v 1 T. fossilibus is also similar to T. japonica Ehara, 1975 andto T. kumaonensis Gupta, 1979, but these species have three pairs of comparatively small leaf-like setae in h-series.
Holotype container -Amber fragment with holotype, sized 2x1x0.5 mm, is mounted in the Buehler EpoThin epoxy resin between two 1 cm coverslips. The preparation is glued with a low-temperature hot-melt adhesive to a 1.5 mm-thick transparent plastic frame (Fig. 5A).

Discussion
With the restricted number of known fossils of the phytophagous mites, the discovery of a tuckerellid specimen in Baltic amber appears lucky. Unlike saprophagous mites, whose recently redescribed species are the same for the two amber deposits (Sidorchuk and Norton 2010;2011a, b;Sidorchuk and Klimov 2011), the tuckerellids from Rovno and Baltic ambers are not conspecific. Their dissimilarity is indicative of a morphological diversity in Eocene (ca 40 mya) tuckerellids comparable to that found among Recent species. It is difficult, though, to discuss the groups of species in Tuckerellidae, for the family is in much need of a thorough revision, as most species are only cursorily illustrated. We can only attempt some general implications based on the morphological similarities discussed in Remarks, as follows.
Ethological implication: the defensive posture. Holotype of T. weiterschani is captured in the posture, first described as defensive by Ochoa (1989) in Tuckerella knorri Baker and Tuttle, 1975. When disturbed, a tuckerellid mite makes a few quick movements, alternately raising and lowering its 'tail' of the h-setae. The whip-like setae of the H-row are often long enough to cover the whole body, legs and gnathosoma of the mite, possibly disturbing an intruder. Our finding shows (Figs 1A-C) the defensive adaptation of a peacock mite dates back at least to the Eocene.
Ecological and geographical implications. Tuckerellidae are suctorial plant feeders, like other tetranychoid mites. Recent species are known to exploit both aboveground and underground parts of the plants, including bark and roots (Jorgensen 1967, McDaniel et al. 1975, Beard et al. 2013. The host range of even an individual species may include multiple plant species, not necessarily closely related: for example, T. japonica is known from Ulmaceae, Theaceae, Actinidiaceae, Ebenaceae, Ericaceae, Rhamnaceae and even coniferous Juniperaceae (Beard et al. 2013). Tuckerellidae are found in climes from humid tropical through Mediterranean and hot desert, but also in cold arid areas. Their presence in amber, therefore, is not informative with regard to the amber forest composition or climate.
Evolutionary implications: setal variation. Our specimens appear to have asymmetrical setation at least on tarsi I (Figs 4, 7, 8): KF-2237 (deutonymph) does not show seta pl' on the left (Figs 7B, 8C), while this seta is present on the right (Figs 7A, 8D), and GPIH 4598 (tritonymph or adult) has at least an alveolus in l 1 " position on the right leg (Figs 4A', 8A), while there is no such seta or alveolus on the left (Figs 4A, 8B). This encouraged us to review what is known on the tuckerellid tarsi (see Material and Methods). Below, we only consider tarsi I, and only the setae (pl), (ft) and l 1 ", for they are the best visible and the least likely overlooked. It appeared that these setae vary within modern tuckerellid species. In T. ornata, seta l 1 " is present in the specimen examined by E. Lindquist, absent in Baker and Pritchard 1953 Fig. 6b. In T. revelata, variable setae appear to be ft' (absent in female, Fig. 4A by Beard and Walter 2005, present in all other illustrated immature and adult legs) and pl' (absent in protonymph, Fig. 12A by Beard and Walter 2005). Note that although Beard and Walter 2005 give setal formulae 14(2) for adult and tritonymph and 13(2) for deutonymph tarsus I in the text, they illustrate 12 setae plus 2 solenidia for the adults and tritonymph and 11(2) for the deutonymph, which is in accordance with the otherwise known maximum complement in this family. The discrepancy is due to the authors' counting for the solenidia twice: (1) in the total count of phaneres and (2) in parentheses (J. Beard, pers. comm., June 2017). Seta pl' also is variable in T. anommata: present according to the counts, illustrated for the male, but lacking in the illustrated female left tarsus, where the remaining seta pl" has an almost dorsal position (Smith Meyer and Ueckermann 1997, Figs. 89 and 85, respectively).
Each of them being admittedly questionable, these observations considered together evoke a long-standing discussion on the possible link between the variation and evolution of leg setae in mites. There are two notions on this link. Based on his 46-year studies on Acariformes, including rearing experiments on parthenogenetic oribatid mite Platynothrus peltifer, Grandjean (1974) formulated the hypothesis of the RVA -regressive evolution (of a seta in the simple row within a lineage) by the alternating vertitions. Unilateral absence of a seta, according to this hypothesis, is an indication of its future disappearance, and the frequency of such absences should grow, albeit very slowly, in a lineage of many generations. Wauthy et al. (1991) brought to acarology another notion, that of the fluctuating asymmetry (Ludwig 1932, Van Valen 1962 -the phenomenon common for bilateral organisms, variation resulting from the random developmental accidents and having no evolutionary implications. In this model, the frequency of absence of a particular seta is not indicative of its future loss, and this parameter for a particular seta is not inheritable. The latter notion is developed and substantiated with new observations and statistical analysis (see Wauthy et al. 1998, Leponce et al. 2001, studying three strains of the haplodiploid prostigmatan Tetranychus urticae, although each with two generations (mothers and sons) only. Although their results were statistically significant, this evolutionary experiment was very far from analyzing 'hundreds, maybe hundreds of thousands of generations' that Grandjean (1974) considered necessary for a seta to become lost by alternating vertitions. He (Grandjean, 1974: 779) has actually predicted that no notable change of the vertition frequency would be statistically observable within ten or twenty generations of a parthenogenetic lineage -thus, he indicated that the RVA hypothesis cannot be rejected by a feasibly long artificial experiment.
Paleontological evidence somehow shows us the result of an unfeasibly long evolutionary experiment. Although they are not linked by a direct lineage from mother to daughter, the two fossils described here on one hand and the modern tuckerellids described in the literature on the other hand are random specimens within the same genus, separated by at least 40 million Figure 8 Leg I setation of two Eocene Tuckerella specimens (layered DIC micrographs). A-B, T. weiterschani holotype: A -right leg abaxially, note at least alveolus of tarsal seta l 1 " present; B -left leg I abaxially, note setae visible but l 1 " apparently absent; C-D, T. fossilibus holotype deutonymph: C -left leg ventrally, note setation well visible but tarsal pl' apparently absent; C' part of C -individual focal plane showing tibial solenidion and seta l"; D -right leg, tarsus dorsally, note seta pl' present.
generations. These specimens seem to show variations of the same, eustatic, setae, which may mean that the frequency of their absences, or vertitions, is high in Recent tuckerellids and was already high in the Eocene ones. In the light of RVA, suggesting a slow but steadily augmenting growth of the frequency of vertitions in the course of evolutionary time, presence in the Recent Tuckerella of the setae commonly hit by vertitions 40 million generations ago in the same genus would be extraordinary. In the light of the fluctuating asymmetry notion, such presence does not seem overly surprising: high frequency of the asymmetrical expression of some leg setae may be part of the stability of the tuckerellid habitus itself. Fossils show us that the body plan of Tuckerella was already fully formed forty million years ago, and it apparently kept existing without notable change since that time. This remarkable stasis possibly included the degree of imperfection of the mite body's symmetry. To learn whether this is true or not, and whether this is widespread in Acari, careful studies of leg setations in Recent and fossil mites are necessary.