Ichnotaxonomic Review of Large Ornithopod Dinosaur Tracks: Temporal and Geographic Implications

Background Large ornithopod tracks are known from the Upper Jurassic to the uppermost Cretaceous rocks of all continents but Antarctica. They include the tracks historically called Iguanodon footprints, iguanodontid footprints, hadrosaur/hadrosaurid footprints, and other large ornithopod tracks that have been used to define ichnotaxa. More than 40 ichnospecies based on large ornithopod tracks have been defined, but the validity of many of them is questionable. Methodology/Principal Findings 34 ichnogenera and 44 ichnospecies have been analysed in this work. Many of them are considered to be invalid because they have been defined on the basis of poorly preserved tracks without diagnostic features, have an inadequate diagnosis, or are based on temporal and/or geographical criteria. Only eight ichnospecies belonging to the ichnogenera Caririchnium, Iguanodontipus and Hadrosauropodus are here regarded as valid. Conclusions/Significance The monospecific ichnogenus Iguanodontipus (I. burreyi) is characterized by a small, rounded heel and elongate, narrow digit impressions. Its distribution is limited to the Berriasian-Valanginian of Europe. Caririchnium consists of four ichnospecies (C. magnificum [type ichnospecies], C. kortmeyeri, C. billsarjeanti and C. lotus) with a large, rounded heel and short, wide digit impressions. This ichnogenus ranges from the Berriasian-Hauterivian to the Aptian-Albian of South America, North America, Asia and Europe. Finally, Hadrosauropodus (three ichnospecies: H. langstoni [type ichnospecies], H. leonardii and H. kyoungsookimi) shows a large, bilobed heel and short, wide digit impressions. It is known from the Aptian-Albian to the Maastrichtian of North America, Asia and Europe. The ichnofamily Iguanodontipodidae includes large iguanodontian tracks characterized mainly by mesaxonic, tridactyl and subsymmetrical pes tracks that are as wide as (or wider than) long and have one pad impression in each digit and one in the heel. Its distribution is confidently limited to the Cretaceous of Europe, Asia, North America and South America.

In Europe, large ornithopod tracks were regarded as Iguanodon footprints for a long time (e.g., [36][37] in England; [38][39] in Spain). Outside Europe, Zhen et al. [40] used this terminology for an ichnite from China. With the aim of giving a formal ichnotaxonomical name to this kind of footprint, Sarjeant et al. [23] published a synthetic work on the European tracks assigned to Iguanodon, and defined the ichnogenus Iguanodontipus.
The term "hadrosaur track" is more recent than "Iguanodon track". It was used to describe the large ornithopod tracks found in the Upper Cretaceous rocks of North America. The first author to use the term was Langston [9] after studying a large Maastrichtian footprint from Alberta, Canada. He regarded it as a hadrosaurian ichnite on the basis of morphological differences relative to Iguanodon tracks and because of the discovery of a hadrosaurian skeleton near the site. Subsequently, other researchers described hadrosaur tracks (e.g., [24,[41][42]).
The first ichnotaxa defined on the basis of large ornithopod tracks were the ichnogenera Amblydactylus, Gypsichnites and Dinosauropodes, the former two described by Sternberg [8] and the latter one by Strevell [45]. Subsequently, Kuhn [46] defined Wealdenichnites and Sinoichnites on the basis of material previously studied by Dietrich [47] and Young [48] respectively.

Material and Method
In the present paper, the ichnotaxa assigned to large ornithopod tracks (Iguanodon and hadrosaur tracks, iguanodont tracks, etc.) or associated with large ornithopod trackmakers have been revised. In total, 34 ichnogenera and 44 ichnospecies of large ornithopod tracks have been studied (Fig. 1, Table 1).
Ornithopodichnites has not been considered in this study since 19th-century researchers classified all bipedal dinosaur tracks in this way in accordance with the approach of Hitchcock [29]. The proposal put forward by Strevell [45] has not been followed because the author defined one ichnogenus and eight ichnospecies on the basis of differently shaped natural casts removed from mines and without giving a diagnosis (see [49]).
b) The holotype must be an elite track, a very well-preserved true track or a natural cast that reproduces the anatomical features of the sole of the dinosaur pes and /or manus. Thus, ichnotaxa defined from tracks on the basis of extramorphological features, or from tracks that are poorly preserved or not impressed in the layer where the animal stepped (undertracks, underprints, etc.), should be avoided.
c) The holotype should not be the only track assigned to the ichnotaxon and, if possible, it should belong to a trackway. In the ichnological literature there are many ichnotaxa that are based on unique and isolated material, whose validity is uncertain. Tracks from the same trackway or other tracks of the type series show the intraichnotaxonomic variation. d) New ichnotaxa should not be described until all the bibliography has been revised. Geographical or temporal distribution should not be considered discriminatory criteria.
e) The diagnosis of a new ichnotaxon must be as accurate as possible to avoid ambiguity. Moreover, it is important to add a detailed description (qualitative and quantitative) of the tracks that are used in the type series.
f) The diagnosis should be based on the morphological features of the tracks. Trackway data such as pace length, degree of rotation, external and internal width, etc. are not valid ichnotaxobases because they reflect dinosaur behaviour, and may be variable.
g) The description of the tracks must include good, perpendicular photographs and drawings of the holotype and paratypes. It is also recommended to publish drawings and photographs of other tracks and trackways assigned to the ichnotaxon.
h) The nomenclature of the trace fossils follows the International Code of Zoological Nomenclature (Fourth edition 1999; ICZN).
Large ornithopod ichnotaxa have been identified from the Late Jurassic to the Maastrichtian ( Fig. 2A). The temporal distribution shows that there are five stages in which the joint presence of several ichnogenera has been cited. In the Berriasian, tracks are assigned to five different ichnogenera, four of which started their distribution in this stage. Eigth ichnogenera have been cited in the Barremian, and four of them occur for the first time in this stage. There are nine ichnogenera in the Aptian, and three of them have their origin in this stage. In the Albian, 17 ichnogenera have been cited, seven of which start in this stage. Finally, in the Maastrichtian there are nine ichnogenera, five of which have been described only in this stage. At least 19 out of the 34 ichnogenera of large ornithopod footprints are grouped in the Barremian-Albian period.
Of the 34 ichnogenera, 23 have been cited only once (in the paper where they are defined), five in two articles, two in three works, and one in seven papers (see S1 Table). Only three ichnogenera have over 10 citations: Iguanodontipus and Amblydactylus 13, and Caririchnium 29.
Concerning the distribution by continents (Fig. 2B, S1 Table), ten ichnogenera have been identified in South America, and five in North America. Six of these ichnotaxa have only one citation, five have two citations and one ichnotaxon has five citations. The ichnotaxa most Almost all the ichnogenera have been identified only in one continent (Fig. 2B). Nevertheless, Caririchnium has been cited in Europe, North America, South America and Asia (from Berriasian to Maastrichtian), Amblydactylus in Europe, North America, Asia and Oceania (from Barremian to Maastrichtian), and Hadrosauropodus in North America, Asia and Europe (from Campanian to Maastrichtian).
Of the 44 ichnospecies studied in this work, more than half (27) were described in the 1980s and 1990s (Table 1). In the 1990s alone 17 ichnospecies were erected. Large ornithopod ichnotaxa have been described in 15 different countries. China and Argentina are the countries where most ichnospecies have been described (six in each). Five ichnospecies have been described in both Spain and Canada, and four in Tajikistan. Brazil, Korea and Japan have three ichnospecies, Chile and the United States two, and the other countries (Germany, Italy, Peru, the United Kingdom and Switzerland) only one.

Validity of large ornithopod ichnotaxa
Akmechetosauropus makhkamovi Dzhalilov and Novikov, [56] This was described as a middle-sized track from the Albian of Tajikistan that belongs to the ichnofamily Hadrosauripodidae (sensu [56]). The diagnosis is not precise, the figure of the holotype (Fig. 3A) is very schematic and the authors suggest that the footprints are poorly preserved. For these reasons, we consider this ichnotaxon to be a nomen dubium.
Amblydactylus gethingi Sternberg, [8] This ichnospecies was established with a footprint found in the Aptian-Albian of Canada [57], which is currently under the waters of the W.A.C. Bennett Dam [41]. A cast of the holotype (plastotype) was deposited in the National Museum of Canada [8]. The track (Fig. 3B) is quite deep, has no pad impressions, the distal end of the digits is acuminate, and there is a depression in the proximal part of the heel interpreted as the metatarsal impression [8]. Lockley et al. [18] suggested that this ichnotaxon is poorly defined. Currie and Sarjeant [57] have pointed out that after several surveys in the area where A. gethingi was defined they have not found any footprints attributable to this ichnotaxon. In later works, Currie [41,43] assigned bipedal trackways to this ichnospecies.
Currie and Sarjeant [57] emended the diagnosis of Amblydactylus when defining the ichnospecies A. kortmeyeri. They proposed as diagnostic features the heel pad and the digit pad impressions. These features are evident in A. kortmeyeri, but Sternberg [8] noted that there were no pad impressions in the footprint of A. gethingi. Moreover, Sternberg [8] and Currie and Sarjeant [57] suggested that Amblydactylus had interdigital webs. According to Lockley et al. [24], there is no real evidence of these structures in ornithopods. Consequently, some footprints from New Mexico with web impressions have been reinterpreted as mud structures [58]. Thus, this feature should be interpreted as an extramorphological structure. Currie [43] referred quadruped trackways with bilobed pes to Amblydactylus isp., but this character was not previously assigned to either A. gethingi or A. kortmeyeri.
Taking into account the above data on A. gethingi, we consider it not valid. The holotype shape is likely conditioned by extramorphological features (metatarsal mark, collapsed digit impressions, interdigital webs and absence of pad marks). Currie and Sarjeant [57] affirmed that the differences between the A. gethingi and A. kortmeyeri could be explained as differences in the circumstances of track formation. Moreover, although a plastotype is preserved in the National Museum of Canada, the holotype is under the waters of a dam. As suggested by Gangloff et al. [61], a neotype found close to the type locality would be advisable for further systematic discussions. Therefore, pending further works confirming the validity of A. gethingi, it is provisionally considered a nomen dubium.
Amblydactylus kortmeyeri Currie and Sarjeant, [57] This ichnotaxon was defined in the same formation as A. gethingi (Gething Formation, Aptian-Albian, Canada). The type series, tracks and trackways, are well preserved. The tracks (Fig. 3C) are tridactyl, with pointed digits, a rounded heel impression, one pad in each digit [57], and the diagnosis reflects the main features of the tracks. Therefore, we consider that this ichnotaxon is valid.
Babatagosauropus bulini Dzhalilov and Novikov, [56] This ichnotaxon was described in the Albian of Tajikistan and assigned to the ichnofamily Hadrosauripodidae [56]. Due to its insufficient diagnosis, its poor state of preservation (according to the authors) and the schematic drawing of the holotype (Fig. 3E) figure, we consider it a nomen dubium.
Bonaparteichnium tali Calvo, [66] This was found in the Albian-Cenomanian? of Argentina. Bonaparteichnium is mainly characterized by a long, wide and robust heel impression [66]. Calvo [67] stated that Bonaparteichnium tracks are conditioned by the dinosaur gait, which impressed the metatarsus when walking. He considered this ichnotaxon a nomen vanum and junior synonym of Limayichnus. The metatarsal impression of Bonaparteichnium implies that the footprint shape (Fig. 3F) is conditioned by the behaviour of the dinosaur and/or features of the substrate. Therefore, this ichnotaxon is a taphotaxon (sensu [55]) and should not be considered ichnotaxonomically valid. We regard it as a nomen dubium.
Camptosaurichnus fasolae Casamiquela, [70] This was defined on the basis of Early Cretaceous tracks from Chile. Casamiquela and Fasola [70] classified Camptosaurichnus within the family Iguanodontidae. The footprints (Fig. 3H) are long and narrow, and show manus impressions. The heel impressions are acuminate and the digit impressions are narrow and sinuous, with claw marks [70]. The authors suggested that these features are due to the action of the mud. Moreno and Rubilar [71] assigned several tracks to this ichnotaxon. Sarjeant et al. [23] interpreted the morphology of the footprints as typically theropod. Nevertheless, Moreno and Pino [72] and later Moreno and Benton [73] maintained the ornithopod affinity of this ichnotaxon. We agree with Lockley et al. [24] in suggesting that the tracks of C. fasolae are poorly preserved and, therefore, in considering the ichnotaxon a nomen dubium.
Camptosauropus vialovi Gabunia and Kurbatov, [74] This ichnotaxon was described from the Late Jurassic of Tajikistan [75], and was assigned to the ichnofamily Iguanodontopodidae [74]. We consider Camptosauropus vialovi a nomen dubium because the diagnosis is insufficient and the holotype (Fig. 3I) has not been adequately figured.
Caririchnium magnificum Leonardi, [11] This was defined on the basis of a quadruped trackway from the Early Cretaceous of Brazil. Leonardi [11] proposed different diagnoses for the ichnogenus and the ichnospecies. The tracks (Fig. 3J) are well preserved. The pes tracks are large, tridactyl, with one pad impression in each digit and one in the heel, and they have short, wide digits. The manus tracks are smaller than the pes tracks. The ichnospecies diagnosis is very detailed and accurately reflects the morphology of the tracks. We consider C. magnificum to be a valid ichnotaxon.
Caririchnium leonardii Lockley, [76] This ichnotaxon (Fig. 3K) is based on a quadruped trackway from the "mid"-Cretaceous of the USA. The pes tracks are tridactyl, with one pad impression in each digit and one in the heel. The heel is not well marked (dashed line in Fig. 3K). The manus track is smaller than the pes track. In terms of the preservation of the type series, C. leonadii could be considered doubtful. Nevertheless, taking into account that these tracks can be related to other well-preserved tracks referred to Caririchnium or Caririchnium leonardii (e.g., [16,42,77]), we consider that this ichnotaxon is valid and that the diagnosis can be emended on the basis of the type series and referred tracks.
Caririchnium protohadrosaurichnos [78] This ichnotaxon ( Fig. 3L-M) was based on six trackways from the Cenomanian of the USA. C. protohadrosaurichnos was defined by the position of the manus impression, which was different from that of other Caririchnium ichnospecies [78]. Hunt and Lucas [79] and Lim et al. [80] have suggested that the position of the manus track is not a valid ichnotaxonomic character because it is conditioned by the dinosaur's behaviour. Therefore, we propose that C. protohadrosaurichnos is a nomen dubium.
Caririchnium lotus Xing, Wang, Pan and Chen, [81] This ichnospecies was defined by Xing et al. [81] from about 200 tracks found in the "mid"-Cretaceous of China. The tracks (Fig. 3N) are tridactyl, with one pad impression in each digit and one in the heel, which is rounded. The tracks are well preserved, and the diagnosis describes the the track shape. Therefore, we consider that this ichnotaxon is valid.
Caririchnium kyoungsookimi Lim, Lockley and Kong, [80] This ichnotaxon (Fig. 3O) is based on two manus-pes pairs found in a slab of stone from the "mid"-Cretaceous of Korea. The pes tracks are tridactyl, have a bilobed heel and one pad impression in the heel and one in each digit. Lim et al. [80] suggested that the morphology of the manus track is diagnostic. Accordingly, they defined C. kyoungsookimi mainly on the basis of differences in the manus shape relative to other Caririchnium ichnospecies. The tracks are well preserved, and the diagnosis conforms to the track shape. Thus, we consider that this ichnotaxon is valid.
Gigantoshiraminesauropus matsuoi Azuma and Takeyama, [82] The isolated and poorly-preserved track (Fig. 3P) used by Azuma and Takeyama [82] to define G. matsuoi was found in the Early Cretaceous of Japan. Matsukawa et al. [83] noted that this ichnotaxon was described without making a comparison with the footprints from other localities. Lockley and Matsukawa [84] considered Gigantoshiraminesauropus a nomen dubium, and we agree with this conclusion.
Goseongosauripus kimi Kim, [85] This ichnotaxon is based on a track from the Early Cretaceous of Korea. It was defined in an abstract without a diagnosis [86]. Subsequently, the same author changed the name Goseongosauripus kimi to Koseongosauripus onychion [87], but still failed to provide any diagnosis. Lockley et al. [86] considered the ichnotaxa proposed by Kim to be invalid. According to Lockley et al. [86], Goseongosauripus could be a junior synonym of Amblydactylus or Caririchnium. Moreover, Kim's [87] work is unpublished, and the footprint that he figures shows several pad impressions in each digit, so it does not belong to an ornithopod. We agree with Lockley et al. [86] that Goseongosauripus is a nomen nudum.
Gypsichnites pacensis Sternberg, [8] This was defined from an isolated track ( Fig. 3Q) from the "mid"-Cretaceous of Canada [8]. Some authors have defended the ornithopod affinities of this ichnotaxon [88][89]. Nevertheless, the footprint morphology (longer than wide) and the acuminate distal end of the digits (claw marks) suggest that the trackmaker is a theropod [61,[90][91]. After analysing the figures and descriptions of G. pacensis, we agree with the second interpretation and regard it as a theropod ichnotaxon.
Hadrosaurichnoides igeensis Casanovas, Ezquerra, Fernández, Pérez-Lorente, Santafé and Torcida, [92] This ichnotaxon ( Fig. 3R) was described on the basis of more than 200 tracks from the Early Cretaceous of Spain [92]. It is primarily characterized by interdigital web impressions. Casanovas et al. [92] suggested that the trackmaker was a transitional form between iguanodontids and hadrosaurids. Lockley et al. [24] interpreted the trackmaker as a theropod and classified Hadrosaurichnoides as a nomen dubium. Díaz-Martínez et al. [69] considered this ichnotaxon to be one of four morphotypes of Iguanodon-like tracks found in the Enciso Group (Cameros Basin). According to Lockley et al. [24], there is no evidence of an interdigital web in the ornithopod foot. As noted by Lockley and Hunt [58], the supposed web may be the result of extramorphological factors. Hadrosaurichnoides was impressed in calcareous sediment with a planar lamination produced by algal mats parallel to the stratification [92]. The algal mats are occasionally broken by the weight of the dinosaur, creating a discontinuity in the mat (I.D.-M., pers. obs.). For this reason, Hadrosaurichnoides is a taphotaxon, and we regard it as a nomen dubium.
Hadrosaurichnus australis Alonso, [93] This ichnotaxon (Fig. 3S) is based on four trackways from the Maastrichtian of Argentina [93]. Lockley et al. [24] suggested that H. australis shows theropod features (elongate footprint with relatively slender, tapering digit impressions). According to these authors, the paratype illustrated by Psihoyos and Knoebber [94] corresponds to a theropod. Sarjeant et al. [23] considered that Kuwajimasauropus, described as a theropod by Azuma and Takeyama [82], is a junior synonym of Hadrosaurichnus. Huh et al. [95] compared some tracks that they considered to be theropod with Hadrosaurichnus. Nevertheless, Thulborn [19] has suggested that H. australis is a hadrosaur ichnotaxon. In sum, we have classified it as a nomen dubium due to the ambiguity of the diagnosis and the poor quality of the holotype figure in the original article.
Hadrosaurichnus titicaensis Ellenberger, [96] Jaillard et al. [96] defined H. titicaensis from the Campanian-Maastrichtian of Peru as tracks that are longer than wide and without digital pad impressions (Fig. 3T). They related this ichnotaxon with a hadrosaur because the tracks preserve impressions of rigid interdigital tissues (web) and have blunt digits. As occurs with Amblydactylus and Hadrosaurichnoides, the web impressions can be explained as extramorphological features [58]. The holotype exhibits a similar morphology to that of an ornithopod track. Nevertheless, the ambiguity of the diagnosis and the lack of diagnostic characters in the holotype figure raise doubts regarding the validity of this ichnotaxon. Therefore, we propose that H. titicaensis be considered a nomen dubium.
Hadrosauripeda hauboldi Vialov, [97] This ichnotaxon was erected from a footprint found in the Late Cretaceous of Canada. No diagnosis is given. Vialov [97] based his proposal on a track ( Fig. 3U) figured by Haubold [98] and regarded as a hadrosaur footprint by Langston [9]. We consider Hadrosauripeda to be a nomen nudum.
Hadrosauropodus langstoni Lockley, Nadon and Currie, [24] This ichnotaxon ( Fig. 3V) was defined to propose a valid name for the hadrosaur tracks found in the Maastrichtian of Canada [24]. The tracks of H. langstoni are quadrupedal, with large, tridactyl pes tracks that have one pad impression in each digit and one in the heel, which is bilobed. This ichnotaxon is defined on the basis of well-preserved material [18] and an adequate diagnosis. We consider it a valid ichnotaxon.
Hadrosauropodus nanxiongensis Xing, Harris, Dong, Lin, Chen, Gou and Ji, [99] H. nanxiongensis is based on footprints ( Fig. 3W) from the Maastrichtian of China. According to Xing et al. [99], the tracks differ from those of H. langstoni in the size of digit II (larger in H. nanxiongensis) and in the divarication between digits II and IV (higher in H. nanxiongensis). Xing et al. [99] suggested that H. nanxiongensis has a notch in the proximal part of digit II that is absent in H. langstoni. They also stated that the size of the digit II impression is not an extramorphological feature. Nevertheless, Xing et al. [99] assigned to this ichnotaxon the isolated cast of a track that has the digit IV impression narrower than that of the other digits, less divarication than in the holotype and with notches in the proximal part of digits II and IV. The holotype is based on a rather poorly-preserved trackway [18,100]. The digital and heel pads are not well marked. The isolated cast is more similar to H. langstoni than to H. nanxiongensis. Based on the data and figures of Xing et al. [99], all these footprints can be assigned to the ichnogenus Hadrosauropodus. However, none of these tracks is sufficiently well preserved to allow the erection of a new ichnotaxon. For this reason, we consider that H. nanxiongensis is a nomen dubium.
Iguanodonichnus frenkii Casamiquela, [70] This ichnotaxon (Fig. 3X) is based on a long trackway from the Berriasian of Chile. Moreno and Rubilar [71] considered that Iguanodonichnus is a nomen dubium and assigned it to cf. Parabrontopodus. Sarjeant et al. [23] suggested that the footprints are similar to sauropod tracks. Moreno and Benton [73] proposed a new combination and denominated it Parabrontopodus frenki (with one i). Casamiquela and Fasola's [70] descriptions are ambiguous, and the illustrations are oblique photographs, without scale diagrams to illustrate the footprint morphology adequately [23]. As the tracks may belong either to a sauropod or an ornithopod, in this work I. frenkii is considered to be a nomen dubium due to its insufficient diagnosis and the poor quality of the holotype figures.
Iguanodonipus cuadrupedae Moratalla, [68] This ichnotaxon (Fig. 3Y), based on Early Cretaceous tracks from La Rioja in Spain, was described by Moratalla [68] in his unpublished doctoral thesis. Sarjeant et al. [23] referred part of these footprints to Iguanodontipus. Díaz-Martínez et al. [69] considered I. cuadrupedae to be one of the four morphotypes of Iguanodon-like footprints found in the Enciso Group (Cameros Basin). Pascual-Arribas et al. [101] regarded Iguanodonipus as not valid. We consider this ichnotaxon to be a nomen nudum because it has not yet been described formally.
Iguanodonopus xingfuensis Zhen, Li and Hang, [102] This ichnotaxon was described from two isolated tracks (Fig. 4A) from the Early Cretaceous of China. Sarjeant et al. [23] considered that these footprints are too long in relation to their width to be included among large ornithopod footprints. We agree with Xing et al. [99] that the diagnosis and discussion of Iguanodonopus are insufficient, and we therefore consider it a nomen dubium.
Iguanodontipus burreyi Sarjeant, Delair and Lockley, [23] Iguanodontipus burreyi (Fig. 4B) was erected in order to group the tracks found in Europe (England, Germany and Spain) that had previously been assigned to Iguanodon [23]. The type series of I. burreyi consists of several natural casts that are not well preserved (see 23, [figs. [13][14]). The diagnosis adequately reflects the morphology of these tracks. Subsequently, other ornithopod tracks found in Europe have been referred to Iguanodontipus (e.g., [103][104]), but their shape is not identical to the holotype and they have not corresponded to the diagnosis for I. burreyi. Lockley et al. [18,23] noted that the type material for I. burreyi is moderately well preserved and is properly described, so they considered it a valid ichnotaxa. However, Meyer and Thüring [105] stated that the outline of the footprints is poorly-preserved, and Diedrich [103] suggested that the type trackway of Iguanodontipus is similar to that of Megalosauropus. Moreover, Pascual-Arribas et al. [101] stated that the diagnosis of Iguanodontipus should be revised.
As regards the preservation of the type series, I. burreyi could be considered a nomen dubium. Nevertheless, it should be noted that these tracks can be related to other well-preserved tracks referred to as Iguanodontipus or "Iguanodon tracks" (e.g., [16,106]), so we consider that I. burreyi is valid and propose to amend the diagnosis on the basis of these tracks.
Iguanodontipus billsarjeanti Meyer and Thüring, [105] This ichnotaxon (Fig. 4C) was defined on the basis of three quadruped trackways from the Early Cretaceous of Switzerland. The tracks are well preserved and exhibit one pad impression in the heel and one in each digit. Meyer and Thüring [105] noted that I. billsarjeanti differs from I. burreyi in having a well-defined contour line and a higher divarication. The holotype of I. billsarjeanti is well preserved, and the diagnosis is accurate. Therefore, we consider it a valid ichnotaxon.
Jiayinosauropus johnsoni Dong, Zhou and Wu, [107] This ichnotaxon is based on an incomplete footprint (Fig. 4D) found in the Late Cretaceous of China. Xing et al. [99] and Lockley et al. [100] suggested that Jiayinosauropus is similar to Hadrosauropodus. For these authors the ichnotaxon is valid but it should be revised in the future. Lockley et al. [18] noted that the morphology of Jiayinosauropus is insufficiently known. It was defined on the basis of a single, poorly-preserved track and the diagnosis is inadequate. Therefore, we consider it a nomen dubium.
Orcauichnites garumniensis Llompart, Casanovas and Santafé, [111] This ichnotaxon (Fig. 4G) was described from several poorly-preserved tracks found in the Maastrichtian of Spain. Lockley et al. [24] suggested that the tracks are probably theropod footprints because they are longer than wide. These authors considered that Orcauichnites is a nomen dubium, and we accept this interpretation.
Ornithopodichnites magna Llompart, Casanovas and Santafé, [111] Ornithopodichnites (Fig. 4H) was described in a site close to Orcauichnites. According to Lockley and Meyer [112] and Lockley et al. [24], Ornithopodichnites has the same problems as Orcauichnites. For these authors, the tracks are poorly preserved, the original description is inaccurate, and the tracks belong to theropods. Therefore, O. magna is considered to be a nomen dubium.
Ornithopodichnus masanensis Kim, Lockley, Kim, Lim and Kim, [20] This ichnotaxon (Fig. 4I) was described from the Early Cretaceous of Korea on the basis of robust, tridactyl, slightly mesaxonic footprints that are wider than long, with short, U-shaped digit impressions that have a blunt distal end, and with a smoothly rounded heel. Kim et al. [20] argued that the tracks are suitable for defining a new ichnotaxon because they are true footprints (presence of extruded rims) and the general shape is the same in more than 100 footprints. Nevertheless, Lockley et al. [113] assigned to Ornithopodichnus footprints with a morphology that differs from that of the type series and does not conform with the original diagnosis. The presence of extruded rims is not an unequivocal character of true tracks (cf. [114]). Lockley et al. [18] noted that the preservation of these trackways is suboptimal, with some tracks still partially filled. Kim et al. [20] affirmed that preservational factors likely play a role in the morphology. The footprints described by Kim et al. [20] and Lockley [115] (including photographs and drawings) have no indisputable diagnostic features. The general morphology is similar to that of other tracks assigned to Hadrosaurichnoides that were defined on the basis of extramorphological features. Pending a revision of the validity of O. masanensis, in this work we consider that this ichnotaxon is a nomen dubium.
Shiraminesauropus reini Azuma and Takeyama, [82] Azuma and Takeyama [82] described the ichnospecies Shiraminesauropus reini (Fig. 4J) and S. hayashidaniensis (Fig. 4K) on the basis of two isolated footprints from the Early Cretaceous of Japan. Matsukawa et al. [83] noted that the material is probably inadequate and undiagnostic, it was not adequately compared with other footprints, and the differences in preservation were not taken into consideration. Lockley and Matsukawa [84] considered that both ichnospecies of Shiraminesauropus are nomina dubia. We agree with this interpretation.
Sinoichnites youngi Kuhn, [46] This ichnotaxon is based on a single track (Fig. 4L) from the Late Jurassic of China [46], which was originally described by Teilhard and Young [6]. S. youngi has been assigned to Iguanodontidae by Kuhn [46] and Zhen et al. [116]. The diagnosis by Kuhn [46] is very imprecise. The exact age and origin of the track is unknown [117]. The footprint is currently lost, and only a cast is preserved in the Museum of Natural History of Beijing [99]. For these reasons, we consider that Sinoichnites youngi is a nomen dubium.
Sousaichnium pricei Leonardi, [10] Based on tracks from the Early Cretaceous of Brazil, this ichnotaxon (Fig. 4M) has been related to Iguanodontidae [10]. Sousaichnium shows elongate heel impressions, mud collapsed inside the track, and it lacks claw impressions. Pérez-Lorente [118] pointed out that the elongate impressions represent metatarsal marks. The tracks that Leonardi [10] assigned to S. pricei are not well preserved (e.g., metatarsal impressions, mud collapsed inside the track, etc.). Therefore, we consider it a nomen dubium.
Sousaichnium monettae Calvo, [66] This ichnotaxon (Fig. 4N) was defined on the basis of tracks from the "mid"-Cretaceous of Argentina, and related to Iguanodontidae [66]. Meyer [109] suggested that Sousaichnium monettae and Bonaparteichnium tali were junior synonyms of Limayichnus major, and regarded them as theropod tracks. Calvo [66, fig. 3] figured the holotype of S. monettae as being mainly characterized by a metatarsal impression. This impression can be caused by a special gait, sloping ground, the action of mud, etc. [118]. Irrespective of the affinities of the trackmaker (likely an ornithopod), we consider it a nomen dubium.
Staurichnium diogenis Leonardi, [10] This ichnotaxon (Fig. 4O) was based on tracks from the Early Cretaceous of Brazil. The footprints are only faintly impressed, which is probably due to the fact that the mud was dry when the dinosaur walked on it [10]. Leonardi [10] related Staurichnium with ornithopods such as Hadrosauridae. This ichnotaxon is characterized by the elongate heel impression. As the tracks are not well preserved and the ichnotaxon was defined on the basis of the metatarsal impressions, we consider that Staurichnium is a nomen dubium.
Taponichnus donottoi Alonso and Marquillas, [119] This ichnotaxon is based on an isolated track (Fig. 4P) from the Late Cretaceous of Argentina. The tracks are twice as long as wide, with short digit impressions and interdigital web marks. Due to the age (Late Cretaceous) and the interdigital web, Alonso and Marquillas [119] related Taponichnus with a medium to large-sized hadrosaur. The diagnosis of T. donottoi is imprecise and the material is a single, poorly-preserved footprint, so we consider it a nomen dubium.
Telosichnus saltensis Alonso and Marquillas, [119] This ichnotaxon from the Late Cretaceous of Argentina was based on a large, rounded track (Fig. 4Q), with no claw impressions, blunt digits II and III and extruded rims [119]. Alonso and Marquillas [119] suggested that Telosichnus is an ornithopod because the tracks show the features listed by Thulborn and Wade [120]. Moreover, they related the tracks to Hadrosauridae on the basis of their age and morphology. The figures are incomplete (dashed line in the drawing) and do not provide useful information. As with Taponichnus, the diagnosis is ambiguous and the material is poorly preserved and isolated. Therefore, we consider that Telosichnus is a nomen dubium.
Wealdenichnites iguanodontoides Kuhn, [46] The footprint (Fig. 4R) is a natural cast found in the Late Jurassic of Germany [46]. It is mainly characterized by having a hallux impression. The diagnosis is imprecise and the holotype is based on a single, isolated specimen. Therefore, we consider Wealdenichnites a nomen dubium.
Yangtzepus yipingensis Young, [7] This ichnotaxon (Fig. 4S) is based on tracks from the Early Cretaceous of China. Young [7] related it to an ornithopod. However, Xing et al. [99] interpreted the tracks as theropod footprints, similar to tracks found in the USA [121][122]. On the basis of the morphology of the tracks (longer than wide, with elongate, narrow digit impressions, and claw marks), we have classified them as theropod tracks.  (Fig. 5A). Group 2: characterized by a large, rounded heel, with short, wide digits. Caririchnium magnificum, Caririchnium lotus, "Amblydactylus" kortmeyeri, Iguanodontipus billsarjeanti and some tracks of Caririchnium leonardii are included in this group (Fig. 5B). Group 3: characterized by a large, bilobed heel, and short, wide digits. Hadrosauropodus langstoni, Caririchnium kyoungsookimi and some tracks of Caririchnium leonardii are included in this group (Fig. 5C).
Caririchnium has two ichnospecies in Group 2 (rounded heel) and two in Group 3 (bilobed heel). As regards Iguanodontipus, I. burreyi shows a small heel and elongate, narrow digits (Group 1), whereas I. billsarjeanti has a large heel and short, wide digits (Group 2). Finally, some tracks assigned to the ichnospecies Caririchnium leonardii have a rounded heel (Group 2), whereas other tracks show a bilobed heel (Group 3).
The presence of the same ichnospecies in two groups (Fig. 5) could be explained by: a) an inadequate diagnosis that has led to tracks with different characters being classified together; or b) geographical and temporal conventions (see below).
In order to simplify the ichnotaxonomy of large ornithopod tracks and to group the valid ichnospecies in ichnogenera with a stable morphology, the following systematic considerations have been made: 1. The diagnoses of the valid ichnogenera and ichnospecies have been amended; 2. Some ichnospecies have been assigned to other ichnogenera, and new combinations are here proposed.
Comments As recently suggested by Lockley et al. [18], we consider it necessary to propose a suprageneric ichnotaxon to include the valid large ornithopod ichnotaxa, as well as the tracks that share the same main features but are not well enough preserved to be assigned to a particular ichnogenus or ichnospecies. Vialov [97] proposed the ichnofamilies Iguanodontipodidae and Hadrosauripodidae. However, he defined the ichnofamilies without providing a diagnosis and without presenting the differences between them. Vialov included the ichnogenus Hadrosauripeda within Hadrosauripodidae, but did not include any ichnotaxa within Iguanodontipodidae. Subsequently, Gabunia and Kurbatov [74] and Dzhalilov and Novikov [56] respectively assigned Camptosauropus and Kharkushosauropus to Iguanodontopodidae instead of Iguanodontipodidae (probably a typographic mistake), and Dzhalilov and Novikov [55] assigned Akmechetosauropus and Babatagosauropus to Hadrosauripodidae. These proposals have not been used by other researchers in subsequent papers. Hadrosauripodidae is composed of ornithopod (Hadrosauripeda sensu [97]) and theropod (Babatagosauropus sensu [107]) morphotypes, and a morphotype of uncertain affinity (Akmechetosauropus). The ichnotaxa included in Iguanodontipodidae have theropod (Kharkushosauropus sensu [108]) and uncertain (Camptosauropus) affinities. Moreover, all these ichnogenera are considered in the present paper to be nomina dubia or nomina nuda. On the other hand, Lockley et al. [18] proposed a new ichnofamily using the nomenclature of Vialov [96] and Dzhalilov and Novikov [55]. In the text, they used the ichnofamily name Iguanodontipodidae, but in the ichnotaxonomical proposal, just before the diagnosis, used Iguanodontopodidae (probably a typographic mistake). According to the International Code of Zoological Nomenclature (art. 64) "the choice of type genus determines the stem of the name of the nominal family-group taxon". In this case, the type ichnogenus is Iguanodontipus, and therefore the correct name of the ichnofamily is Iguanodontipodidae. We emend the diagnosis of Iguanodontipodidae on the basis of the shared morphology of the ichnotaxa here considered valid.
Comments Iguanodontipus was described by Sarjeant et al. [23] in order to include within a formal group the tracks that had previously been assigned to Iguanodon. The type series comprises seven casts made from tracks found in the Berriasian of England (Fig. 6A-C). The tracks have a large, rounded heel and three short, wide digits; pad and claw impressions are lacking. Meyer and Thüring [105] noted that the contour line of the tracks from the type series is poorly preserved. The diagnosis proposed by Sarjeant et al. [23] reflects the shape of the tracks of the type series, but these authors assigned to Iguanodontipus other footprints with a more complex morphology that did not correspond to these diagnostic features. In this context, researchers such as Gierlinski et al. [60] and Lucas et al. [16] have illustrated Iguanodontipus using a footprint of Ornithoidichnites that was studied originally by Beckles [3] (Fig. 6D) and was subsequently referred to Iguanodontipus by Sarjeant et al. [23] (fig. 3). This footprint is not deformed and, in contrast to those of the type series, has a narrow, rounded heel, and long and independent digit impressions with acuminate distal ends. The wide, rounded heel impression shown by the holotype of I. burreyi is due to the loss of the notch of the proximal part of digits II and IV. The triangular shape of the digits is likely the result of the distal displacement of the hypex in poorly-preserved footprints. Tracks with these features from the Berriasian of Britain [23], Germany [125] and Spain [101,126] have been assigned to Iguanodontipus. Other footprints found in the Berriasian of Spain [67,[127][128] that exhibit the same features have been classified as Therangospodus oncalensis. Recently, Castanera et al. [129] referred them as Iguanodontipus? oncalensis and discussed their relation with the ichnogenus Iguanodontipus. These footprints show diagnostic features of Iguanodontipus and probably belong to this ichnogenus [126].
On the other hand, several tracks from the Late Jurassic-Early Cretaceous transition of Europe that lack the diagnostic features of Iguanodontipus have been assigned to this ichnogenus [104,[130][131][132], probably because of geographical and temporal conventions.
On the basis of tracks from the Aptian of Switzerland, Meyer and Thüring [105] described a second ichnospecies of Iguanodontipus, I. billsarjeanti. However, in the present paper we consider that the morphology of the heel and digits of these tracks is consistent with the diagnostic features of Caririchnium (see above).
Diedrich [103] proposed as paratypes of Iguanodontipus burreyi two quadruped trackways from Spain and Germany, and claimed that the ichnotaxon could be quadruped. Nevertheless, he did not emend the diagnosis of I. burreyi, and the proposed paratypes come from outside the type locality; therefore, this proposal has no ichnotaxonomic validity.
Lockley [59] suggested that Iguanodontipus is a junior synonym of Amblydactylus and Caririchnium. According to Lucas et al. [16], Iguanodontipus is a junior synonym of Amblydactylus and differs from Caririchnium in having graceful digits directed laterally and a narrower heel. Finally, Lockley et al. [18] noted that Amblydactylus and Iguanodontipus are not synonyms. As discussed above, we consider that the shape of the heel and digits allow the ichnogenera of Iguanodontipodidae to be differentiated from one another. Iguanodontipus is characterized by a narrow, rounded heel and long, narrow digit impressions. In contrast, Caririchnium has a wide, rounded heel and short, wide digit impressions, and Hadrosauropodus shows a wide, bilobed heel and short, wide digit impressions.
A few quadruped tracksites could be assigned to Iguanodontipus. Lockley et al. [133] studied some tracksites from the Berriasian of Germany that present manus tracks smaller than the pes tracks. The manus tracks are wider than long and are situated in front of the pes track close to digit IV.
In sum, we consider that I. burreyi is the only ichnospecies that belongs to Iguanodontipus; I. billsarjeanti is assigned to Caririchnium.
Ichnogenus Caririchnium Leonardi, [11] Emended diagnosis Pes tracks belonging to Iguanodontipodidae, with a large heel impression that is rounded, centred and wide (wider than the width of the proximal part of the digit III impression); short, wide digit impressions.
Comments When Leonardi [11] defined Caririchnium he proposed two separate diagnoses, one for the ichnogenus and the other for the ichnospecies. Leonardi [11] characterized the pes tracks of Caririchnium as large, tridactyl, with a pad impression in the heel, and short, wide digits. Moreover, the manus tracks are small and elliptical. Subsequently, Lockley [76] emended the ichnogeneric diagnosis, although basically he translated into English the original diagnosis in Italian by Leonardi. Lee [78] also emended the diagnosis of Caririchnium. He proposed different features for bipedal and quadrupedal trackways. Lee's diagnosis is inaccurate and scarcely takes into account the morphology of the tracks. The three diagnoses of Caririchnium mostly used ichnotaxobases that depend on the values obtained from trackway analysis. It should be noted that trackway data depend mainly on dinosaur behaviour and are of little value for ichnotaxonomy [53]. The important features are those obtained from the shape of the pes and manus sole impressions.
Leonardi [11] and Lockley [76] claimed that Caririchnium is mainly characterized by quadruped trackways with the pes tracks much larger than the manus tracks, the pes tracks having one pad impression in each digit and one in the heel, and the pes digits being short and wide. Nevertheless, these authors do not consider the heel shape. In the present paper, two kinds of heel impression have been identified: rounded and bilobed (see [126]). Caririchnium and Iguanodontipus show rounded heel impressions whereas in Hadrosauropodus the heel is bilobed. Caririchnium has a large heel that is wider than the maximum width of the proximal part of digit III. On the other hand, Iguanodontipus has a small heel that is no wider than the maximum width of the proximal part of digit III. Moreover, the digit impressions of Iguanodontipus are elongate and narrow whereas Caririchnium and Hadrosauropodus have short, wide digit impressions. The differences in the impressions of both the heel and digits allow us to differentiate between the three ichnogenera of large ornithopod footprints.
Currie [43] suggested that Caririchnium, Hadrosaurichnus and Ornithopodichnites were junior synonyms of Amblydactylus. Nevertheless, in this work Hadrosaurichnus, Ornithopodichnites and Amblydactylus are regarded as nomina dubia (see discussion above). On the other hand, Lucas et al. [16] noted that the diagnoses of Caririchnium and Hadrosauropodus were similar and that there are no features that differentiate between them. Consequently, they proposed that Caririchnium is a senior synonym of Hadrosauropodus. Several authors (e.g., [79][80]) have considered that Caririchnium shows a bilobed heel impression. Nevertheless, as discussed above, Caririchnium has a rounded heel whereas the heel impression of Hadrosauropodus is bilobed. We support Lockley et al. [18] and do not accept the synonymy proposed by Lucas et al. [16].
Several trackways of Caririchnium are quadrupedal. Lockley [76] considered that the shape and position of the manus tracks are diagnostic features. Subsequently, Lim et al. [80] accepted only the manus track shape as diagnostic, because the position of the manus impressions in all quadrupedal ornithopod trackways is variable [21]. This proposal is considered valid here, but further studies on the variability of the manus track shape in relation to the trackmaker and the action of the mud are needed. Assignment to an ichnospecies should be possible even without manus tracks, so it is here postulated that pes tracks should be given greater ichnotaxonomic importance.
In the present study, we propose that only two of the ichnospecies currently assigned to Caririchnium can in fact be so assigned: C. magnificum and C. lotus. The ichnotaxa C. leonardii and C. kyoungsookimi are referred to Hadrosauropodus, whereas C. protohadrosaurichnos is regarded as a nomen dubium. "Ambydactylus" kortmeyeri and Iguanodontipus billsarjeanti are here assigned to Caririchnium on the basis of the digit and heel shape.
Other large ornithopod tracks share the diagnostic characters of Caririchnium, but they cannot be related accurately with a specific ichnospecies. This is the case for many tracks classified as iguanodontid footprints, ornithopod footprints, Hadrosaurichnoides, Brachyguanodonipus and Iguanodonipus from the lowermost part of the Early Cretaceous of Spain, especially from the Urbión [147] and Enciso Groups (basal Valanginian-middle Albian, Spain) (see [126], p.126, table 9.4). Moreover, tracks found in the Areniscas de Camarillas Formation from the Barremian of Spain [131], the Salema Formation from the Hauterivian-Barremian of Portugal [130], and the Wessex Formation from the Barremian of England [145] can also be assigned to Caririchnium. The presence of Caririchnium in Europe was first cited recently by Díaz-Martínez [126] and has been subsequently supported by Lockwood et al. [145]. Several large ornithopod ichnotaxa considered not valid in the present study display a similar morphology to Caririchnium, but they need to be revised: Sousaichnium, Staurichnium, Limayichnus and Ornithopodichnus.
Caririchnium magnificum Leonardi, [11] Emended diagnosis Pes tracks belonging to Caririchnium with very large heel pad impressions, approximately as wide as or wider than long; blunt distal end of digit impressions; subtriangular distal part of the heel pad; manus tracks elliptic and wider than long.
Holotype Trackway about 25 m long. Cast of the first manus and pes impressions in the Museu Câmara Cascudo Federal University of Rio Grande do Norte, Natal, Brazil [11] p. 177, fig. 8  (Fig. 7A-B).
Description All the information on the type series is in Leonardi [11]. Comments Leonardi [11] defined C. magnificum on the basis of a quadruped trackway from the Early Cretaceous of Brazil. He figured as holotypes a pair of manus and pes impressions (Fig. 7A). The pes track belongs to the first pair and the manus track to the third pair. The author proposed in the ichnospecific diagnosis that the heel contour is rounded. Moreover, the pes tracks present a wide heel impression. Therefore, according to the emended diagnosis its presence in Caririchnium is justified. The pes tracks of the type trackway have a variable morphology, but most of them show heel pad impressions that are rounded and wider than long (Fig. 7B). A wider than long heel is also present in C. kortmeyeri and C. billsarjeanti, though not in C. lotus, which has longer than wide heel pad impressions. The presence of blunted claw marks is also shared with C. billsarjeanti and C. lotus, whereas C. kortmeyeri presents pointed claw marks. The distal part of the heel pad impression is subtriangular in C. magnificum, C. kortmeyeri and C. lotus, but rounded in C. billsarjeanti. The manus track shape of C. magnificum is variable, but generally it is elliptic with the lateromedial axis larger than the posteromedial one.
The ichnospecies C. magnificum was originally described in the Antenor Navarro Formation of Brazil [11], which is Berriasian-Barremian [154] or Berriasian-Hauterivian in age [146]. In the present work, we have assigned to this ichnospecies four trackways (including part of the material classified as Brachyguanodonipus by [68]) from the Enciso Group of La Rioja in Spain (see discussion in [126]). The age of these tracks is basal Berriasian to middle Albian (sensu [124]).

Ichnotaxonomy of Large Ornithopod Dinosaur Tracks
Emended diagnosis Pes tracks belonging to Caririchnium with very large heel pad impressions, approximately as wide as or wider than long; sharp distal end of digit impressions; subtriangular distal part of the heel pad.
Description All the information on the type series is in Currie and Sarjeant [57]. Comments Currie and Sarjeant [57] defined this ichnotaxon on the basis of a well-preserved natural cast (holotype) and several tracks and trackways (paratypes) found in the Aptian-Albian of western Canada (Fig. 7C-D). They suggested that the main differences between A. gethingi and A. kortmeyeri are the length/width ratio of the track (greater in A. gethingi), its contour shape, and the distal end of the digits (more tapered in A. gethingi). Currie and Sarjeant [57] suggested that these differences may depend on the circumstances of track formation. Currie [41] noted that it was possible to distinguish the ichnospecies in well-preserved tracks, but this identification was not possible for the majority of the Amblydactylus tracks. Currie [41,43] considered that A. gethingi and A. kortmeyeri seem to represent the same general type of animal, and discussed both ichnospecies together. Gangloff et al. [61] suggested that A. gethingi could be a senior synonym of A. kortmeyeri. In the present work, the type ichnospecies of Amblydactylus, A. gethingi, is provisionally considered a nomen dubium (see above). Therefore, A. kortmeyeri could be assigned to another ichnogenus, used to define a new ichnogenus, or become the type ichnospecies of Amblydactylus after a formal request to the ICZN. This ichnospecies is mainly characterized by a wide and rounded heel impression, as is typical in Caririchnium. Consequently, we propose to assign A. kortmeyeri to Caririchnium.
There are no manus tracks in the type series of C. kortmeyeri. Currie and Sarjeant [57] assigned some trackways as paratypes, but the track drawings are somewhat different in form from the holotype. The holotype specimen has as its only autapomorphy the sharp claw impressions that are different from other ichnospecies of Caririchnium. Further studies should discuss whether this character is of ichnotaxonomic significance or is due to preservational biases. C. kortmeyeri also differs from C. lotus in the width of the heel pad impression, from C. billsarjeanti in the shape of the distal part of the heel pad, and from C. magnificum only in the shape of the claw impressions.

Emended diagnosis
Pes tracks belonging to Caririchnium with very large heel pad impressions, approximately as wide as or wider than long; blunt distal end of digit impressions; rounded distal part of the heel pad; manus tracks elliptic and wider than long.

Comments
This ichnotaxon was defined on the basis of a well-preserved quadruped trackway from the Aptian of Switzerland (Fig. 7F). The pes tracks have one pad impression in each digit and one in the heel, and the manus tracks are small and elliptic [105] (Fig. 7E). The ichnospecies, which was originally referred to Iguanodontipus, is regarded as valid (see above). However, the observed features do not correspond with those of Iguanodontipus but with Caririchnium. Therefore, the ichnospecies billsarjeanti is placed in Caririchnium as C. billsarjeanti nov. comb.
The general shape of the holotype is very similar to that of C. magnificum and C. kortmeyeri. Moreover, the manus track is also similar to that of C. magnificum. Nevertheless, C. billsarjeanti differs from other ichnospecies of Caririchnium in that the distal part of the heel pad is rounded (instead of subtriangular). Future studies might resolve whether this character is part of the variability in the preservation of C. magnificum or C. kortmeyeri.
Caririchnium lotus Xing, Wang, Pan and Chen, [81] Emended diagnosis Pes tracks belonging to Caririchnium with very large heel pad impressions that are longer than wide; blunt distal end of digit impressions; subtriangular distal part of heel pad; manus tracks rectangular, smoothly concave in the proximal part and wider than long.
Description All the information on the type series is in Xing et al. [81] and Xing et al. [161]. Comments C. lotus was defined by Xing et al. [81] on the basis of biped and quadruped trackways from the "mid"-Cretaceous of China. The material consists of about 200 well-preserved tracks of different sizes. Xing et al. [81] classified the pes tracks in three groups according to their length: 37-40 cm (adults); 25-30 cm (subadults); and 19-23 cm (juveniles). This ichnotaxon has large, rounded heel impressions [81,161] (Fig. 7G-H), so its assignment to Caririchnium is justified. C. lotus is mainly characterized by having a heel pad impression that is longer than wide, in contrast to other ichnospecies of Caririchnium, which are wider than long. Xing et al. [81] suggested that the manus tracks had digit impressions in their distal part (Fig. 7I). Subsequently, Xing et al. [161] stated that the manus track is rectangular in shape, with rounded edges and a slightly concave proximal surface. This feature of C. lotus is also different from the other ichnospecies of Caririchnium.
The ichnospecies C. lotus was originally described in the Jiaguan Formation in China which is Barremian-Albian in age, (sensu [148]). In the present work, we have assigned to this ichnospecies tracks of four tracksites (including part of the material classified as Iguanodonipus by [68]) from the Enciso Group of La Rioja in Spain (see discussion in [126]). The age of these tracks is basal Barremiean to middle Albian (sensu [124]).
Ichnogenus Hadrosauropodus Lockley, Nadon and Currie, [24] Emended diagnosis Tracks belonging to Iguanodontipodidae with a large heel impression that is bilobed, centred and wide (wider than the width of the proximal part of the digit III impression); pad of digit III shorter than those of digits II and IV; short, wide digit impressions with blunt distal ends.
Comments Hadrosauropodus was defined by Lockley et al. [24] on the basis of tracks from the Maastrichtian of Canada previously regarded as "hadrosaur footprints" [42]. These footprints are mainly characterized by a bilobed heel impression and short, wide digits. As noted above, these features allow Hadrosauropodus to be distinguished from Caririchnium and Iguanodontipus.
To date, there are only a few citations of Hadrosauropodus. In the original paper, Lockley et al. [24] proposed as the type series several footprints from the St. Mary River Formation (Maastrichtian, Canada), and assigned to Hadrosauropodus isp. footprints from the Lance Formation (Maastrichtian, USA). In a preliminary study, Suñer et al. [169] assigned some casts from the Maastrichtian of Spain to Hadrosauropodus langstoni. Xing et al. [99] described Hadrosauropodus nanxiongensis from several tracks found in the Zhutian Formation (Maastrichtian, China), and assigned to Hadrosauropodus isp. additional tracks from the same formation. As discussed above, H. nanxiongensis is considered to be a nomen dubium, but the presence of a bilobed heel impression suggests that the tracks belong to Hadrosauropodus. Recently, Vila et al. [168] classified as Hadrosauropodus isp. several footprints found in the late Maastrichtian of the southern Pyrenees in Spain. All these data suggest that the ichnogenus Hadrosauropodus could present a wide geographical distribution (North America, Asia and Europe), but a very limited temporal distribution (Maastrichtian). Nevertheless, Hadrosauropodus is not the first ichnotaxon to which footprints with bilobed heel impressions have been assigned. Lockley [77] classified one footprint with skin and a bilobed heel impression from the Dakota Formation (Albian-Cenomanian) as Caririchnium. Since then, all the footprints with this kind of heel impression have been assigned to Caririchnium, except the Maastrichtian ones related to Hadrosauropodus. Examples include the ichnospecies C. leonardii and C. kyoungsookimi from the mid-Cretaceous of the USA and Korea, respectively (see [21,80,170]). Based on this, Lucas et al. [16] suggested that Hadrosauropodus is a junior synonym of Caririchnium. In the present paper, we consider that Caririchnium has a rounded and not bilobed heel impression, and is thus different from Hadrosauropodus. Consequently, C. leonardii and C. kyoungsookimi are referred to Hadrosauropodus.
Currie [43] assigned to Amblydactylus isp. several quadrupedal tracks from the Aptian-Albian of Canada with bilobed heel impressions. Moreover, Carpenter [165] described bilobed tracks from the Campanian of the Mesa Verde Group (USA) and interpreted them as hadrosaur footprints. Pending revision, these tracks are here provisionally regarded as belonging to Hadrosauropodus.
Hadrosauropodus langstoni Lockley, Nadon and Currie, [24] Emended diagnosis Pes tracks belonging to Hadrosauropodus with a heel impression much wider than the width of the proximal part of the digit III impression; proximal part of the pads of digits II and IV situated close to the proximal part of the heel pad; manus tracks are obtuse-isosceles-triangle-shaped.
Holotype TMP 87.76.7 [24] 240, fig. 12A (Fig. 8A).  12). Description All the information on the type series is in Lockley et al. [24]. Comments Lockley et al. [24] defined this ichnotaxon on the basis of several tracks from the Maastrichtian of Canada (Fig. 8A-B). H. langstoni is mainly characterized by having pes tracks that are much larger than the manus tracks, one pad impression in each digit and one in the heel, which is bilobed. Before the study by Lockley et al. [24], these tracks had been assigned to hadrosaurs (e.g., [21,43,138]). The tracks are very well preserved and have manus prints, skin impressions and tail marks. This ichnospecies is characterized above all by a very wide heel impression. Moreover, the notches of digits II and IV are positioned far back, close to the proximal part of the heel. Other ichnospecies of Hadrosauropodus show a narrower heel and the notches of digits II and IV are close to the proximal part of the digit III pad impression. On the other hand, the manus tracks of H. langstoni are different (triangular) from those of H. leonardii (rectangular) and H. kyoungsookimi (crescent-shaped).

Emended diagnosis
Pes tracks belonging to Hadrosauropodus with a heel impression as wide as or slightly wider than the width of the proximal part of the digit III impression; proximal part of the pads of digits II and IV situated in the medial-distal part of the heel pad; manus tracks are ovoid to rectangular, with the digit I impression in the proximal part and directed medially.
Description All the information on the type series is in Lockley [76].

Comments
Lockley [76] defined the ichnospecies C. leonardii on the basis of a trackway from the "mid"-Cretaceous of the USA. He noted that the main difference relative to C. magnificum is the shape of the manus tracks. Several researchers (e.g., [80,[173][174]) have assigned footprints with a bilobed heel to Caririchnium leonardii. In the original paper, Lockley [76] described a trackway in which the footprints have a poorly-preserved heel impression (shown by a dashed line or unclosed contour line) (Fig. 8C) and the heel shape is not mentioned. Lockley [77] (fig.  6) assigned to Caririchnium a bilobed footprint. This footprint was figured by Lucas et al. [16] as having a Caririchnium-like morphology.
The bilobed heel has been used as a diagnostic feature of Caririchnium (e.g., [95,170,173) and Caririchnium leonardii (e.g., [42,175]). In the present work, we have emended the diagnosis of Hadrosauropodus, and we consider it to be the only ichnogenus with a bilobed heel impression. Therefore, C. leonardii is assigned to Hadrosauropodus and not to Caririchnium.
As occurs with Iguanodontipus, several authors (e.g., [16,21]) have used the referred footprints of H. leonardii as a model for comparison instead of those that form part of the type series ( Fig. 8B-C). One example is a well-preserved quadrupedal trackway from the Dakota Group of Colorado described by Lockley [77] (fig. 6) and subsequently by Currie et al. [42]. This trackway consists of eight pes tracks with skin impressions and several well-preserved manus tracks. Taking into consideration that the outline of the footprints from the type series of C. leonardii is unreliable (see [76-77, 106, 176]), in the present paper the diagnosis has been completed with data from the second trackway (Fig. 8D-E).
The footprints of H. leonardii are characterized primarily by having a narrower heel than H. langstoni, with the notches of digits II and IV placed more distally. Nevertheless, there are no clear differences with respect to the pes tracks of H. kyoungsookimi. The rectangular manus tracks of H. leonardii are different from those of H. langstoni and H. kyoungsookimi (see above).

Emended diagnosis
Pes tracks belonging to Hadrosauropodus with a heel impression as wide as or slightly wider than the width of the proximal part of the digit III impression; proximal part of the pads of digits II and IV situated in the medial-distal part of heel pad; manus tracks are crescent in outline, with three circular digit impressions of about the same size.
Type horizon Jindong Formation, Lower Cretaceous, upper Aptian [80]. Type locality Duhori area, Goseong County, Korea [80]. Distribution Jindong Formation, Aptian, Korea [80]. Synonymy Caririchnium kyoungsookimi Lim, Lockley and Kong, [80] Referred material Tracks from the Jindong Formation of Korea [80] p. 103, figs. 2-3, NHCG 10194. Description All the information on the type series is in Lim et al. [80]. Comments Lim et al. [80] defined this ichnospecies on the basis of two pairs of manus-pes tracks preserved in a block of stone from the Aptian of Korea (Fig. 8F). They compared C. kyoungsookimi with ichnospecies of Caririchnium, Iguanodontipus and Amblydactylus, and suggested differences in the shape of the manus tracks. However, Lim et al. [80] did not compare the material with Hadrosauropodus, with which it shares a bilobed heel impression. In the present work, we propose to assign C. kyoungsookimi to Hadrosauropodus.
The bilobed heel impression of H. kyoungsookimi is similar to that of H. leonardii, and narrower than that of H. langstoni. Bearing in mind the view propounded by Lockley [76] and Lim et al. [80] that the manus track shape is a diagnostic feature, H. kyoungsookimi is different from the other ichnospecies. Nevertheless, further studies should analyse the variability in the manus track impressions observed in the quadruped trackways of H. leonardii in order to discuss possible synonyms.

Geographical and temporal distribution
The ichnofamily Iguanodontipodidae is mainly related to tracks found in the Cretaceousfrom the Berriasian to the Maastrichtian-of Europe, Asia, North America and South America (Fig. 9). This distribution could yet be extended in time, since large ornithopod tracks have been described from the Late Jurassic (e.g., [177][178][179][180]), or geographically, if we take into consideration some tracks found in Australia [12] and Africa (Cameroon [181]; Morocco [13][14]). We consider that these tracks could have ornithopod affinities but do not show diagnostic characters of the ichnofamily Iguanodontipodidae.
Iguanodontipodid tracks are present in all the Cretaceous stages, but there is a stratigraphic hiatus spanning from the Turonian to the Coniacian (Fig. 9), which may be correlated with transgressive sea-level phases [182]. This hiatus appears to be more marked in Europe than in North America and South America, with the apparent absence of tracks and trackways in the Cenomanian and even in the Santonian if Apulosauripus is considered to be a non-ornithopod ichnotaxon (see above).
The geographical and temporal distribution of Iguanodontipus is very limited. All the tracks referred to I. burreyi come from the earliest Cretaceous (Berriasian-Valanginian) of Europe, including England, Germany and Spain (Fig. 10).
The ichnogenus Caririchnium has a wider distribution, both temporally and geographically, than Iguanodontipus. Caririchnium has been identified in the Early Cretaceous of South America, North America, Asia and Europe (Fig. 10). The type ichnospecies C. magnificum is known from the Berriasian-Hauterivian of Brazil to the Barremian-Albian of Spain (Fig. 10B). C. lotus also has an intercontinental distribution: this ichnospecies is known in the Barremian-Albian of China and Spain (Fig. 10B). As regards C. kortmeyeri (previously referred to Amblydactylus), this has only been cited in the Aptian-Albian of Canada (Fig. 10B). Finally, C. billsarjeanti (originally assigned to Iguanodontipus) is known exclusively from the Aptian-Albian of Switzerland (Fig. 10B).
Tracks of Hadrosauropodus have been identified in Asia, North America and Europe (Fig. 10). The oldest records (previously referred to Caririchnium) are those of H. kyoungsookimi from the Aptian-Albian of South Korea and H. leonardii from the Albian-Cenomanian of the United States. H. langstoni has only been cited in its type locality from the Maastrichtian of Canada, but Hadrosauropodus tracks have been identified in latest Cretaceous formations in Europe and Asia (Fig. 10B).
The data presented here show the temporal and geographical trends of the ichnotaxa in question. This analysis is based on the large ornithopod tracks previously assigned to a particular ichnogenus or ichnospecies (S1 Table). Tracks classified informally as "large ornithopod footprints" or referred to indeterminate iguanodonts and hadrosaurs will be studied in detail in a further paper with the aim of ascertaining the precise temporal and geographical distribution of Iguanodontipodidae.

Possible identity of trackmakers
Ornithopoda is a taxon defined as all ornithischians more closely related to Edmontosaurus than to Triceratops [183]. Ornithopoda ("bird feet") was the name used by Marsh [184] to designate bipedal, unarmoured herbivorous dinosaurs. It has been used for a long time as a taxonomic wastebasket, into which almost all bipedal ornithischians have been placed ( [185][186], and references). Recent phylogenetic studies have restricted Ornithopoda to a clade that includes a paraphyletic assemblage of "hypsilophodontids" and iguanodontians comprising tenontosaurs, rhabdodontids, dryosaurids, "camptosaurids", "iguanodontids" and hadrosauroids. Ornithopods are known in the fossil skeletal record from the Middle Jurassic (e.g., the earliest dryosaurid Callovosaurus; see [187]) to the end of the Cretaceous.
Large ornithopods consist mostly of iguanodontian forms, with the exception of dryosaurids (2-4 m in length) and rhabdodontids (up to 5 m length). Well-preserved skeletal material of iguanodontians has been recorded from the Late Jurassic to the Late Cretaceous in fossiliferous sites in all continents [188]. By the Early Cretaceous, large iguanodontians were widely distributed, being present in Europe, North America, Asia, Africa and Australia. Hadrosauroids reached a near-cosmopolitan distribution during the Late Cretaceous, with records in all landmasses except Africa, Australia and India.
Based on their size and morphology, the tracks assigned to Iguanodontipodidae would correspond to iguanodontian ichnites. The apparent absence of iguanodontipodid tracks in the Late Jurassic could be an artefact due to taphonomic or ecological biases. Moreover, it cannot be ruled out that Late Jurassic iguanodontians produced different tracks from the Cretaceous ones [189].
The ichnotaxon Iguanodontipus burreyi was proposed by Sarjeant et al. [23] to accommodate "Iguanodon footprints" and those of typical "iguanodontids" (commonly regarded as a paraphyletic assemblage; [190] and references; but see [191] for a different interpretation). Tracks of Iguanodontipus burreyi are limited to the basal Cretaceous (Berriasian-Valanginian) of Europe (Fig. 10B). Consequently, the trackmaker cannot be Iguanodon, because I. bernissartensis-the only currently recognized species of the genus-is known from the Barremian-early Aptian of England and Belgium (see [192] for a revised taxonomy of Wealden iguanodontians and references). Iguanodontipus tracks might be associated with basal members of Ankylopollexia or Styracosterna from the Berriasian-Valanginian of Europe (e.g., Barilium dawsoni and Hypselospinus fittoni, both from the Valanginian Lower Wealden Group of England, see [193]; or Owenodon hoggii from the Berriasian Purbeck Beds of England; see [194][195]), but these ichnites cannot be reliably assigned to particular taxa in the absence of detailed knowledge of the foot anatomy of these iguanodontian ornithopods [196].
Caririchnium, and particularly the type ichnospecies C. magnificum, spans all the Early Cretaceous: the oldest records correspond to the basal Cretaceous of South America; other ichnospecies of Caririchnium are known in the Barremian-Aptian of Europe, North America and Asia (Fig. 10). The C. magnificum trackway in Brazil could have been made by a basal iguanodontian, perhaps a basal ankylopollexian or styracosternan (although there is no skeletal record of large representatives of these clades in the Early Cretaceous of South America; [146,197]). The same interpretation can be applied to the other ichnospecies of Caririchnium, namely C. billsarjeanti, C. kortmeyeri and C. lotus. Large-sized representatives of basal styracosternans and "iguanodontids" are known in the late Early Cretaceous of North America (Hippodraco, Iguanacolossus, Theiophytalia; see [190,198], and references), Europe (Iguanodon, Mantellisaurus; see [192,199]) and, tentatively, Asia (Lanzhousaurus; [200]). Iguanodontids (sensu [191]) are also known in the late Early Cretaceous of Africa (Ouranosaurus), though there is still no evidence of Caririchnium or large ornithopod tracks in this part of Gondwanaland.
The temporal distribution of the ichnogenus Hadrosauropodus (Aptian-Maastrichtian) is quite coherent with the skeletal fossil record of Hadrosauroidea. Hadrosauroidea consists of all taxa more closely related to Edmontosaurus than to Iguanodon [186], and includes the hadrosaurids, which became the most diverse and abundant large vertebrates of Laurasia during the second half of the Late Cretaceous [201]. The earliest hadrosauroids are known from the late Early Cretaceous, and were particularly diversified in Asia [202] and to a lesser extent in North America ( [186], and references). On stratigraphical grounds, the tracks of Hadrosauropodus kyoungsookimi from the Barremian-Albian of Asia and H. leonardii from the Albian-Cenomanian of North America could have been made by non-hadrosaurid hadrosauroids (such as Altirhinus, Bolong, Equijubus, Jinzhousaurus, Probactrosaurus and Xuwulong from the Barremian-Albian of Asia, and Eolambia and Protohadros from the Cenomanian of North America), whereas those of H. langstoni from the Maastrichtian of North America and Hadrosauropodus isp. from the Maastrichtian of Europe could have been made by derived hadrosaurids.
However, the data currently available do not permit confident identification of the trackmakers on the basis of morphological features.

Conclusions
The ichnotaxonomy of large ornithopod tracks has been revised: of 44 ichnospecies described in the literature only eight are here considered to be valid. These ichnospecies are grouped into three ichnogenera: Iguanodontipus, Caririchnium and Hadrosauropodus, mainly on the basis of the size and shape of the heel and digit pad impressions. The manus track shape is of ichnotaxonomic value, but the diagnosis of large ornithopod ichnotaxa should be primarily based on the features of the pes tracks. The monospecific ichnogenus Iguanodontipus is mainly characterized by a narrow, rounded heel, and long, thin digit impressions. The distribution of I. burreyi is limited to the basal Cretaceous (Berriasian-Valanginian) of Europe. Caririchnium is characterized by a large, rounded heel, and short digit impressions. Four ichnospecies of Caririchnium have been recognized in this work: C. magnificum (type ichnospecies), C. lotus, C. kortmeyeri (formerly referred to Amblydactylus) and C. billsarjeanti (previously referred to Iguanodontipus). Caririchnium is known in the Early Cretaceous (Berriasian-Albian) of South America, North America, Asia and Europe. Hadrosauropodus is characterized by a wide, bilobed heel, and short, wide digit impressions. Hadrosauropodus consists of the type ichnospecies H. langstoni and two ichnospecies that were formerly assigned to Caririchnium: H. leonardii and H. kyoungsookimi. Hadrosauropodus ranges from the Aptian to the Maastrichtian, and is known in North America, Asia and Europe.
Iguanodontipus, Caririchnium and Hadrosauropodus are included within the ichnofamily Iguanodontipodidae. This ichnofamily is characterized mainly by mesaxonic, tridactyl, subsymmetrical pes tracks that are as wide as (or wider than) long and have one pad impression in each digit and one in the heel. Tracks belonging to Iguanodontipodidae can be assigned to iguanodontian ornithopods: those of Iguanodontipus and Caririchnium could have been made by basal representatives of Ankylopollexia or Styracosterna, whereas those of Hadrosauropodus could have been made by hadrosauroids. Iguanodontipodid tracks can confidently be said to be distributed in the Cretaceous of Europe, Asia, North America and South America.
Supporting Information S1 Table. Citations of large ornithopod ichnogenera studied in this work. Ã paper in which the ichnotaxon was described. (DOCX) S1 Text. Data on large ornithopod ichnotaxa (in alphabetical order): diagnosis, holotype, type horizon and type locality. The diagnoses are in the original language. (DOCX) into English. The present work is based on a chapter of the doctoral thesis of the first author; the Ph.D. dissertation was read in November 2013 at the Universidad de La Rioja (Spain).