The historically-famous Lotus Fortress site, a deep 1.5–3.0-meter-high, 200-meter-long horizonal notch high up in near-vertical sandstone cliffs comprising the Cretaceous Jiaguan Formation, has been known since the 13th Century as an impregnable defensive position. The site is also extraordinary for having multiple tetrapod track-bearing levels, of which the lower two form the floor of part of the notch, and yield very well preserved asseamblages of ornithopod, bird (avian theropod) and pterosaur tracks. Trackway counts indicate that ornithopods dominate (69%) accounting for at least 165 trackmakers, followed by bird (18%), sauropod (10%), and pterosaur (3%). Previous studies designated Lotus Fortress as the type locality of Caririchnium lotus and Wupus agilis both of which are recognized here as valid ichnotaxa. On the basis of multiple parallel trackways both are interpreted as representing the trackways of gregarious species. C. lotus is redescribed here in detail and interpreted to indicate two age cohorts representing subadults that were sometimes bipedal and larger quadrupedal adults. Two other previously described dinosaurian ichnospecies, are here reinterpreted as underprints and considered nomina dubia. Like a growing number of significant tetrapod tracksites in China the Lotus Fortress site reveals new information about the composition of tetrapod faunas from formations in which the skeletal record is sparse. In particular, the site shows the relatively high abundance of Caririchium in a region where saurischian ichnofaunas are often dominant. It is also the only site known to have yielded Wupus agilis. In combination with information from other tracksites from the Jiaguan formation and other Cretaceous formations in the region, the track record is proving increasingly impotant as a major source of information on the vertebrate faunas of the region. The Lotus Fortress site has been developed as a spectacular, geologically-, paleontologically- and a culturally-significant destination within Qijiang National Geological Park.
Citation: Xing L, Lockley MG, Marty D, Zhang J, Wang Y, Klein H, et al. (2015) An Ornithopod-Dominated Tracksite from the Lower Cretaceous Jiaguan Formation (Barremian–Albian) of Qijiang, South-Central China: New Discoveries, Ichnotaxonomy, Preservation and Palaeoecology. PLoS ONE 10(10): e0141059. https://doi.org/10.1371/journal.pone.0141059
Editor: Matt Friedman, University of Oxford, UNITED KINGDOM
Received: May 14, 2015; Accepted: October 1, 2015; Published: October 22, 2015
Copyright: © 2015 Xing et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited
Data Availability: All relevant data are within the paper and its Supporting Information files.
Funding: This research was supported by a special project grant of the Qijiang District Bureau of Land Resources, Chongqing (No. QDBLR-2007-2015) (LX); the Research of Paleoenvironment in Early Cretaceous Qijiang Dinosaur Assemblage (No. CQGT-KJ-2014057) (HD, LX) and the National Natural Science Foundation of China (No. 41402017) (YW).
Competing interests: The authors have declared that no competing interests exist.
Abbreviations: MWC, Museum of Western Colorado, U.S.A; SMU, Southern Methodist University, Texas, USA; QI–VII, Qijiang Layer I–VII, Lotus tracksite Chongqing Municipality, China; UCM, University of Colorado Museum of Natural History, USA
There has long been an absence of Cretaceous dinosaur fossils in south-central China, although the Late Jurassic record is well represented by the rich Shunosaurus-Mamenchisaurus fauna. Early discoveries of theropod and ornithopod tracks in Lower Cretaceous strata of south-central China offered a small glimpse of the Cretaceous fauna , but a more significant ichnological evidence was not reported until after 2007 . Since then, multiple other Lower Cretaceous (Jiaguan Formation) tracksites have been found [3, 4].
Xing et al. [2,4–6] described dinosaur/pterosaur ichnoassemblages from the Lotus tracksite, Qijiang National Geological Park located in Qijiang District, south of Chongqing Municipality near the southeastern border of the Sichuan Basin. The Lotus tracksite includes over 300 tracks of ornithopods, non-avian theropods, birds, pterosaurs and sauropods . Xing et al. [7–10] reported roughly 1000 theropod, sauropod, and ornithopod tracks from the Zhaojue tracksite, Zhaojue Region, near the southern border of the Sichuan Basin. These assemblages helped to fill the gap in the tetrapod fossil record, revealing a distinct change in the ecology of south-central China after the Late Jurassic epoch that was dominated by the Shunosaurus-Mamenchisaurus fauna .
Because of the well-preserved dinosaur tracks and the unique setting underneath a waterfall in a historic fortress, in the heart of the Danxia landscape, the Lotus tracksite has become a national and international tourist attraction (Fig 1). However, despite the site's fame, many of its fossil tracks have remained poorly described. In November 2012, an international team investigated the Lotus tracksite, mapped the entire site on transparent plastic film (cataloged as CUGB-Q), and measured and photographed selected tracks for 2D and 3D analyses. Here we offer a re-description of these tracks and document new aspects of their morphology, preservation history, and paleoecology.
History of Research
In 2006, Qijiang Land and Resources Bureau and South-East Sichuan Province Geological Team discovered over 100 dinosaur tracks at the historically-famous Lotus Fortress (GPS: 29° 1'11.62"N, 106°45'26.20"E), Hongyan Village in Laoying Mountain area, Qijiang, Chongqing. Xing et al.  described these tracks and attributed them to four vertebrate ichnotaxa: Caririchnium lotus, Wupus agilis, Laoyingshanpus torridus and Qijiangpus sinensis. Based in part on the importance of the tracksite, the Ministry of Land and Resources of PRC established Qijiang National Geological Park, in 2009, which includes the track-bearing areas within its protection zone, along with an extensive Jurassic petrified forest.
Xing et al. [12–13] discussed the often surprisingly intimate relationship between dinosaur tracks and Chinese folktales. The name “Lotus” tracksite reflects the local belief that the track site represented lotus leaf veins (the mud cracks) and petals (the ornithopod tracks) submerged in water (the ripple marks). Lotus Fortress is famous as a castle stronghold dating back to the time of the Mongol invasions of the late 13th century (Southern Song Dynasty Baoyou 4th Year, A.D. 1256), and humans have been living at Lotus tracksite for over 700 years (Fig 1). During this period, most tracks were covered with soil to make castle grounds more comfortable and, thus, the tracks were largely protected, despite the abundant human traffic. Today the site, previously very difficult of access, is now approachable by a steep series of about 800 steps designed to help visitors reach the site with relative ease.
In 2011, pterosaur tracks were first recognized by Daqing Li, from the Geological Museum of Gansu, and one of us (FW). In 2012, one 3D Caririchnium lotus pes track from the Lotus tracksite was described . Xing et al.  gave the first detailed description of the Pteraichnus tracks and evaluation of their paleoecological significance. Xing et al.  reviewed Wupus agilis from the Lotus tracksite. Wupus, originally identified as the trace of a small non-avian theropod track-maker , is now considered to be the track of a large avian and referred to the ichnofamily Limiavipedidae.
1 Jiaguan Formation
Qijiang National Geological Park is situated in the eastern part of the Yantze Platform and at the southeastern border of Sichuan Basin. From bottom to top, the exposed strata include the Middle Jurassic Shangshaximiao and Suining formations, the Upper Jurassic Penglaizhen Formation, the Lower Cretaceous Jiaguan Formation and unconsolidated Quaternary deposits mainly exposed along river banks and hillsides . The Penglaizhen Formation and the Jiaguan Formation are separated by a non-angular unconformity. Strata at the Lotus tracksite are more than 700 meters thick, with the Upper Jurassic Pengliazhen Formation (about 340 m) at the base and the Lower Cretaceous Jiaguan Formation (about 390 m) on top (Fig 2). The lithological association of the Jiaguan Formation consists of massive sandstones intercalated with thinner mudstone intervals, exposed at the Lotus site in an impressive near-vertical cliff face. The track-and wrinkle structure-bearing levels first occur in dark purple redquartz sandstone in the lower part of the Jiaguan Formation about 30–40 m above the base of the unit [2, 4, 6]. The beds are near-horizontal with the result that notches have been eroded horizontally into the steep vertical cliff faces by removal of the soft siltstones and mudstones. The main tracksite (levels QI and QII) comprise the floor of one of the notches (Figs 1A and 3), developed as the fortress, and reveals tracks that are particularly well preserved. Other track-bearing levels (QIII–QVII) occur within this notch at higher levels (Fig 2).
Based on ostracod distributions, Li et al.  referred the lower part of the Jiaguan Formation to the Lower Cretaceous, the middle part to the Middle Cretaceous and the upper part to the Upper Cretaceous. Based on total magnetochronology and ESR dating, the Jiaguan Formation was formed between 117–85 Ma (Aptian–Santonian)  and 140–85 Ma (Valanginian–Santonian) . However, recent pollen studies indicate a Barremian–Albian age for the Jiaguan Formation  and this latter age assignment is adopted here.
2 Depositional environment
At the research area, the Lower Cretaceous Jiaguan Formation in mainly composed of alternating thick purple red sandstone layers and thin purple red mudstone and siltstone layers, and bottom layers of thick conglomerate. The maturity of the Lotus tracksite area sediments is quite high, the rocks largely consist of quartz and feldspar (mainly potassium feldspar) with a little of debris and limestone. The sediments are divided into different zones based on increasing grain sizes from top to bottom. Various bedding plains are present within the purple red sandstone layers, including convolute beddings, tabular cross-bedding, wedge cross-bedding, current bedding, and parallel bedding. Many of the sandstones are lenticular and contain rip-up clasts of the underlying siltstones and mudstones. Some of the sandstone surfaces display current ripples, and deep desiccation cracks are common in the siltstones .
Dai et al.  analyzed grain size in sandstone samples from the Lotus tracksite and found that the cumulative grain size curve showed a bi-modal pattern that is inferred to represent a moderate slope “bouncing” grain population and low slope suspension population, with the former being dominant. The cut-off points of the bouncing population and the suspension population are between 3 to 3.5Φ. This evidence suggests a meandering river as the likely depositional environment .
3 Invertebrates traces
The Lotus tracksite also preserves many invertebrate traces, including Scoyenia gracilis, Beaconites antarcticus, and Planolites beverleyensis , among which Scoyenia dominates. All of these ichnogenera pertain to the Scoyenia ichnofacies  and are fodinichnia type traces. Scoyenia and Beaconites reflect intermittent emergence in a low-energy ultra-shallow water environment [20–21]. In river systems, the Scoyenia ichnofacies typically appears in over-bank deposits, such as floodplains, ponds, and flood fans [22–23]. Planolites, however, is seen in all kinds of sedimentary environments . The trace makers of Scoyenia and Beaconites were probably arthropods [22, 25]. Buatois and Mangano  suggested that the trackmakers of Planolites in nonmarine environments were also arthropods.
Invertebrate traces from Emei Region in the western Sichuan Basin also come from the Jiaguan Formation and include at least twelve ichnogenera and two identified ichno-assemblages: (1) Scoyenia-Steinichnus-Rusophycus and (2) Skolithos-Arenicolites. These traces formed in frequently drought-prone fluvial environments, mostly in flood plain deposit [24, 27]. By studying invertebrate traces in the same area, Chen  identified five ichnofabrics: Arenicolites, Skolithos, Scoyenia, Planolites and Palaeophycus. Invertebrate traces from the Lotus tracksite are similar to those from Emei Region, reflecting a river environment with periodic flooding and frequent droughts.
4 Microbial mats
From a macroscopic point of view, Dai et al.  identified and described two different wrinkle structure types from the Lotus tracksite. By applying microstructure analysis with a scanning electron microscope (SEM) with energy dispersive spectroscopy (EDS) and other high-resolution instrumentation that detected sheath-like and globular organic matter, Dai et al.  inferred the microbial origin of the observed wrinkle structures.
The sandy substrate was covered with a thin (0.5–3 mm) microbial mat with wrinkle structures. When a trackmaker stepped on the microbial mat, a well-defined small displacement rim formed all around the track. The microbial mat served as a water-resisting layer which may have kept underlying sediment relatively moist even while the mat itself was dry. The superficial microbial mat led to cracking around the foot rather than an outward transmission of the applied force; in such case, a relatively deep and well-preserved footprint was formed when a trackmaker walked on this kind of substrate. The presence of a microbial mat appears, therefore, to have been an important, if not the crucial, factor for the exquisite preservation of the vertebrate tracks at the Lotus tracksite. The microbial mat may have enhanced the stabilization and/or early precipitation of carbonate and hence have consolidated the tracks.
Material and Methods
In November 2012, the entire Lotus site was mapped on transparent plastic film (Fig 3, S1 Fig). The tracks of Layers 1 and 2 were measured and photographed for 2D and 3D analyses. The original tracings on plastic film have been reposited at the Qijiang National Geological Park. Replicas were made of several sets using latex for the initial molds and plaster of Paris and fibreglass replicas. They are housed in the Qijiang National Geological Park Museum, with additional replicas in the University of Colorado collections.
The maximum track length (L), maximum width (W), maximum depth (D), pace length (PL), stride length (SL), pace angulation (PA), rotation (R), trackway width (TW) and the angle between digits II–III, III–IV were measured for the ornithopod trackways.
Photogrammetric images were produced from multiple digital photographs (Canon EOS 5D Mark III) which were converted into scaled, highly accurate 3D textured mesh models using Agisoft Photoscan Professional. The mesh models were then imported into Cloud Compare where the models were rendered with accurately scaled color topographic profiles.
Distribution of Dinosaur Tracks
The tracks occur on at least seven surfaces, referred to as layers QI to QVII (Figs 2 and 4). Due to changes of sediment thickness, we measured three sections (A–C) at the Lotus tracksite. These show a thinning of the section from the NNW (section A) to the SSE (section C).
An ornithopod track cast (A) and sauropod track cast (B) from QIII, ornithopod track mold and cast (C and D respectively) from QIV, ornithopod track casts (E and F) from QV, ornithopod track casts (G and H) from QVI, ornithopod track casts (F, I–K) from QVII.
QI and II: The lowest surface (QI) is dominated by trackways of the small tridactyl ichnospecies Wupus agilis and Pteraichnus pterosaur trackways. The second layer (QII), is 10 cm higher and contains the trackways of the large ornithopod Caririchnium lotus and a few invertebrate traces. These two different ichnoassemblages are present within only a thin stratigraphic interval, and therefore, the depth of the C. lotus tracks on QII is sufficient to leave undertracks on the QI surface.
QIII: The third layer, about 50 cm above the QII layer. Several infilled tracks are present on the undersides of overlying, bench-forming sandstone layers . The Qijiang District Bureau of Land Resources collected ornithopod casts, including two complexly overprinted series. Additionally, there are seven small ornithopod trackway molds, which most likely belong to Caririchnium lotus, on a collapsed sandstone slab.
QIV: The fourth layer is found in both sections B and C and is about 1.2 m above QIII. It preserves mud cracks, ripple marks, and only 2–3 Caririchnium lotus molds in poor preservation.
QV and QVI: The fifth and sixth layers are only found in section A and are about 0.5 m and 1m above the fourth layer, respectively. Tracks include both sauropod and ornithopod morphotypes. All sauropod tracks are deep casts. Ornithopod sandstone casts occur with extremely large mud cracks which are regarded as "lotus leafs" by locals .
QVII: The seventh layer is identified in sections B and C and is about 1 m higher than the underlying layer. A high density of invertebrate traces co-occurring with dinosaur tracks is a distinguishing feature of this layer. Most of the tracks were left by ornithopods while one is an isolated sauropod track.
1 Systematic Ichnology
Ornithischia Seeley, 1888
Ornithopoda Marsh, 1881
Iguanodontipodidae Lockley et al 2014
Ichnogenus Caririchnium Leonardi, 1984
Ichnospecies C. lotus, Xing et al. 2007
A complete manus-pes set of natural-mold tracks, catalogued as QII-O20-RP2 and RM2 (former specimen number: QJGM-T37-3) from the Lotus tracksite (Figs 5 and 6, Table 1). The original specimens remain in the field.
Holotype shown in box of trackway QII-O20.
Specimens, QII-O20-RP1–RM1, and LP2–LM2 comprise two manus-pes sets of natural mold tracks in the same trackway as the holotype (Fig 5, Table 1). Specimens, QII-O9-LP4–LM4, RP4–RM4, and LP5–LM5 comprise three manus-pes sets. As with the holotype, these specimens remain in the field. However, one manus-pes set is preserved as a rubber mold and replica in the University of Colorado collections as UCM 214.256.
Locality and horizon.
Lotus tracksite, Qijiang, Chongqing, Lower Cretaceous (Barremian–Albian), Jiaguan Formation, China.
Large size (~35 cm) quadrupedal ornithopod tracks. Pes trace mesaxonic, functionally tridactyl, with quadripartite morphology, consisting of impressions of three digits and a heel pad separated by pronounced ridges. Mean length /width ratio 1.1. Mean value of mesaxony, measured as L/W of anterior triangle 0.37. Manus trace suboval to semicircular, situated anterolaterally to pes trace, sometimes with faint traces of anteromedially positioned digit. Typical heteropody (ratio of manus to pes size) 1:6.1.
The Lotus tracksite reveals thirty-seven ornithopod trackways, catalogued as QI-O1‒O6, QII-O1‒O24 and QIII-O1‒O7; at least thirty-eight isolated ornithopod tracks were also preserved on the second layer. All these tracks are natural molds (concave epireliefs). Lotus ornithopod pes traces are ~15 cm to ~48 cm in length. Trackways QII-O1, O2, O4, O7, O9, O11, O13, O14 and O17‒O21 are longer than 30 cm and are referred to as Type A. Other pes traces shorter than 30 cm are referred to Type B. The authors take two sets of representative tracks from Type A and Type B for detailed description.
Type A (Figs 5 and 6; Tables 1 and 2): QII-O20 is the holotype trackway, revealing the most complete sequential manus-pes sets. The average L/W ratio is 1.2 for the pes and 0.7 for the manus. The manus impression is rotated approximately 46° outward from the trackway axis. This outward rotation is much larger than the inward rotation of the pes impressions (approximately 5°). The average manus PA is 149°, while the average pes PA is 165°.
The holotype pes-manus couple QII-O20-RP2-RM2 is the best-preserved example (Fig 6). The pes trace RP2 is mesaxonic, functionally tridactyl and plantigrade with a length of 30.5 cm, and shows quadripartite morphology, consisting of impressions of three digits and a heel pad separated by pronounced ridges, which, in life, represented well defined concave-up creases that separated the convex-down pads, as seen in natural casts which represent close approximations of foot replicas. The L/W ratio is 1.2 while the anterior triangle L/W ratio is 0.29. The digit III trace is the shortest, but most anteriorly-situated, while traces of digits II and IV are longer and almost equal in length. Each digit trace has a strong and blunt claw or ungual mark, which in the lateral digit is sharper than in digit III. The heel is sub-triangular. There is a distinct border between the heel and the three digits. The interdigital divarication II–IV is 57°. The divarication angle between digits II and III (28°) is almost equal to the one between digits III and IV (29°). The manus trace is oval, with no discernable digit or claw marks. The short axis of the oval manus trace aligns with the antero-lateral margins of the pes (i.e. aligned almost with the tip of digit III trace). The ratio of manus centre-pes center distance/pes trace length (mc-pcD/p’L) is 0.8. The heteropody (ratio of manus to pes size) is 1:6.1.
QII-O9 is a paratype trackway, andone of the best-preserved examples of Lotus ornithopod tracks. RP4-RM4 is the best preserved set. QII-O9 tracks are basically consistent with the holotype morphology, with an average L/W ratio of 1.1, and an average anterior triangle L/W ratio of 0.38 for the pes. A digit II claw trace can be seen in RM4. The mc-pcD/p’L ratio is 1.0. The heteropody is 1:5.6
In other large-sized sets, LM4, a manus track of QII-O14, has three depressions which may correspond to traces of digits II, III and IV. It is similar to the manus impression seen in a specimen from Lamar, Colorado (Denver Museum of Natural History #1608) . RP2, a pes trace of QII-O14, has two associated manus tracks which might be formed by two consecutive steps when the trackmaker lost balance. Several tracks (e.g. O11-RP2, O21) show considerable morphological variation, which is probably due to the original substrate being wet and slippery.
Type B (Figs 6–8; Tables 1 and 2): these small-sized tracks are only half as deep as Type A or less. Most of them are poorly preserved, however, some display excellent morphological features. QII-O3 and O6 are the best preserved and most representative. The pes-only QII-O3 trackway represents one of the smallest ornithopod individuals from the Lotus tracksite. The mean length of the QII-O3 trackway is 20.6 cm; the average ML/MW ratio is 1.2; the pes impression is rotated approximately 21° inward from the trackway axis. QII-O3-RP1is best preserved with an average ML/MW ratio of 0.9; digits II–IV are similar in length and digit II has the most developed ungual mark; the interdigital divarication II–IV is 63°; the divarication angle between digits II and III (32°) is almost equal to the one between digits III and IV (31°); the anterior triangle L/W ratio is 0.33.
Note that specimens UCM 214.258–214.260 are preserved as replicas in the University of Colorado collections.
The pes-manus association QII-O6-LP2–LM2 is also one of the best-preserved examples. Digit III is the shortest, digit III and IV are almost the same in length, and digit II has the most developed ungual mark. The anterior triangle L/W ratio is 0.27. The manus trace is almost semicircular in shape. The manus trace is situated anterior to pes digit III. The mc-pcD/p’L ratio is 1.3, the heteropody 1:8.3.
All seven trackways of QIII belong to type B, among which QIII-O1–O4 are true well-preserved tracks and O5–O7 are undertracks. The walking direction of the former trackmaker group was from east to west and that of the latter group from north to south. These tracks are morphologicallysimilar to Type B tracks (QII-O3, O6) from level II.
OI17 is the smallest pes. OI17 shows only one complete lateral digit and most part of digit III, while the remaining portion is overprinted by QII-O20-RP2. Based on complete type B specimens, OI17 is probably 11–12 cm in length, less than a third the size of type A, and is likely a juvenile.
- In well-defined ichnotaxa, sizes of tracks can reflect the size and age of the individual trackmakers [34–35]. The strong similarity in morphology suggests that type B tracks probably represent juveniles or subadults of type A. The scatter diagram (Fig 9, Table 3) of length and width of the pes tracks shows that most tracks fall in the ranges of 20–24 cm and 33–37 cm, and this likely reflects two age cohorts, although other explanations are possible (e.g., sexual dimorphism).
- Generally, the manus impressions are strongly rotated outward from the trackway axis, and the pes impressions are rotated slightly inward.
- The L/W and PL/L of all pes tracks are similar, with consistent averages (1.1) and medians (2.6).
- In type B, only QII-O3, O16, and OIII-O1 lack manus traces, if not due to bipedal gait, possibly because the original manus tracks were too shallow to be preserved. Xing et al.  considered the possibility that subadult trackmakers usually walked only on their hind feet [36–37], but there is no unequivocal ichnological or osteological evidence to support this interpretation.
- In type A tracks, the axis of the manus trace always aligns with the antero-lateral margins of the pes trace. However, in type B tracks, the positions of the manus traces appear more variable. In addition to those with similar configurations to type A, others align more with the anterior margins of the pes trace: i.e., anterior to the distal end of digit III.
- Mesaxony (0.33–0.35) of pes tracks in type B is slightly smaller than that (0.37–0.52) of type A (Fig 9, Table 4). Thus, there is a slight tendency towards elongation of the anterior triangle (stronger mesaxony) in the smaller tracks in group B. This is consistent with observations by Lockley  and references therein.
- The area ratio of pes and manus tracks in type B is generally smaller than type A, so adult trackmakers had higher heteropody.
- Most Lotus ornithopod trackmakers from levels I and II went from west to east, suggesting that this area was a possible thoroughfare for trackmakers perhaps controlled by some physical landscape features . The west-to-east direction may have been parallel to a river course or shoreline . The bird and pterosaur tracks have the same general orientation. Bird trackways run parallel to the interpreted direction of the river’s flow . This suggests that the trackmakers may have been foraging, as has been observed in Goseongornipes isp.  and Koreanaornis cf. hamanensis trackways .
3 Comparisons and discussion
Lower Cretaceous ornithopod trackways are well-known from Europe, North America, and East Asia. To date, six valid ornithopod ichnogenera have been named from the Cretaceous: Amblydactylus (two ichnospecies), Caririchnium (four ichnospecies), Iguanodontipus, and Ornithopodichnus from the Lower Cretaceous, and Hadrosauropodus (two ichnospecies) and Jiayinosauropus  from the Upper Cretaceous. For historical reasons, three ichnogenera first named Amblydactylus, Caririchnium and Iguanodontipus were inferred to have been made by ornithopods of the genus Iguanodon, or similar iguanodontian trackmakers. Lockley et al.  referred them to the ichnofamily Iguanodontipodidae based on morphological similarity.
The holotype trackway of Caririchnium was originally named by Leonardi . C. magnificum is based on a well-preserved trackway of a quadruped from the Antenor Navarro Formation (Fig 10), in the lower part of the Rio do Peixe Group (pre Aptian) of Brazil . The mesaxony of the type specimen of C. magnificum is 0.31; the heteropody is 1:3.7. The former value is lower than that measured for the Lotus specimens, whereas the latter is higher. The manus traces of C. magnificum are irregular in size and shape, ranging from a large and irregular ‘L’ shaped trace to oval or subcircular.
The second ichnospecies of Caririchnium, C. leonardii from the upper part of the Dakota Group (Albian–Cenomanian) of Colorado, USA, [31, 33, 44], is similar to C. magnificum in overall morphology, however, C. leonardii differs from C. magnificum in the configuration of the manus and the shape of the heel . Some specimens later referred to C. leonardii are well preserved, with skin impressions, such as trackway MWC 201.1 from the South Platte Formation (Late Aptian-Early Cenomanian) of the Dakota Group of Colorado, USA . MWC 201.1 has a bilobed (2-lobed) heel. Moreover, the mesaxony of the type specimen of C. leonardii is 0.46 and the heteropody is 1:8.1. As such, C. leonardii has stronger mesaxony and weaker heteropody than the Lotus specimens.
The holotype of the ichnospecies Caririchnium protohadrosaurichnos  comes from the Woodbine Formation (Cenomanian) of Texas, USA. C. protohadrosaurichnos has a less defined quadripartite pes and a more elongate manus . The mesaxony of the type specimen of C. protohadrosaurichnos is 0.39 (based on the holotype SMU 74653), and the heteropody is 1:14.6. The latter suggests an extremely small manus for C. protohadrosaurichnos. All these features differ to some degree from to those in the Lotus specimens.
Caririchnium kyoungsookimi is a quadrupedal ornithopod trackway from the Jindong Formation (Late Aptian) of Korea . The manus shows three unusual subcircular indentations arranged in an elongate arc, a pattern which is unlike any seen in other Caririchnium ichnospecies, although it is somewhat similar to ornithopod tracks from the basal Cretaceous of Germany .
Additionally, Caririchnium can be found in the Antlers Formation (Aptian-Albian) of Oklahoma , the Mesa Rica Sandstone and Pajarito formations (late Albian) of New Mexico [50–51] and the Patuxent Formation (Aptian) of Virginia from USA;, the Jindong Formation (Late Aptian) of Korea  and the Amagodani Formation (Barremian) of Japan . Lockley et al.  noted that, all holotype Caririchnium trackways preserve both manus and pes tracks, unlike other ornithopod holotype ichnotaxa, which are often based on single pes tracks. Thus, the Caririchium trackmakers were typically quadrupedal. The approximate age range of Caririchnium (similar to Amblydactylus) is ~Barremian–Cenomanian, which corresponds to the age of the Lotus specimens (Barremian–Albian).
4 Caririchnium from China
Besides the Lotus tracksite, Caririchnium has already been found at five other tracksites of China (Fig 11).
A, Caririchnium isp. from Jiufotang Formation, Luanping tracksite, Hebei Province [55–56]; B, Caririchnium type from Tongfosi Formation, Tongfosi tracksite, Jilin Province ; C, Caririchnium from Hekou group, Yanguoxia tracksite No. II, Gansu Province; D, Caririchnium type from Hekou group, Yanguoxia tracksite No. I ; E, Caririchnium from Hekou group, SS1 site, Gansu Province ; F, Caririchnium from Jiaguan Formation, Longjing tracksite, Sichuan Province ; G and H, Caririchnium from Feitianshan Formation, Zhaojue tracksites, Sichuan Province ; I and J, this text. Scale bar = 10 cm.
- Caririchnium isp. tracks from the Early Cretaceous Jiufotang Formation, Luanping tracksite, Hebei Province. You and Azuma  first reported these ornithopod trackways in a large exposure along a railway. Matsukawa et al.  mapped this tracksite and referred the tracks to Caririchnium. These tracks represent quadrupedal ornithopods and are mostly not very well preserved. The pes tracks show quadripartite morphology, and the mesaxony value is 0.29. The manus traces are oval, the short axis aligns with the antero-lateral margins of the pes trace. Slender-toed theropod tracks (Asinodopodus) are also found at this tracksite.
- Caririchnium type tracks from the Early Cretaceous or early Late Cretaceous Tongfosi Formation, Tongfosi tracksite, Yanji City, Jilin Province . These tracks are poorly preserved. Manus track are absent and the pes track shows possible quadripartite morphology. Gracile theropod tracks are also found at this tracksite.
- Caririchnium type tracks from the Early Cretaceous Hekou group, Yanguoxia tracksites, Gansu Province [58–59]. At the Yanguoxia tracksites, ornithopod tracks are preserved as in situ trackways and as natural casts. The trackways from site II and site VI are typical Caririchnium. The mesaxony value is 0.30. These tracks show that the trackmakers were gregarious. The inter-trackway spacing is fairly regular at about 1.3 m. Thus, only about 4 m separates the four trackways . These tracks represent bipedal and quadrupedal ornithopods.
GDM-Y-SS1-1 from Yanguoxia site SS1 and a natural cast from site I  have relatively strong mesaxony values, reaching 0.38 and 0.41, respectively. However, the natural casts are similar to Caririchnium in overall morphology, including quadripartite morphology and oval or triangular heel pad with a bilobed posterior margin. Such difference may arise from variable individual development or preservation factors and requires further study.
- Caririchnium from the Early Cretaceous Jiaguan Formation, Longjing tracksite, Sichuan Province . All trackways from Longjing tracksite occur on a sandstone bedding plane in a river bed, and, consequently, are subject to continued erosion. The ornithopod trackway that lacks a manus imprint is assigned to Caririchnium . The best preserved LJ-O1-R1 is 22 cm long and has a mesaxony value of 0.34. Generally, it is quite similar to Caririchnium lotus type B. Digit II is the shortest but thickest, but this feature may be the result of weathering.
- Caririchnium from the Lower Cretaceous Feitianshan Formation, Zhaojue tracksites, Sichuan Province . The Zhaojue Caririchnium pes tracks have lengths ranging between about 20–30 cm, and indicate quadrupeds or facultative bipeds, among which the best preserved ZJII-O98 and O99 are generally similar to Caririchnium lotus with mesaxony values of 0.34–0.38. The Caririchnium trackmakers were almost certainly gregarious. Besides Caririchnium, Ornithopodichnus corresponding to smaller bipeds are also found at the Zhaojue tracksites.
So, the large ornithopod tracks from the Lotus, Longjing and Zhaojue sites of the Lower Cretaceous Sichuan-Yunan Basin are morphologically similar to, and most likely belong to, Caririchnium lotus. Numerous tracksites demonstrate that ornithopods were flourishing in the basin during the Early Cretaceous. Contemporary large ornithopod tracks from Lanzhou-Minhe Basin and Northeast China require further comparison.
In this regard it is worth noting the extensive samples of Caririchnium reported from the Cretaceous of Korea . In many cases the abundant Korean assemblages show up to several dozen parallel trackways, with regular inter-trackway spacing, which strongly indicate gregarious behavior. Most of the Korean trackways indicate bipedal progression.
5 Speed estimates
Thulborn  suggests that for large ornithopods (p’L > 25 cm) h = 5.9* p’L and that for small ornithopods (p’L < 25 cm) h = 4.8* p’L. The relative stride length (SL/h) may be used to determine whether the animal is walking (SL/h≤ 2.0), trotting (2<SL/h<2.9), or running (SL/h≥2.9) [62–63]. The SL/h ratios calculated for the Lotus ornithopod trackways type A range from 0.41 to 0.98 and accordingly suggest a walking gait. Using the formula of Alexander , the speed of the trackmakers ranges between an estimated 1.12‒4.14 km/s. The type B trackways also indicates a walking speed, the SL/h ratios ranges from 0.98 to 1.72, and 2.59 and 6.91 km/s (Table 5). Obviously, minor Type B walked much faster than adult type A.
Caririchnium lotus tracks from the Lotus tracksite are preserved in different ways, including typical impressions or molds (concave epireliefs), natural casts, (convex hyporeliefs) deep casts and undertracks. This variable preservation can help to give insight into morphological difference and variation between ornithopod tracks formed under different substrate conditions. However, most of the trackways of surface QII show exceptionally good preservation.
6. 1 Toe-only tracks.
The second layer preserves about five isolated toe-only Caririchnium natural pes molds, with QII-OI10 being the most distinct (Fig 12). OI10 has three separated, rounded distal digit impressions but lacks a heel impression. Analyzing the well preserved Caririchnium trackway QII-O5 helps us to understand such toe-only tracks. QII-O5 RP1-RP4 show a well preserved quadripartite morphology consisting of impressions of three digits and a heel pad separated by pronounced ridges, while RP5 is overlapped by a sauropod pes undertrack which compressed the RP5 bearing sediment and lead to external morphological distortions of RP5. In this way, the relatively shallow part RP5 is flattened, and the tracks were turned from quadripartite tracks to toe-only pes prints. However, OI10 is not covered by a sauropod undertrack and was likely left when the substrate was relatively firm. Therefore, only the distal ends of the digits, which were generally more deeply impressed than the rest of the print, were registered. Alternatively, it could be the undertrack of an overlying track which disappeared due to denudation of sediment.
These extramorphological variants of C. lotus formed the basis for two ichnospecies, which we reject here as nomina dubia. See text for details.
Tracks similar to OI10 were also discovered at the Early Cretaceous Yanguoxia tracksite in Yongjing, Gansu  and the Early Cretaceous Huangyanquan tracksite in Xijiang, China . The Yanguoxia specimens are thought to have been made by an ornithopod, trackmaker on a partially submerged substrate, that propelled itself by toe-only steps, leaving a subaqueously registered trail [64, 66]. However, this interpretation is unduly complex, and we prefer to consider the toe-only traces as evidence of animals walking on the overlying bed (QII) while parts of their feet (toes) penetrated to the underlying layer (QI). The Huangyanquan specimens are interpreted as thyreophoran undertracks (possibly Deltapodus curriei) . The discovery of Lotus toe-only Caririchnium natural pes prints may help to understand the origin (behavior or preservation) of such morphology.
6.2 Misleading undertracks.
Caririchnium natural molds are well-preserved on the second layer, due to suitable, even near optimal sediment and the presence of a microbial mat. Caririchnium, from the second layer, left undertracks on the first layer 10 cm below. These undertracks register different “morphological” (i.e. extramorphological) characteristics (Fig 12). OI-O6-RP3 has narrower digits II–IV, which are similar to theropod tracks. The heel of OI-O7-LP2 is quite shallow, while lateral digits (especially digit IV) have deep distal and proximal ends, forming a pentadactyl outline with digit III. This undertrack slightly resembles typical ankylosaur track morphology. Xing et al.  referred OI-O6-RP3 to ornithopod tracks naming it Laoyingshanpus torridus, and referred OI-O7-LP2 to ankylosaur tracks naming it Qijiangpus sinensis. Based on the present detailed study we conclude that such designations are not ichnotaxonomically useful or valid, and that these two ichnotaxa are nomina dubia , better interpreted as ornithopod undertracks, presumably transmitted to layer QI by the registration of C. lotus tracks on the overlying QII surface.
6.3 Complexly overprinted series.
Xing et al.  described the deep (three-dimensional) large ornithopod sandstone cast QIII-OI20 (former specimen number: QJGM-C1) from the third layer of the Lotus tracksite. This specimen provides unique insights into the locomotor mechanics of the trackmakers and permits reconstruction of the footfall, weight-bearing, and kick-off phases of the step cycle. The third layer also preserves two complex overprinted series, including nine and seven tracks, respectively (QIII-OI1-9 and QIII-OI10-16) (Figs 13 and 14). These two track series were made by multiple individuals travelling in different directions.
The 3D color topographic profiles help to sort out the sequences of QIII-OI1–9 and QIII-OI10–16. For example, OI1 and 2 were likely made first. Then OI3 and 4 were made and the former covered OI1 and 2. Then, OI5, 6, 8 and 9 were probably made, which overlapped and resulted in external morphological distortion of the earlier tracks. OI7 destroyed the edges of OI3 and 4. OI7 and OI8 are the best preserved casts. The former shows the principal Caririchnium lotus morphology, while the latter displays increased space between the lateral digit (right one) and digit III, probably attributable to the digits splaying outward in slippery sediment . In addition, both OI7 and OI8 show digit III impressed considerably deeper than digits II and IV, a phenomenon that has been observed in some other hadrosaurid tracks .
QIII-OI10–16 and the especially shallow OI10–13 were likely made first, followed by overlapping OI14–16. Interestingly, the sediment bearing these tracks may have been compressed such that the tracks are flattened, especially OI14 and 15. During compaction, two layers of concentric outlines were formed on OI14. The flattened heel pad of OI16 is separated from the digital pad impression region by a shallower area.
Such interesting extramorphological variation shows that even tracks of the same ichnospecies can differ widely in shape due to different substrate conditions and various diagenetic processes.
Sauropod tracks from the Lotus tracksite have two preservation patterns: undertracks on the second layer and deep casts on the third and fifth–seventh layers (Fig 15, Table 6). The sediment of the third layer may have been soft and therefore most tracks are deep, such as the 37.1 cm deep Caririchnium lotus cast . The feet of the trackmakers from the third layer caused strong distortion , leaving undertracks on the second layer. There are at least 20 broad, shallow depressions interpreted as undertracks transmitted onto the second layer. Three of these depressions form a pes only trackway catalogued as QII-S1, and four depressions form clear pes and manus traces, catalogued as QII-SI1p-1m–3p-3m.
Photograph (C, D, E and G) of sauropod casts from QVI. Close-up photographs (F, H) show details with striation marks. Dotted line indicates outline of undertracks. Arrow indicates moving direction of tracks.
The QII-S1 pes impression is oval with an average length of 60.3 cm and a L/W ratio of 1.2. The track turns outward by about 22°. The digit traces of QII-S1 are too indistinct to recognize with confidence, and the metatarso-phalangeal region is smoothly curved. QII-SI1p is oval and 65.0 cm in length, similar to QII-S1, and has a L/W ratio of 1.1. QII-SI1m is slightly U-shaped and has a L/W ratio of 0.6. Though these are only undertracks, they show the typical morphology of sauropod pes-manus prints [69–70].
For the trackways with both manus and pes traces, gauge (trackway width) was quantified for pes and manus tracks by using the ratio between the width of the angulation pattern of the pes (WAP) and the pes length (L) [71–72]. If the ratio is smaller than 1.0, tracks intersect the trackway midline, which corresponds to the definition of narrow-gauge . Accordingly, a value of 1.0 separates narrow-gauge from medium-gauge trackways, whereas the value 1.2 has been used to distinguish between medium-gauge and wide-gauge trackways. The WAP/ML ratio of QII-S1 is 1.4, which is between medium-gauge and wide-gauge trackway. On the other hand, the factors affecting gauge may include the speed of the trackmaker [74–75] and the quality of preservation. In reference to this latter factor, it is important, as noted in the following section, to differentiate between true tracks with steep walls and well-defined outlines and undertracks (transmitted tracks are usually wider) with very low angle margins, which may reduce the inner trackway width and thus estimations of gauge . If based on the displacement rim of QII-S1, the WAP/ML ratio is 1.1, more close to narrow-gauge.
Most sauropod trackways in China are wide- (or medium-) gauge and are therefore referred to the ichnogenus Brontopodus . The wide-gauge of Brontopodus-type trackways suggests that the tracks were left by titanosauriform sauropods [70, 77]. The Lotus sauropod tracks are also consistent with the characteristics of Brontopodus type tracks from the Lower Cretaceous of the USA [69, 78]. These characteristics include medium-gauge/wide-gauge trackways, outwardly-directed pes tracks that are longer than wide, and a high degree of heteropody (ratio of manus to pes size). The heteropody ratio of 1:2.6 in QII-SI1 is close to 1:3 in Brontopodus birdi. Unfortunately, effective statistical evaluation is difficult due to the low relief and indistinct margins of the Lotus sauropod undertracks. Therefore, despite being measured as narrow-gauge, these tracks are provisionally assigned to the Brontopodus-type.
Theropod (with bird) and ornithopod tracks from the Jiaguan Formation were the only track types first described by Zhen et al.  from Emei Region. Sauropod tracks were not reported from that assemblage, a conclusion confirmed by Matsukawa et al. . Although a local geologic report  mentioned a quadruped trackway, the tracks were not re-located in a 1982 field survey conducted by the Chongqing Natural Museum , or during the 2001 field investigation leading to the report of Matsukawa et al. . Subsequently, narrow- and medium-gauge sauropod tracks were found at Hanxi  and Xinyang  tracksites. Hanxi specimens have WAP/ML ratios of 0.9 to 1.1, and the ratio of Xinyang specimens is 1.4. Therefore, sauropod tracks from the Jiaguan Formation are mostly narrow- and medium-gauge, consistent with Early Cretaceous sauropod tracks from other sites in south-central China, such as the Zhaojue tracksites in Sichuan Province , but differ from wide-gauge tracks of the Yanguoxia tracksite in Gansu Province . This may imply that titanosauriform sauropods from different Early Cretaceous basins were distinct.
2 Natural casts
The sauropod track casts are deep natural tracks left in soft and moist substrates with a relatively high cohesiveness. They offer a glimpse into the three-dimensional foot morphology of the sauropod trackmakers and their foot movement (locomotion). Among the deep sauropod casts on the third and fifth–seventh layers, casts on the sixth layer are the best preserved and include at least four sauropod casts, catalogued as QVI-SI1–4.
For these deep sauropod casts, Xing et al.  suggested a more specific measurement and descriptive approach. However, QVI-SI1–4 is not fully exposed and thus some data are inaccessible. The upper surface of QVI-SI1 is about 55 cm in length and its lower surface is about 41 cm in length. QVI-SI1 has three clear digit traces in the anterior area, which are spaced by two grooves. The lateral digit is the largest and most likely to be digit I while the other two are likely digits II and III. The 23 cm deep digit area of QVI-SI1 is shallow, while the heel is 28 cm in depth. Obvious concave deformation can be seen below the track, forming a 14 cm deep undertrack area. Vertical to the track (especially digit I area) are several grooves and invertebrates trace casts. The former are spaced about 2–3cm apart and are probably traces made by the polygonal skin texture of sauropods which typically consists of an integument with a tightly-packed tubercle-like mosaic of polygons each up to 2–3 cm in diameter. Similarly wide striation traces have been observed in association with sauropod tracks at other localities . The filled invertebrate traces suggest that, after the track was left, invertebrates lived or foraged in the depressions.
QVI-SI2 is behind QVI-SI1 and is oval shaped. Its upper surface is about 29 cm in length and its lower surface is about 24 cm in length and 17cm in depth. Based on position and size, QVI-SI2 is probably a manus track, and belonging to the same trackway as QVI-SI1. The lower surface of QVI-SI2 is crossed by two large mud cracks (about 2.5 cm deep). It is worth noting that about nine to ten 1.8–2.5 cm wide grooves are distributed transversely on the flanks of the track cast, forming small angles (~10°) with the upper and lower surfaces, and are not aligning with the common grooves, which are more vertical to the casts. They may be formed by the trackmaker’s polygonal skin ornament when turning its feet transversely in the sediment. In addition, at least two invertebrates traces run vertically to the track.
QVI-SI3 and SI4 are incomplete. The former is about 19 cm deep, similar to QVI-SI1 in morphology and size, most likely to be a pes cast, and has an undertrack about 5 cm deep. Partially preserved, QVI-SI4 is about 23cm deep and has well-preserved striation marks (about 1–2 cm wide) vertical to the cast. A cast is lost between QVI-SI3 and SI2 but there is an undertrack area about 24 cm deep. The undertack area likely formed when the sediment was affected by another pes trace.
The sixth layer also has about six casts of ornithopod tracks. These ornithopod tracks are about 7 cm deep and are generally much shallower than the sauropod casts; they co-occur with developed mud cracks. The sauropod trackmakers likely left tracks on relatively wet substrate, then tracks and substrate dried, and large mud cracks formed before the ornithopod tracks were left. In future, more sauropod tracks are likely to be discovered at the base of the sandstone ledges, ribbons, fins, and within the mudstone beds of the Jiaguan Formation. Such tracks are particularly conspicuous due to their large size.
Trace fossils provide the only records of Early Cretaceous birds from many parts of the world. Previously, bird tracks from the Jiaguan Formation were rare. Specimens from the Emei tracksite were named Aquatilavipes sinensis , but A. sinensis was later reassigned to Koreanaornis sinensis . Wupus from the Lotus tracksite, was originally identified as the trace of a small theropod track-maker  (Fig 16). It is similar in both footprint and trackway characteristics to the Early Cretaceous (Albian) large avian trace Limiavipes curriei  from western Canada, and Wupus is reassigned to the ichnofamily Limiavipedidae . Reanalysis of Wupus agilis indicates that it represents the traces of a relatively large avian track-maker, analogous to extant herons. The analysis also reveals that, despite the current lack of body fossils, large wading birds were present globally in both Laurasia and Gondwana during the Early Cretaceous [4, 85–87]. A notable feature of the Wupus track assemblage on surface QI is that the trackways are nearly all parallel with an eastward orientation.
Non-Avian Theropod Tracks
Lotus tracksite lacks identifiable non-avian theropod tracks and only has two isolated tracks QI-BI48 and QII-OI12 (they were initially recognized as ornithopod undertracks and thus given a prefix of "O") from the first and second layer (Fig 17).
QI-BI48 is 20.7 cm in length, tridactyl, and has a L/W ratio of 1.16. It presents three slender digits with sharp claw marks at the distal ends, and an oval shaped metatarsophalangeal pad. Digit III has three digit pads, and the digit pad of the lateral digits is unclear. QI-BI48 co-occurs with Wupus agilis and is nearly identical to the latter in morphology, although it is much larger (~10 cm in W. agilis). QII-OI12 is a tridactyl track of 19 cm length and has a L/W ratio of 1.39. Digit III shows a sharp claw mark, while lateral digits are poorly preserved with indistinct digit pads. The developed metatarsophalangeal region is smoothly curved.
QI-BI48 and QII-OI12 meet morphological features of non-avian theropod track . The mesaxony of QI-BI48 and QII-OI12 are 0.47 and 0.58, respectively, which is close to the footprints of the ichno- or morphofamily Eubrontidae . Although QI-BI48 and QII-OI12 show affinity with non-avian theropod tracks, the small quantity and poor preservation complicates further comparison and discussion.
The Early Cretaceous was the height of pterosaur radiation. The number of known pterosaur tracksite reports from China has increased substantially in recent years. It compares favorably with the growing record from Korea and Japan, and adds significantly to the record from East Asia, such as Jimo, Shandong Province , Lotus, Chongqing Municipality , Xinjiang [91–92], Zhaojue, Sichuan Province , and Yanguoxia, Gansu Province .
The Lotus tracksite is one of the most important pterosaur tracksites in the Cretaceous of China due to the large number of imprints and their co-occurrence with bird tracks (Wupus). Xing et al.  described thirty tracks from five Pteraichnus trackways from the Lotus tracksite, and interpreted them as left by the same kind and similarly sized pterosaurs.
As in other Pteraichnus specimens from China, the Pteraichnus trackmaker from the Lotus tracksite was most likely a small to medium-sized pterodactyloid. However, a non-pterodactyloid trackmaker cannot be ruled out, because the impression of pedal digit V is rarely impressed clearly and unambiguously .
Tetrapod ichnofaunas from different layers at the Lotus tracksite vary in composition. Pterosaur and bird tracks were only registered on the first layer (QI), while ornithopod and sauropod tracks co-occur on some layers, but occur separately on others. Xing et al.  pointed out that small-sized theropods and pterosaurs probably preferred a stable sand bed, rather than highly-saturated muddy silt where they might sink in deep and expend more energy when walking . After the second layer was covered by 10 cm of sandy sediment, a group of ornithopod track makers left tracks, some of which were transmitted to level QI as undertracks. Most tracks from the other layers are deep casts. This indicates the substrates represented by these layers were very soft.
If one trackway or one isolated track is regarded as reflecting an individual trackmaker, the abundance of trackmakers and proportions of different groups from the Lotus tracksite can be estimated (Table 7). Ornithopods dominate (69%) accounting for at least 165 trackmakers, followed by bird (18%), sauropod (10%), and pterosaur (3%). These proportions are similar to those at the Zhaojue tracksites, that are also located in the Sichuan-Yunan Basin. Ornithopods dominate (42%) in 76 trackmakers from Zhaojue tracksites, followed by theropods (25%), sauropods (24%) and pterosaurs (9%) [7–9]. However, Early Cretaceous ornithopod-dominated tracksites are unusual in China. Other important dinosaur-dominated footprint assemblages in China are mainly composed of saurischians, such as the theropod-dominated (~90%) Houzuoshan Dinosaur Park site from the Lower Cretaceous Tianjialou Formation at Junan County, Shandong Province . The Yanguoxia tracksites are sauropod-dominated (38%), followed in abundance by theropods (32%), ornithopods (18%), birds (6%) and pterosaurs (3%) . The Lower Cretaceous Jingchuan Formation represented at the Chabu sites in Inner Mongolia  lacks any confidently identified ornithopod tracks and is sauropod-theropod-bird dominated . Besides, the Jehol Fauna, the most important and diversified Early Cretaceous dinosaur fauna in China , has produced countless specimens which are held in many institutions, therefore making a complete statistics nearly impossible. Therefore, based on only the Shandong Tianyu Museum of Nature which holds the biggest collection (about 3500 specimens of dinosaur-pterosaur material), ornithopods only account for 1%, while birds and non-ornithopod/theropod animals account for 65% and 19%, respectively. These differences may be explained in many ways, but most interpretations are necessarily speculative. In any event, the samples are very different in composition, in part due to sampling different sites, different regions and different facies all of which contribute certain biases to the fossil record. However, from the unusual ornithopod domination in the Qijiang track assemblage, in low latitude Early Cretaceous fluvial-lacustrine facies environments of the Sichuan-Yunan Basin, we can conclude that ornithopods were sometimes at least as commonas sauropods in these settings.
- The Caririchnium lotus track assemblages associated with levels I, II and III are among the best-preserved and most significant of any known Cretaceous track assemblages, comprising at least 28 measurable trackways and an equal number of isolated specimens.
- The trackways are morphologically similar to other Caririchnium ichnospecies from Brazil, North America and Korea, but differ in the configuration of the manus in most cases.
- As is the case in most of these other regions the parallel trackways indicate gregarious behavior.
- The assemblages indicate that at least two distinct cohorts were present indicating larger adults (Type A) and smaller sub adults (Type B). The adults all appear to have progressed quadrupedally, perhaps as obligate quadrupeds, whereas the smaller sub-adults progressed both as bipeds and quadrupeds: i.e., as facultative bipeds.
- The Caririchnium assemblages are associated with bird- like tracks assigned to the ichnogenus Wupus, which is currently only known from this locality, and described in detail elsewhere , and the pterosaur tracks Pteraichnus also described in detail elsewhere . Both these ichnogenera are confined to level QI.
- Tracks named Laoyingshanpus torridus and Qijiangpus sinensis by Xing et al.  are here revaluated and regarded as undertracks transmitted from level II to level I. They are therefore considered extramorphological ichnites and here referred to as nomina dubia.
- Casts of large sauropod tracks and ornithopod tracks occur at other levels (especially IV-VII).
- Collectively the seven track-bearing levels indicate the presence of ornithopods, birds, pterosaurs and sauropods, with the former two groups being numerically dominant on the basis of raw trackway counts.
- Collectively such ichnofaunas add vastly to our knowledge of the Cretaceous faunas known from the Jiaguan Formation and from the Lower Cretaceous of this region, which is otherwise very poorly represented by body fossils.
S1 Fig. Map of track-bearing level at QI and II of the Lotus tracksite with trackway numbers.
Thanks for Luis Alcalá and an anonymous reviewer for their helpful reviews of the manuscript. This research was supported by a special project grant of the Qijiang District Bureau of Land Resources, Chongqing (No. QDBLR-2007-2015); the Research of Paleoenvironment in Early Cretaceous Qijiang Dinosaur Assemblage (No. CQGT-KJ-2014057) and the National Natural Science Foundation of China (No. 41402017).
Conceived and designed the experiments: LX ML. Performed the experiments: LX ML DM JZ YW HK RM LB MB OM WP FW HR HD XX GG LP. Analyzed the data: LX ML DM. Contributed reagents/materials/analysis tools: LX ML DM. Wrote the paper: LX ML DM JZ HK RM LB MB OM WP.
- 1. Zhen S, Li J, Zhang B. Dinosaur and bird footprints from the Lower Cretaceous of Emei County, Sichuan, China. Mem. Beijing Nat Hist Mus. 1994; 54: 105–120.
- 2. Xing LD, Wang FP, Pan SG, Chen W. The Discovery of Dinosaur Footprints from the Middle Cretaceous Jiaguan Formation of Qijiang County, Chongqing City. Acta Geol Sin Chin Ed. 2007; 81: 1591–1602.
- 3. Lu TQ, Zhang XL, Chen L. Dinosaur tracks in vertical sections from the Upper Cretaceous Jiaguan Formation of Emei, Sichuan Province. Acta Pal Sin. 2013; 52:518–525.
- 4. Xing LD, Buckley LG, McCrea RT, Lockley MG, Zhang JP, Piñuela L, et al. Reanalysis of Wupus agilis (Early Cretaceous) of Chongqing, China as a large avian trace: differentiating between large bird and small theropod tracks. PLoS ONE 10(5): e0124039. pmid:25993285
- 5. Xing LD, Bell PR, Harris JD, Currie PJ. An unusual, three–dimensionally preserved, large hadrosauriform pes track from “mid”–Cretaceous Jiaguan Formation of Chongqing, China. Acta Geol Sin Engl Ed. 2012; 86: 304–312.
- 6. Xing LD, Lockley MG, Piñuela L, Zhang JP, Klein H, Li DQ, et al. Pterosaur trackways from the Lower Cretaceous Jiaguan Formation (Barremian–Albian) of Qijiang, Southwest China. Palaeogeogr Palaeoclimatol Palaeoecol. 2013; 392: 177–185.
- 7. Xing LD, Lockley MG, Zhang JP, Milner ARC, Klein H, Li DQ, et al. A new Early Cretaceous dinosaur track assemblage and the first definite non–avian theropod swim trackway from China. Chin Sci Bull Engl Ed. 2013; 58: 2370–2378.
- 8. Xing LD, Lockley MG, Zhang JP, Klein H, Persons WS IV, Dai H. Diverse sauropod–, theropod–, and ornithopod–track assemblages and a new ichnotaxon Siamopodus xui ichnosp. nov. from the Feitianshan Formation, Lower Cretaceous of Sichuan Province, southwest China. Palaeogeogr Palaeoclimatol Palaeoecol. 2014; 414: 79–97.
- 9. Xing LD, Lockley MG, Marty D, Piñuela L, Klein H, Zhang JP, et al. Re–description of the partially collapsed Early Cretaceous Zhaojue dinosaur tracksite (Sichuan Province, China) by using previously registered video coverage. Cretac Res. 2015; 52: 138–152.
- 10. Xing LD, Lockley MG. First report of small Ornithopodichnus trackways from the Lower Cretaceous of Sichuan, China. Ichnos. 2014; 21: 213–222.
- 11. Peng GZ, Ye Y, Gao YH, Shu CK, and Jiang S. Jurassic Dinosaur Faunas in Zigong. Chengdu: People’s Publishing House of Sichuan; 2005.
- 12. Xing LD, Mayor A, Chen Y, Harris JD, Burns ME. The folklore of dinosaur trackways in China: Impact on Paleontology. Ichnos. 2011; 18: 4, 213–220.
- 13. Xing LD, Mayor A, Chen Y. Lianhua Baozhai (Lotus Mountain Fortress, Qijiang County of Chongqing City): A direct evidence of co–existing ancient Chinese and dinosaurtracks. Geol Bull Chin. 2011; 30: 1530–1537.
- 14. Liu Y, He ZW, Long XJ, Li XQ, Li NJ, Gong LM. Characteristics and geological significance of geological relics in Qijiang Geopark, Chongqing City. Chin J Geol Haz Contr. 2010; 21: 118–124.
- 15. Li YW, Wang XH, Gao YR. The ostracods and the age of Jiading Group, Sichuan. Proc Chin Acad Geol Sci. 1983; 6:107–124
- 16. Li YL. Daxi conglomerate and its geological time. J Chengdu Univ Techn. 1995; 22: 11–14.
- 17. Gou ZH, Zhao B. The Cretaceous and Tertiary systems in Dayi and Chongzhou Regions, Sichuan. J Stratigr. 2001; 25: 28–33.
- 18. Chen HX. Research of paleoenvironment and paleoclimate of Cretaceous in Ya'an area of western Sichuan Basin. M.Sc. Thesis, Chengdu University of Technology, China. 2009.
- 19. Dai H, Xing LD, Marty D, Zhang JP, Persons WS IV, Hu HQ, et al. Microbially-induced sedimentary wrinkle structures and possible impact of microbial mats for the enhanced preservation of dinosaur tracks from the Lower Cretaceous Jiaguan Formation near Qijiang (Chongqing, China). Cretac Res. 2015; 53: 98–109.
- 20. Yang S., Zhang J, Yang M. Trace Fossils of China. Beijing: Science Press; 2004.
- 21. Wang CZ, Hu B, Yang K. Ichnofossils and sedimentary environments of the Upper Cretaceous Qiupa Formation in Tantou Basin of western Henan. Acta Sedimentol Sin. 2014; 32: 1007–1015.
- 22. Frey RW, Pemberton SG, Fagerstrom JA. Morphological, ethological and environmental significance of the ichnogenera Scoyenia and Anchorichnus. J Paleontol. 1984; 58: 511–528.
- 23. Buatois LA, Mangano MG. Trace fossils from Carboniferous flood-plain deposits in western Argentina: implication for ichnofacies models of continental environments. Palaeogeogr Palaeoclimatol Palaeoecol. 2002; 183: 71–86
- 24. Hu B, Wu XT, Pan LM, Ichnocoenoses of the Late Paleozoic and Mesozoic fluvial deposits of Emei Area, Western Sichuan, China. Acta Sedimentol Sin. 1991; 9: 128–135
- 25. Graham JR, Pollard JE. Occurrence of the trace fossil Beaconites antarcticus in the Lower Carboniferous fluviatile rocks of County Mayo, Ireland. Palaeogeogr Palaeoclimatol Palaeoecol. 1982; 38: 257–268.
- 26. Buatois LA, Mangano M G. Trace fossils from a Carboniferous turbiditic lake: implication for the recognition of additional nonmarine ichnofacies. Ichnos. 1993; 2: 237–258
- 27. Hu B, Wu XT. Ichnocoenosis of alluvial Jiaguan Formation (Upper Cretaceous), Emei, Sichuan, China. Acta Pal Sin. 1993; 32: 478–489.
- 28. Chen L. Ichnofabrics Research of the Upper Cretaceous Jiaguan Fomation in E'mei, Sichuan. M.Sc. Thesis, Southwest Petroleum University. 2014.
- 29. Seeley HG. On the classification of the fossil animals commonly named Dinosauria. Proc R Soc London. 1888; 43: 165–171.
- 30. Marsh OC. Principal characters of American Jurassic dinosaurs, Part V. Am J Sci. 1881; 21: 417–423.
- 31. Lockley MG, Xing LD, Lockwood JAF, Pond S. A review of large Cretaceous ornithopod tracks, with special reference to their ichnotaxonomy. Biol J Linn Soc. 2014; 113: 721–736.
- 32. Leonardi G. Le impronte fossili di dinosauri. In: Bonaparte JF, Colbert EH, Currie PJ, de Rocles A, Kielan-Jaworowska Z, Leonardi G, Morello N, Taquet P, editors. Sulle ormi de dinosauri. Venice: Editio Editrice; 1984. pp. 165–186.
- 33. Lockley MG. Dinosaur footprints from the Dakota Group of Eastern Colorado. Mt Geol. 1987; 24: 107–122.
- 34. Lockley MG. Dinosaur ontogeny and population structure: Interpretations and speculations based on footprints. In: Carpenter K, Hirsch K, Horner J, editors. Dinosaur Eggs and Babies, Cambridge University Press; 1994. pp. 347–365.
- 35. Matsukawa M, Lockley MG, Hunt AP. Three age groups of ornithopods inferred from footprints in the mid Cretaceous Dakota Group, eastern Colorado, North America. Palaeogeogr Palaeoclimatol Palaeoecol. 1999; 147: 39–51.
- 36. Shipman P. How a 125-million year-old dinosaur evolved in 160 million years. Discover Mag. 1986; 10: 94–102.
- 37. Norman D. The illustrated encyclopedia of Dinosaurs. New York: Gramercy Books; 1988.
- 38. Lockley MG. New perspectives on morphological variation in tridactyl footprints: clues to widespread convergence in developmental dynamics. Geol Q. 2009: 53: 415–432.
- 39. Ostrom J. Were some dinosaurs gregarious? Palaeogeogr Palaeoclimatol Palaeoecol. 1972; 11: 287–301.
- 40. Paik IS, Kim HJ, Lee YI. Dinosaur track-bearing deposits from the Cretaceous Jindong formation, Korea: occurrence, paleoenvironments and preservation. Cretac Res. 2001; 22: 79–92.
- 41. Xing LD, Harris JD, Jia CK, Luo ZJ, Wang SN, An JF. Early Cretaceous Bird–dominated and dinosaur footprint assemblages from the northwestern margin of the Junggar Basin, Xinjiang, China. Palaeoworld. 2011; 20: 308–321.
- 42. Anfinson OA, Lockley MG, Kim SH, Kim KS, Kim JY. First report of the small bird track Koreanaornis from the Cretaceous of North America: implications for avian ichnotaxonomy and paleoecology. Cretac Res. 2009; 30(4): 885–894.
- 43. Leonardi G. Inventory and statistics of the South American dinosaurian ichnofauna and its paleobiological significance. In: Gillette D, Lockley MG, editors. Dinosaur tracks and traces. New York: Cambridge University Press; 1989. pp. 165–179.
- 44. Lockley MG, Hook N, Taylor A. A brief history of paleontological research and public education on dinosaur ridge. In: Lockley MG, Taylor A, editors. Dinosaur Ridge: celebrating a decade of discovery. Mt Geol. 2001; 38: 87–89.
- 45. Currie PJ, Nadon G, Lockley MG. Dinosaur footprints with skin impressions from the Cretaceous of Alberta and Colorado. Can J Earth Sci. 1991; 28: 102–115.
- 46. Lee YN. Bird and dinosaur footprints in the Woodbine Formation (Cenomanian) Texas. Cretac Res. 1997; 18: 849–864.
- 47. Lim JD, Lockley MG, Kong DY. The trackway of a quadrupedal ornithopod from the Jindong Formation (Cretaceous) of Korea. Ichnos. 2012; 19: 101–104.
- 48. Lockley MG, Wright JL, Thies D. Some observations on the dinosaur tracks at Münchehagen (Lower Cretaceous), Germany. Ichnos. 2004; 11: 261–274.
- 49. Lockley MG. Cretaceous dinosaur-dominated footprint assemblages: their stratigraphic and palaeoecological potential. In: Mateer NJ, Chen PJ, editors. Aspects of nonmarine Cretaceous geology. Beijing: China Ocean Press; 1992. pp. 269–282.
- 50. Hunt AP, Lucas SG. A reevaluation of the vertebrate ichnofauna of the Mesa Rica Sandstone and Pajarito formations (Lower Cretaceous: late Albian), Clayton Lake State Park, New Mexico. New Mex Geol. 1996; 18: 57.
- 51. Kappus EJ, Cornell WC. A new Cretaceous dinosaur tracksite in Southern New Mexico. Paleonto Electron. 2003; 6: 1–6.
- 52. Weems RE, Bachman JM. A dinosaur-dominated ichnofauna from the Lower Cretaceous (Aptian) Patuxent Formation of Virginia. Geological Society of America Abstracts with Programs, Denver 2004; 116.
- 53. Huh M, Hwang KG, Paik IS, Chung CH, Kim BS. Dinosaur tracks from the Cretaceous of South Korea: Distribution, occurrences and paleobiological significance.The Island Arc. 2003; 12: 132–144.
- 54. Matsukawa M, Shibata K, Kukihara R, Koarai K, Lockley MG. Review of Japanese dinosaur track localities: Implications for ichnotaxonomy, paleogeography and stratigraphic correlation. Ichnos. 2005; 12: 201–222.
- 55. You H, Azuma Y. Early Cretaceous dinosaur footprints from Luanping, Hebei Province, China. Sixth Symposium on Mesozoic Terrestrial Ecosystems and Biotas. Beijing: China Ocean Press; 1995.
- 56. Matsukawa M, Lockley M, Li JJ. Cretaceous terrestrial biotas of East Asia, with special reference to dinosaur-dominated ichnofaunas: towards a synthesis. Cretac Res. 2006; 27: 3–21.
- 57. Matsukawa M, Futakami M, Lockley MG, Chen PJ, Chen JH, Cao Z, et al. Dinosaur footprints from the Lower Cretaceous of eastern Manchuria, northeast China: evidence and implications. Palaios. 1995; 10: 3–15.
- 58. Zhang J, Li D, Li M, Lockley MG, Bai Z. Diverse dinosaur-, pterosaur- bird track assemblages from the Hakou Formation, Lower Cretaceous of Gansu Province, Northwest China. Cretac Res. 2006; 27: 44–55.
- 59. Xing LD, Li DQ, Lockley MG, Marty D, Zhang JP, Persons WS IV, et al. Dinosaur natural track casts from the Lower Cretaceous Hekou Group in the Lanzhou–Minhe Basin, Gansu, Northwest China: ichnology track formation, and distribution. Cretac Res. 2015; 52: 194–205.
- 60. Xing LD, Peng GZ, Lockley MG, Ye Y, Klein H, Zhang JP, et al. Early Cretaceous sauropod and ornithopod trackways from a stream course in Sichuan Basin, Southwest China. New Mex Mus Nat Hist Sci Bull. 2015; 68: 319–325.
- 61. Lockley MG, Houck K, Yang SY, Matsukawa M, Lim SK, Dinosaur dominated footprint assemblages from the Cretaceous Jindong Formation, Hallayo Haesang National Park, Goseong County, South Korea: evidence and implications. Cretac Res. 2006; 27, 70–101.
- 62. Thulborn RA. Dinosaur Tracks. London: Chapman and Hall; 1990.
- 63. Alexander RM. Estimates of speeds of dinosaurs. Nature. 1976; 26: 129–130.
- 64. Li D, Azuma Y, Fujita M, Lee YN, Arakawa Y. A preliminary report on two new vertebrate track sites including dinosaurs from the early Cretaceous Hekou Group, Gansu province, China. J Paleont Soc Korea. 2006; 22: 29–49.
- 65. Xing LD, Lockley MG, McCrea RT, Gierliński GD, Buckley LG, Zhang JP, et al. First Record of Deltapodus tracks from the Early Cretaceous of China. Cretac Res. 2013; 42: 55–65
- 66. Fujita M, Lee YN, Azuma Y, Li D. Unusual tridactyl trackways with tail traces from the Lower Cretaceous Hekou Group, Gansu Province, China. Palaios. 2012; 27: 560–570.
- 67. Currie P, Badamgarav D, Koppelhus E. The first Late Cretaceous footprints from the Nemegt locality in the Gobi of Mongolia. Ichnos. 2003; 10: 1–13.
- 68. Milàn J, Bromley RG. True tracks, undertracks and eroded tracks, experimental work with tetrapod tracks in laboratory and field. Palaeogeogr Palaeoclimatol Palaeoecol. 2006; 231: 253–264
- 69. Farlow JO, Pittman JG, Hawthorne JM. Brontopodus birdi. Lower Cretaceous sauropod footprints from the US Gulf coastal plain. In: Gillette DD, Lockley MG, editors. Dinosaur Tracks and Traces; 1989, pp. 371–394.
- 70. Lockley MG, Wright J, White D, Li JJ, Feng L, Li H. The first sauropod trackways from China. Cretac Res. 2002; 23: 363–381.
- 71. Marty D. Sedimentology, taphonomy, and ichnology of Late Jurassic dinosaur tracks from the Jura carbonate platform (Chevenez–Combe Ronde tracksite, NW Switzerland): insights into the tidal-flat palaeoenvironment and dinosaur diversity, locomotion, and palaeoecology. GeoFocus. 21. PhD Thesis, University of Fribourg, Fribourg. 2008.
- 72. Marty D, Belvedere M, Meyer CA, Mietto P. Paratte G, Lovis C, et al. Comparative analysis of Late Jurassic sauropod trackways from the Jura Mountains (NW Switzerland) and the central High Atlas Mountains (Morocco): implications for sauropod ichnotaxonomy. Hist Biol. 2010; 22: 109–133.
- 73. Farlow JO. Sauropod tracks and trackmakers: integrating the ichnological and skeletal record. Zubia. 1992; 10: 89–138.
- 74. Xing LD, Harris JD, Jia CK. Dinosaur tracks from the Lower Cretaceous Mengtuan Formation in Jiangsu, China and morphological diversity of local sauropod tracks. Acta Pal Sin. 2010; 49(4): 448–460.
- 75. Castanera D, Pascual C, Razzolini NL, Vila B, Barco JL, Canudo JI. Discriminating between mediumsized tridactyl trackmakers: Tracking ornithopod tracks in the base of the Cretaceous (Berriasian, Spain). PLoS ONE. 2013; 8: e81830. pmid:24303075
- 76. Xing LD, Peng GZ, Lockley MG, Ye Y, Klein H, McCrea RT, et al. Saurischian (theropod-sauropod) track assemblages from the Jiaguan Formation in the Sichuan Basin, Southwest China: Ichnology and indications to differential track preservation. Historical Biology. in press.
- 77. Wilson JA, Carrano MT. Titanosaurs and the origin of “wide-gauge” trackways: a biomechanical and systematic perspective on sauropod locomotion. Paleobiology. 1999; 25: 252–267.
- 78. Lockley MG, Farlow JO, Meyer CA. Brontopodus and Parabrontopodus ichnogen. nov. and the significance of wide- and narrow-gauge sauropod trackways. Gaia. 1994; 10: 135–145.
- 79. The No.2 Territorial Survey Team of Sichuan Geological Bureau. Report on survey of geological and mineral resources in Emei. Bejing: Geological Publishing House; 1971.
- 80. Yang XL, Yang DH. The Dinosaur Footprints from Mesozoic of Sichuan Basin (in Chinese). Chengdu: Sichuan Scientific and Technological Publishing House; 1987.
- 81. Xing LD, Lockley MG, Zhang JP, Klein H, Marty D, Peng GZ, et al. The longest theropod trackway from East Asia, and diverse sauropod–, theropod–, and ornithopod–track assemblages from the Lower Cretaceous Jiaguan Formation, southwest China. Cretac Res. 2015; 56: 345–362.
- 82. Garcia-Ramos JC, Lires J, Pinuela L. Dinosaurios: rutas por el Jurásico de Asturias. La Voz de Asturias, Lugones, Asturias, Spain, 2002.
- 83. Lockley MG, Li JJ, Li RH, Matsukawa M, Harris JD, Xing LD. A review of the tetrapod track record in China, with special reference to type ichnospecies: implications for ichnotaxonomy and paleobiology. Acta Geol Sin Engl Ed. 2013; 87: 1–20.
- 84. McCrea RT, Buckley LG, Plint AG, Currie PJ, Haggart JW, Helm CW, et al. A review of vertebrate track-bearing formations from the Mesozoic and earliest Cenozoic of western Canada with a description of a new theropod ichnospecies and reassignment of an avian ichnogenus. New Mex Mus Nat Hist Sci Bull. 2014; 62: 5–93.
- 85. Lockley MG, Yang SY, Matsukawa M, Fleming F, Lim SK. The track record of Mesozoic Birds: Evidence and Implications. Philos Trans R Soc London. 1992; 336: 113–134.
- 86. Lockley MG, Lim JD, Kim JY, Kim KS, Huh M, Hwang KG. Tracking Korea’s early birds: ichnological insights into avian evolution and behavior. Ichnos.2012; 19: 17–27.
- 87. Lockley MG, Harris J. On the trail of early birds: a review of the fossil footprint record of avian morphological evolution and behavior. Trends in Ornithological Research, Novapublishers, 2010.
- 88. Lockley MG, Hunt AP. Dinosaur tracks and other fossil footprints of the western United States. New York: Columbia University Press; 1995.
- 89. Lull RS. Fossil footprints of the Jura-Trias of North America. Mem Bost Soc Natl Hist. 1904; 5: 461–557.
- 90. Xing LD, Harris JD, Gierliński GD, Gingras MK, Divay JD, Tang YG, et al. Early Cretaceous Pterosaur tracks from a "buried" dinosaur tracksite in Shandong Province, China. Palaeoworld. 2012; 21: 50–58.
- 91. Xing LD, Lockley MG, Klein H, Zhang JP, He Q, Divay JD, et al. Dinosaur, bird and pterosaur footprints from the Lower Cretaceous of Wuerhe asphaltite area, Xinjiang, China, with notes on overlapping track relationships. Palaeoworld. 2013; 22:42–51.
- 92. He Q, Xing LD, Zhang JP, Lockley MG, Klein H, Persons WS IV, et al. New Early Cretaceous pterosaur–bird track assemblage from Xinjiang, China–palaeoethology and palaeoenvironment. Acta Geol Sin Engl Ed. 2013; 87: 1477–1485.
- 93. Li DQ, Xing LD, Lockley MG, Piñuela L, Zhang JP, Dai H, et al. A manus dominated pterosaur track assemblage from Gansu, China: implications for behavior. Sci Bull. 2015; 60: 264–272.
- 94. Lockley MG, Harris JD, Mitchell L. A global overview of pterosaur ichnology: tracksite distribution in space and time. Zittel B. 2008; 28:187–198.
- 95. Li RH, Lockley MG, Matsukawa M, Liu MW. Important Dinosaur-dominated footprint assemblages from the Lower Cretaceous Tianjialou Formation at the Houzuoshan Dinosaur Park, Junan County, Shandong Province, China. Cretac Res. 2015; 52:83–100.
- 96. Li JJ, Bai ZQ, Wei QY, On the Dinosaur Tracks from the Lower Cretaceous of Otog Qi, Inner Mongolia. Beijing: Geological Publishing House; 2011.
- 97. Zhou Z, Wang Y. Major features of the vertebrate diversity of the Early Cretaceous Jehol Biota and their paleoecological implications. J Earth Sci. 2010; 21: 228–230.