J. Paleont., 70(6), 1996, PP. 1004-1010
Copyright © 1996, The Paleontological Society
0022-3360/96/0070-1004$03.00
VARIATIONS IN SALAMANDER TRACKWAYS RESULTING FROM
SUBSTRATE DIFFERENCES
LEONARD R. BRAND
Department of Natural Sciences, Lorna Linda University,
Lorna Linda, California 92350
ABSTRACT-As part of a study of vertebrate trackways in the cross-stratified Coconino Sandstone (Permian) of northern Arizona,
trackways made by living salamanders under different substrate conditions were compared, The sample of 230 trackways of the
western newt, Taricha torosa, included ten combinations of the following substrate characteristics: I) sediment: muddy or of fine
sand; 2) attitude: level or sloped (25 degrees); 3) moisture content: dry, damp, wet, or submerged.
Trackways in wet mud produced the most accurate representation of the number of toes per foot and the arrangement of toes,
All other conditions studied yielded a reduced average number of toes per foot, and a large sample was needed before the data had
the potential to indicate the true structure of the trackmaker's feet. Trackways made on sloped, submerged mud or sand, sloped,
dry sand, and sloped, damp sand rarely included the full complement of toes. The positions and orientations of the toe marks were
distorted if the animals were walking underwater or on sloped, damp sand. Trackways on the slopes of cross-stratified deposits make
reliable identification of the trackmaking animals exceptionally difficult.
INfRODUCTION
and invertebrate traces are the only
fossils found in the cross-stratified Coconino Sandstone
(Permian) of northern Arizona. Previous workers disagreed on
the identity of the animals that made the vertebrate trackways.
Lull (1918) identified them as amphibian tracks; Gilmore (1926,
1927) concluded that they are either reptile or amphibian; McKee
(1947) stated that they could be reptile, amphibian, or both;
Peabody (1959) called them trackways of primitive reptiles;
Haubold (1971, 1974, 1984) classified the ichnogenus Baropezia
as the trackways of an amphibian and Laoporus (the most common ichnogenus in the Coconino Sandstone) as caseid pelycosaur trackways. Baird (cited in Spamer, 1984) concluded that
the Coconino trackways were made by caseids, but he did not
complete further systematic analysis.
Many factors can alter animal tracks before they are preserved
as fossils. After tracks are impressed into the substrate, taphonomic alteration of the tracks has an important bearing on the
potential for preservation of the tracks and the potential for
correctly determining the identity of the trackmakers (Cohen et
al., 1991; Laporte and Behrensmeyer, 1980). Tracks are also
affected at the time of production by the nature of the sediment,
Peabody (1959) compared salamander tracks on various types
of mudflat substrates, and Lockley (1986, 1987) summarized
some of the variation in fossil tracks that resulted from substrate
differences, pointing out that three- and four-toed tracks could
be made by the same animal, but on different substrates, Padian
and Olsen (1984a, 1984b) emphasized the importance of understanding animal limb mechanics and substrate characteristics in correctly interpreting the animal that made a given set
of fossil tracks. They concluded that the features of a trackway
are affected by three factors: anatomy of the foot, kinematics of
the limb, and conditions of the substrate. Padian and Olsen
(1984a, 1984b) also emphasized the importance of systematic
experimental study ofthe trackways ofliving animals to provide
a better basis for understanding fossil trackways. Even though
experimental study ofthe locomotion ofPermian tetrapods cannot be done, study of modern animals with similar locomotor
adaptations can offer helpful insights,
In the course of field and laboratory research on the paleoecology of the vertebrate trackways in the Coconino Sandstone
(Brand, 1979, 1992; Brand and Tang, 1991), it became evident
that modern animal trackways made on sloped surfaces differ
V
ERTEBRATE TRACKWAYS
from trackways made on level surfaces, and this could hinder
attempts to identify the trackmaking animals responsible for
modern or fossil trackways. Baird (cited in Spamer, 1984) noted
the same phenomenon and stated that, "footprints made in
sloping sand are much more difficult to interpret than those
made on mudflats,"
As a modern analogue for studying this phenomenon, trackways made by living salamanders under different substrate conditions were compared to determine the influence oflocomotor
patterns and substrate on trackways of a single species of small
vertebrate. A previous study (Brand, 1979) compared the quality of trackway preservation as indicated primarily by the presence or absence of toemarks. The current paper systematically
compares a larger number of trackway characteristics, and reports a condition under which trackways can be formed on damp
sand slopes,
METHODS
Western newts, Taricha torosa (Figure 1), were allowed to
walk on substrates in a sedimentation tank or in aquaria, A
random sample often newts were measured and weighed (Table
1). Ten combinations of the following substrate characteristics
were used in the observations: 1) sediment: muddy or of fine
sand; 2) attitude: level or sloped (25 degrees); 3) moisture content: dry, damp, wet, or submerged (Table 2). The wet substrate
was saturated, but not submerged. In the submerged observations, the animals were under at least 0.1 m of water. All trackways on 25 degree slopes were made by animals moving upslope.
Two types of fine sand were used, and there were no apparent
differences in the tracks made on the two types. Sand derived
from the Navajo Sandstone was collected near Mt. Carmel Junction, Utah. The other sand was collected in the desert near Indio,
California. Mud was collected in a wash in the Mecca Hills,
near the Salton Sea, California. Table 3 gives the grain size
distribution for the two types of sand and the mud.
Many different damp sand combinations were tried in an
attempt to find a set of conditions that would form tracks with
good preservation of detail. Sand on a slope of 25 degrees was
lightly sprayed with water from a 1 liter spray bottle with a
built-in pressure pump. The bottle was held about 30 cm (1 ft)
above the sand surface and moved quickly back and forth to
spray the whole surface as evenly as possible. This was done
from one to 10 times, depending on the desired level of damp-
1004
BRAND-SALAMANDER TRACKWA YS
1005
TABLE 1-Weight and measurements of a random sample of 10 western
newts.
Weight
Snout to vent length
Glenoacetabular length
Mean
9.39 g
7.05 em
3.78 em
Median
10 g
7.15 em
3.81 ern
Range
7-11 g
6.67-7.30 em
3.49-4.13 em
ness. The animals were then either allowed to walk on the sand
immediately, or, in some cases, after the sand was allowed to
dry for several hours.
Photographs were made of all western newt trackways (N =
230), and the number of toes in pes and manus prints were
counted. In a sample of 11 trackways (mean of 7.55 pes and
manus pairs per trackway; range = 4-11) from each of six experimental conditions (Table 2), and 11 Coconino Sandstone
fossil trackways (mean of 7.20 pes and manus pairs per trackway; range = 4-11) from sloped cross-bed surfaces for comparison, the following additional measurements were made: stride
(of pes), pace angulation (of pes), glenoacetabular length, width
of trackway, and mean divergence of the middle three toe prints
(of both pes and of manus; Figure 2). Ifonly two toe prints were
present, the divergence of these was used. The same measurements were also made for four other experimental conditions,
with smaller sample sizes, but these were not used in the statistical analysis. Number of toes in fossil trackways from level
surfaces across the ends of truncated cross-beds (at a bounding
surface) in the Coconino Sandstone were also counted from
photographs.
Because dry, sloped sand tracks almost never had toe marks,
the number of manus toe marks and toe divergence could not
be determined. Toe divergence for level, wet sand tracks could
not be measured because ofthe lack ofclarity ofthe impressions.
A usable sample from level subaqueous mud was not obtained
because of practical difficulties in working with this substrate
condition. The mud rose up into the water easily, hindering
photography. It then settled down on the trackways and partly
obscured them.
Differences among six of the experimental conditions and the
fossil tracks on cross-beds were analyzed by stepwise discriminant analysis, using Wilk's routine, in SPSS/PC (Norusis, 1988).
FIGURE 1- Photographs of the western newt, Taricha torosa, showing
body form and (inset) shape of manus (left) and pes (right) of right
feet. Scale bar = 2 em.
Discriminant analysis is a multivariate technique designed to
maximally discriminate among known groups (Norusis, 1988).
It calculates its own canonical variables, which are mutually
uncorrelated, linear combinations of the original variables designed to maximize statistical distances among groups, and uses
these for comparing the groups. A scattergram, using the first
two canonical variables as the two axes, provided a visual representation of the degree of similarity or difference among the
groups. The comparison of dry sand trackways with the other
six substrate conditions would have been pertinent to the study,
but dry sand trackways could not be included in this analysis
because of the missing data resulting from the absence of toe
marks.
Significance of differences between the six experimental groups
was tested with a one-way analysis of variance (ANOVA) and
with a Kruskal-Wallis test, a nonparametric analogue ofthe oneway ANOVA. Differences between pairs ofgroups for individual
variables were tested with the Tukey's HSD test, a post test to
the parametric ANOVA (Norusis, 1988). Normality of the data
sets was tested with the Martinez and Iglewicz (1981) normality
test, in the Number Cruncher Statistical System, version 5.03.
TABLE 2-Characteristics of laboratory trackways of living western newts (Taricha torosa), on level surfaces or 25 degree slopes, and Coconino
Sandstone fossil tracks on sloping cross-beds or on level, truncated ends of cross-beds. F statistics from one-way ANOVA (* = p < 0.00001;
NS = P > 0.05).
Toes per print
Toe divergence
Pace
GlenoPes
Manus
Pes
Manus
Stride
Stride/width angulation acetabular L.
F = 44.00* F = 38.06*
F = 46.58* F=45.61* F = 1.88N S
F = 16.50* F = 9.79* F = 0.98N S
Substrate
it
SD
condition
it
R
it
R N'
it
SD
it
SD
SD
it
SD
it
it
SD
Sloped subaq. sand
2.65 2-4 II
3.34 2-4
6.9 4.85
5.9 3.76 5.0 1.12 1.03 0.28 69.0 16.62
6.4 1.29
Sloped subaq. mud
3.29 3-4
2.22 2-3 19
4.8 2.16 10.8 9.18 4.6 0.53 0.88 0.13 63.7 9.03
5.7 0.67
Sloped wet mud
4.76 3-5
3.69 2-4 19 26.4 7.71 28.2 3.65 4.8 0.29 1.27 0.08 85.2 3.73
5.9 0.31
Level wet mud
4.86 4-5
3.79 2-4 37 35.4 4.88 40.2 5.46 5.1 0.54 1.33 0.13 87.1 7.81
5.9 0.49
Level dry sand
4.14 3-5
2.75 2-4 26 26.4 10.03 17.0 10.82 5.1 0.63 l.l8 0.09 84.9 8.16
6.0 0.56
Sloped dry sand
73
0.24 0-2
3.9 0.61 0.77 0.15 69.3 6.90
Sloped damp sand
3.54 3-4
2.72 2-4 11
4.4 5.01 4.4 0.68 0.89 0.13 74.5 1l.l6
5.8 2.96
6.0 0.42
Sloped wet sand
4.24 3-5
3
5.7 3.04
3.00 2-4
5.9 10.50 3.3 0.59 0.79 0.10 63.5 4.45
5.0 0.30
Level wet sand
4.05 3-5
5.7 0.89 1.31 0.19 91.2 3.52
2.92 2-4
5
6.7 0.65
Level subaq. sand
3.60 3-5
2.92 2-4 13
7.2 6.84 15.8 12.45 6.1 1.04 1.21 0.26 82.3 14.69
7.4 0.85
Fossil tracks, slopes
3.71 3-5
2.98 2-4 II
1.8 2.40
4.5 3.93 8.1 1.38 1.25 0.24 80.2 10.08 10.2 1.81
4
Fossil tracks, level
4.24 2-6?
42
, Total number of trackways in sample, for count of number of toes.
2Number of trackways analysed for toe divergence, stride, stride/trackway width, pace angulation, and glenoacetabular length.
3 Could not identify pes or manus.
4 Number of prints; these tracks were almost all isolated prints, not in trackways; could not identify pes or manus.
N2
II
II
II
II
II
7
11
3
5
7
II
JOURNAL OF PALEONTOLOGY, V. 70, NO.6, 1996
1006
TABLE 3--Grain size distributions of the substrates used in the laboratory trackway experiments.
Grain size categories
Coarse sand (500-1,000 11m)
Medium sand (250-500 11m)
Fine sand (125-250 11m)
Very fine sand (62-125 11m)
Silt and clay «62 11m)
Substrates
California
sand
Utah sand
0.07%
18
51
29
2
0.003%
44
45
10
1
Mud
5%
3
2
>-
~~
i~s
~ ~~
I
PACE
ANGULATION
oQ
~~
f)~
f-
II
79
RESULTS
There was considerable variation among experimental conditions in the trackway characteristics (Figure 3 and Table 2;
Figure 4 shows fossil trackways for comparison). All variables
except stride length and glenoacetabular length showed significant differences (Table 2) when tested with one-way ANOVA.
The data in only four out of 48 subgroups were not normally
distributed (Martinez and Iglewicz normality test: P < 0.10). I
also tested overall variation with a Kruskall-Wallis nonparametric ANOVA, and the same variables were still significant (P
< 0.00001). The greatest variation occurred in the angle of
divergence between the toe marks and in the number of toes
per foot (Figure 5 and Table 2). Trackways made on wet mud
produced the most accurate representation of the number of
toes per foot. All other conditions studied yielded a reduced
average number of toes per foot, and, under some experimental
conditions (sloped, dry sand, sloped, subaqueous sand or mud,
and sloped, damp sand) none of the trackways indicated the full
complement of toes for both manus and pes.
Differences in stride length and in glenoacetabular length were
not significant; differences in gait were significantly reflected in
pace angulation (Table 2). Although the western newts appeared
to have a very consistent walking speed and pattern, their trackways indicated a narrower stance and/or longer stride on level
surfaces and on the firm surface of the wet mud slopes. On
softer, sloping substrates, pace angulation was higher, reflecting
a wider stance and/or shorter stride.
The consistency of level wet sand was not suitable for preserving the details of the foot impression that were preserved
in wet mud. Level, wet sand tracks were generally just the impressions of the tips of the digits. Subaqueous level sand tracks
varied from impressions of the digit tips only to elongated toe
marks that did not reflect the true divergence of the western
newt toes.
The capacity of damp sand to record tracks varied considerably, depending on the specific conditions (Figure 6). Virtually
no details of the tracks were preserved if the animals walked on
the sand soon after it was dampened. If the sand was given at
least three sprays with the spray bottle, a wet crust was formed
that broke up into pieces (Figure 6.2). With a larger number of
sprays the wet sand was too firm for western newts to make a
significant impression (Figure 6.3). However, if the sand was
moderately dampened (four to 10 sprays) and then left to dry
overnight or for at least eight hours, good trackways were produced (Figure 6.4-6.6). These were the damp sand trackways
analyzed in this paper.
Subaqueous tracks were affected by the mechanics of walking
underwater. The animals were partially buoyed by the water,
and consequently they were not resting their weight fully on the
substrate (Figure 7). Their feet were not placed flat on the surface, as they were under subaerial conditions, but rather, when
viewed from the side, were held approximately in line with the
TOE
DIVERGENCE
f«-~~-----~../
~
STRIDE
FIGURE 2-Measurements used in this study.
lower limbs and were pressed into the substrate at an angle as
the animal pushed back against the substrate (Figure 7). The
resulting toe marks were not impressions of the animals' feet,
but were the scratches produced by the backward motion of the
digits. These scratches usually did not represent all of the toes,
and they did not show the normal divergence of the western
newt toes, but were closer to parallel with each other. A similar
result was evident in trackways on damp sand slopes, perhaps
because the loose sand allowed the feet to slide backward. This
produced nearly parallel scratch marks instead of the true toe
impressions seen in wet mud trackways. This relationship is
shown in Figure 5, which compares the foot structure ofwestern
newts with their trackways made on seven of the experimental
substrate conditions.
In the scattergram (Figure 8), sloped damp sand and the sloped
subaqueous sand and mud groups cluster together, and the subaerial, wet mud groups (sloped and level) cluster together. Level,
dry sand falls between these two clusters. The fossil tracks are
associated with the damp sand and subaqueous sand and mud
cluster, as could be predicted from the appearance of the tracks.
Tables 4 and 5 give the canonical discriminant function coefficients and the within-groups correlations between discriminating variables for the canonical variables used in Figure 8.
The Tukey's HSD test indicated homogenous (P > 0.05) subsets for each variable that consistently mirrored the clusters
indicated on the scattergram, with level, dry sand sometimes
being grouped with one of the two clusters (Figure 9). One
exception (pace angulation) combined sloped, damp sand with
the wet mud cluster as a homogeneous subset.
Trackways made on sloped, dry sand slopes were usually
featureless depressions because the sand slumped and covered
all or most track details. These trackways could not be included
in the statistical analysis (because of missing data for most variables), but they are very distinct from the other types and can
be consistently and unambiguously identified by visual inspection.
DISCUSSION
Trackways in wet mud produced the most accurate representation of the number of toes per foot and the arrangement of
toes. All other conditions studied yielded a reduced average
number of toes per foot, and a large sample was needed before
the data had the potential to indicate the true structure of the
trackmaker's feet; tracks made under certain conditions did not
have this potential regardless of sample size. Precise identification of trackmakers is based on maximum number of toes
clearly present in the fossil trackways, not on the mean number
of toes. For identification to be correct, the full complement of
toes must be present in some ofthe individual tracks, and other
BRAND-SALAMANDER TRACKWA YS
1007
3-Photographs of representative trackways from each of 10 experimental conditions. 1, sloped subaqueous sand; 2, sloped subaqueous
mud; 3, sloped wet mud; 4, level wet mud; 5, level dry sand; 6, sloped dry sand; 7, sloped damp sand dried overnight; 8, sloped wet sand; 9,
level wet sand; 10, level subaqueous sand. Scale bar = I em. Lighting in all photos is from top or (in 6) right side of photo.
FIGURE
trackway features must reasonably reflect the animal's foot
structure.
Mean number of toes, although not a useful statistic for systematics, was significant for comparing the appearance of tracks
on different substrates. The reduced number of toes in sub-
mergence or on damp sand slopes and the near absence of toe
marks on dry sand slopes are consistent features of these conditions. The position and orientation of the toe marks were
distorted if the animals were walking underwater and consequently were partially buoyed up by the water, or if they were
1008
JOURNAL OF PALEONTOLOGY, V. 70, NO.6, 1996
PES
MANUS
PES
MANUS
NEWT FOOT OUTLINE
25° SUBAQUEOUS SAND
25° SUBAQUEOUS MUD
25° WET MUD
LEVEL WET MUD
LEVEL DRY SAND
25° DRY SAND"
FIGURE 4- Photographs of representative fossil tracks in the Coconino
Sandstone.1, a trackway ascendinga cross-bed;2-4, individual prints
from a level surface on the ends of truncated cross-beds. Scale bar =
I em. Lightingis from top of all photos.
o
25° DAMP SAND
=
C>
FOSSIL TRACKS, SLOPE
walking on loose, damp sand on a slope. Toe marks produced
underwater tend to be more parallel on level surfaces, and this
effect is exaggerated on submerged, sloped surfaces.
The glenoacetabular lengths measured on laboratory trackways (Table 2) were longer than those measured on the western
newts (Table 1). This may result from the undulating motion
of the newt body, producing a stride longer than it would be if
the body were held in a straight line.
The trackways made by western newts under different substrate conditions contained as much implied variation in foot
structure as can be found in fossil trackways ranging from the
Permian Laoporus (Gilmore, 1926; Haubold, 1974, 1984), which
are alleged to be made by caseid reptiles, to Pliocene newt trackways (Peabody, 1959). Laboratory trackways of western newts
made in sloped, subaqueous mud, sloped sand (regardless of
moisture or submergence), and level, subaqueous sand were less
accurate representations offoot structure than those made under
other conditions.
These data further emphasize the importance of considering
the nature ofthe substrate when drawing systematic conclusions
from trackways. Trackways on the slopes of cross-bedded deposits are particularly unsuitable for reliable identification of
the trackmaking animals. Gilmore (1927) may have been correct
in his conclusion that there is not much likelihood of determining whether the Coconino Sandstone tracks were made by
reptiles or by amphibians. In any case, extreme caution is called
for in attempts to identify the makers of fossil tracks on crossbedded deposits, as these tracks may not contain much of the
necessary information. The Coconino Sandstone tracks made
TABLE 4-Canonical discriminant function coefficients for the first two
discriminant functions used in constructing the scattergram in Figure8.
Toe divergence; manus
Toes per print; manus
Toe divergence; pes
Stride/width
Function I
Function 2
0.44602
0.36375
0.62861
0.33115
-0.19620
0.80975
-0.48970
0.52777
FOSSIL TRACKS, LEVEL"
o
5
0
5
TOES/PRINT
FIGURE 5-Drawings showingthe shape of the manus and pes ofwestem
newts compared with representative prints made on seven of the
experimental conditions and compared with fossil Coconino Sandstone prints made on sloping cross-beds and on level surfaces of
truncated cross-beds. Bar graphs indicate the mean number of toes
present on prints from the same seven experimental conditions. * It
was not possible to identify manus or pes prints, or left or right, for
these tracks. All others are left prints.
on the level surfaces of the truncated cross-beds may give a
more correct picture of the number of toes of the trackmakers.
However, they are oflimited systematic usefulness because they
do not contain much other reliable information and occur as
mostly isolated prints.
The western newts used in my experiments were smaller than
the Coconino Sandstone trackmakers (Table 2), and the possibility that this difference could influence the quality of their
trackways must be considered. However, I have also studied
trackways of modern animals equivalent in size to the Laoporus
trackmakers of the Coconino Sandstone, on the same sand subTABLE 5-Pooled within-groups correlations between discriminating
variables for the first two canonical discriminant functions used in
constructing the scattergram in Figure 8.
Toe divergence; pes
Toe divergence; manus
Toes per print; manus
Stride/width
Pace angulation
Toes per print; pes
Function I
Function 2
0.73922
0.64906
0.42411
0.27649
0.13973
0.27099
-0.47812
-0.15169
0.64898
0.39903
0.19486
0.16829
BRAND-SALAMANDER TRACKWAYS
1009
('
6-Photographs of western newt trackways on damp sand, on 2S degree slopes, 1-3, trackways made soon after moistening the sand; 1,
one spray; 2, three sprays; 3, five sprays; 4-6, trackways made after the sand was sprayed and allowed to dry overnight; 4, 5, four sprays; 6,
10 sprays.
FIGURE
4
strates reported in this paper (Brand, 1979). The trackways of
the larger animals were not qualitatively different from western
newt trackways.
Although some authors (e.g., McKeever, 1991) have suggested
that tracks made on sand not enriched with clay could still be
preserved if the sand was damp, previous efforts to confirm this
* 25°SAND
SUBAQ.
.. FOSSIL TRACKWAYS
•
25° WET MUD
2
N
Z 1
0
SUBAERIAL
MANUS
PES
TOP
VIEW
SIDE
VIEW
q
~
SUBAQUEOUS
MANUS
PES
q
~
~~ ~~
~.::::::--...
------.::::.::::::--...
~~
~~
7- Tracings from photographs of newts walking on a subaqueous surface, and a subaerial surface. Note that under water the feet
are not resting flat on the surface and are not bearing the animal's
weight.
t5
~ 0
u,
-1
0
-2
0
0
-3
o
o
l1li 25° DAMP SAND
'*
-4
-4
25° SUBAQ. MUD
-2
0
2
4
LEVEL WET MUD
LEVEL DRY SAND
6
8
FUNCTION 1
FIGURE
8 -Scattergram showing the amount of similarity or difference
between the experimental western newt tracks and the fossil tracks
on sloping cross-beds.
FIGURE
JOURNAL OF PALEONTOLOGY, V. 70, NO.6, 1996
1010
EXPERIMENTAL GROUPS
VARIABLES
TOES PER PRINT; PES
1
3
2
4
5
6
TOES PER PRINT; MANUS
2
3
1
4
5
6
TOE DIVERGENCE; PES
1
2
3
4
5
6
TOE DIVERGENCE; MANUS
3
2
4
5
6
STRIDE/WIDTH
1
2
3
4
5
6
PACE ANGULATION
1
3
2
4
5
6
FIGURE 9-Homogenous subsets for each variable (Tukey's HSD test;
P > 0.05). Each horizontal line indicates a subset of experimental
groups within which the variable shows no significant difference.
Groups are arranged approximately in the left-right sequence in which
they appear on Figure 8. 1, sloped subaqueous mud; 2, sloped damp
sand; 3, sloped subaqueous sand; 4, level dry sand; 5, sloped wet mud;
6, level wet mud.
experimentally were not successful (Brand, 1979). The results
reported here now confirm that good trackways can be produced
in damp sand. When the sand was moderately moistened and
left overnight, it then had enough cohesion to produce good
foot impressions (Figure 6).
The fossil tracks on sloped cross-beds of the Coconino Sandstone are most similar to the experimental trackways in sloped,
subaqueous sand or mud or subaerial, damp sand. The close
similarity between these experimental trackways and the Coconino Sandstone trackways does not necessarily indicate anything about the systematic relationships ofthe trackmakers, but
may be caused by similarity of substrate conditions. The laboratory trackways were made by modern amphibians and the
fossil trackways were made by Permian tetrapods; the data in
this study do not necessarily imply that the Permian tetrapods
were amphibians. It is quite possible that the trackways of'aquatic and terrestrial reptiles would be affected by substrate differences in much the same way as newt tracks are affected. If that
is correct, then the results reported here would likely apply at
least partially to both reptiles and amphibians. The data do
suggest that the Coconino Sandstone fossil trackways may have
been produced in either subaqueous sand or subaerial damp
sand. The choice between these alternatives depends on further
study of criteria that can indicate buoyancy of the trackmakers
(Brand, 1992; Brand and Tang, 1991), and continued study of
the sedimentological evidence.
ACKNOWLEDGMENTS
H. T. Goodwin's and D. Cowles's assistance with statistical
analysis in this project was much appreciated. M. G. Lockley
and an anonymous individual provided helpful critiques of an
earlier draft of the manuscript.
REFERENCES
BRAND, L. 1979. Field and laboratory studies on the Coconino Sandstone (Permian) vertebrate footprints and their paleoecological implications. Palaeogeography, Palaeoclimatology, Palaeoecology, 28:
25-38.
- - . 1992. Reply (to comments) on "Fossil vertebrate footprints in
the Coconino Sandstone (Permian) of northern Arizona: evidence for
underwater origin". Geology, 20:668-670.
- - , AND T. TANG. 1991. Fossil vertebrate footprints in the Coconino
Sandstone (Permian) of northern Arizona: evidence for underwater
origin. Geology, 19:1201-1204.
COHEN, A., M. LoCKLEY, J. HALFPENNY, AND A. E. MICHEL. 1991.
Modern vertebrate track taphonomy at Lake Manyara, Tanzania.
Palaios, 6:371-389.
GILMORE, C. W. 1926. Fossil footprints from the Grand Canyon.
Smithsonian Miscellaneous Collections, 77(9):1-41.
- - . 1927. Fossil footprints from the Grand Canyon: second contribution. Smithsonian Miscellaneous Collections, 80(3): 1-78.
HAUBOLD, H. 1971. Ichnia Amphibiorum et Reptiliorum fossilium.
Handbuch der Palaoherpetologie, Teil 18. Gustav Fischer Verlag,
Stuttgart, 124 p.
- - . 1974. Die Fossilen Saurierfahrten, Die Neue Brehrn-Bucherei,
A. Ziemsen Verlag, Wittenberg Lutherstadt, 168 p.
- - . 1984. Saurierfahrten, Die Neue Brehm-Bucherei, A. Ziemsen
Verlag, Wittenberg Lutherstadt, 232 p.
LAPORTE, L. F., AND A. K. BEHRENSMEYER. 1980. Tracks and substrate
reworking by terrestrial vertebrates in Quaternary sediments of Kenya. Journal of Sedimentary Petrology, 50:1337-1346.
LOCKLEY, M. G. 1986. The paleobiological and paleoenvironmental
importance of dinosaur footprints. Palaios, 1:37-47.
- - . 1987. Dinosaur tracks symposium signals a renaissance in vertebrate ichnology. Paleobiology, 13:246-252.
LULL, R. S. 1918. Fossil footprints from the Grand Canyon of the
Colorado. American Journal of Science, 45:337-346.
MARTINEZ, J., AND B. IGLEWICZ. 1981. A test for departure from normality based on a biweight estimator of scale. Biometrika, 68:331333.
MCKEE, E. D. 1947. Experiments on the development of tracks in fine
cross-bedded sand. Journal of Sedimentary Petrology, 17:23-28.
MCKEEVER, P. J. 1991. Trackway preservation in eolian sandstones
from the Permian of Scotland. Geology, 19:726-729.
NORUSIS, M. J. 1988. SPSS/PC+ Advanced Statistics V2.0, p. BlB39. SPSS Inc., Chicago.
PADIAN, K., AND P. E. OLSEN. 1984a. The fossil trackway Pteraichnus:
not Pterosaurian, but crocodilian. Journal of Paleontology, 58:178184.
- - , AND - - . 1984b. Footprints of the Komodo Monitor and the
trackways of fossil reptiles. Copeia, 1984:662-671.
PEABODY, F. E. 1959. Trackways of living and fossil salamanders.
University of California Publications in Zoology, 63:1-71.
SPAMER, E. E. 1984. Paleontology in the Grand Canyon of Arizona:
125 years of lessons and enigmas from the late Precambrian to the
present. The Mosasaur, 2 (December):45-128.
ACCEPTED 13 FEBRUARY 1996