Fossil footprint discoveries at John Martin Reservoir, Bent County, Colorado

Material Information

Fossil footprint discoveries at John Martin Reservoir, Bent County, Colorado new insights into the paleoecology of the cretaceous dinosaur freeway
Kukihara, Reiji
Publication Date:
Physical Description:
xiv, 112 leaves : ; 28 cm


Subjects / Keywords:
From 65 to 140 million years ago ( fast )
Footprints, Fossil -- Colorado -- John Martin Reservoir ( lcsh )
Dinosaur tracks -- Colorado -- John Martin Reservoir ( lcsh )
Paleontology -- Cretaceous ( lcsh )
Cretaceous Geologic Period ( fast )
Dinosaur tracks ( fast )
Footprints, Fossil ( fast )
Paleontology ( fast )
Colorado -- John Martin Reservoir ( fast )
bibliography ( marcgt )
theses ( marcgt )
non-fiction ( marcgt )


Includes bibliographical references (leaves 102-112).
General Note:
Integrated Sciences Program
Statement of Responsibility:
by Reiji Kukihara.

Record Information

Source Institution:
University of Colorado Denver
Holding Location:
Auraria Library
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
71753178 ( OCLC )
LD1193.L584 2006m K84 ( lcc )

Full Text
Reiji Kukihara
B.S., Kyushu University, 2002
M.E., Tokyo Gakugei University, 2004
A thesis submitted to the
University of Colorado at Denver
in partial fulfillment
of the requirements for the degree of
Master of Integrated Science

The thesis for the Master of Integrated Science
degree by
Reiji Kukihara
has been approved
Nqffra Matthews


Kukihara, Reiji (M.S., Integrated Science)
Fossil Footprint Discoveries at John Martin Reservoir, Bent County, Colorado:
New Insights into the Paleoecology of the Cretaceous Dinosaur Freeway
Thesis directed by Professor Martin G Lockley.
The John Martin Reservoir tracksites from the Dakota Group of Bent County,
southeastern Colorado are part of the Dinosaur Freeway characterized by
abundant omithopod footprints (Caririchnium). Over 350 tracks (331
Caririchnium, one Magnoavipes, 22 crocodiles, and one pterosaur) were
discovered from 10 tracksites. All tracks were found as natural casts, and
some Caririchnium trackways were found as natural casts in their original
positions. Some of these in-situ trackways are oriented in the same direction
and spaced regularly, suggesting the organized gregarious behavior of
omithopods which is also suggested by many other Dinosaur Freeway
tracksites. Most crocodile tracks are swim tracks that consist of three or four
scratch marks. In contrast to previous studies, these were interpreted not as
dinosaurs but crocodilians from their size and morphological characters. One
pterosaur track consists of one pes footprint and scratch marks suggesting the
pterosaurs swimming or floating activity in shallow water. It is the first
pterosaur evidence from the Dakota Group. The Caririchnium size structure
from John Martin Reservoir is consistent with the size structure tendency of
the whole Dinosaur Freeway which shows larger track size in the north. This
tendency can be interpreted as evidence for more than one omithopod species
spread out through the Dinosaur Freeway area with the northern species being
larger. The alternative, that there was one omithopod species that migrated
north and south seasonally is less likely. The John Martin paleoecosystem is
interpreted from the track evidence as a well vegetated coastal plain
environment with many omithopods and a few theropods on land, pterosaurs

in the sky, and crocodiles in the water.
This abstract accurately represents the content of the candidates thesis. I
recommend its publication.
Martin G Lockley l

My thanks to advisor, Martin G. Lockley, for his patient advise, help, and
support. I thank Don Headlee and Mark Stark (US Army Corps of
Engineers) for their efforts to show me sites and support my research
investigations. I also thank Bruce Schumacher and John Abbott for tracksite
information. Partial funding for this project was provided by contract of US
Army Corps of Engineers.

1. Introduction.......................................................1
2. Geological Setting................................................6
3. Previous Studies.................................................13
3.1 Early discoveries (1930s 1980s)................................13
3.2 The Dinosaur Freeway Concept (late 1980s 2000s)...............14
4. Study Methods....................................................16
4.1 General Methods.................................................16
4.2 Trackmaker Identification.......................................16
5. Results..........................................................25
5.1 Tracksite Descriptions...........................................25

5.1.1 Plover Point.......................................................25
5.1.2 Site 1..............................................................26
5.1.3 Site 2..............................................................32
5.1.4 Site 2E.............................................................35
5.1.5 Site 2W.............................................................35
5.1.6 Site 3..............................................................40
5.1.7 Site 4..............................................................40
5.1.8 Site 5..............................................................44
5.1.9 Site 6..............................................................44
5.1.10 Site 7.............................................................48
5.1.11 Site 8.............................................................48
5.1.12 Site 9.............................................................48
5.2 Track Measurement Data................................................50
6. Discussion.............................................................55
6.1 Crocodile Swim Tracks.................................................55
6.2 The Pterosaur Swim Track..............................................65

6.3 Social Behavior and Migration of Dinosaurs
7. Conclusions....................................................89
A. Significance of tracksites and recommendations................91
B. Photographic views of tracksites..............................95

1.1 The distribution of Dakota Group dinosaur tracksites in portions of
Colorado, New Mexico, and Oklahoma. After Lockley and Hunt
2.1 Relationship of track-bearing zone to Dakota sequences from north-central
Colorado to east-central New Mexico. After Holbrook and Wright
Dunbar (1992)........................................................7
2.2 Tracksite map of John Martin Reservoir.................................9
2.3 Stratigraphic columns of John Martin Reservoir tracksites. A: Northern
part of John Martin Reservoir. B: Southern part of John Martin Reservoir.
Scales are valid for both boxes.....................................11
4.1 Track measurements. A, B: after Lockley and Hunt (1995). In C; TL= track
length, TW= track width, DL=digit length, DW=digit width, DIV=
interdigital angle......................................................17
4.2. Typical Caririchnium leonaridii........................................20
4.3 Typical crocodile swim track...........................................21
4.4 Swimming crocodile in shallow water....................................22

4.5 Pterosaur track from Site 2W
5.1 Caririchnium with above sandstone layer (CU-MWC 209.64). A: front
view, B: cross section view.....................................27
5.2 Overview of Plover Point........................................28
5.3 Crocodile tracks from Site 1...................................30
5.4 Crocodile swim tracks from Site 1..............................31
5.5 Trackway map of Site 2.........................................33
5.6 Track occurrence view of Site 2................................34
5.7 Trackway map of Site 2E.........................................36
5.8 Trackway map of Site 2W.........................................38
5.9 Crocodile swim track from Site 2W..............................39
5.10 Trackway map of Site 3.........................................41
5.11 Replicas of a trackway from Site 3............................42
5.12 Tracks from Site 4. A: Caririchnium, B: crocodile swim tracks.43
5.13 Trackway like features of Site 6...............................45
5.14 Crocodile swim tracks from Site 6..............................47
5.15 Caririchnium from Site 8......................................49

5.16 Trackway maps of Site 9......................................51
5.17 Crocodile tracks from Site 9.................................52
6.1 Comparison between Hatcherichnus and a crocodile track (CU-MWC
209.44) from Site 1. A: Hatcherichnus, B: CU-MWC 209.44. A after
Foster and Lockley (1997)...................................56
6.2 Comparison with modem alligator and Dakota crocodile pes. A: modem
alligator pes, B: CU-MWC 209.122 from Site 1. A is offered by J.O.
6.3 Swim tracks from Kansas. After McAllister (1989b)................59
6.4 The Dakota Group tracks. All tracks are shown in a same scale. A:
Kansas swim tracks, B: John Martin swim tracks, C: John Martin crocodile
tracks, D: Magnoavipes, E: Caririchnium, F: Ankylosaurid tracks. A: after
McAllister (1989b), D above: after Schumacher (2003), below: after
Lockley et al. (2004), and F: after Kurtz, et al. (2001)...........61
6.5 Cross section comparison of assumed swim tracks of the Dakota Group
trackmakers. All tracks are shown in a same scale. A: cross sections of
real crocodile swim tracks and assumed swim tracks (right: CU-MWC
209.37, left: CU-MWC 209.43), B: assumed crocodile swim track from
CU-MWC 209.122, C: assumed Caririchnium swim track from CU-MWC
209.48, D: assumed Magnoavipes swim track from Schumacher (2003), E:
assumed ankylosaurid swim track from Kurtz et al. (2001)...........62
6.6 Dinosaur swim tracks from the Middle Jurassic of Yorkshire, UK. A:
Normal footprints. B: Swim track Characichnos tridactylus. A and B
occurred on the same slab. After Whyte and Romano (2001)...........64

6.7 Albertasuchipes type specimen from the Upper Paleocene of Alberta,
Canada. After McCrea et al. (2004)....................................66
6.8 Pteraichnus type specimen from the Morrison Formation of Arizona. After
Stokes (1957).....................................................67
6.9 Pteraichnus photos and tracings. A: CU-MWC 187.14, B: CU-MWC
188.1, C: CU-MWC 188.4, D: CU-MWC 188.33, E: CU-MWC 188.39 F:
6.10 Pteraichnus with swim scratches. After Lockley and Wright (2003.73
6.11 Two patterns of reconstructed pterosaur swim activity...........76
6.12 Paleogeography of the marine embayment of the middle to late Jurassic,
showing the occurrence of Pteraichnus tracksites and John Martin
Reservoir tracksite. After Lockley and Hunt (1995).............78
6.13 A: Plots of the proportion of Caririchnium of different sizes from various
outcrops of the Dinosaur Freeway. B: The Caririchnium size proportion of
John Martin Reservoir. A after Matsukawa et al. (1999).........82
A.l Map (top) and cross section (below) showing the vertical displacement or
settling down of large track casts that resulted when the outcrop (left)
was dissolved by long term submergence in the reservoir. It is possible to
trace a sequence of 12 tracks belonging to an omithopod trackway for a
distance of about 12 meters from where they are still more or less in situ in
their normal vertical orientation location onshore to their more offshore
location where they are displaced both laterally and vertically. Tracks

appear to have been displaced to the east possibly as a result of the result
of the prevailing wind and waves...............................93
B.l Photographic views of Plover Point and Site 1.....................96
B.2 Photographic views of Site 2......................................97
B.3 Photographic views of Site 2W and 3...............................98
B.4 Photographic views of Site 4 and 5................................99
B.5 Photographic views of Site 6 and 7...............................100
B.6 Photographic views of Site 8 and 9...............................101

1.1 Numbers of recorded track and trackway types at John Martin tracksites...5
5.1 Measurements of selected loose Caririchnium from Site 1, Site 2, Site 2E,
Site 2W, and Site 9................................................53
5.2. Measurements of crocodile swim tracks from Site 1, Site 2W, Site 4, Site
7, and Site 9. *: track or digit length could be longer, p: digits are

1. Introduction
The Lower Cretaceous Dakota Group, which consists of coastal plain
sediment, is well known for its abundant dinosaur tracks and other trace fossils.
It is widely distributed from northern Colorado to northeastern New Mexico and
the Oklahoma panhandle. It had been known that the Dakota Group has tracks
since the late 19th century and the discovery of the large dinosaur tracksite, which
is now called Dinosaur Ridge, along Alameda Parkway in 1930s made it into a
famous site (Markman, 1938). After that, some tracks were reported from the
Dakota Group (McKenzie, 1968,1972, Chamberlain, 1976), however, these
dinosaur tracks were not studied in detail until the 1980s. Lockley (1985,1987)
reported dinosaur tracks from the Dakota Group and described them in detail and
systematically for the first time. Since then, more than 60 tracksites have been
reported from the Dakota Group, and more new discoveries are still being made
(Kukihara et al. 2005). These tracksites vary from small ones with only one
track to big ones with more than two or three hundred tracks. The total number of
trackways that have been counted is more than one thousand. This huge
complex of track-bearing coastal plain sediments is called the Dinosaur
Freeway not just as a catchy label for an ancient trampled coast but because it is
a coherent facies complex that can be understood in terms of sequence
stratigraphy (Lockley et al., 1992, Fig. 1.1). This remarkable abundance of

Figure 1.1

tracks has given us information about not only the fact that dinosaurs existed
abundantly in this region but also the data provides many interesting insights into
social behavior and population structures which is difficult to understand only
from sparse skeletal remains (i.e. Lockley and Hunt, 1995; Matsukawa et al.,
Recent drought conditions at John Martin Reservoir, Bent County,
southeastern Colorado revealed abundant Dakota Group outcrops and the first
discovery of a new Dinosaur Freeway tracksite with 79 dinosaur tracks at this
reservoir was reported by Schumacher (2003). Since then, nine more tracksites
with more than two hundred dinosaur footprints and other vertebrate tracks have
been found and many specimens were rescued during research in summer 2004-
summer 2005. Essentially, all tracks collected from John Martin are natural casts
although some were collected from out of their original impressions. Most of the
tracks are Caririchnium which is a three-toed omithopod track. One theropod
track, many crocodile swimming tracks and the first probable pterosaur track from
the Dakota Group were also found. Stratigraphic sections were measured at all
tracksites and some tracksites with in situ tracks were mapped. 63 in-situ
Caririchnium trackways, more than 260 loose Caririchnium, 1 loose
Magnoavipes, 22 loose crocodile walk and swim tracks, and one loose pterosaur

track were discovered (Table 1.1). Sizes of 181 Caririchnium were measured for
statistical analysis. More than 50 Caririchnium tracks, 21 crocodile tracks and
one pterosaur track all in good condition, were collected. These specimens were
traced, given
specimen numbers and housed in the CU-Denver Dinosaur Track Museum.
Replicas of some tracks and trackways were also made. Additional collected
tracks are also held at the US Army Corp of Engineers visitors center at John
Martin Reservoir. CU Denver has also provided this center with selected
replicas that are most representative of the tracks from the study area.
The purpose of this study is to report these new dinosaur track sites,
compare them with other tracksites from the Dakota Group and summarize and
interpret the whole Dakota Group tracks database through various methods of data
synthesis. The project was supported by the Army Corps of Engineers and the
offices of the Dinosaur Tracks Museum (CU Denver). These institutions
provided permission for access to sites, financial support and access to
comparative collections and research archives.

Table 1.1.
Tracktype Tracksltes Total
PP 1 2 2E 2W 3 4 5 6 7 8 9
Cartrlchnlum Trackway . . 13 20 7 10 . . . . . 13 63
Loose track 77 - 45 111 6 4 1 - - 7 17 268
Magnoavlpaa 1 1
Crocodile Walk track - 2 2
Swim track - 12 - - 1 . 2 - 2 1 - 2 20
Pterosaur - - - - 1 - - - - - - - 1

2. Geological Setting
The Cretaceous Dakota Group is distributed throughout much of the
Colorado Plateau and High Plains area. The age is estimated as Albian to
Cenomanian for the eastern part of the Group, from northeastern Colorado to
northeastern New Mexico and the Oklahoma Panhandle, including the John
Martin Reservoir area (Lockley et al., 1992; Fig. 1.1). Although there are
various interpretations that show different lithostratigraphic divisions, they all
show that the group is made up of three sequences throughout the whole area of
the Dakota distribution (Holbrook and Wright Dunbar, 1992; Fig. 2.1). It is
known that the track bearing strata of the Dakota Group are mostly restricted to
the stratigraphic interval of Sequence 3 (Lockley et al., 1992). Sequence 3
consists of terrestrial and marginal marine deposits, which accumulated during the
transgression of the Western Interior Seaway. There were many cycles of
transgression and regression affecting this seaway including the Kiowa-Skull
Creek regression (R5) and the following transgression (T6) which influenced the
deposition of the track-bearing beds in the upper part of Sequence 3 (Kauffman,
Dinosaur tracks of southeastern Colorado, the Oklahoma Panhandle, and
northeastern New Mexico are restricted to the uppermost layers of the Mesa Rica
Sandstone and its equivalent strata and the Pajarito Formation (Lockley et al.,

Figure 2.1
North-Central Colorado
Modified from
Gustason & Kauffman
South-Central Colorado
Kues and Lucas
Northern New Mexico
Kues and Lucas (1987)
Holbrook et al. (1987)
East-Central New Mexico
Graneros Shale
Graneros Shale
Graneros Shale
Graneros Shale
Mowry Shale
Mowry Shale
Muddy (J) Formation
Skull Creek Shale
Plainview Sandstone
Lytle Formation
Upper Transitional
Lower Channel
Sandstone Member
Romeroville Sandstone
Pajarito Formation
Glencairn Formation
Mesa Rica Sandstone
Glencairn Formation
Plainview Formation
Lytle Sandstone
Long Canyon
Sandstone bed
Lytle Sandstone
Romeroville Sandstone
Pajarito Formation
Mesa Rica Sandstone
Tucumcari Shale
Sandstone bed
Jurassic Morrison Formation
Jurassic Morrison Formation
Jurassic Morrison Formation
Jurassic Morrison Formation

1992). The Mesa Rica, which accumulated during R5, represents a widespread,
continuous sheet of channel sandstones deposited by lateral migration and
avulsion of rivers under conditions of relatively stable base level. The basal
Pajarito strata represent marine-influenced coastal/delta floodplain deposits which
accumulated by backfilling of Mesa Rica channels and, together with the Mesa
Rica Sandstone, reflect the deposits of an extensive, very-slowly-aggrading
coastal plain during T6 (Holbrook and Wright Dunbar, 1992). This coastal plain
is considered to have covered most of northeastern New Mexico and much of
southeastern Colorado. Other parts of the Pajarito Formation and equivalent
section are also considered to be coastal and delta plain deposits (Lockley et al.,
1992). Although body fossils of either vertebrates or invertebrates are scarce in
the Dakota Group, plant remains, dinosaur and other vertebrate tracks and
invertebrate trace fossils are common. These plant fossils suggest that
paleoecosystems of the Dakota Group were heavily vegetated lowland, coastal
plains in warm temperate to subtropical climates in which the newly evolved
angiosperms were a significant component (Lockley 1985, Lockley et al. 1992).
John Martin Reservoir which was constructed in 1948 as a part of a
comprehensive plan for flood damage reduction and water resources management
in the Arkansas River Basin is located near Hasty, Bent County, southeast
Colorado (Fig. 2.2). This reservoir was constructed by damming the Arkansas

Figure 2.2

River which cuts into the Dakota Group bedrock. Although this reservoir is the
largest body of water in southeast Colorado, the water body is presently very
small and restricted to an area near the dam because of the recent drought
(2004-2005). However, this drought revealed many Dakota outcrops with
dinosaur tracks. Only Sequence 3 of the Dakota Group, namely the Mesa Rica
Sandstone, Pajarito Formation, and Romeroville Sandstone can be seen at John
Martin Reservoir. The overlaying Graneros Shale, which is not well exposed,
was not investigated in this study because it does not contain known fossil
footprints. Dinosaur tracks and other trace fossils were found in the Pajarito
Formation which consists of alternating beds of thin sandstone and shale (Fig. 2.3).
Ripple marks can be seen commonly in each sandstone stratum. There are few
trace fossils but plenty of current ripple marks in the Mesa Rica Sandstone and the
Romeroville Sandstone. This type of restricted track occurrence is consistent
with many other tracksites of the Dakota Group in its eastern outcrop. However,
the type of track preservation is different from other the Dakota megatracksites
such as Dinosaur Ridge and Clayton Lake State Park. Although tracks occur as
molds on sandstone beds at these other megatracksite locations, the John Martin
dinosaur tracks are mostly all natural casts and sometimes they were found in situ
sitting in the soft ground which represents the coincidence of the present
weathered land surface with soft mudstone beds.

2W 2 2E
massive sandstone
Xbedded sandstone
i l u one set of Xbedded
, V\i sandstone
alternating beds of
sandstone and shale
dinosaur track
crocodile track
5 1 Plover Pt.


These soft lithologies are particularly susceptible to dissolution and erosion when
flooded by reservoir water. In addition, when under only a few meters of water,
they are above wave base. This increases erosion and we noticed the difference
in erosion of these soft layers between summer 2004 and 2005 after several sites
(e.g. 2, 2E, 2W) had been under water again for about six months (Dec 2004 -May
2005). Although we could not quantity this rate of erosion, it appeared that the
exposed surface of fine sediment making up horizontal terraces at some sites (e.g.
2, 2E, 2W) had been lowered several cm between 2004 and 2005. This exposed
new sandstone track casts and left others more eroded and protruded above the
surface that had been part of the lake bottom in previous months (winter and
spring of 2005).

3. Previous studies
3.1 Early discoveries (1930s 1980s)
Tracks from the Dakota Group have been known since the late 19th
century. One large dinosaur tracksite with more than one hundred fifty tracks
which is called the Dinosaur Ridge was found in the 1930s when Alameda
Parkway was under construction. Because of the good accessibility and
visibility of Dinosaur Ridge, dinosaur tracks from the Dakota Group received
much publicity. However, these dinosaur tracks were not studied systematically
until 1980s though several earlier reports mentioned them (McKenzie 1968, 1972,
Chamberlain 1976). Only their existence was recognized and their significance
was ignored. Lockley (1985,1987) studied these tracks in detail for the first
time and assigned them to iguanodontids and theropods. The iguanodontids
tracks are the most abundant tracks in the Dakota Group and have been named
Caririchnium leonaridii. In the Early 1990s a major excavation was conducted
at Dinosaur Ridge. This increased the number of known tracks to about 350
including 23 trackways of C. leonaridii and 14 trackways of Magnoavipes earneri
(Lockley and Hunt 1994, Lockley et al. 2001). After this, many Dakota
tracksites were found one after another along the Front Range (Lockley et al. 1992,
Lockley and Hunt 1995). Now sixty two tracksites with more than one thousand
tracks have reported from the Dakota Group. The biggest tracksites are Clayton

Lake State Park (249 trackways: Gillette and Thomas 1985), Richardson Ranch
(152 tracks: Lockley 1987), and Mosquero Creek (82 trackways: Cotton et al.,
1998; Lockley and Hunt 1995; Lucas et al., 1987; Matsukawa et al., 2001).
3.2 The Dinosaur Freeway Concept (late 1980s 2000s)
This large dinosaur tracksite distribution from northern Colorado to
northeastern New Mexico and the Oklahoma panhandle allowed researchers to
develop the concept of the Dinosaur Freeway, which is a huge dinosaur track
assemblage with billions of dinosaur tracks estimated for large areas (^
100,000km2) in well-defined strata units of the coastal plain along the shores of
the Cretaceous Western Interior Seaway (Lockley et al., 1992; Lockley and Hunt,
1995). These studies also showed that the track census structure of the Dinosaur
Freeway is characterized by overwhelming dominance of large omithopod
dinosaur tracks and some theropod tracks, crocodile swim tracks, bird tracks and
invertebrate trace fossils. Researchers have found and interpreted not only
simple dinosaur presence-absence data but also other facts that are more difficult
to find from normal size tracksites within the Dinosaur Freeway area. For
example, the omithopod dinosaurs gregarious migrant behavior along the coastal
plain was inferred through this notable abundance of tracks (Lockley and Hunt
1995; Cotton et al., 1998). Matsukawa et al. (1999) studied the track sizes

statistically, and suggested that there are more than three age groups of omithopod
from the Dinosaur Freeway, and that their herd structures show that juveniles are
accompanied by adults or sub-adults. McAllister (1989 a, b) reported some
dinosaur swim tracks from the Dakota Group of Kansas, from the other eastern
coast of the Western Interior Seaway. These so called swim tracks have been
reinterpreted as tracks of crocodilians (see discussion below).
There are different kinds of approaches to the study of the Dinosaur
Freeway. These have included stratigraphic and sedimentological studies as well
as investigations of the invertebrate ichnofacies (Chamberlain, 1985) and
vertebrate tracks of birds (Mehl, 1931), dinosaurs, and crocodilians. Some
dinosaur tracksites such as Dinosaur Ridge are very accessible and visible
outcrops. Therefore such tracksites are very useful for public education, and
community resource projects (Lockley et al., 2001). For example, Dinosaur
Ridge which was just a road cut with dinosaur tracks when it was found is now a
good education area with roofs, fences, interpretative signs, and educational
volunteer programs. There is a visitors center and shop near the site where
people can get information, guidebooks and so on. Thus the study on tracks of
the Dinosaur Freeway has contributed quite broadly to science and science

4. Study Methods
4.1 General Methods
Stratigraphic sections were measured at all tracksites (Fig. 2.3). Tracks
which are in good condition were collected regardless of whether they are in-situ
or loose. These collected specimens were recorded by tracing, measuring and
photographs. Those that were collected were reposited in the CU Denver
Dinosaur Tracks Museum Collection. Some of collected specimens were latexed
and replicated and also given numbers in the CU Denver Dinosaur Tracks
Museum Collection. In the case that in-situ trackways on the ground were found,
these tracksite areas were mapped and some trackways were traced. These
collected data were compared with data from other tracksites of the Dakota Group
and specimens made by similar track makers. Measured data were processed
4.2 Track maker identification
Track identification methods follow standard procedure as outlined in
Lockley and Hunt (1995). Tracks can be classified as 1,2, 3, 4, 5 toed (i.e.
mono-, di-, tri-, tetra-, pentadactyl). Each track is measured for length, width,
and sometimes interdigital angles (Fig.4.1 a, b). Depth was not measured
because it does not show any clues to trackmakers identify but just the hardness of

Figure 4.1

the ground. In addition, each digit length and width were measured in crocodile
swim tracks to discuss trackmakers in later section. It is also standard practice to
find whether trackmakers were bipeds or quadrupeds by identifying front (manus)
and back or hind (pes) tracks. If tracks constitute trackways, there will be more
information available such as step lengths, stride lengths, trackway widths, pace
angulations, and angle of rotation from midline (i.e., toe in or toe out pattern).
By these criteria, four kinds of vertebrate tracks were identified at John Martin
Reservoir. Vertebrate tracks include three-toed omithopods which walked in
both bipedal and quadrupedal manner, three-toed bipedal theropods, three or four
toed crocodile tracks of normal and swim tracks type, and a probable four toed
pterosaur footprint. Non-dinosaurian tracks were measured the same way,
however, swim tracks were measured in a different way. Because swim tracks
are not foot prints but scratch marks, they seldom leave whole palm impressions.
Therefore each digit length and width, and total length and width must be
measured (Fig. 4.1c). Digit length (DL) was measured from tip to tip regardless
of digit curvature. Digit width (DW) was measured perpendicular to the DL line
at the point where it is widest, and interdigital angle (DIV) was measured between
DL lines. Track length (TL) was measured parallel to the maximum DL of digit
III, and track width (TW) was measured perpendicular to the DL line of digit III at
its maximum width.

Omithopod pes tracks are the most common tracks at John Martin
Reservoir and they are characterized by a bilobed heel and three broad and
rounded digit impression. Each digit pad is clear and broad or swelled up, if
well preserved, but claw or ungual impressions are absent or indistinct (Fig. 4.2).
Size is varied and lengths vary from 17 cm to 43 cm. Rarely they are
accompanied by an elliptical manus impression. These characters corresponding
with those of the iganodontid dinosaur track Caririchnium leonaridii which is the
most common dinosaur track in the Dakota Group (Lockley 1987). Only one
theropod track was reported from John Martin Reservoir by Schumacher (2003).
It is characterized by three slender toes, pointed tips, lack of a heel, a sinuous
middle digit, and an inter-digital angle between digits II and IV of slightly less
than 90 It could be assigned to Magnoavipes which is found at some of the
Dakota tracksites (Lee, 1997; Lockley et al., 2001; Schumacher, 2003, fig. 7).
Crocodile swim tracks look like scratch marks made by three or perhaps in a few
cases four digits. Sometimes pes and manus occur together. Length varies
from about 5 to 39 cm, and width varies from about 4 to 20 cm. Cross sections
look like a skewed half ellipse (Fig. 4.3) due to the crocodiles swim behavior in
shallow water (Fig. 4.4). Crocodile swim tracks have been found at other
Dakota tracksites (e.g. McAlister 1989a, b). Although McAllister (1989a, b)
reported them as dinosaur swim tracks, they may be inferred to be crocodile swim

Figure 4.2

Figure 4.3

Figure 4.4

tracks because these swim tracks are too small compared with dinosaur tracks
found from the Dakota Group and too common for terrestrial animals swim
tracks (see discussion below). It is natural to consider them to be
swim tracks of smaller aquatic animals such as crocodiles rather than dinosaurs
because they occur in subaqueous deposits representing fluvial systems. One
unusual track was found from site 2W at John Martin Reservoir. It consists of
one pes impression and its scratch marks (Fig. 4.5). It is similar to pterosaur
tracks such as Pteraichnus found from other formations. The discussion about
this track will be given in a later section. All track bearing strata have various
invertebrate trace fossils. Arenicolites, which is a U-shaped burrow, is
particularly common.

Figure 4.5

5. Results
5.1 Tracksite Descriptions
In addition to the Plover Point site previously studied by Schumacher
(2003), we studied nine other sites, designated Sites 1-9, in order of discovery.
Because reservoir levels fluctuated shoreline position change. Therefore
reference to shorelines in the section that follows are either general (e.g., north +
south reservoir shorelines) or they are relative to the location of the Plover Point
and 1-9 sites at the time they were studied and located using GPS coordinates.
5.1.1 Plover Point
Plover Point is the first reported tracksite by Schumacher (2003) from
John Martin Reservoir (Fig. 2.2). It is situated on the northern shoreline of the
reservoir, 6km west of the dam. The Pajarito Formation and the Romeroville
Sandstone can be seen, though the upper part of the Pajarito is covered by soil and
vegetation. There are probably two track bearing layers and they are situated in
the middle of the Pajarito (Fig. 2.3). The site consists of sandstone track casts
and minimally 79 tracks were found (Schumacher, 2003). Most of them are
Caririchnium and one Magnoavipes was reported. Although Schumacher (2003)
suggested the possibility of one Ankylosaur track, preservation of the specimen
was poor and it was not relocated because the water level rose in spring 2005;

thus it was difficult to confirm in the track census. Tracks originate from thick
sandstone layers (five to more than 20 cm) that occur in between thin mudstone or
shale layers. The tracks are sometimes more than 20 cm deep. The tracks were
made in soft mud and sand filled the track molds later. Track casts are often
attached to overlying sandstone layers; thus overlying sandstone layers may make
track casts appear deeper than the original tracks (Fig. 5.1). The tracksite has
been washed and eroded by waves of the reservoir for a long time. The Pajarito
sandstone layers are broken, and many track-bearing sandstone blocks have been
restricted (Fig. 5.2; see Appendix A for more tracksite views). Therefore most of
the tracks are loose and often moved and overturned by wave action in the
reservoir. Track bearing sandstone layers and most of the other Pajarito
sandstone layers are burrowed heavily. These invertebrate traces are dominated
by Arenicolites and they often contain plant debris and iron concretions. These
characters are common in dinosaur track bearing layers of John Martin Reservoir.
Some current ripple marks can be observed on layers below the track bearing
5.1.2 Site 1
Site 1 is situated about 30m west from Plover Point (Fig. 2.2). The
Mesa Rica Sandstone and lower most part of the Pajarito Formation are covered.

Figure 5.1

Figure 5.2

Most of the Pajarito and the Romeroville Sandstone can be observed though the
upper part of the Pajarito is poorly exposed. In spite of the short distance from
Plover Point, the track census of Site 1 is quite different from Plover Point Only
one well preserved Caririchnium cast was found from the middle to lower part of
the Pajarito Formation. Two crocodile pes (CU-MWC 209.44,122) and many
crocodile swim tracks were found from the upper part of the Pajarito (Fig. 5.3,
5.4). Most of the swim tracks are three toed and some of them are four toed
scratch marks. Length varies from about 6-25cm, and width varies from about
5-15cm. This part of the Pajarito shows alternating beds of thicker sandstone
(3-15cm) and thinner mudstone (1 -5cm). These beds start from the middle part
of the Pajarito and continue up to about l-2m below the Romero ville sandstone.
The upper most part of the Pajarito Formation at Site 1 is thick black shale and
there is no sign of tracks. Crocodile swim tracks occur as casts on the underside
of each these sandstone layers. Track bearing sandstone layers are relatively
thinner than at Plover Point and broken out as large-sized slabs. Tracks can be
found by turning over these sandstone slabs. Each swim track was found
separately as an isolated specimen, and no trackway has been recognized. Track
bearing sandstones were also burrowed heavily with Arenicolites and other traces.

Figure 5.3

Figure 5.4

5.13 Site 2
Site 2 is situated 2 or 3 km west of Site 1 (Fig. 2.2). This is the biggest
tracksite at John Martin Reservoir. Only about 2m thickness of the middle of the
Pajarito section can be observed here. There are two dinosaur track bearing
layers, as at Plover Point, and the upper track bearing layer is the top surface of a
horizontal terrace where trackways can be observed. There are many loose
Caririchnium casts along the reservoir shore. This situation is also similar to
Plover Point. Fourty-seven Caririchnium were found along the shoreline but no
other track types. Many in-situ Caririchnium trackways were found on the
terrace 3m or more inward from the reservoir on shoreline as marked by the
boundary between the reservoir bottom mud flats and the lower rock outcrops (Fig.
5.5). These trackways are sandstone casts sitting in the ground in the positions
they were made (Fig. 5.6). However, most of the track bearing sandstone except
for the track casts are missing and therefore the tracks look like they were placed
in the present topsoil like decorations in cake icing. However they are in fact
in-situ. Probably soft mudstone layers underneath and thinner sandstone layers
above were eroded by weathering and solution by lake water and the hard, thicker
track fills were left
All trackways from Site 2 are heading to the same direction, northeast.
These trackways are somewhat regularly spaced, l-2m so they look like a

Figure 5.5

Figure 5.6

gregarious group (see discussion below). The track casts are all very fragile and
unfortunately, these trackways were disturbed by water action as observed
between 2004 and spring 2005, and some parts of trackways were lost. Each
sandstone layer has Arenicolites and other invertebrate traces and some current
ripple marks were observed on some layers including track bearing layers.
5.1.4 Site 2E
Site 2E is close to Site 2, about 130m east (Fig. 2.2). Outcropping
stratigraphic sections and track occurrence are almost the same as Site 2. 103
loose Caririchnium casts were found along the reservoir shoreline of the site and
some in-situ Caririchnium tracks and trackways were found on the terrace about
2m away from the shoreline between mud flat and outcrop (Fig. 5.7).
Trackways from Site 2E are not as well organized or ordered as at Site 2. No
other tracks were observed but many invertebrate trace fossils including
Arenicolites occur.
5.1.5 Site2W
Site2W which is situated about 10m west from Site 2 is also similar to
Site 2. The Pajarito Formation is exposed for only about 2m of thickness much

Figure 5.7

the same as Site 2 but the top layer of the terrace is not a dinosaur track bearing
layer. Loose Caririchnium casts are not so common around Site 2W, less than
10 were found, and the number of in-situ trackways is also very limited (Fig. 5.8).
Two trackways which go parallel about 2m apart can be recognized and they are
heading northeast and east. This is consistent with Site 2 trackway orientations.
Although the in-situ track occurrence situation is the same, track preservation is
relatively poor compared to Site 2 and 2E. Great efforts were made to collect
several broken tracks and to restore them. However, one pair of Caririchnium
manus and pes was found from Site 2W and this is the only Caririchnium manus
so far found from John Martin Reservoir. Site 2W also yields some other
vertebrate trace fossils. Some crocodile swim tracks were found from the top
layer (Fig. 5.9). These swim tracks at Site 2W are a little different from Site 1.
Each digit is more robust and scratch length is longer than those of Site 1. One
pterosaur pes track was found from a small loose slab of Site 2W (Fig. 4.5).
This specimen was tentatively considered to be a crocodile track at first sight; it is
much smaller than any other crocodile swim tracks from John Martin Reservoir.
Its characteristic shape does not look like a crocodile track but is quite similar to
pterosaur pes tracks such as Pteraichnus. It is evidently the first pterosaur track
form the Dakota Group (see discussion below).

Figure 5.8

Figure 5.9

5.1.6 Site 3
Site 3 is situated about 30 m west from Site 2W (Fig. 2.2). Tracks occur
from one layer in the same manner as described for Site 2 trackways, namely
tracks look like they are sitting in the present topsoil, and stratigraphic sections
are hard to observe because of minimal vertical thickness. However, probably
the track-bearing layer can be assigned to the same layer as trackway bearing
layers at Site 2. Three in situ Caririchnium trackways and isolated tracks were
found from Site 3 (Fig. 5.10). Track preservation is excellent at this site,
especially trackway A in Fig. 5.10 is the best trackway from John Martin
Reservoir. Tracks were collected and made into a replica that faithfully recreates
the trackway pattern (Fig. 5.11). This replica makes an excellent exhibit piece
which has already been procured by the U.S. Army Corps for their visitor center.
5.1.7 Site 4
Site 4 is situated on the opposite shoreline from Sites 1-3, i.e., on the east
side of Rule Creek where it discharges into the reservoir on the south side (Fig.
2.2). The upper part of the Mesa Rica Sandstone and the lower to middle part of
the Pajarito Formation can be seen. One sandstone slab with a set of
well-preserved pes and manus crocodile swim tracks was found from the lower
part of the Pajarito (Fig. 2.3, 5.12B). Some leaf fossils were also found here.

Figure 5.10

Figure 5.11

Figure 5.12

The lower Pajarito is partially covered by vegetation and rocks from the track
discovery point to the middle part, then well exposed sandstone layers (3-30cm)
which are mainly burrowed by Arenicolites, and appear for about 1-2 m in
thickness. The top of these layers is a 4WD wheel drive road and some poorly
preserved Caririchnium casts were found from the roadside (Fig. 5.12A). The
upper part of the Pajarito is covered.
5.1.8 Site 5
Site 5 is situated along the northern shoreline, in between Site 1 and Site
2 (Fig. 2.2). The site condition is similar to Site 1 and only one poorly preserved
loose Caririchnium cast was found.
5.1.9 Site 6
Site 6 is situated on the north shore, across from Site 4 and 2km west of
Site 3 (Fig. 2.2). The upper most part of the Pajarito and the overlying
Romeroville can be observed. The ground surface consists of a sandstone bed in
the Pajarito and here is bumpy and rugged. There are many ironstone
concretions covering irregular, contorted parts of the basal sandstone (Fig. 5.13).
Although these parts look like dinosaur tracks, they are not tracks because then-
sizes and arrangements vary very irregularly and we did not see any characteristic


patterns of trackways that can be seen at other tracksites. However, some
crocodile swim tracks and other small sole marks were found from a sandstone
layer which is lm above the lowest exposed sandstone. These crocodile swim
tracks are relatively large and show robust digits which are similar to swim tracks
of Site 2W (Fig. 5.14).

Figure 5.14

5.1.10 Site 7
Site 7 is on the opposite (south) side of the reservoir across from Site 1
(Fig. 2.2). Site characteristics are similar to Site 1 and some small crocodile
swim tracks were found from the lower part of the Pajarito Formation (Fig. 2.3).
5.1.11 Site 8
Site 8 is situated along the east banks of Rule Creek, about 1km upstream
from the confluence point with the reservoir. The lower to middle part of the
Pajarito Formation can be observed here. Some well-preserved Caririchnium
casts were found but no trackways were found because the track bearing layer is
exposed as a part of a low cliff so that no large horizontal surfaces are exposed
(Fig. 5.15).
5.1.12 Site 9
Site 9 is situated 2km east from Plover Point (Fig. 2.2). The middle and
upper part of the Pajarito Formation and the Romeroville Sandstone can be
observed. The middle part of the Pajarito has two dinosaur track {Caririchnium)
bearing layers which are similar to those of Site 2 (Fig. 2.3). Probably they are
equivalent to Site 2 dinosaur track bearing layers. Most tracks near the low
water shoreline are loose tracks but tracks from more inland ground are in-situ.


This track occurrence situation is also similar to Site 2. Although one of the
track bearing layers is exposed as the ground surface like Site 2, well preserved
long trackways were not found. However, some isolated in-situ tracks and some
poorly preserved short trackways were recorded (Fig. 5.16). Thick shale (1.5m)
can be seen at the contact point of the Pajarito and the Romeroville and a couple
of fallen crocodile swim tracks were found around there (Fig. 5.17). These
crocodile swim tracks have robust digits and long scratch marks which are the
same in character as crocodile tracks from Site 2W and Site 6.
5.2 Track measurement data
Table 5.1 and 5.2 show track measurement data from John Martin
Reservoir. Table 2 is data of length and width of Caririchnium natural casts.
Caririchnium loose tracks from Site 1, 2, 2E, 2W, 3 and 9 were measured. Data
show track length (TL), width (TW), and site name. Units are in cm.
Crocodile track measurement data is shown in Table 2. Data shows specimen
number, TL, TW, digit length (DL), digit width (DW), interdigital angles or
divariation (DIV) and site name. Units are in mm, except for DIV in degrees.
P in DIV columns means digits (scratch marks) are parallel and asterisk marks
beside values of TL and DL mean that measurements may not be correct because
of incomplete specimens.

Figure 5.16

Figure 5.17

Table 5.1

Table 5.2
Specimen # Sit* TL TW DL DW DIV
1 T 11 III IV | V 1 1 " "i| !V | V Ml tl-lll lll-IV IV-V Total
208.116 1 130* 130 90* 100 65 25 23 15 . 8 P - 8
209.116 1 80 S3 65 25 65 10 7 12 - P P - P
208.117 1 175* 125 75* 140* 175* 10 15 18 P - P
208.37 1 90 105 85 55 55 18 20 15 10 9.5 19.5
209.38 1 100 110 45 100 45 17 25 20 P P .
209.39 1 85 45 50 40 85 5 4 4 P 12 12
209.40 1 85 125 80 70 55 35 30 22 16 10 32
209.41 1 240* 150 210* 215* 200* 25 30 22 P . P
209.42 1 190 110 180 165 160 30 20 20 8 P 8
209.43 1 95 140 85 95 75 45 40 25 8 6 14
209.50 1 65 75 60 65 60 15 20 10 P 16 16
209.51 1 170 160 105 100 120 22 30 25 P 11 11
209.12 2W 287* 30 267* - 30 - .
208.82 4 135* 135 - 50* 80* 100* 15 20 40 P P P
209.78M 4 95 140 60 76 95 50 35 35 P P P
209.78P 4 165 195 135 185 165 55 45 30 P P P
209.111 7 250* 165 210* 200* 170* 33 35 25 5 4 9
208.73 7 275* 140 200* 200* 165* - 20 30 25 P P P
209.118 8 155* 110 95* 155* 110* - 25 35 20 P P
209.119 9 390* 80 390* 390* - - 20 35 - - P E

6. Discussion
6.1 Crocodile Swim Tracks
Some crocodilian tracks are reported from formations which date from
before and after the age of the Dakota Group. The crocodymorph ichnogenus
Batrachopus (Lull, 1904) which is known from Early Jurassic strata of the United
States and Europe is an earlier crocodile track which is considered to be made by
protosuchian crocodilians (Olsen and Padian, 1986; Foster and Lockley, 1997;
Lockley and Meyer, 2004). These terrestrially made walking tracks consist of
five toed manus and four toed pes and the size is very small, 2-6cm pedal length.
Actually Batrachopus has never been reported from Cretaceous strata.
Therefore it is difficult to compare Batrachopus with the Dakota swim tracks.
Foster and Lockley (1997) reported the crocodilian ichnogenus Hatcherichnus
sanjucmensis from the Salt Wash Member of the Upper Jurassic Morrison
Formation in eastern Utah. The Hatcherichnus pes (CU-MWC 197.1) is
tetradactyl and its size is 11.8 cm in length and 16.8cm in width. This track is
similar to one of the normal walking footprints from Site 1 (CU-MWC 209.44) in
size, proportion of overall length and width, and curved digit shape (Fig. 6.1).
Probably CU-MWC 209.44 can be interpreted as a crocodile pes track which is
close to Hatcherichnus. CU-MWC 209.122 shows similar characters with
modem crocodile and alligator tracks in digit morphology. For example, the

Figure 6.1

modem alligator pes has four digits including short digit I and long digits II-IV
(Fig. 6.2) as CU-MWC 209.122 also does. Although the heel morphology is
different (the CU-MWC heel impression is broader), these tracks show strong
similarity as a whole. Besides skeletal remains of the Cretaceous crocodile
Dakotasuchus kingi were discovered from the Dakota of Kansas (Mehl, 1941;
Vaughn, 1956), and probably Dakotasuchus is the trackmaker of these tracks.
McAllister (1989 a, b) reported many vertebrate swim tracks from the
Dakota Formation of Kansas. These tracks are characterized by a manus with
3-5 digits with width ranges from 6-11.5cm and a pes with three digits with width
ranges 9.5-18.5 cm (Fig. 6.3). Their shape suggests evidence of scratching the
bottom with their digits. This track type is very similar to swim tracks from John
Martin Reservoir and so suggests similar behavior. McAllister interpreted these
tracks as omithschian dinosaur swim tracks based on deduced foot morphology
(digit counts and digit length) and mode of locomotion (discontinuous propulsion).
However, McAllisters interpretation can be considered improbable for several
reasons. First, size can be shown as the primary problem. Two kinds of
omithschian dinosaurs that are considered to be swim trackmakers by McAllister
are known from the Dakota Group. They are omithopods and ankylosaurs and
their track sizes are generally too large for these swim tracks. For example, the
omithopod track size at John Martin Reservoir ranges from 17-43 cm in length

Figure 6.2


and 20-48 cm in width. These size ranges are much larger than swim tracks
from both Kansas and John Martin Reservoir. Ankylosaur tracks are also much
larger than the swim tracks (Fig. 6.4). It is unlikely that all swim tracks were
made by juvenile dinosaurs. Besides the largest swim track is smaller than the
smallest omithopod track which is considered to be a juvenile. Second, there are
some digit morphology problems. Fig. 6.5 shows the comparison of tracks and
cross sections of assumed swim tracks (Fig. 6.5A) of the Dakota large vertebrates
(Fig. 6.5B-E). To construct hypothetical cross-sections of swim tracks from
walking tracks, the casts of walking tracks were set downward to touch the ground
with their tips (Fig. 6.5). Cross-section diagrams of the ground surfaces beneath
the track tips were then drawn. These hypothetical cross-sections were compared
with cross-sections of actual swim tracks from John Martin Reservoir. Although
all swim pes tracks from Kansas and most swim tracks from John Martin have
three digits, ankylosaurs pes tracks have four digits and their lengths are almost
the same. Thus, there is only a slight possibility that these tracks were made by
ankylosaurs. On the other hand, an omithopod pes track has three digits,
however, these digits are blunt, with no claw impression, and have wide
interdigital angles. Swim tracks from the Dakota have streamlined digit
scratches with pointed ends that suggest the trackmaker had sharp clawed digits,
and also digit spacing is narrow. Besides well-preserved omithopod pes tracks


Figure 6.5

show a prominent digit III that suggest it would make only a digit III scratch mark
or a three digit scratch mark with the center of the track very deep. However,
there are no significant differences in digit depth in the swim tracks. Therefore it
is natural to consider that these swim tracks were not made by omithopods but
crocodiles that have relatively uniform digit lengths.
Then how about theropods as trackmakers? Theropods have slender
and pointed toes that could leave such swim tracks. However, this possibility is
also denied by consideration of track size and digit length. As shown in Fig. 6.4,
the Dakota theropod track Magnoavipes is much larger. Magnoavipes also has
wide interdigital angles and prominent digit III that would not leave such swim
tracks with uniform digit lengths. Dinosaur swim tracks named Characichnos
tridactylus by Whyte and Romano (2001) do have three digits, but the middle
digit trace is longer as is consistent with footprint morphology and size (Fig 6.6).
Thus it can be considered that the possibility of any dinosaur leaving swim tracks
in the Dakota is much less than for crocodiles.
As mentioned in section 5, there are two types of swim tracks at John
Martin Reservoir, smaller and short scratch mark types, and larger and long
scratch mark types. Site 1 has only the small type and probably they were made
by the same trackmaker type as CU-MWC 209.44, as size is consistent with this
conclusion. McCrea et al. (2004) reported Albertasuchipes russellia, which is a

Figure 6.6

crocodilian trackway from the Lacombe Member of the Upper Pal eocene
Paskapoo Formation in Alberta (Fig. 6.7). This trackway is characterized by
tridactyl pes and manus. The manus track shape of Albertasuchipes looks like
the small type of swim tracks. The pes track shape, which is characterized by
parallel, elongate scratch marks very similar to those of the large type of swim
track from John Martin, although Albertasuchipes is a little smaller than the John
Martin swim tracks. These Albertasuchipes characters suggest that small and
large swim tracks were not made by different kinds of animals. Probably
differences in character between swim tracks are due to manus-pes differences
and size difference among individuals. For example, crocodile tracks in Fig. 5.4
are manus and Fig. 5.9 and 5.17 are pes. Thus, all these swim tracks from John
Martin Reservoir can be interpreted as crocodile swim tracks.
6.2 The Pterosaur Swim Track
The history of pterosaur tracks research is a repetition of claims of
identification and counterclaims. The first supposed pterosaur trackway was
reported by Stokes (1957) from the Morrison Formation of Arizona. This
trackway consisted of a peculiar manus with three digits and plantigrade pes with
four or five toes and named Pteraichnus saltwashensis (Fig. 6.8). Later this
interpretation was denied by Padian and Olsen (1984) who interpreted the tracks

Figure 6.7

Figure 6.8

as crocodile tracks. They compared Pteraichnus with small crocodilian tracks
and claimed some similarities. They also thought pterosaurs were not
quadrupedal, plantigrade animals but bipedal, digitigrade animals. However,
Lockley et al. (1995) overturned the crocodile interpretation. They showed that
characters of Pteraichnus, namely plantigrade, semi-erect, and quadrupedal
walking pattern, with anterior emphasized weight-bearing posture with deeper
manus tracks. They showed that pes and manus track morphology is strikingly
consistent with pterosaur anatomy and thus Pteraichnus was attributed to be
pterosaurian again. Now it is the consensus that Pteraichnus is a pterosaurian
track. In North America they are mainly known from the Late Jurassic
Summerville Formation and the Sundance Formation. But they also occur in
some scattered Cretaceous locations in North America and around the world.
The first supposed pterosaur tracks from the Dakota Group were reported
by Gillette and Thomas (1989). The trackway supposedly consisted of six tracks
considered to be left manus prints of a large pterodactyloid pterosaur accelerating
on the ground just before taking flight. However, this interpretation was
supported by few workers, and there were various arguments that the trackway
did not match the predicted morphology of pterosaurian tracks (e.g. Lockley,
1991). Then Bennett (1992) re-examined the tracksite and found some
additional tracks, trackways and tail traces and he concluded that the trackways

including the supposed pterosaur trackway are crocodilian tracks that were formed
in shallow water and may record the transition from walking to swimming. Thus,
no true pterosaur track had previously been confirmed from the Dakota Group.
Specimen CU-MWC 209.83 from Site 2W of John Martin Reservoir
consists of a well-preserved pes trace and its scratch marks (Fig. 4.5). The pes
impression seems to be made after the scratching because the pes depth is deeper
than the scratch mark and it looks like the pes impressed upon the upper part of
the scratch mark. The pes part is characterized by clear phalangeal pads, four
slightly curved digits uniform in length (two center digits are a little longer), and
overall elongated shape with heel impression. The scratch mark consists of four
parallel lines and their widths and intervals are almost the same. Such patterns
have been reported for Jurassic swim tracks of pterosaurs (Lockley and Wright,
2003). This specimen was assumed to be crocodile track at first sight simply
because crocodilian tracks are so common here. However, this specimen is
probably not crocodilian for several reasons. First reason is the four parallel line
scratch marks of almost the same length. This character can not be seen in
crocodile swim tracks because crocodile swim tracks usually have three scratch
marks due to the short digits I and V of crocodiles. Second, the proportion of
scratch length to track size is much longer than normal crocodile swim track
length, which is almost the same or less than the width. Although some large

crocodile swim tracks have larger length than width, this is rare and their scratch
marks are slightly curved, however, the scratch marks of CU-MWC 209.83 are
straight. The size is also different, this specimen is smaller than any other
crocodile swim tracks from John Martin Reservoir. Third, the pes part of the
specimen also suggests that this track does not belong to crocodiles. The whole
shape of the footprint is elongate and digit lengths are uniform in length while
other crocodile footprints are much more transverse. Although the four visible
digit lengths of the specimen are almost the same, the crocodile pes is
characterized by a shorter digit I which often does not leave a trace (Fig. 6.7).
Clear phalangeal pads are not seen in crocodile tracks either. Therefore this
specimen can not be a crocodile track. Also it is not any other dinosaur known
from the Dakota Group because the size is too small and the pes shape characters
such as four digits which are uniform in length are not seen in any of the Dakota
dinosaurs. Actually, characters of the specimen such as the elongate oblong
shape that suggests plantigrade stance and four digits which are uniform in length
are very similar to those of pterosaur pes tracks.
Although well preserved Pteraichnus may be considered to show five
toes. Digit V is often missing in many specimens because it is very much shorter
than other digits and situated in the rear part of the track as a pad or bulge not as a
separate digit trace. Comparison of the CU-MWC 209.83 and Pteraichnus

specimens shows morphological similarities between them in their oblong shape
and digit numbers (Fig. 6.8, 6.9). These similarities strongly suggest that
CU-MWC 209.83 is a pterosaur track. Some features of the specimen also
suggest the swimming activity of pterosaurs. Swimming or at least floating in
the water has been interpreted based on some Pteraichnus tracks by Lockley and
Wright (2003). They showed some swimming scrape marks of pterosaurs,
especially Pteraichnus specimen CU-MWC 186.46 from the late Jurassic
Summerville Formation shows strong similarities with CU-MWC 209.83 (Fig.
6.10). CU-MWC 186.46 which consists of well-preserved isolated pes print and
some sets of scratch marks. The digit parts are clear and deep but the heel pad
impression is shallow. Lockley and Wright (2003) interpreted this character as
the track may have been made by a buoyant or partially buoyant animal
impressing most of the distal part of the foot on the substrate while progressing
in/on shallow water. They also indicated that the track is much deeper and
clearer than any known walking pes tracks and concluded that this is due to the
soft, ductile, but cohesive substrate which can be found in a shallow water
conditions. Basically, the Pteraichnus pes is shallow due to the pterosaurs
unique walking posture that has weight on the front of the body (Lockley et al.,
1995). Therefore, the deeper pes of CU-MWC186.46 could suggest not only the
subaqueous substrate setting, but also what is eventually bipedal posture that

Figure 6.9

Figure 6.10

allows deeper pes impressions due to swimming or floating activities. Scratch
marks are parallel to the pes but probably most of them do not belong to the pes.
Thus, specimen CU-MWC186.46 is quite similar to CU-MWC 209.83. Their
common characters are; pes and scratch mark associations, deeper pes than
walking pes, and deeper digits and shallower heel. Sometimes the pes trace
consists only of four toe impressions. Therefore the interpretation of CU-MWC
209.83 as a pterosaur track is confirmed more strongly. However, there are also
some differences between CU-MWC 186.46 and 209.83 and they suggest different
interpretations of pterosaur swimming activities. The scratch marie of CU-MWC
209.83 which was made before the pes, clearly belongs to the pes, even though it
is not parallel to the pes as in CU-MWC 186.46. Digit traces of CU-MWC
209.83 are slightly curved while those of CU-MWC 186.46 are straighten. So
the activity of CU-MWC 209.83 can not be interpreted quite like CU-MWC
186.46 as swimming progression evidently was straight ahead the latter. The
CU-MWC 186.46 pes is interpreted as a swim or walk track because it is parallel
to scratch marks. If the animal was floating or swimming, scratch marks could
be parallel to the water flow direction. However, the CU-MWC 209.83 pes is
not parallel to the scratch mark but situated diagonally at about 45. So the
diagonal pes position relative to the curved digit traces give us a picture that a
floating pterosaur in a gentle flow touched the bottom with its digits to stop or

turn and finally stopped by standing firmly on the bottom. In particular, the
different directions of the pes and the scratch mark can be interpreted as the
animal wanting to stop against a current flow, and the curved digits are considered
the result of bracing in the mud. There are three or four drip-shaped scratch
marks which are slightly curved and impressed on the straight scratch marks.
They can be interpreted as evidence that the animal pulled out its foot from mud
while twisting its leg. Therefore the part of the track deeper digit traces is not
due to kicking the bottom for progression but bracing in the mud to stop or turn.
Thus two leg movement pattern hypotheses can be pictured by interpretation of
this activity (Fig. 6.11). Fig. 6.11A is a simple interpretation; a floating
pterosaur scratched half of the scratch mark with the back of its digits (0), then it
stuck its digits into the bottom a little to brake (0). And it kept floating a little
more () and finally pressed its foot on the bottom to stop (). Fig. 6.1 IB is a
little more complicated; a floating pterosaur scratched the bottom with the back of
its digits (0) and scraped back the bottom by lifting and twisting its leg to change
the foot direction for stopping his body against the water flow direction (,(§)).
Then it pressed its foot on the bottom to stop its body (). In either hypothesis,
this pterosaur track shows us not only the pterosaurs swim or float activity but
also a pterosaurs interesting movement in water. In both cases, the different
directions of the pes and the scratch marks can be interpreted as the animal

Figure 6.11

wanting to stop against a current flow, and the slightly curved digits are
considered the result of bracing in the mud.
Although other fossil evidence that proves Dakota pterosaurs has not
been found, it is known that pterosaurs were diverse in the Late Jurassic and early
Cretaceous (Lockley et al., 1995). Pteraichnus tracksites were often found in
inland areas in the United States and so Stokes (1957) argued that pterosaur
habitats were distributed not only in marine shorelines as is traditionally inferred
but also in more inland areas. However, by plotting tracksites and comparing
lithofacies, Lockley et al. (1995) found tracksite occurrences are distributed along
a Late Jurassic paleogulf (Fig. 6.12). Therefore Pteraichnus habitats were
similar for the Eastern Dakota Group that was facing to the Western Interior
Seaway. It is natural to expect that pterosaurs continued to live in the Dakota
Paleoecosystem which was similar to the Pteraichnus habitats of the late Jurassic
63 Social Behavior and Migration of Dinosaurs
There are many reports that suggest social or gregarious behavior of
dinosaurs. For example, Homer and Gorman (1988) reported that the
hadrosaurian dinosaur Maiasaura made nesting colonies and parented babies until
they became lm long. Such behavior is like that of altricial birds such as

500 km
Pteraichnus tracksite
John Martin Resevoir
Figure 6.12

sparrows and swallows. The nest arrangement was organized; they were roughly
spaced 7m apart which is same as an adult Maiasaura body length, suggesting
that this dinosaur developed a social behavior system. Monospecific bone beds
also can prove gregarious behavior of dinosaurs (e.g. Currie and Dodson, 1984).
These reports have helped to change the image of dinosaurs from clumsy
reptile-like animals to active bird-like animals. This change has been called the
Dinosaur Renaissance (Bakker, 1975).
Dinosaur tracks take on a significant role in the study of dinosaur social
behavior. Tracks are superior to the fossil bones of dead bodies for social
behavior study because tracks are direct in-situ records of dinosaurs active
behavior, and are much more abundant than bones. Many dinosaur social
behaviors have been interpreted from in-situ trackways. Bird (1939, 1941)
reported parallel sauropod and theropod trackways along the Paluxy River at
Dinosaur State Park, Texas (Lockley and Hunt, 1995, fig. 5.4). This tracksite is
famous because it has been suggested that theropods were stalking or attacking
sauropods though the attack possibility is discounted by some authors (Lockley
and Hunt, 1995). The sauropod trackway distributions, which are parallel to
each other and spaced somewhat regularly, suggest sauropods were herding or
walking as a group along shorelines. Three parallel theropod trackways also
suggest that theropods could stalk or hunt in a group, or at least could walk

together without fighting, maybe like wolves. Bird (1944) found another
sauropod trackway assemblage from Davenport Ranch, Texas. At this site, there
are at least twentythree sauropod trackways oriented in the same direction.
Bakker (1968) claimed that these trackways show sauropod herd structure which
consists of smaller individuals in the center surrounded by larger individuals.
However, Lockley (1987,1995) studied the site in detail and concluded that the
trackways do not show the protecting herd structure that Bakker claimed, but the
animals walked more in line with many of the small individuals following after
larger ones to pass over a relatively narrow section of substrate. This example
suggests that trackways give us much interesting information about dinosaur
behavior, which may seem simple, but needs careful analysis. Gregarious
behavior has also been reported from the Dakota Dinosaur Freeway. One of the
best examples is the Mosquero Creek site in northern New Mexico. There are at
least 82 omithopod trackways (Caririchnium) and most of them are parallel or
subparallel trending to the northwest Matsukawa et al. (2001) researched these
trackways statistically, and showed that trackways were not made by the random
movement of individuals.
At John Martin Reservoir, only trackways from Site 2 show gregarious
behavior (Fig. 5.5). These trackways are all parallel, spaced somewhat regularly,
and trending N74E on average. The trackway bearing layer has some current

ripple marks about 8m south of the trackways. The trends of these current
ripples are N20-30E. Thus it can be interpreted that these omithopods were
perhaps herding north eastern along a river shore line. But as such organized
trackways were not found from other sites, except for two trackways from Site
2W, this interpretation is tentative. Tracks from other sites show random
orientation probably indicating random feeding or moving activity. Therefore
we can not infer whether the omithopod herd was a well organized group headed
by a leader, such as in the wolf pack, or a simple flock like a sheep herd.
However, it is certain that John Martin omithopods were gregarious animals and
they sometimes moved together in a line. Many Dakota sites show evidence of
gregarious behavior, whereas others show more random movement. This might
be expected if the animals sometimes traveled together but at other times stopped
to mill around.
What then is the evidence for social behavior across the whole entire
Dinosaur Freeway? Did these omithopod groups migrate across this area or
were they just distributed widely, but with limited individual habitats?
Matsukawa et al. (1999) studied track size-frequency distributions of
Caririchnium trackways statistically and interpreted a growth curve for
Caririchnium tracks that can be subdivided into three groups corresponding to
track lengths, 16.5-21.7,21.7-29.3, and greater than 29.3 cm. They interpreted

Figure 6.13

these groups as juvenile, subadults, and adult. Assemblage age structure of
Caririchnium on the Dinosaur Freeway shows an interesting trend of juvenile
distribution (Fig. 6.13). Juveniles are the most dominant in the south, accounting
for almost half of the sample. The proportion reduces sharply in the middle
region, and almost disappear in the north. John Martin Caririchnium
assemblages also support this picture. Age structure of John Marin, which is
situated in the middle part of Dinosaur Freeway, shows a strong similarity with
other middle-region track sites. However, these track size variations also could
be interpreted as a difference of species. Therefore two hypotheses can be
considered from these data. One is that the Dinosaur Freeway was dominated
mostly by one kind of omithopod and they migrated north and south and bleeded
in south seasonally, and the other is that there were two or more omithopod
species and larger species were dominant in the north and smaller species were
dominant in the south.
Migration behavior is varied and many animals on the earth migrate in
some manner. Bird and insect migrations are especially famous and well studied,
however, probably large terrestrial mammal migration examples should be most
appropriate for this study from the viewpoint of comparison with dinosaur
ecology. Here dinosaur migration is assumed to have moved north and south
seasonally probably with eggs laid in the south in this hypothesis (though many

omithopod egg sites are situated in the north and birds also migrate to the north).
This can be classified as the seasonal migration which involves movements
between two or more areas that are occupied in different seasons during the
annual cycle (Dingle, 1996). Similar seasonal migrations can be seen in many
hoofed mammal ecologies but found only in three habitat types, tundra, desert and
grassland (Baker, 1981). There are some reasons for seasonal migration; they
are mainly physical factors such as water, food, temperature, and sometimes
biological factors such as the avoidance of tsetse flies. For example, wildebeests
(Connochaetes taurinus) and other ungulates of the Serengeti National Park
migrate seasonally seeking for water and food (Kruuk, 1972). Reindeer and
caribou (both are subspecies of Rangifer tarandus) living in the transition zones of
tundra and coniferous forest migrate long distances, up to 1000km in each
direction. They move to southern forest before winter to avoid frozen land,
attended by the lack of food, and then back to northern tundra in the spring to give
birth (Baker, 1981). However, wildebeests in the Ngorongoro Crater where the
dry and wet seasons are less extreme are largely sedentary and reindeer and
caribou which live in exclusively one kind of habitat migrate no more than a few
dozen kilometers. Thus it is necessary that places that have these resources are
available in a walkable distance.
Water temperatures of the Dakota paleoecosystem throughout the

Western Interior Seaway were measured by fossil biotas related to those
characterizing modem marine climatic zones (Kauffman, 1977). They ranged
from Subtropical in the southern part that is Texas, New Mexico and Arizona, to
Warm-Temperate levels in the middle as far north as Wyoming, southern Montana
and South Dakota, and to Mid-Temperate levels in the north through Montana and
most of Canada. Thus the study area experienced a mild climate regime and
probably damp conditions as can be interpreted from the topography facing the
sea. Rich plant fossils also support this interpretation. If the seasonal
migration model is correct, omithopods may have traveled for over 600km, one
way, from north eastern New Mexico to the Denver area. As shown above,
however, seasonal migration over a long distance for more than hundreds of
kilometers usually occurs in severe conditions such as cold tundra and dry inland
savanna. Therefore it is unlikely that omithopods migrated seasonally through
this area. Of course there may have been some migrations but probably they
were much smaller, regional ones. Probably omithopods spread out for a long
distance over a long time period the way ceratopsians migrated from Asia to
North America (Forster and Sereno, 1997). However, this must not be confused
with seasonal migrations because this type of migration, which is characterized by
persistent movement of greater duration than occurs during local movements, is
classified differently from seasonal migration (Dingle, 1996).

Let us consider the other hypothesis that these omithopods were not
migratory animals. This is the other hypothesis that different track size structure
distributions through the Dinosaur Freeway show the distribution of regionally
different species of omithopod rather than the age structure of one omithopod
species. Probably it is natural to think that there was more than one species of
omithopods along the Dinosaur Freeway. Actually, the coexistence of several
related species is normal in the natural environment. For example, more than 40
species of mammals can be observed in the Serengeti ecosystem and more than 20
of them are ungulates (Kruuk, 1972). There are more than 10 species of even the
most dominant even-toed ungulates that can leave similar tracks. The assumed
mild, damp, and well vegetated Dakota paleoecology seems to be a much more
comfortable setting than the Serengeti ecosystem that requires animals to make
seasonal migrations and so it is not hard to imagine there were various omithopod
species there. Besides the area of the Dinosaur Freeway reaches at least
80,000km2 (Lockley et al., 1992) whereas the Serengeti ecosystem extends for
only 25,000km2 (Kruuk, 1972). Matsukawa et al. (1999) considered them as the
same species because these different size trackways were found in a same
geological group and this reason is good enough to consider a reasonable
possibility. The rules of ichnotaxonomy are also a factor because they require
that similar or identical track morphologies be included in the same ichnotaxa.