Extending the footprint record of Pareiasauromorpha to the Cisuralian: earlier appearance and wider palaeobiogeography of the group

Pareiasauromorpha is one of the most important tetrapod groups of the Permian. Skeletal evidence suggests a late Kungurian origin in North America, whereas the majority of occurrences come from the Guadalupian and Lopingian of South Africa and Russia. However, Pareiasauromorpha footprints include the ichnogenus Pachypes, which is unknown from strata older than late Guadalupian. A revision of several Pachypes‐like footprints from the Cisuralian–Guadalupian of Europe and North America confirm the occurrence of this ichnogenus and of the ichnospecies Pachypes ollieri comb. nov. beginning in the Artinskian. This is the earliest known occurrence of Pachypes and it coincides with the Artinskian reptile radiation. Based on a synapomorphy‐based track–trackmaker correlation, P. ollieri can be attributed to nycteroleter pareiasauromorphs such as Macroleter. Therefore, the earliest occurrences of pareiasauromorph footprints precede by at least 10 myr the earliest occurrence of this group in the skeletal record. Moreover, the palaeobiogeography of the group is extended to the Cisuralian and Guadalupian of western Europe.

Pachypes and it coincides with the Artinskian reptile radiation. Based on a synapomorphy-based track-trackmaker correlation, P. ollieri can be attributed to nycteroleter pareiasauromorphs such as Macroleter. Therefore, the earliest occurrences of pareiasauromorph footprints precede by at least 10 myr the earliest occurrence of this group in the skeletal record. Moreover, the palaeobiogeography of the group is extended to the Cisuralian and Guadalupian of western Europe.  (Lucas 2006(Lucas , 2018Tsuji & M€ uller 2009;Tsuji et al. 2012). However, few occurrences of this group are known from low Pangaean palaeolatitudes (i.e. western Europe, North Africa and the USA), and only a single occurrence is known from Cisuralian strata (Lucas 2002(Lucas , 2006, although different interpretations of the age of these strata exist (Reisz & Laurin 2001). Also, the origin of the group is poorly understood and constrained (Tsuji & M€ uller 2009). An independent source of information about this group may come from the study of Pareiasauromorpha tracks. The only ichnogenus with a reliable attribution to Pareiasauromorpha is track-trackmaker correlation with the possible pareiasauromorph producers, following the methodology of Voigt et al. (2007) and Marchetti et al. (2017a) for a synapomorphybased trackmaker attribution. The aim of this study is to provide an anatomy-consistent ichnotaxonomic revision of this material with the evaluation of previously undescribed specimens from Arizona, Texas, Spain and France. This includes a detailed track-trackmaker correlation and the discussion of the implications for the evolution and palaeobiogeography of Pareiasauromorpha.

MATERIAL AND METHOD
The material includes both tetrapod footprints currently in museum collections and in situ specimens in the field. All the specimens were studied first-hand and photographed perpendicular to the trampled surface and with adequate light conditions (oblique light, artificial or natural) and a metric scale. The photographs included trackway sections, manus-pes couples and single imprints. Outline drawings on transparent films were also obtained for some specimens. The best-preserved footprints were also photographed to obtain 3D models by using the photogrammetry technique. The 3D models were built using the software Agisoft Photoscan Professional (v.1.4.0); meshes were edited with the software MeshLab (v.2016.12), and contour lines and colour depth maps were obtained using the software Cloud Compare (v.2.8 beta) and ParaView (v.5.0.1). 3D models were also obtained using a Next Engine 3D Scanner. Specimens were selected based on the quality of morphological preservation and the preservation scale sensu Marchetti et al. (2019c). The 3D models were uploaded to the digital repository MorphoSource (Marchetti et al. 2020b). The selected material was digitally measured with Gimp (v.2.8.16), and the measurements are reported in Tables 1-3.

SYSTEMATIC PALAEONTOLOGY
Ichnogenus PACHYPES Leonardi et al., 1975 Type ichnospecies. Pachypes dolomiticus Leonardi et al., 1975 Type strata and locality.  Description. Semiplantigrade and pentadactyl footprints of a quadrupedal tetrapod. The pes imprint is about 30-100 mm long, it is about as long as wide, and clearly ectaxonic (digit IV impression is the longest; 19-60 mm long). Digit impressions are thick and end in rounded tips. The digit I-IV impressions are closely-grouped, superimposed at their bases and may be distally bent inwards (medially-oriented). The digit V impression is relatively short (10-22 mm long, about as long as I, 6-29 mm long), thin, in a proximal position and may be distally bent outwards (laterally oriented). The sole impression is elliptical and wider than long. Digit terminations and the medial (inner) side of the footprint are more deeply impressed than the digit III-V bases, sometimes resulting in a crescent-moon shape in incompletely impressed tracks. The manus imprint, wider than long, is slightly smaller and more deeply impressed than the pes imprint. The manus footprint depth is more equally distributed than in the pes imprints. Digit impressions are extremely short and tightly-grouped and superimposed for most of their length, and they have rounded terminations. Manus imprints are slightly ectaxonic to mesaxonic with external digit imprints slightly shorter (especially digit V impression), but the lengths of digits II-IV are remarkably similar. The digit I imprint may show a rounded basal pad. The digit V imprint is in a proximal position. The palm impression is elliptical and wider than long. The trackway pattern is characterized by a simple alternating arrangement of manus-pes couples with a short manus-pes distance and long stride length and width of pace compared to the pes length (SLp/FL and WPp/FL ratios of 7.5 and 4.1 on average, respectively). Despite the long stride, the pace angulation is relatively low (about 70-100°for the pes and 80-110°for the manus). The pes imprint is aligned with the midline, and the manus imprint is in front of the pes imprint, slightly medially positioned and slightly rotated inward respect to the pes. No tail/body impressions have been reported for this ichnogenus, and digit drag traces are rarely observed.
Remarks. The first ichnotaxonomic description of this material was by Ellenberger (1983a), who introduced the ichnotaxon 'Paranomodontipus' ollieri with a short description and drawings of five different trackways from the La Lieude site in southern France (La Lieude Formation; Ellenberger 1983a, fig. 2). These trackways were later indicated with trackway numbers 3, 4, 9, 10 and 11 by Gand et al. (2000). A holotype was not designated, so these five trackways are syntypes, according to the ICZN (1999). Trackway 9 is relatively better-preserved (maximum grade 2.5 for both the pes and the manus tracks), it is also the holotype of the ichnospecies 'Lunaepes ollierorum' Gand et al., 2000 and a plaster cast of the best-preserved pes-manus couple ( Fig. 1B) is stored in a repository (UM-LOD 16). We consider trackway 9 as the lectotype ( Fig. 1) and trackways 3, 4, 10 and 11 as paralectotypes (maximum grade 2.0 for both the pes and the manus tracks) (Figs 2, 3). Another more digitigrade ichnotaxon, 'Theriopodiscus lieudensis', was based on a morphologically-similar trackway from the same locality (Ellenberger 1983a, fig. 10b). We consider 'Theriopodiscus lieudensis' to be a junior synonym of 'Paranomodontipus' ollieri because the overall morphology and trackway pattern are identical in these two ichnotaxa, whereas the digitigrade impressions are a common feature in the type material of both ichnotaxa. In fact, transitions between these two differently impressed morphologies were observed along the same individual trackway (Fig. 2G).
Because these two ichnotaxa were clearly described and illustrated (Ellenberger 1983a, pp. 12-13, 15, 17, 20), we think that the requirements of the ICZN (1999) were fulfilled, so these names are not nomina nuda as suggested by Gand et al. (2000). Indeed, as part of his description of P. ollieri, Ellenberger (1983a, pp. 13-14) provided statements that compare and contrast the ichnotaxon that we take to state 'characters that are purported to differentiate the taxon'; thus, P. ollieri is available according to the mandates of ICZN Article 13.1.1.
Considering P. ollieri to be a nomen nudum, Gand et al. (2000) used one of the syntypes of P. ollieri as the holotype of their new ichnospecies 'Lunaepes ollierorum'. However, given that P. ollieri Ellenberger, 1983a was not a nomen nudum, but an available name, the ichnospecies 'Lunaepes ollierorum' is simply an objective junior synonym of P. ollieri.
The ichnotaxon 'Paranomodontipus' ollieri shares numerous features (such as the thick parallel digits and the short digit V impression) with the ichnotaxon Pachypes dolomiticus Leonardi et al., 1975 (Fig. 4A), introduced for an isolated but complete and well-preserved pes imprint from the Lopingian Arenaria di Val Gardena Formation of the Dolomites (Southern Alps, Italy). Subsequent studies refined the diagnosis of this ichnotaxon, based on a large number of additional specimens from the type locality, including well-preserved pes-manus couples, isolated footprints, a few incomplete step cycles and a trackway (Conti et al. 1977;Valentini et al. 2008Valentini et al. , 2009Marchetti et al. 2019a). The overall morphology and proportions of 'Paranomodontipus' ollieri are consistent with Pachypes dolomiticus (Fig. 4A) with the exception of the digit thickness and superimposition, the trackway pattern and the maximum footprint size (Tables 1-3 We consider these differences adequate for a differentiation at the ichnospecies level, but not sufficient to distinguish two ichnogenera (note also that size is not a valid ichnotaxobase). Therefore, we synonymize the two ichnogenera and propose the new combination P. ollieri (Ellenberger, 1983a) for the material previously assigned to 'Paranomodontipus' ollieri and its junior synonyms 'Theriopodiscus lieudensis' and 'Lunaepes ollierorum'.
This ichnospecies differs also from Pachypes loxodactylus (Fig. 4C), described from the Lopingian Locharbriggs Formation of Scotland and revised by Marchetti et al. (2019b). In fact, P. ollieri comb. nov. is characterized by relatively shorter and thicker digit impressions of the pes and a different trackway pattern compared to P. loxodactylus.
Some specimens from the Cisuralian of the USA (Hermit and San Angelo formations) and Europe (Rabejac Formation, southern France and Peranera Formation, Lower Red Unit, Catalan Pyrenees, Spain) and the Guadalupian of France (Le Muy Formation) show features similar to P. ollieri comb. nov.
The specimen GRCA 3172-3173, from the Hermit Formation of Grand Canyon National Park (Arizona, USA), includes a trackway with three consecutive pes-manus couples (Fig. 5A-E;Marchetti et al. 2020a). The morphology of the pes is similar to Amphisauropus because of the medial-lateral decrease in relief and the thick and straight digit impressions with rounded terminations, but the manual tracks are evidently different, because the digit impressions are extremely short, and the central digit impressions show a very similar length. The morphology of the manus imprint is similar to incompletely-impressed Dimetropus tracks (Voigt 2005), because the semi-circular arrangement of digits resembles the semi-circular arrangement of basal pads of Dimetropus. However, no claw or digit-tip impressions were registered, so they probably represent complete digits, not just basal digital pad impressions. Moreover, the pes impression is completely different from Dimetropus stout, short and superimposed digits and the trackway pattern are instead consistent with P. ollieri comb. nov.
An incomplete step cycle, including two consecutive pes-manus couples from the San Angelo Formation of Texas (Lucas & Hunt 2005, fig. 8e-f), was tentatively assigned to cf. Amphisauropus isp. (Fig. 5F, G). However, the digit impressions of the manus are very short, and the pes digit impressions are thick, closely-grouped and distally bent inward (medially). Also, the typical marked impression of the digit I base of Amphisauropus (Voigt 2005) is absent, and the trackway pattern is different, because the pes imprints are not markedly rotated outwards. Therefore, this specimen is not assignable to Amphisauropus. All the morphological traits and the trackway pattern are instead consistent with P. ollieri comb. nov.
Specimens from the Peranera Formation (Lower Red Unit) of the Catalan Pyrenees (north-eastern Iberian Peninsula) (Fig. 6) previously referred to cf. Amphisauropus (including IPS73723, the mould and replica IPS82605 and other in situ tracks) and others identified as indeterminate tracks (Mujal et al. 2016, figs 5, S2d-l, S3, S4) also resemble P. ollieri comb. nov. These tracks are generally wider than long, and display very short and round digit impressions (increasing in length from I to IV, with a shorter digit V, though not always impressed), which are often deeper than the sole/palm impressions, different from Amphisauropus. We re-assign all these footprints to P. ollieri comb. nov.
The specimen UM-LOD 96 from the Rabejac Formation (Lod eve Basin) of southern France (Fig. 7A, B), including an isolated left pes-manus couple, was assigned to the supposed therapsid track 'Eotheriopodiscus lenis' by Ellenberger (1983b). Because this ichnotaxon was not described but only illustrated and listed, we consider it to be nomen nudum in agreement with the ICZN (1999). Some similar material from the same formation (MNHN.F.LOD142-145) was later listed and illustrated (Heyler & Gand 2000, fig. 30), but not assigned to an ichnotaxon. These specimens include a left pes-manus couple (MNHN.F.LOD142) (Fig. 7C, D) and some plaster  NN 1;  Fig. 7F, G) includes an incomplete step cycle with a manus and two consecutive pes-manus couples, and a larger pes-manus couple assignable to Tambachichnium isp. The incomplete step cycle on MF-NN 1 and the pesmanus couples of the specimens UM-LOD 96 and MNHN.F.LOD142-145 show morphological traits consistent with P. ollieri comb. nov., including stout, short and superimposed digit imprints and the trackway pattern.
Some specimens found in the Le Muy Formation (Fig. 7H), close to the transition between the Le Mitan and Le Muy formations, initially assigned to Limnopus (Demathieu et al. 1992), are also assignable to P. ollieri comb. nov., because of the closely-grouped pes digit imprints and the morphology of manus digit imprints, which are very short and subequal in length (Fig. 7H).
Remarks. The ichnotaxon 'Lunaepes fragilis' Gand et al., 1995, based on a trackway from the Guadalupian Pradineaux Formation in the Provence basins, is characterized by a peculiar pes imprint morphology resembling a crescent moon. Although the holotype is poorly-preserved (the digit impressions are mostly not visible, and the footprints are not completely impressed) a few couples (Marchetti et al. 2020c) show the typical morphological traits and trackway pattern of P. ollieri comb. nov., such as the thick, parallel and superimposed digits (more than P. loxodactylus and less than P. dolomiticus) and the wide trackways with well-spaced pes-manus couples. Due to the poor preservation of the holotype, we consider 'Lunaepes fragilis' a nomen dubium and assign this material to cf. Pachypes isp.

TRACKMAKER ATTRIBUTION
The material herein assigned to P. ollieri comb. nov. has received different trackmaker attributions. Ellenberger The markedly ectaxonic pes of these specimens with a relatively short sole is not consistent with a dicynodont therapsid producer, because they are characterized by a broad tarsus and mesaxony. Moreover, the deeper medial part of the pes impression is clearly inconsistent with all the widely-recognized synapsid tracks of the Permian (Brontopus, Capitosauroides, Dicynodontipus, Dimetropus, Dolomitipes and Karoopes) that show instead a deeper lateral part of the tracks (Marchetti et al. 2019e; Mujal et al. 2020). Also, the characteristic semi-circular basal pad arrangement of digits well-separated from the digit tips and forming paw-like impressions observed in therapsid tracks does not occur in material here assigned to P. ollieri comb. nov. Therefore, an assignment to synapsid producers is excluded.
In contrast, the Lopingian material assigned to the ichnogenus Pachypes has been consistently attributed to pareiasaurian producers, because of the parallel/superimposed digit impressions and the marked reduction of external digits (Leonardi et al. 1975;Conti et al. 1977;Valentini et al. 2008Valentini et al. , 2009Marchetti et al. 2019a, b). However, no pareiasaurs are known from Laurasia during the late Cisuralian to Guadalupian interval. Guadalupian pareiasaurs are known only from South Africa (e.g. Bradysaurus, Embrithosaurus, Nochelosaurus;Day et al. 2015) and Brazil (e.g. Provelosaurus;Ara ujo, 1985) but, with the exception of Provelosaurus (body length of 1.0 m), their size is too large (body length of 2.5-3.0 m) to have been the producers of P. ollieri comb. nov. Also, the phalangeal formula of pareiasaurs is very reduced (manus 2-3-3-3-2; pes 2-3-3-4-3; e.g. Turner et al. 2015) and this is inconsistent with the material assigned to P. ollieri comb. nov.
Nevertheless, several small-sized non-pareiasaur pareiasauromorphs are known from the upper Cisuralian to lower Lopingian of Laurasia (Sennikov & Golubev 2017; Lucas 2018) matching relatively well the stratigraphic distribution and size of P. ollieri comb. nov. These taxa are generally known as nycteroleters and include: an indeterminate nycteroleter from the Cisuralian of Texas ( Only Emeroleter, Macroleter (Figs 8,9) and Rhipaeosaurus have well-preserved and relatively complete appendicular skeletons. The pes of Rhipaeosaurus is relatively similar to the pareiasaur pes because of the robust digits and the ectaxony (Efremov 1940), although the phalanges are more elongated, and the phalangeal formula is more primitive: 2-3-4-5-4. The ectaxonic condition is marked, and digit V is shorter than digit III, and in a proximal position. Digits I-IV are closely-packed, and the metatarsals are overlapped. The metatarsals decrease in thickness from digit I to digit V. The distal tarsals are robust and angular, especially distal tarsals I-IV. The astragalus and calcaneum are compact but not fused in a single element (the astragalocalcaneum), as in pareiasaurs (Tsuji et al. 2012). The pes is about 100 mm long. The limb bones are robust but slenderer and more elongate than in pareiasaurs. In a similar way, the vertebral column is also relatively longer. The tail is very thin and short, the manus is not known (Efremov 1940).
The pes of Macroleter (Figs 8A-D, 9B) is similar to the pes of Rhipaeosaurus. The ectaxony is marked, and the digit V is very short and in a proximal position. Digits I-IV are closely-packed, and the metatarsals are overlapped. The metatarsals decrease in thickness from digit I to digit V and increase in length between digits I-IV. The phalangeal formula is 2-3-4-4-2 (Fig. 9B). The distal tarsals are robust and angular, especially distal tarsals I-IV. There is also a fused astragalocalcaneum, as in pareiasaurs. The pes is about 60 mm long.
The manus of Macroleter (Figs 8E-H, 9B) is relatively smaller than the pes (about 40 mm long). The metacarpals are thick and increase in length between metacarpals I and IV and show a marked overlap. Digits II-IV are larger and more robust than digit I; digit V was not observed. The phalangeal formula is probably similar to that of the pes: 2-3-4-4-?2 (Fig. 9B). The distal carpals are robust and angular, so that the carpus structure is compact.
Emeroleter presents a pedal phalangeal formula of 2-3-4-5-3. As in pareiasaurs and Macroleter, the astragalus and the calcaneum are fused in adults and the pes is considerably larger than the manus (Tsuji et al. 2012). The metapodium and acropodium elements are slenderer than in Macroleter and therefore differ from those observed in pareiasaurs. Metapodial elements are longer and slenderer than the phalanges. The manus probably has the same number of phalanges as the pes (Tsuji et al. 2012).
The hind limb bones of Emeroleter and Macroleter are slenderer and more elongate than in Rhipaeosaurus, while the front limbs of Macroleter are quite robust (Fig. 9A). Also, the vertebral column of Emeroleter and Macroleter (Fig. 9A) is longer than in Rhipaeosaurus, compared to the total body length. The glenoacetabular length/pes length ratio of Macroleter is 4.6 (Fig. 9A), which is close to the average value of 4.9 for calculated body length/foot length ratio measured from the trackways of P. ollieri comb. nov. (Table 3), whereas Emeroleter and Rhipaeosaurus have lower ratios (3.54 and 3.65, respectively). The ectaxonic pes of Rhipaeosaurus, Macroleter and Emeroleter is consistent with P. ollieri comb. nov. The phalangeal formula of Macroleter (2-3-4-4-2) is more consistent with the manus digit imprint proportions of P. ollieri comb. nov. than that of Emeroleter (Fig. 9B). The thick, overlapping metatarsals and metacarpals, increasing in length and decreasing in thickness between digits I-IV, and the compact structure of the tarsus and carpus observed in Rhipaeosaurus and Macroleter are consistent with P. ollieri comb. nov., because the digit imprints are separated only distally, and the medial part of the footprints is the most deeply impressed (Fig. 9C). The pes size, the relative length of pes digit V, the more robust fore limb and the elongated limbs of Macroleter are also consistent with P. ollieri comb. nov., because of the digit proportions of the pes imprints, the very high SLp/FL and WPp/FL ratios and the more deeply impressed manus imprints ( Fig. 9;  Tables 1, 3). Also, the stratigraphic distribution of Macroleter . Therefore, we attribute P. ollieri comb. nov. to nycteroleter pareiasauromorphs such as Macroleter, although presently we cannot exclude forms such as Rhipaeosaurus as producers of some of the analysed tetrapod footprint material. Because of the slender and elongated pes structure and the same stratigraphic distribution (lower Wuchiapingian), Emeroleter is instead a possible producer of Pachypes loxodactylus.

TRACK BIOSTRATIGRAPHY
The ichnotaxonomic revision of tracks here assigned to P. ollieri comb. nov. substantially extends the stratigraphic record of the ichnogenus, previously known exclusively from units of Guadalupian or Lopingian age.  et al. 2003;Surkov et al. 2007;Buchwitz et al. 2017;Marchetti et al. 2019a;Mujal et al. 2017). All these units are considered to be Guadalupian or Lopingian in age (Schneider et al. 2020). Therefore, these occurrences belong to the late Erpetopus biochron (mostly or entirely post-Brontopus sub-biochron) and to the Paradoxichnium biochron (Voigt & Lucas 2018;Marchetti et al. 2019b). Conversely, the stratigraphic range of P. ollieri comb. nov. is from the Artinskian (represented by the Hermit Formation, Arizona; the Rabejac Formation, France; and the Peranera Formation, Spain) to the lower Capitanian (Le Muy Formation; France), therefore it belongs to the older part of the Erpetopus biochron and to the Brontopus sub-biochron (Voigt & Lucas 2018;Marchetti et al. 2019b;Schneider et al. 2020) (Fig. 10). The FAD (first appearance datum) of P. ollieri comb. nov. is in the upper Artinskian of the Rabejac Formation of France, age constrained by radiometric ages (Schneider et al. 2006;Michel et al. 2015). This is in agreement with the late Artinskian age of the Hermit Formation of Arizona inferred from marine biostratigraphy of laterally-correlated and bracketing units (lower Leonardian;Blakey 1990;Marchetti et al. 2020a) and the Artinskian age of the Peranera Formation of Spain inferred from radiometric ages from the laterallycorrelated Lower Red Unit and Ignimbrite Enclave of Castellar de n'Hug (Gisbert 1981;Pereira et al. 2014;Mujal et al. 2018). Further age constraints on the Peranera Formation are provided by Voigt & Haubold (2015), Mujal et al. (2016) andMujal et al. (2018) through biostratigraphy and lithostratigraphic correlation.
Importantly, the first occurrence of Pachypes coincides with (or even slightly precedes) the beginning of the Erpetopus biochron, in agreement with the reptile radiation of parareptile and eureptile captorhinomorph tracks that began during the Artinskian (Marchetti et al. 2019d).

PAREIASAUROMORPHA OCCURRENCES
The Pareiasauromorpha is the most diverse and abundant parareptile group of the Guadalupian and Lopingian. The majority of the Guadalupian occurrences are from South Africa and Russia, and a form is also known from Brazil   (Sennikov & Golubev 2017). Moreover, the tetrapod remains come from the lower part of the Chickasha Formation, which is laterally-correlated with the San Angelo Formation, constrained by marine biostratigraphy to the upper Leonardian (=Kungurian; e.g. Lucas & Golubev 2019). So, a Kungurian age for the lower Chickasha and San Angelo formations is the best supported age assignment, and is followed here.

CONCLUSION
The revision of several potential pareiasauromorph tracks from the Cisuralian and Lopingian of USA and western Europe allows, for the first time, the recognition of the pareiasauromorph ichnogenus Pachypes in units older than late Capitanian (Valentini et al. 2009;Voigt & Lucas 2018). We propose the new combination P. ollieri (Ellenberger, 1983a) for this material. A track-trackmaker correlation, based on synapomorphy and stratigraphy, suggests nycteroleter pareiasauromorphs similar to Macroleter as the most likely producers of this ichnotaxon (Tsuji et al. 2012). We recognize the earliest known occurrence of Pachypes in Artinskian units of Arizona (Hermit Formation), Spain (Peranera Formation, Lower Red Unit) and France (Rabejac Formation). This suggests a fast and palaeobiogeographically-extensive radiation of the trackmakers, which coincides with the late Artinkian reptile radiation known from both the track and the skeletal records (Marchetti et al. 2019d; and references within). The skeletal evidence of Pareiasauromorpha suggests a late Kungurian origin of the group in North America (Reisz & Laurin 2001, 2002Lucas 2002), whereas the largest part of the skeletal occurrences comes from the Guadalupian and Lopingian of South Africa and Russia (Lee 1997;Tsuji et al. 2012;Sennikov & Golubev 2017). Therefore, the earliest occurrences of Pareiasauromorpha footprints precede by at least 10 myr the earliest known occurrence from the skeletal record. Moreover, the tetrapod footprints highlight for the first time Cisuralian and Guadalupian occurrences of Pareiasauromorpha in western Europe, currently unknown from the skeletal record (Lucas 2018).
Moscow. We extend our thanks to Colleen Hyde, Kim Besom and Klara Widrig, Grand Canyon National Park, for their assistance obtaining photographs of the track specimen from the Hermit Shale in the park collection. We acknowledge Sergio Ll acer, Institut Catal a de Paleontologia (ICP), for his assistance in the creation of several 3D photogrammetric models. We would like to thank Lara Sciscio, Massimo Bernardi, Marc J. Van