Energetic costs of locomotion in bears: Is plantigrade locomotion energetically economical?

Abstract

Ursids are the largest mammals to retain a plantigrade posture. This primitive posture has been proposed to result in reduced locomotor speed and economy relative to digitigrade and unguligrade species, particularly at high speeds. Previous energetics research on polar bears (Ursus maritimus) found locomotor costs were more than double predictions for similarly sized quadrupedal mammals, which could be a result of their plantigrade posture or due to adaptations to their Arctic marine existence. To evaluate whether polar bears are representative of terrestrial ursids or distinctly uneconomical walkers, this study measured the mass-specific metabolism, overall dynamic body acceleration, and gait kinematics of polar bears and grizzly bears (Ursus arctos) trained to rest and walk on a treadmill. At routine walking speeds, we found polar bears and grizzly bears exhibited similar costs of locomotion and gait kinematics, but differing measures of overall dynamic body acceleration. Minimum cost of transport while walking in the two species (2.21 J kg-1 m-1) was comparable to predictions for similarly sized quadrupedal mammals, but these costs doubled (4.42 J kg-1 m-1) at speeds ≥5.4 km h-1 Similar to humans, another large plantigrade mammal, bears appear to exhibit a greater economy while moving at slow speeds.

Full text

RESEARCH ARTICLE
Energetic costs of locomotion in bears: is plantigrade locomotion
energetically economical?
Anthony M. Pagano
1,2,
*, Anthony M. Carnahan
3
, Charles T. Robbins
3
, Megan A. Owen
4
, Tammy Batson
5
,
Nate Wagner
5
, Amy Cutting
6
, Nicole Nicassio-Hiskey
6
, Amy Hash
6
and Terrie M. Williams
2
ABSTRACT
Ursids are the largest mammals to retain a plantigrade posture. This
primitive posture has been proposed to result in reduced locomotor
speed and economy relative to digitigrade and unguligrade species,
particularly at high speeds. Previous energetics research on polar
bears (Ursus maritimus) found locomotor costs were more than
double predictions for similarly sized quadrupedal mammals, which
could be a result of their plantigrade posture or due to adaptations to
their Arctic marine existence. To evaluate whether polar bears are
representative of terrestrial ursids or distinctly uneconomical walkers,
this study measured the mass-specific metabolism, overall dynamic
body acceleration, and gait kinematics of polar bears and grizzly
bears (Ursus arctos) trained to rest and walk on a treadmill. At routine
walking speeds, we found polar bears and grizzly bears exhibited
similar costs of locomotion and gait kinematics, but differing
measures of overall dynamic body acceleration. Minimum cost of
transport while walking in the two species (2.21 J kg
−1
m
−1
)was
comparable to predictions for similarly sized quadrupedal mammals,
but these costs doubled (4.42 J kg
−1
m
−1
) at speeds ≥5.4 km h
−1
.
Similar to humans, another large plantigrade mammal, bears appear
to exhibit a greater economy while moving at slow speeds.
KEY WORDS: Acceleration, Cost of transport, Metabolism, Overall
dynamic body acceleration, Ursus arctos,Ursus maritimus
INTRODUCTION
A plantigrade posture in which the heel makes contact with the
ground during a step is considered to be an ancestral form of
locomotion (Lovegrove and Haines, 2004). This posture has been
shown to enhance locomotor economy while walking in humans,
despite a reduced economy while running relative to digitigrade or
unguligrade postures, which enable greater stride length and elastic
storage (Carrier, 2016). Members of the family Ursidae represent
the largest mammals to have retained a plantigrade posture (Brown
and Yalden, 1973), which likely increases their dexterity for digging
and climbing and enhances support for their large body mass
(McLellan and Reiner, 1994), but may impose a reduced energetic
economy during locomotion (Lovegrove and Haines, 2004; Shine
et al., 2015).
Ursids represent a small family of large-bodied terrestrial
mammals with a diverse range of diets from specialist carnivores
to specialist herbivores and generalist omnivores. Energetics research
on ursids has largely focused on their ability to reduce metabolism
during hibernation (e.g. Watts et al., 1987; Watts and Cuyler, 1988;
Watts and Jonkel, 1988; Tøien et al., 2011). Resting metabolic rates
(RMRs) have also been examined in many ursids (Fei et al., 2016;
Hurst, 1981; McNab, 1992; Tøien et al., 2011; Watts et al., 1987).
Giant pandas (Ailuropoda melanoleuca) (Fei et al., 2016) and sloth
bears (Melursus ursinus) (McNab, 1992) exhibit RMRs that are 18%
and 41% less than predictions for similarly sized mammals (Kleiber,
1975), while polar bears (Ursus maritimus) (Hurst et al., 1991;
Pagano et al., 2018; Watts et al., 1991) and black bears (Ursus
americanus) (Tøien et al., 2011) exhibit RMRs that are 62% and 23%
greater than predictions. This increased maintenance cost in polar
bears, and to a lesser extent in black bears, is likely a result of their
carnivorous diet, whereas giant pandas are a specialist herbivore and
sloth bears an insectivore, both of which impose a lower energetic
cost than carnivory (McNab, 1986). Despite this understanding of
baseline energetic costs in ursids, the energetic costs of locomotion
have received less attention and have only been examined in polar
bears. In polar bears, the energetic cost of walking is more than twice
that predicted for similarly sized quadrupedal mammals (Hurst et al.,
1982a; Øritsland et al., 1976; Watts et al., 1991). Yet, it remains
unknown whether this high cost of transport is found across the
Ursidae, potentially as a result of plantigrade locomotion, or whether
polar bears are distinctly uneconomical walkers as a result of their
carnivorous, marine and semi-aquatic lifestyle (Pagano et al., 2018;
Williams, 1999; Williams et al., 2002).
Despite the paraphyletic relationship between polar bears and
grizzly bears (Ursus arctos) (Talbot and Shields, 1996), polar bears
exhibit a number of physiological and behavioral adaptations
distinct from grizzly bears, likely as a consequence of their marine
existence. In addition to being the most carnivorous of the bear
species (Stirling and Derocher, 1990), polar bears have larger paws
(potentially as an adaptation for swimming; DeMaster and Stirling,
1981), reduced forelimb dexterity (Iwaniuk et al., 2000) and exhibit
distinct running kinematics using a transverse gallop compared with
the rotary gallop of grizzly bears (Renous et al., 1988). Additionally,
a study using tri-axial accelerometers to test the ability of data
from grizzly bears to serve as proxies for discriminating basic
behaviors in polar bears found that data from grizzly bears failed to
reliably discriminate polar bear behaviors (Pagano et al., 2017).
This suggests differences in morphology and body movements
between the two species while performing similar behaviors
(Pagano et al., 2017).
To evaluate whether polar bears have uniquely high energetic
costs of locomotion among ursids, we examined the metabolic rates
Received 2 December 2017; Accepted 21 April 2018
1
US Geological Survey, Alaska Science Center, 4210 University Dr., Anchorage,
AK 99508, USA.
2
Department of Ecology & Evolutionary Biology, University of
California, Santa Cruz, CA 95060, USA.
3
School of the Environment and School of
Biological Sciences, Washington State University, Pullman, WA 99164, USA.
4
Institute for Conservation Research, San Diego Zoo Global, San Diego, CA 92027,
USA.
5
San Diego Zoo Global, San Diego, CA 92027, USA.
6
Oregon Zoo, Portland,
OR 97221, USA.
*Author for correspondence (apagano@usgs.gov)
A.M.P., 0000-0003-2176-0909
1
© 2018. Published by The Company of Biologists Ltd
|
Journal of Experimental Biology (2018) 221, jeb175372. doi:10.1242/jeb.175372
Journal of Experimental Biology

of resting and locomotion in polar bears and grizzly bears. To do
this, we measured the oxygen consumption, overall dynamic body
acceleration (ODBA), stride length and stride frequency of captive
polar bears and grizzly bears while at rest in a metabolic chamber
and walking on a motorized treadmill. We tested the hypotheses that
polar bears differ from grizzly bears in their relationships between
speed and oxygen consumption, ODBA, stride length and stride
frequency. We compared the costs of locomotion of polar bears and
grizzly bears with respect to other plantigrade mammals and
digitigrade carnivores, and with estimates based on allometric
relationships. We further evaluated the relationship between oxygen
consumption and ODBA in polar bears and grizzly bears as a proxy
for energy expenditure. In other species, ODBA is strongly
correlated with energy expenditure because of the relationship
between acceleration and muscle contraction (Gleiss et al., 2011;
Wilson et al., 2006), enabling the use of accelerometers to measure
energy expenditure in wild animals (e.g. Gómez Laich et al., 2011;
Halsey et al., 2009a, 2011; Williams et al., 2014; Wilson et al.,
2006, 2012). For example, ODBA has been used to measure
instantaneous energetics (e.g. Williams et al., 2014) and to evaluate
the energy landscapes of wild animals (e.g. Shepard et al., 2013;
Wilson et al., 2012). This is based on the assumption that movement
is the primary factor influencing variability in energy expenditure
(Costa and Williams, 1999; Gleiss et al., 2011; Wilson et al., 2006).
If such relationships are similar in ursids, it could provide a method
to remotely measure their energy expenditure. Lastly, we evaluated
the locomotor speeds of polar bears walking and running on the sea
ice to assess whether preferred locomotor speeds in the wild
conform to our energetic predictions.
MATERIALS AND METHODS
Experimental design
We measured oxygen consumption (V
̇
O
2
) via open-flow
respirometry, and stride frequency, stride length and ODBA via
kinematic and accelerometry analyses in polar bears and grizzly
bears. Measurements were made within a sealed metabolic chamber
(2.7 m×0.9 m×1.2 m) constructed of polycarbonate walls that were
reinforced with a steel frame (Technical Services, Washington
State University, Pullman, WA, USA) and mounted on the surface
of a variable-speed treadmill (T1 Trotter horse treadmill, Horse
Gym USA, LLC, Wellington, FL, USA). We further measured the
movement rates of wild female polar bears while walking and
running on the sea ice of the Beaufort Sea.
Animals
One polar bear (Ursus maritimus Phipps 1774) at the San Diego
Zoo and seven grizzly bears (Ursus arctos Linnaeus 1758) at
Washington State University were used for metabolic, acceleration
and gait kinematic measurements (Table 1). Additionally, one polar
bear at the Oregon Zoo was used for acceleration and gait kinematic
measurements (Table 1). The polar bear at the San Diego Zoo was
trained over 5 months and conditioned to rest while lying in sternal
recumbency and to walk on the moving treadmill while receiving
food (i.e. meat and fish) every 20 s. The polar bear at the Oregon
Zoo was trained over 8 months to walk on the moving treadmill
while receiving food every 20 s. The grizzly bears were similarly
trained over 2 months and conditioned to rest while lying in sternal
recumbency and walk on the moving treadmill while receiving food
every 10–20 s. The research was approved by the Animal Care and
Use Committees of the University of California, Santa Cruz, the US
Geological Survey, Alaska Science Center, the San Diego Zoo
Global, Oregon Zoo and Washington State University (protocols
04780 and 04952). Polar bear research was further approved under
US Fish and Wildlife Service Marine Mammal Permit MA77245B.
To measure locomotor speed in wild bears, we captured one
subadult and five adult female polar bears without dependent young
on the sea ice of the Beaufort Sea in April 2015 and 2016. Polar
bears were located from a helicopter and immobilized with a rapid-
injection dart (Palmer Cap-Chur Equipment, Douglasville, GA,
USA) containing zolazepam-tiletamine (Telazol
®
) (Stirling et al.,
1989). Procedures were approved by the Animal Care and Use
Committees of the University of California, Santa Cruz, and the US
Geological Survey, Alaska Science Center. Field research was
approved under US Fish and Wildlife Service Marine Mammal
Permit MA690038.
Metabolic measurements
V
̇
O
2
measurements were collected over 6–13 min intervals with a
minimum of 5 min of steady-state behaviors to ensure equilibration.
For both species, at least one resting measurement was taken
following an overnight fast to ensure a post-absorptive state. For the
grizzly bears, a subsequent resting measurement was taken 3 h after
feeding to evaluate the potential effects of specific dynamic action
on V
̇
O
2
measurements. Food intake per session ranged from 728 to
963 g (polar bear) and 2000 to 2300 g (grizzly bears).
We used a vacuum pump (FlowKit Mass Flow Generator –2000,
Sable Systems International, Inc., Las Vegas, NV, USA) to draw air
in along the lower edge of the treadmill at 700 l min
−1
during
measurements. We monitored flow rates continuously and
maintained oxygen levels ≥20% to avoid hypoxic conditions.
Sub-samples of air from the exhaust port of the chamber were drawn
through a series of six columns, filled with desiccant (Drierite,
W. A. Hammond Drierite, Xenia, OH, USA), and scrubbed of
carbon dioxide (Sodasorb, W. R. Grace & Co, Chicago, IL, USA)
before entering the oxygen analyzer (Sable Systems International,
Inc.). We monitored the percentage of oxygen in the expired air
continuously and recorded values once per second using Expedata
Analysis software (Sable Systems International, Inc.). Air
temperature within the chamber ranged from 22.2 to 24.6°C
(mean 23.9°C) for polar bears and from 18.6 to 34.3°C (mean 28.9°C)
for grizzly bears. We converted values to V
̇
O
2
using eqn 4B from
Withers (1977), assuming a respiratory quotient of 0.78. All values
were corrected to standard temperature and pressure, dry. We
calibrated the entire system prior to measurements with dry ambient
air (20.95% O
2
) and periodically with dry N
2
gas (Fedak et al.,
1981). Body mass was measured using a platform scale. We
estimated net minimum cost of transport (COT
min
) as the slope and
postural cost of activity as the y-intercept of the relationship between
V
̇
O
2
(ml O
2
kg
−1
s
−1
)andspeed(ms
−1
) (Taylor et al., 1982). We
estimated total cost of transport (COT
tot
) by dividing V
̇
O
2
by speed.
Table 1. Summary of animals used in this study
Species and
individual Sex Age
Body
mass (kg) Location
Polar bear 1 Female 31 242 Oregon Zoo
Polar bear 2 Female 16 235 San Diego Zoo
Grizzly bear 1 Male 15 253 Washington State University
Grizzly bear 2 Male 15 239 Washington State University
Grizzly bear 3 Female 14 164 Washington State University
Grizzly bear 4 Female 12 143 Washington State University
Grizzly bear 5 Female 12 142 Washington State University
Grizzly bear 6 Male 2 126 Washington State University
Grizzly bear 7 Female 2 95 Washington State University
2
RESEARCH ARTICLE Journal of Experimental Biology (2018) 221, jeb175372. doi:10.1242/jeb.175372
Journal of Experimental Biology

Gait kinematics
We measured stride frequency (strides s
−1
) and stride length (m) at
each speed using video from a high-speed camera (Panasonic,
Lumix FZ300, 120 frames s
−1
) and a high-definition video camera
(Sony, Tokyo, Japan; HDR-CX260V, 1080 HD, 60p) positioned
perpendicular to the treadmill. Video images were analyzed with
video-editing and motion analysis software (Corel Video Studio Pro
X5, Corel Corp., Ottawa, ON, Canada; ProAnalyst, Xcitex, Woburn,
MA, USA). Stride frequency was measured as the average interval
for 25 cycles of the front right foot (Heglund and Taylor, 1988).
ODBA
We bolted archival loggers (TDR10-X-340D, Wildlife Computers,
Inc., Redmond, WA, USA) to the side of collars such that they were
on the left side of the bear’s neck (see fig. 1 in Pagano et al., 2017).
Archival loggers measured tri-axial acceleration (m s
−2
)at16Hz
(range ±20 m s
−2
) while bears were resting and walking within
the metabolic chamber. We also included acceleration and
V
̇
O
2
measurements collected from the same polar bear (264 kg) at
the San Diego Zoo while she rested during a previous study (Pagano
et al., 2018). We estimated the V
̇
O
2
of the polar bear at the Oregon
Zoo based on the relationship between speed and V
̇
O
2
derived below.
We converted accelerometer measures from m s
−2
to g
(1 g=9.81 m s
−2
). We used a 2 s running mean of the raw
acceleration data to calculate static acceleration (gravitational
acceleration) and subtracted the static acceleration from the raw
acceleration data to calculate dynamic acceleration (Wilson et al.,
2006; Shepard et al., 2008). ODBA was calculated as the absolute
sum of dynamic acceleration across the three axes (Wilson et al.,
2006).
Preferred locomotor speeds
We measured the movement rates (km h
−1
) of six female polar bears
over 3–13 days while walking or running on the sea ice. Movement
rates were derived from global positioning system (GPS) collars
(Exeye, LLC, Bristow, VA, USA) with a GPS fix rate every 5 or
10 min. Location data were transmitted via the Iridium satellite
system. We used a continuoustime correlated random walk (CRAWL)
model (https://CRAN.R-project.org/package=crawl; Johnson et al.,
2008) in program R (http://www.R-project.org/) to predict locations
on a 10 min interval based on GPS locations. The CRAWL model
accounts for variable location quality and sampling intervals.
We assigned GPS location data an accuracy of 30 m (Frair et al.,
2010). We calculated the minimum distance traveled between two
successive predicted locations asthe great-circle distance (i.e. distance
accounting for the Earth’s curvature), and calculated movement rate
by dividing distance by the duration between predicted locations (i.e.
10 min) in SAS (version 9.3, SAS Institute Inc., Cary, NC, USA). We
identified walking and running movements based on archival loggers
(TDR10-X-340D, Wildlife Computers, Inc.) attached to the GPS
collars, which measured tri-axial acceleration (m s
−2
) continuously at
16 Hz (range ±20 m s
−2
). Walking and running were discriminated
within the accelerometer data using a Random Forest model
(Breiman, 2001) in program R (RandomForest package, https://
CRAN.R-project.org/package=randomForest) as described by
Pagano et al. (2017). We linked these accelerometer-derived
behaviors with their corresponding predicted location data by
calculating the percentage time spent walking or running between
predicted locations (i.e. 10 min) in SAS. If ≥95% of the time
between predicted locations was classified as walking or running,
we considered the movement rate during this interval to be
indicative of walking or running.
Analyses
We combined our polar bear V
̇
O
2
measurements while walking
with V
̇
O
2
measurements similarly recorded using open-flow
respirometry from seven sub-adult polar bears (two females and
five males) that ranged in body mass from 110 to 235 kg, walking
and running–walking on a treadmill (Hurst et al., 1982a,b;
Øritsland et al., 1976; Watts et al., 1991). We used least-squares
linear regression to evaluate the relationship between V
̇
O
2
and
speed. Although Hurst et al. (1982a) proposed a curvilinear
relationship between V
̇
O
2
and speed in polar bears as a result of
measurements at speeds ≥5.4 km h
−1
,weevaluatedV
̇
O
2
measurements
at speeds ≥5.4 km h
−1
separately as data from wild polar bears
indicate they rarely walk this fast (Whiteman et al., 2015) and the
predicted gait transition speed for 100–250 kg animals is 5.7–
5.3 km h
−1
(Heglund and Taylor, 1988). We used analysis of
covariance (ANCOVA) to evaluate whether the relationships between
V
̇
O
2
and speed differed between speeds <5.4 and ≥5.4 km h
−1
.For
grizzly bears, we similarly used least-squares linear regression to
evaluate the relationship between V
̇
O
2
and speed. We used
ANCOVA to evaluate whether the intercepts and slopes differed
between polar bears and grizzly bears in their relationships
between V
̇
O
2
and speed. We further used least-squares linear
regression to evaluate the relationship between V
̇
O
2
and ODBA and
speed and ODBA, and used ANCOVA to evaluate whether the
relationship between V
̇
O
2
and ODBA differed between species.
ANCOVA was also used to evaluate whether the relationship
between stride frequency and speed as well as stride length and
speed differed between species. We calculated the mean and
distribution of walking and running speeds measured in wild
female polar bears on the sea ice. All analyses were conducted in
program R and differences of P≤0.05 were considered significant.
RESULTS
Metabolic rates
RMR of the adult female polar bear averaged 0.27±0.01 ml O
2
g
−1
h
−1
(mean±s.e.m., n=5), with a low of 0.25 ml O
2
g
−1
h
−1
. In combination
with measures previously collected from sub-adult male and female
polar bears (Hurst, 1981; Watts et al., 1991), the post-absorptive
RMR of polarbears averaged 0.23±0.02 ml O
2
g
−1
h
−1
(n=6). Grizzly
bears remained active during resting measurements (e.g. head and
limb movements) and, thus, their RMRs are akin to zero-velocity
measurements (i.e. y-intercept), relating to the postural effect of
activity (Schmidt-Nielsen, 1972; Taylor et al., 1970). Zero-velocity
metabolic rates of the grizzly bears while post-absorptive averaged
0.55±0.11 ml O
2
g
−1
h
−1
(n=5) with a low of 0.30 ml O
2
g
−1
h
−1
.
Zero-velocity metabolic rates of the grizzly bears 3 h post-
prandial averaged 0.50±0.04 ml O
2
g
−1
h
−1
(n=5) with a low of
0.36 ml O
2
g
−1
h
−1
.
We found a significant difference in the slope (F
1,107
=6.87, P=0.01)
and intercept (F
1,108
=58.21, P<0.001) in the relationship between
V
̇
O
2
and speed for bears walking at <5.4 km h
−1
(Fig. 1A) and bears
walking at ≥5.4 km h
−1
(Fig. 2A). Polar bear metabolic rates while
walking at <5.4 km h
−1
exhibited a linear relationship between
V
̇
O
2
(ml O
2
g
−1
h
−1
) and speed (km h
−1
): V
̇
O
2
=0.44+0.12×speed
(r
2
=0.42, P<0.001, n=35), and were on average 1.5 times greater than
rates predicted for terrestrial carnivores based on body mass and
speed (Taylor et al., 1982). At speeds ≥5.4 km h
−1
, polar bear
V
̇
O
2
exhibited a linear relationship with speed: V
̇
O
2
=0.41+0.22×speed
(r
2
=0.32, P<0.001, n=37; Fig. 2A). At speeds ≤4.6 km h
−1
, grizzly
bear V
̇
O
2
similarly exhibited a linear relationship with speed:
V
̇
O
2
=0.50+0.13×speed (r
2
=0.82, P<0.001, n=39), and metabolic
rates averaged 1.7 times greater than rates predicted for terrestrial
3
RESEARCH ARTICLE Journal of Experimental Biology (2018) 221, jeb175372. doi:10.1242/jeb.175372
Journal of Experimental Biology

carnivores based on body mass and speed (Taylor et al., 1982). We
found no difference in the slope (F
1,70
=0.06, P=0.80) or intercept
(F
1,71
=3.56, P=0.06) in the relationship between V
̇
O
2
and speed for
the two species at speeds <5.4 km h
−1
. Combining data from the two
species, at speeds <5.4 km h
−1
we found a linear relationship
between V
̇
O
2
and speed: V
̇
O
2
=0.50+0.11×speed (r
2
=0.64, P<0.001,
n=74; Fig. 1A). Postural cost of activity (i.e. y-intercept) was
0.50 ml O
2
g
−1
h
−1
or 2.2 times greater than predictions based on
body mass (Taylor et al., 1982). Net COT
min
was 0.11 ml O
2
kg
−1
m
−1
(2.21 J kg
−1
m
−1
), or 1.1 times greater than predictions based on body
mass (Fig. 3) (Taylor et al., 1982). COT
tot
waslowestat1.2ms
−1
(4.3 km h
−1
) (Fig. 4). At speeds ≥5.4 km h
−1
, net COT
min
was
0.22 ml O
2
kg
−1
m
−1
(4.42 J kg
−1
m
−1
) (Fig. 3).
Gait kinematics
Bears exhibited plantigrade gaits with the toes and metatarsals flat on
the ground (Fig. 5; Movies 1, 2). We found no difference in the slope
(F
1,28
=0.93, P=0.34) or intercept (F
1,29
=2.43, P=0.13) in the
relationship between stride frequency and speed or stride length and
speed (F
1,28
=2.26, P=0.14; F
1,29
=2.08, P=0.16, respectively) between
the two species. Stride frequency (strides s
−1
) increased linearly with
speed: stride frequency=0.21+0.16×speed (r
2
=0.88, P<0.001, n=32;
Fig. 1B). Stride length (m) increased linearly with speed: stride
length=0.71+0.15×speed (r
2
=0.76, P<0.001, n=32; Fig. 1C).
ODBA
The relationship between V
̇
O
2
(ml O
2
g
−1
h
−1
) and ODBA (g)
differed in the slope (F
1,29
=5.49, P=0.03) and intercept (F
1,30
=4.92,
P=0.03) between the species. This difference appeared to be
predominantly driven by differences in dynamic body acceleration
in the sway (z) dimension (Fig. 6). Polar bear V
̇
O
2
increased linearly
as a function of ODBA: V
̇
O
2
=−0.90+12.33×ODBA (r
2
=0.84,
P<0.001, n=18; Fig. 7A). Polar bear speed was also strongly
predicted by ODBA: speed=−2.92+16.25×ODBA (r
2
=0.92,
P<0.001, n=18). Grizzly bear V
̇
O
2
increased linearly as a function
of ODBA: V
̇
O
2
=−0.05+2.03×ODBA (r
2
=0.76, P<0.001, n=15;
Fig. 7B). Grizzly bear speed was also strongly predicted by ODBA:
speed=−4.62+16.12×ODBA (r
2
=0.81, P<0.001, n=15).
Preferred locomotor speeds
Walking and running speeds of female polar bears on the sea ice
over 10 min intervals averaged 3.4±0.04 km h
−1
(n=533, Fig. 2B)
and ranged from 0.4 to 10.0 km h
−1
. Only 3% of these movements
were at ≥5.4 km h
−1
(Fig. 2B).
DISCUSSION
Contrary to previous energetic studies on polar bears, our results
indicate that polar bears and grizzly bears are energetically similar to
other quadrupedal mammals while walking at preferred speeds. In
humans, a plantigrade posture while walking has been shown to
reduce the cost of transport relative to a digitigrade posture, but
incurs a 61% increase in cost of transport while running
(Cunningham et al., 2010). Our results similarly indicate that, at
routine walking speeds, both polar bears and grizzly bears exhibit
costs of transport that are comparable to predictions from other
quadrupedal mammals based on their body mass (Taylor et al.,
1982), but at speeds ≥5.4 km h
−1
the cost of transport doubles,
greatly exceeding predictions.
Hurst et al. (1982a) proposed a curvilinear relationship between
speed and energy expenditure in polar bears as a result of these
disproportionately high energetic costs at speeds ≥5.4 km h
−1
.
However, data from wild polar bears indicate they rarely walk this
fast (Fig. 2B; Whiteman et al., 2015), which suggests these speeds
are likely non-preferred and may require an uneconomical gait. We
found COT
tot
was lowest at 4.3 km h
−1
, which is almost 1 km h
−1
greater than the mean walking speed measured in polar bears on the
1.6
Oxygen consumption (ml O2 g–1 h–1)
Stride frequency (strides s–1)
Stride length (m)
1.2
0.8
0.4
0
01
2345
012345
012
Speed (km h–1)
345
1.2
1.0
0.8
0.6
0.4
0.2
0
1.6
1.4
1.2
1.0
0.8
0.6
A
B
C
Fig. 1. Relationship between oxygen consumption, gait kinematics and
locomotor speed in polar bears and grizzly bears. (A) Least-squares
regression (solid line) of mass-specific oxygen consumption in relation to
locomotor speed for polar bears and grizzly bears on a treadmill. Points
represent individual steady-state measurements for polar bears (yellow circles,
present study; orange circles, Hurst et al., 1982a; dark-orange circles,
Øritsland et al., 1976; red circles, Watts et al., 1991) and grizzly bears (black
circles) (see Results for regression statistics). (B) Least-squares regression
(solid line) between stride frequency and speed in polar bears (yellow circles)
and grizzly bears (black circles) (see Results for regression statistics).
(C) Least-squares regression (solid line) between stride length and speed in
polar bears (yellow circles) and grizzly bears (black circles) (see Results for
regression statistics).
4
RESEARCH ARTICLE Journal of Experimental Biology (2018) 221, jeb175372. doi:10.1242/jeb.175372
Journal of Experimental Biology

sea ice over 10 min periods. Additionally, field movements would
be expected to impose greater energetic costs relative to movements
on a treadmill (Bidder et al., 2017). Shine et al. (2015) documented
the lack of a trotting gait in grizzly bears and reported transition
speeds of ≥7.2 km h
−1
for running walks and ≥10.8 km h
−1
for
canters. Walking involves storing and recovering energy with each
stride via an exchange between gravitational–potential and kinetic
energies through an inverted pendulum (Cavagna et al., 1977).
However, the benefits of these pendulum mechanics decline at both
low and high speeds. At high speeds, animals can trot, run or hop,
which allows energy to be conserved through elastic energy
recovery (Cavagna et al., 1977). Yet, given their plantigrade posture,
bears would be expected to have reduced energy savings from
elastic energy recovery relative to unguligrade or digitigrade
mammals (Cunningham et al., 2010; Reilly et al., 2007). In
humans, plantigrade locomotion enhances pendular mechanics
and reduces ground collisional losses in kinetic energy while
walking, at the expense of reduced elastic storage at higher speeds
(Cunningham et al., 2010). At present, no data exist on the gait
mechanics of polar bears at speeds between 5.4 and 7.2 km h
−1
to
better evaluate the causes of these disproportionate energetic costs,
and V
̇
O
2
of grizzly bears has not been examined at speeds
>4.6 km h
−1
. Although polar bears seldom walk at these speeds
in the wild (Fig. 2B; Whiteman et al., 2015), future research
evaluating the gait kinematics and cost of transport of bears at
speeds ≥5.4 km h
−1
would help to better elucidate the aerobic
performance of ursids compared with other quadrupedal mammals.
At routine walking speeds, polar bears and grizzly bears
exhibited similar energetic costs and gait kinematics. Despite the
evolutionary divergence of polar bears from grizzly bears, which
has enabled polar bears to exist within the Arctic marine
environment and facilitated their ability to swim long distances
(Pagano et al., 2012; Pilfold et al., 2017), these adaptations appear
to have had little effect on their costs of transport while walking
compared with their closest living relative. This result is contrary to
most semi-aquatic mammalsthat have higher costs of transport than
3.2 A
B
2.8
2.4
2.0
1.6
1.2
0.8
0.4
Oxygen consumption (ml O2 g–1 h–1)
% Total measurements
0
40
30
20
10
0
0123456
Speed (km h–1)
78
012345678
91011
Fig. 2. Relationship between oxygen
consumption and locomotor speed for bears
moving on a treadmill and locomotor speed of
wild polar bears while walking and running on
the sea ice. (A) Mass-specific oxygen
consumption in relation to locomotor speed. Points
represent individual steady-state measurements
for polar bears (orange circles, Hurst et al., 1982a;
yellow circles, Hurst et al., 1982b; dark-orange
circles, Øritsland et al., 1976; red circles, Watts
et al., 1991). The solid line is the least-squares
regression from polar bears and grizzly bears at
<5.4 km h
−1
(Fig. 1A) and the dotted line is the
least-squares regression from polar bears at
≥5.4 km h
−1
(see Results for regression statistics).
The dashed line is the predicted relationship
derived from other terrestrial carnivores (Taylor
et al., 1982). (B) Frequency distribution of walking
and running speeds over 10 min intervals from six
female polar bears on the sea ice of the Beaufort
Sea in April 2015 and 2016 (n=533).
0.4
Running
walk
Run
Walk Walk
0.3
0.2
0.1
10
COTmin (ml O2 kg–1 m–1)
Body mass (kg)
100
Fig. 3. Net minimum cost of transport (COT
min
) in digitigrade carnivores
and plantigrade mammals. Digitigrade carnivores: canids (gray squares:
Bryce and Williams, 2017; Taylor et al., 1982) and felids (green triangles:
Taylor et al., 1982; Williams et al., 2014). Plantigrade mammals: primates (blue
diamonds: Cunningham et al., 2010; Taylor et al., 1982) and ursids (yellow
circles: present study). The solid line is the predicted relationship for COT
min
of
quadrupedal mammals (Taylor et al., 1982). Silhouette images are from http://
www.supercoloring.com (https://creativecommons.org/licenses/by/4.0/).
5
RESEARCH ARTICLE Journal of Experimental Biology (2018) 221, jeb175372. doi:10.1242/jeb.175372
Journal of Experimental Biology

strict terrestrial or aquatic mammals (Williams, 1999; Williams
et al., 2002), and suggests that polar bears are primarily adapted for
walking and may incur high energetic costs while swimming
(Durner et al., 2011; Griffen, 2018).
Despite walking costs that were similar to those of other
quadrupedal mammals, we found both polar bears and grizzly
bears have postural costs that are more than double predictions
based on other quadrupedal mammals (Taylor et al., 1982). This
result is consistent with high resting metabolic rates (Hurst, 1981;
Pagano et al., 2018; Watts et al., 1991) and high field metabolic rates
in polar bears (Pagano et al., 2018). Taylor et al. (1970) found
postural costs ranged from 1.3 to 2.1 times RMR and Cavagna et al.
(1977) proposed that this elevated cost may reflect the cost of lifting
the center of mass against gravity. However, the postural costs we
found are greater than those reported in other large terrestrial
mammals. For example, in elephants (Elephas maximus), postural
costs were 1.4 times greater than predictions (Langman et al., 2012),
while in pumas (Puma concolor), postural costs were 1.6 times
greater than predictions (Williams et al., 2014). Hence, this
increased postural cost in polar bears and grizzly bears may in
part be a result of their plantigrade posture as more erect limb
postures (e.g. digitigrade and unguligrade) are known to have lower
muscle mass and greater effective mechanical advantage (Biewener,
1989; Reilly et al., 2007). We recommend further research to
explore the potential causes of these high postural costs in polar
bears and grizzly bears. These high costs of activity have important
energetic implications for wild polar bears, which appear to be
increasing their movement and activity rates in response to climate
change (Durner et al., 2017).
Similar to behavior discrimination using tri-axial accelerometers
(Pagano et al., 2017), we found the relationship between ODBA and
V
̇
O
2
differed between species. This difference appeared to be
primarily driven by differences in the sway (z) dimension between
species (Fig. 6C), which suggests greater side-to-side movement by
the grizzly bears while walking. Yet, such movements did not
appear to influence either gait kinematics or locomotor costs
between species. As our accelerometers were attached to collars on
the neck, these movements may reflect differences in head and neck
motions between species rather than limb or center of mass
movements. Halsey et al. (2009b) found body mass explained most
of the variation in the relationship between V
̇
O
2
and ODBA among
species. Our adult female grizzly bears wearing accelerometers
differed by an average of 89 kg from our adult female polar bears
wearing accelerometers, which may have also influenced their side-
to-side movements. Our results support Halsey et al.’s (2009b)
finding that the relationship between ODBA and V
̇
O
2
is species
specific. We recommend further evaluation of the effect of body
mass on the relationship between ODBA and V
̇
O
2
in ursids. In
particular, ursids are known for extreme seasonal fluctuations in
body mass as a result of changes in food availability and winter
dormancy (Nelson et al., 1983), and such changes may affect the
relationship between ODBA and V
̇
O
2
even on an intraspecific level.
Furthermore, the relationships we derived between ODBA and
V
̇
O
2
resulted in negative intercepts for both species, which suggests
that these relationships need to be further developed in order to use
ODBA as a proxy for energy expenditure in these species.
Polar bears and grizzly bears are known to travel extensive
distances and have large home ranges relative to other mammals
(Ferguson et al., 1999; McLoughlin and Ferguson, 2000;
McLoughlin et al., 1999), yet they are primarily ambush and
opportunistic predators that typically catch prey through sit-and-wait
and stalk behaviors rather than chasing down prey (Garneau et al.,
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
0 0.3 0.6 0.9 1.2
Speed (m s–1)
COTtot (ml O2 kg–1 m–1)
1.5 1.8 2.1 2.4
Fig. 4. Mass-specific total cost of transport (COT
tot
) in relation to
locomotor speed in polar bears and grizzly bears. Data are for polar bears
and grizzly bears walking at <1.5 m s
−1
(gray circles) and polar bears walking
at ≥1.5 m s
−1
(yellow circles). The equation describing the second-order
polynomial relationship between COT
tot
and walking speed at <1.5 m s
−1
is
COT
tot
=0.40×speed
2
−0.96×speed+0.80 (r
2
=0.83). The equation describing
the second-order polynomial relationship between COT
tot
and speed at
≥1.5 m s
−1
is COT
tot
=0.44×speed
2
−1.64×speed+1.78 (r
2
=0.22).
0
A
B
0.35 0.70
Time (s)
1.05 1.40 1.75
0 0.4 0.8 1.2 1.6 2.0
Fig. 5. Plantigrade walking gait of the grizzly bear and polar bear. (A) Single walking stride of an adult female grizzly bear moving on a treadmill at 2.8 km h
−1
over 1.75 s. (B) Single walking stride of an adult female polar bear moving on a treadmill at 2 km h
−1
over 2 s.
6
RESEARCH ARTICLE Journal of Experimental Biology (2018) 221, jeb175372. doi:10.1242/jeb.175372
Journal of Experimental Biology

2007; Pagano et al., 2018; Stirling, 1974; Stirling and Derocher,
1990). Our results provide the physiological basis for these seemingly
contradictory behaviors. Both species exhibit economical costs of
walking, facilitated by their plantigrade posture. However, like
humans, this comes at the expense of a less economical cost while
moving at higher speeds. Observations of polar bears chasing down
flightless geese (Iles et al., 2013) have inspired analyses that found
this hunting strategy to be energetically profitable (Gormezano et al.,
2016). Nevertheless, our results highlight the elevated energetic
demands for polar bears to chase down their prey compared with
traditional sit-and-wait tactics. This reinforces the importance of
Arctic sea ice to enable polar bears to efficiently capture prey.
Acknowledgements
We thank San Diego Zoo polar bear trainers B. Wolf and P. O’Neill, Washington
State University grizzly bear trainer B. E. Hutzenbiler, and C. Dunford. We thank
G. Durner, K. Rode, D. Ruthrauff, and members of the Williams lab for comments on
previous drafts of the manuscript. This research used resources of the Core Science
Analytics and Synthesis Applied Research Computing programat the US Geological
Survey. Any use of trade, firm or product names is for descriptive purposes only and
does not reflect endorsement by the US government.
Competing interests
The authors declare no competing or financial interests.
Author contributions
Conceptualization: A.M.P., T.M.W.; Methodology: A.M.P., T.M.W.; Formal analysis:
A.M.P.; Investigation: A.M.P., C.T.R., T.M.W.; Data curation: A.M.P., A.M.C., C.T.R.,
T.B., N.W., N.N., A.H., T.M.W.; Writing - original draft: A.M.P.; Writing - review &
editing: A.M.P., A.M.C., C.T.R., M.A.O., T.M.W.; Supervision: C.T.R., M.A.O., A.C.,
T.M.W.; Project administration: C.T.R., M.A.O., T.M.W.; Funding acquisition: A.M.P.,
C.T.R., M.A.O., A.C., T.M.W.
Funding
Support was provided by the U.S. Geological Survey’s Changing Arctic Ecosystems
Initiative, Polar Bears International, the North Pacific Research Board, Interagency
Grizzly Bear Committee, fRI Research, the Raili Korkka Brown Bear Endowment,
the Bear Research and Conservation Endowment, the Nutritional Ecology
1.4 A
B
C
1.2
1.0
0.8
0.6
0.4
0.2
0
1.4
1.2
1.0
Oxygen consumption (ml O2 g–1 h–1)
0.8
0.6
0.4
0.2
0
1.4
1.2
1.0
0.8
0.6
0.4
0.2
0
0 0.1 0.2
DBA z (g)
0.3
0 0.1 0.2
DBA y (g)
0.3
0 0.1 0.2
DBA x (g)
0.3
Fig. 6. Relationship between oxygen consumption and dynamic body
acceleration (DBA) in polar bears and grizzly bears. (A–C) Least-squares
regression of mass-specific oxygen consumption and mean absolute DBA in
the surge (x; A), heave (y; B) and sway (z; C) dimension from two adult female
polar bears (yellow circles, solid line) and three adult female grizzly bears
(black circles, dashed line) resting and walking on a treadmill. Points are mean
steady-state measurements.
1.4 A
B
1.2
1.0
0.8
0.6
0.4
0.2
0
0 0.1 0.2 0.3
ODBA (g)
0.4 0.5 0.6
0 0.1 0.2 0.3 0.4 0.5 0.6
1.4
1.2
1.0
Oxygen consumption (ml O2 g–1 h–1)
0.8
0.6
0.4
0.2
0
Fig. 7. Relationship between oxygen consumption and overall dynamic
body acceleration (ODBA) in polar bears and grizzly bears. (A) Least-
squares regression of mass-specific oxygen consumption and mean ODBA
from two adult female polar bears (polar bears 1 and 2; yellow and orange
circles, respectively) resting and walking on a treadmill. Points are mean
(±s.e.m.) steady-state measurements (see Results for regression statistics).
(B) Least-squares regression of mass-specific oxygen consumption and mean
ODBA from three adult female grizzly bears (grizzly bears 3–5; red, blue and
black circles, respectively) resting and walking on a treadmill. Points are mean
(±s.e.m.) steady-state measurements (see Results for regression statistics).
7
RESEARCH ARTICLE Journal of Experimental Biology (2018) 221, jeb175372. doi:10.1242/jeb.175372
Journal of Experimental Biology

Endowment, Washington State University, San Diego Zoo Global, Oregon Zoo,
SeaWorld and Busch Gardens Conservation Fund, University of California, Santa
Cruz, and the International Association for Bear Research and Management.
Funding was also provided by National Science Foundation DBI 1255913-015 (to
T.M.W.).
Data availability
Data reported in this paper are archived in the USGS Science Data Catalog: https://
doi.org/10.5066/F7QR4W91 and https://doi.org/10.5066/F7XW4H0P.
Supplementary information
Supplementary information available online at
http://jeb.biologists.org/lookup/doi/10.1242/jeb.175372.supplemental
References
Bidder, O. R., Goulding, C., Toledo, A., van Walsum, T. A., Siebert, U. and
Halsey, L. G. (2017). Does the treadmill support valid energetics estimates of field
locomotion? Integr. Comp. Biol. 57, 301-319.
Biewener, A. C. (1989). Scaling body support in mammals: limb postureand muscle
mechanics. Science 245, 45-48.
Breiman, L. (2001). Random forests. Mach. Learn. 45, 5-32.
Brown, J. C. and Yalden, D. W. (1973). The description of mammals-2 Limbs and
locomotion of terrestrial mammals. Mamm. Rev. 3, 107-134.
Bryce, C. M. and Williams, T. M. (2017). Comparative locomotor costs of domestic
dogs reveal energetic economy of wolf-like breeds. J. Exp. Biol. 220, 312-321.
Carrier, D. R. (2016). The fight or flight dichotomy: functional trade-off in
specialization for aggression versus locomotion. In Understanding Mammalian
Locomotion: Concepts and Applications (ed. J. E. A. Bertram), pp. 325-348.
Hoboken, NJ: John Wiley & Sons, Inc.
Cavagna, G. A., Heglund, N. C. and Taylor, C. R. (1977). Mechanical work basic
mechanisms in terrestrial locomotion: two basic mechanisms for minimizing
energy expenditure. J. Physiol. 268, 467-481.
Costa, D. P. and Williams, T. M. (1999). Marine mammal energetics. In Biology of
Marine Mammals (ed. J. E. Reynolds and S. A. Rommel), pp. 176-217.
Washington, DC: Smithsonian Institution Press.
Cunningham, C. B., Schilling, N., Anders, C. and Carrier, D. R. (2010). The
influence of foot posture on the cost of transport in humans. J. Exp. Biol. 213,
790-797.
DeMaster, D. P. and Stirling, I. (1981). Ursus maritimus.Mamm. Species 145, 1-7.
Durner, G. M., Whiteman, J. P., Harlow, H. J., Amstrup, S. C., Regehr, E. V. and
Ben-David, M. (2011). Consequences of long-distance swimming and travel over
deep-water pack ice for a female polar bear during a year of extreme sea ice
retreat. Polar Biol. 34, 975-984.
Durner, G. M., Douglas, D. C., Albeke, S. E., Whiteman, J. P., Ben-david, M.,
Amstrup, S. C., Richardson, E. and Wilson, R. R. (2017). Increased Arctic sea
ice drift alters adult female polar bear movements and energetics. Glob. Chang.
Biol. 23, 3460-3473.
Fedak, M. A., Rome, L. and Seeherman, H. J. (1981). One-step N
2
-dilution
technique for calibrating open-circuit VO
2
measuring systems. J. Appl. Physiol.
51, 772-776.
Fei, Y., Hou, R., Spotila, J. R., Paladino, F. V., Qi, D., Zhang, Z., Zhang, Z., Wildt,
D., Zhang, A., Zhang, H. et al. (2016). Metabolic rates of giant pandas inform
conservation strategies. Sci. Rep. 6, 27248.
Ferguson, S. H., Taylor, M. K., Born, E. W., Rosing-Asvid, A. Messier, F. (1999).
Determinants of home range size for polar bears (Ursus maritimus). Ecol. Lett. 2,
311-318.
Frair, J. L., Fieberg, J., Hebblewhite, M., Cagnacci, F., DeCesare, N. J. and
Pedrotti, L. (2010). Resolving issues of imprecise and habitat-biased locations in
ecological analyses using GPS telemetry data. Philos. Trans. R. Soc. B Biol. Sci.
365, 2187-2200.
Garneau, D. E., Post, E., Boudreau, T., Keech, M. and Valkenburg, P. (2007).
Spatio-temporal patterns of predation among three sympatric predators in a
single-prey system. Wildlife Biol. 13, 186-194.
Gleiss, A. C., Wilson, R. P. and Shepard, E. L. C. (2011). Making overall dynamic
body acceleration work: on the theory of acceleration as a proxy for energy
expenditure. Methods Ecol. Evol. 2, 23-33.
Gómez Laich, A., Wilson, R. P., Gleiss, A. C., Shepard, E. L. C. and Quintana, F.
(2011). Use of overall dynamic body acceleration for estimating energy
expenditure in cormorants. Does locomotion in different media affect
relationships? J. Exp. Mar. Bio. Ecol. 399, 151-155.
Gormezano, L. J., Mcwilliams, S. R., Iles, D. T. and Rockwell, R. F. (2016). Costs
of locomotion in polar bears: when do the costs outweigh the benefits of chasing
down terrestrial prey? Conserv. Physiol. 4, cow045.
Griffen, B. D. (2018). Modeling the metabolic costs of swimming in polar bears
(Ursus maritimus). Polar Biol. 41, 491-503.
Halsey, L. G., Green, J. A., Wilson, R. P. and Frappell, P. B. (2009a).
Accelerometry to estimate energy expenditure during activity: best practice with
data loggers. Physiol. Biochem. Zool. 82, 396-404.
Halsey, L. G., Shepard, E. L. C., Quintana, F., Gómez Laich, A., Green, J. A. and
Wilson, R. P. (2009b). The relationship between oxygen consumption and body
acceleration in a range of species. Comp. Biochem. Physiol. A Mol. Integr.
Physiol. 152, 197-202.
Halsey, L. G., White, C. R., Enstipp, M. R., Wilson, R. P., Butler, P. J., Martin,
G. R., Grémillet, D. and Jones, D. R. (2011). Assessing the validity of the
accelerometry technique for estimating the energy expenditure of diving double-
crested cormorants Phalacrocorax auritus.Physiol. Biochem. Zool. 84, 230-237.
Heglund, N. C. and Taylor, C. R. (1988). Speed, stride frequency and energy cost
per stride: how do they change with bodysize and gait? J. Exp. Biol. 138,301-31 8.
Hurst, R. J. (1981). Thermal and energetic consequences of oil contamination in
polar bears. Masters Thesis, University of Ottawa, Ontario, Canada.
Hurst, R. J., Leonard, M. L., Watts, P. D., Beckerton, P. and Øritsland, N. A.
(1982a). Polar bear locomotion: body temperature and energetic cost.
Can. J. Zool. 60, 40-44.
Hurst, R. J., Øritsland, N. A. and Watts, P. D. (1982b). Body mass, temperature
and cost of walking in polar bears. Acta Physiol. Scand. 115, 391-395.
Hurst, R. J., Watts, P. D. and Øritsland, N. A. (1991). Metabolic compensation in
oil-exposed polar bears. J. Therm. Biol. 16, 53-56.
Iles, D. T., Peterson, S. L., Gormezano, L. J., Koons, D. N. and Rockwell, R. F.
(2013). Terrestrial predation by polar bears: not just a wild goose chase. Polar Biol.
36, 1373-1379.
Iwaniuk, A. N., Pellis, S. M. and Whishaw, I. Q. (2000). The relative importance of
body size, phylogeny, locomotion, and diet in the evolution of forelimb dexterity in
fissiped carnivores (Carnivora). Can. J. Zool. 78, 1110-1125.
Johnson, D. S., London, J. M., Lea, M.-A. and Durban, J. W. (2008). Continuous-
time correlated random walk model for animal telemetry data. Ecology 89,
1208-1215.
Kleiber, M. (1975). The Fire of Life: An Introduction to Animal Energetics, p. 454.
New York, NY: John Wiley & Sons, Inc.
Langman, V. A., Rowe, M. F., Roberts, T. J., Langman, N. V. and Taylor, C. R.
(2012). Minimum cost of transport in Asian elephants: do we really need a bigger
elephant? J. Exp. Biol. 215, 1509-1514.
Lovegrove, B. G. and Haines, L. (2004). The evolution of placental mammal body
sizes: evolutionary history, form, and function. Oecologia 138, 13-27.
McLellan, B. and Reiner, D. C. (1994). A review of bear evolution. Int. Conf. Bear
Res. Manag. 9, 85-96.
McLoughlin, P. D. and Ferguson, S. H. (2000). A hierarchical pattern of limiting
factors helps explain variation in home range size. Ecoscience 7, 123-130.
McLoughlin, P. D., Case, R. L., Gau, R. J., Ferguson, S. H. and Messier, F.
(1999). Annual and seasonal movement patterns of barren-groundgrizzly bears in
the central Northwest Territories. Ursus 11, 79-86.
McNab, B. K. (1986). The influence of food habits on the energetics of eutherian
mammals. Ecol. Monogr. 56, 1-19.
McNab, B. K. (1992). Rate of metabolism in the termite-eating sloth bear (Ursus
ursinus). J. Mammal. 73, 168-172.
Nelson, R. A., Folk, G. E., Jr, Pfeiffer, E. W., Craighead, J. J., Jonkel, C. J. and
Steiger, D. L. (1983). Behavior, biochemistry, and hibernation in black, grizzly,
and polar bears. Int. Conf. Bear Res. Manag. 5, 284-290.
Øritsland, N. A., Jonkel, C. and Ronald, K. (1976). A respiration chamber for
exercising polar bears. Nor. J. Zool. 24, 65-67.
Pagano, A. M., Durner, G. M., Amstrup, S. C., Simac, K. S. and York, G. S. (2012).
Long-distance swimming by polar bears (Ursus maritimus) of the southern
Beaufort Sea during years of extensive open water. Can. J. Zool. 90, 663-676.
Pagano, A. M., Rode, K. D., Cutting, A., Owen, M. A., Jensen, S., Ware, J. V.,
Robbins, C. T., Durner, G. M., Atwood, T. C., Obbard, M. E. et al. (2017). Using
tri-axial accelerometers to identify wild polar bear behaviors. Endanger. Species
Res. 32, 19-33.
Pagano, A. M., Durner, G. M., Rode, K. D., Atwood, T. C., Atkinson, S. N.,
Peacock, E., Costa, D. P., Owen, M. A. and Williams, T. M. (2018). High-energy,
high-fat lifestyle challenges an Arctic apex predator, the polar bear. Science 359,
568-572.
Pilfold, N. W., Mccall, A., Derocher, A. E., Lunn, N. J. and Richardson, E. (2017).
Migratory response of polar bears to sea ice loss: to swim or not to swim.
Ecography 40, 189-199.
Reilly, S. M., McElroy, E. J. and Biknevicius, A. R. (2007). Posture, gait and the
ecological relevance of locomotor costs and energy-saving mechanisms in
tetrapods. Zoology 110, 271-289.
Renous, S., Gasc, J.-P. and Abourachid, A. (1988). Kinematic analysis of the
locomotion of the polar bear (Ursus maritimus, Phipps, 1774) in natural and
experimental conditions. Netherlands J. Zool. 48, 145-167.
Schmidt-Nielsen, K. (1972). Locomotion: energy cost of swimming, flying, and
running. Science 177, 222-228.
Shepard, E. L. C., Wilson, R. P., Quintana, F., Gómez Laich, A., Liebsch, N.,
Albareda, D. A., Halsey, L. G., Gleiss, A., Morgan, D. T., Myers, A. E. et al.
(2008). Identification of animal movement patterns using tri-axial -accelerometry.
Endanger. Species Res. 10, 47-60.
Shepard, E. L. C., Wilson, R. P., Rees, W. G., Grundy, E., Lambertucci, S. A. and
Vosper, S. B. (2013). Energy landscapes shape animal movement ecology. Am.
Nat. 182, 298-312.
8
RESEARCH ARTICLE Journal of Experimental Biology (2018) 221, jeb175372. doi:10.1242/jeb.175372
Journal of Experimental Biology

Shine, C. L., Penberthy, S., Robbins, C. T., Nelson, O. L. and McGowan, C. P.
(2015). Grizzly bear (Ursus arctos horribilis) locomotion: gaits and ground reaction
forces. J. Exp. Biol. 218, 3102-3109.
Stirling, I. (1974). Midsummer observations on the behavior of wild polar bears
(Ursus maritimus). Can. J. Zool. 52, 1191-1198.
Stirling, I. and Derocher, A. E. (1990). Factors affecting the evolution and
behavioral ecology of the modern bears. Int. Conf. Bear Res. Manag. 8, 189-204.
Stirling, I., Spencer, C. and Andriashek, D. (1989). Immobilization of polar bears
(Ursus maritimus) with Telazol
®
in the Canadian Arctic. J. Wildl. Dis. 25, 159-168.
Talbot, S. L. and Shields, G. F. (1996). A phylogeny of the bears (Ursidae) inferred
from complete sequences of three mitochondrial genes. Mol. Phylogenet. Evol. 5,
567-575.
Taylor, C. R., Schmidt-Nielsen, K. and Raab, J. L. (1970). Scaling of energetic
cost of running to body size in mammals. Am. J. Physiol. 219, 1104-1107.
Taylor, C. R., Heglund, N. C. and Maloiy, G. M. (1982). Energetics and mechanics
of terrestrial locomotion. I. Metabolic energy consumption as a function of speed
and body size in birds and mammals. J. Exp. Biol. 97, 1-21.
Tøien, Ø., Blake, J., Edgar, D. M., Grahn, D. A., Heller, H. C. and Barnes, B. M.
(2011). Hibernation in black bears: independence of metabolic suppression from
body temperature. Science 331, 906-909.
Watts, P. and Cuyler, C. (1988). Metabolism of the black bear under simulated
denning conditions. Acta Physiol. Scand. 134, 149-152.
Watts, P. D. and Jonkel, C. (1988). Energetic cost of winter dormancy in grizzly
bear. J. Wildl. Manage. 52, 654-656.
Watts, P. D., Øritsland, N. A. and Hurst, R. J. (1987). Standard metabolic rate of
polar bears under simulated denning conditions. Physiol. Zool. 60, 687-691.
Watts, P. D., Ferguson, K. L. and Draper, B. A. (1991). Energetic output of
subadult polar bears (Ursus maritimus): resting, disturbance and locomotion.
Comp. Biochem. Physiol. A Physiol. 98, 191-193.
Whiteman, J. P., Harlow, H. J., Durner, G. M., Anderson-Sprecher, R., Albeke,
S. E., Regehr, E. V., Amstrup, S. C. and Ben-David, M. (2015). Summer
declines in activity and body temperature offer polar bears limited energy savings.
Science 349, 295-298.
Williams, T. M. (1999). The evolution of cost efficient swimming in marine mammals:
limits to energetic optimization. Philos. Trans. R. Soc. Lond. Ser. B Biol. Sci. 354,
193-201.
Williams, T. M., Ben-David, M., Noren, S., Rutishauser, M., McDonald, K. and
Heyward, W. (2002). Running energetics of the North American river otter: do
short legs necessarily reduce efficiency on land? Comp. Biochem. Physiol. A Mol.
Integr. Physiol. 133, 203-212.
Williams, T. M., Wolfe, L., Davis, T., Kendall, T., Richter, B., Wang, Y., Bryce, C.,
Elkaim, G. H. and Wilmers, C. C. (2014). Instantaneous energetics of puma kills
reveal advantage of felid sneak attacks. Science 346, 81-85.
Wilson, R. P., White, C. R., Quintana, F., Halsey, L. G., Liebsch, N., Martin, G. R.
and Butler, P. J. (2006). Moving towards acceleration for estimates of activity-
specific metabolic rate in free-living animals: the case of the cormorant. J. Anim.
Ecol. 75, 1081-1090.
Wilson, R. P., Quintana, F. and Hobson, V. J. (2012). Construction of energy
landscapes can clarify the movement and distribution of foraging animals.
Proc. R. Soc. B Biol. Sci. 279, 975-980.
Withers, P. C. (1977). Measurement of VO
2
,VCO
2
, and evaporative water loss with
a flow-through mask. J. Appl. Physiol. 42, 120-123.
9
RESEARCH ARTICLE Journal of Experimental Biology (2018) 221, jeb175372. doi:10.1242/jeb.175372
Journal of Experimental Biology

Movie 1. High-speed video (120 frames s−1) showing the plantigrade
walking gait of an adult female polar bear walking on a treadmill at 2
km h-1.
Movie 2. High-speed video (120 frames s−1) showing the plantigrade
walking gait of an adult female grizzly bear walking on a treadmill
at 2.8 km h-1.
Journal of Experimental Biology 221: doi:10.1242/jeb.175372: Supplementary information
Journal of Experimental Biology • Supplementary information