Exercise in an aquatic environment may be an effective mode of therapy andtraining due to reduced impact forces. The purpose of this study was to comparethe physiological responses of walking/running on a land treadmill withwater treadmill responses at two different depths. Six subjects completedwalking and running trials on both a land-based and a water-based treadmill.Water-based trials were completed in both thigh- and waist-deep water.Each trial was five minutes in duration. Oxygen uptake (VO2), heart rate (HR),respiratory exchange ratio (RER), stride frequency (SF), and the oxygen costper stride (VO2/stride) were compared between the conditions using a twowayANOVA with repeated measures. Walking and running in water elevatedVO2 (p < 0.02) and HR (p < 0.04) above land treadmill values. When runningin waist-deep water, VO2 and HR failed to increase to the same extent asthigh-deep running. Stride frequency was similar between the three differentdepths during walking but lower in waist-deep water during running. VO2/stride was significantly higher (p < 0.01) in water-based walking and runningcompared to land-based values. Water-based walking and running eliciteda greater physiological cost than land-based exercise, which can be attributedto the elevated cost of moving in water due to increased resistance.When running in waist-deep water, buoyancy may counter the resistance ofthe water and serve to lower the physiological cost of locomotion.
Responses to Running and Walking in Water 63
63Research in Sports Medicine 11: 6378, 2003
Copyright Taylor & Francis Inc.
ISSN:1543-8627 print 1543-8635 online
The Physiological Responses to Running
and Walking in Water at Different Depths
MICHAEL B. POHL
University of Bath
LARS R. MCNAUGHTON
University of Hull
Exercise in an aquatic environment may be an effective mode of therapy and
training due to reduced impact forces. The purpose of this study was to com-
pare the physiological responses of walkingrunning on a land treadmill with
water treadmill responses at two different depths. Six subjects completed
walking and running trials on both a land-based and a water-based tread-
mill. Water-based trials were completed in both thigh- and waist-deep water.
Each trial was five minutes in duration. Oxygen uptake (VO
2), heart rate (HR),
respiratory exchange ratio (RER), stride frequency (SF), and the oxygen cost
per stride (VO
2stride) were compared between the conditions using a two-
way ANOVA with repeated measures. Walking and running in water elevated
2 (p 0.02) and HR (p 0.04) above land treadmill values. When running
in waist-deep water, VO
2 and HR failed to increase to the same extent as
thigh-deep running. Stride frequency was similar between the three different
depths during walking but lower in waist-deep water during running. VO
stride was significantly higher (p 0.01) in water-based walking and run-
ning compared to land-based values. Water-based walking and running elic-
ited a greater physiological cost than land-based exercise, which can be at-
tributed to the elevated cost of moving in water due to increased resistance.
When running in waist-deep water, buoyancy may counter the resistance of
the water and serve to lower the physiological cost of locomotion.
Keywords aquaciser, energy cost, buoyancy, immersion, oxygen consump-
Received 1 October 2002; accepted 10 February 2003.
The authors would like to thank the following people: Jim Grant for the use of the
Aquaciser at the Royal United Hospital; Katie Buck and Jane Hall for sharing their experi-
ences of working in this area of research and Simon Roberts for his technical assistance.
Address correspondence to Lars R. McNaughton, Department of Sport Science, Uni-
versity of Hull. Kingston upon Hull, HU6 7RX, United Kingdom. Email:
64 M. B. Pohl and L. R. McNaughton
It has been well recognized that exercise in water can be an effective and useful
mode of therapeutic exercise (Bishop, Frazier, Smith et al. 1989; DAcquisto,
DAcquisto, and Renne 2001; Dowzer, Reilly, and Cable 1998; Nakazawa, Yano,
and Miyashita 1994; Napoletan, Janes, and Hicks 1991). Patients suffering from
injury or dysfunction of the lower extremities often experience difficulty with the
weight-bearing components of land exercise. This led to the adoption of deep-
water (DW) and shallow-water (SW) walking and running by many therapists, as
the buoyancy provided by the water served to reduce the impact forces, and hence
pain, on the lower limbs (Hall, Skevington, Maddison et al. 1996).
When an athlete has to abstain from training, decrements occur in parameters
that determine aerobic fitness (Coyle, Martin, Sincamore et al. 1984). Participa-
tion in endurance exercises such as water running may help to minimize perfor-
mance decrements in injured athletes (Bushman, Flynn, Andres et al. 1996; Wilber,
Moffatt, Scott et al. 1996). It has also been demonstrated that water running may
be beneficial in enhancing cardiovascular fitness for untrained subjects (Davidson
and McNaughton 2000; Michaud, Brennan, Wilder et al. 1995; Morrow, Jensen,
and Peace 1996).
The initial research in this area was conducted using DW running (Bishop,
Frazier, Smith et al. 1989; Frangolias and Rhodes 1996; Gehring, Keller, and Brehm
1996; Mercer and Jensen, 1998; Svedenhag and Seger 1992), which is performed
in the deep end of a swimming pool and simulates the movement pattern used for
land running. It has been recognized that DW running results in altered technique
when compared to running on land (Dowzer and Reilly 1998; Svedenhag and
Seger 1992). This led to the introduction of SW exercise into rehabilitation and
conditioning programs (Dowzer, Reilly, Cable et al. 1999; Evans, Cureton, and
Purvis 1978; Town and Bradley 1991), as it more closely resembled running on
land. However, SW running and walking can produce a postural distortion due to
frontal resistance (Byrne, Craig, and Willmore 1996). Underwater treadmills en-
able subjects to walk in a more normal ambulatory posture due to diminished
frontal resistance. This is especially important for patients undergoing rehabilita-
tion in which the aim is a return to correct functional gait patterns (Hall, Macdonald,
Maddison et al. 1998).
When the individual is at rest and immersed in water up to the neck, a 30%
35% increase in CO has been reported (Dowzer and Reilly 1998). This has been
attributed to an increase in SV (Christie, Sheldahl, and Tristani 1990; Farhi and
Linnarsson 1977) and is believed to be the result of the increased central blood
volume caused by the increased pressure on the thoracic cavity and abdomen by
the water (Christie, Sheldahl, and Tristani 1990). There are contrasting results
between researchers as to the effect of water immersion on resting heart rate (HR).
Derion, Guy, Tsukimoto et al. (1992) and Christie, Sheldahl, and Tristani (1990)
found no significant difference in HR values for land and water-based conditions.
Responses to Running and Walking in Water 65
Farhi and Linnarsson (1977), on the other hand, found that water immersion be-
tween the level of the hips and the neck decreased HR by as much as eight beats
per minute. They speculated that the level of immersion was associated with changes
in cardiovascular parameters, reporting that SV was higher in water compared to
land values and that SV increased with an increased depth of immersion. Thus it
appears that walking and running in water at different depths may have important
implications for cardiovascular training.
Deep-water running has been reported to elicit lower
O V&2max and maximal
max) values compared to running on a treadmill on land (Bishop, Frazier,
Smith et al. 1989; Frangolias and Rhodes 1996; Michaud, Brennan, Wilder et al.
1995; Svedenhag and Seger 1992). Differences in muscle recruitment patterns
and redistribution of central blood flow have been implicated in these altered re-
sponses. It is believed that confounding factors such as water temperature and the
fitness of the subjects may influence HR and
O V&2 relationships (Buck,
McNaughton, Sherman et al. 2001; Dowzer and Reilly 1998; McArdle, Magel,
Lesmes et al. 1976). During maximal exercise, SW running at waist level has been
illustrated to better simulate land-based running physiological responses in trained
distance runners (Dowzer, Reilly, Cable et al. 1999; Town and Bradley 1991).
Few studies have been conducted to examine the effect of water immersion
on submaximal exercise, and even fewer have investigated the responses on un-
derwater treadmills. The present study compared the physiological responses of
conventional land-based treadmill exercise and underwater treadmill exercise at
two different depths of water. The relationship between the physiological responses
of running and walking also was compared to see if the selected ambulatory mode
had any influence.
Six students(mean SD) age 23.2 2.9 yr, height 179.5 9.9cm, and weight
66.3 11.3 kgvolunteered to participate in this study. Subjects were recruited
on the basis that they were recreationally active and free from any known disease
or orthopaedic dysfunction. In addition, the participants were required to com-
plete a preexercise medical screening questionnaire. Prior to data collection, the
study received ethical approval from the departmental ethics committee. All as-
pects of the procedures, risks, and benefits of the study were explained to the
participants before they provided written informed consent.
A Powerjog JX200 treadmill (Sport Engineering Ltd) was used for land-based
testing of the physiological responses of the subjects. The exercise bouts in water
66 M. B. Pohl and L. R. McNaughton
were performed on an AquaCiser II underwater treadmill (AquaCiser, Ferno UK
Ltd), which allowed runningwalking in varied depths and at different speeds.
Metabolic data, including
O V&2 and respiratory exchange ratio (RER), were col-
lected using a portable metabolic system (Cosmed k4b2 SRL, Italy). Heart rate
also was acquired using the k4b2 Cosmed by attaching a Polar Vantage NV (Polar
Electroy, Kempele, Finland) around the chest of the subject.
All subjects were unfamiliar with underwater treadmill runningwalking and there-
fore each was given a 15-minute practice session prior to the day of testing. After
familiarization, all subjects completed the treadmill tests in both a land-based and
a water-based environment. The participants were asked to abstain from heavy
exercise 48 hours prior to any testing and to refrain from caffeine on the actual day
of testing. The water-based treadmill speed was calibrated to eliminate any dis-
crepancies with the land-based treadmill.
Subjects were required to complete six tests for running and walking in three
different conditions: land, thigh-deep water, and waist-deep water. Thigh-deep
was defined as the point midway between the anterior superior iliac spine and the
central patella, and waist-deep was defined as the level of the umbilicus. Subjects
performed a walking test for 5 minutes at 4.0kmh and a running test for 5 minutes
for each environmental condition. Subjects were instructed to maintain
a reciprocal arm action for runningwalking in all tests. This was defined as having
the elbows flexed to 90 with the forearms in the mid-prone position (Hall,
Macdonald, Maddison et al. 1998).Water temperature was adjusted to 33C prior
to every water-based test.
Prior to each exercise test, the Cosmed was calibrated using gases of known
2 16.0%, CO2 5.0%) as well as room air. After being fitted with
the Cosmed k4b2, the subject was asked to step onto the treadmill and, in the case
of water-based exercise, the required depth of water was added to the chamber.
Subjects then walked or ran for 4 minutes at the required speed (to obtain steady-
state exercise values (Byrne, Craig, and Willmore 1996)) before metabolic and
cardiovascular measurements were collected. Oxygen uptake, HR, and RER were
then collected in the fifth minute of exercise. Stride frequency (SF) was measured
for 30 seconds in the fourth minute, which was then doubled.
All analyses were performed using the Statistical Package for Social Sciences
(SPSS, version 8.0; Chicago, Illinois, U.S.). Submaximal values of O V&2, HR, RER,
O V&2 per stride (the O V&2 per stride was found by simply dividing the
O V&2 by the SF) were analyzed using separate two-way (ambulatory
mode x level of immersion) analyses of variance with repeated measures. Statisti-
Responses to Running and Walking in Water 67
cal significance was set at p
0.05. Where a significant main effect was observed,
post-hoc pair-wise comparisons were calculated using the Bonferroni method.
The two-way ANOVA with repeated measures revealed significant effects for
ambulatory mode and depth. Values for
O V&2 were found to be significantly higher
during running when compared to walking (p
0.001). The highest O V&2 was
associated with thigh-deep exercise, which was significantly higher than waist-
0.01) and land-based values (p 0.01). Waist-deep O V&2 was also sig-
nificantly higher than the land-based trials (p
The relationship between the ambulatory mode and depth, although not sta-
tistically significant (p = 0.054), still displayed a marked interaction effect (Figure
1). It can be seen that both land and waist-deep trials respond in a similar manner
to the ambulatory mode increasing by 13.80 and 13.00 mlminkg, respectively
(Tables 1 and 2). In contrast,
O V&2 increased by 19.23 mlminkg when progress-
ing from walking to running in thigh-deep water.
The two-way ANOVA with repeated measures revealed significant effects for
ambulatory mode, depth, and interaction between the ambulatory mode and depth.
Ambulatory ModeRun Walk
Estimated Marginal Means for VO2 (mlminkg)
Figure 1. Interaction between the ambulatory mode and the level of immersion for
68 M. B. Pohl and L. R. McNaughton
Stride frequency was higher for running compared to walking (p
0.001) for all
The ANOVA revealed that waist-deep exercise had a significantly lower SF
than thigh-deep (p
0.001) and land-based (p 0.003) tests. The ANOVA com-
bines walking and running values, and inspection of Figure 2 indicates that the
main discrepancy in SF occurs during running. Land-based and thigh-deep trials
illustrate a 48% and 49% increase in SF, respectively, between walking and run-
ning (Tables 1 and 2). The increase in SF from walking to running was only 33%
in the waist-deep exercise.Table 2
Comparison of Physiological Variables of Interest
During Running at Three Different Depths
Land Thigh-deep Waist-deep
O V&2 (mlminkg) 23.64 10.84 39.39 18.11 30.48 14.54
HR (beatsmin) 124.00 17.00162.00 10.00130.00 14.00
RER 0.88 10.06 0.85 10.07 0.85 10.04
SF (stridesmin) 149.00 12.00143.00 11.00122.00 12.00
O V&2stride (mlminkg) 0.16 10.03 0.28 10.06 0.25 10.05
Note: Values are expressed as the mean standard deviation.
Comparison of Physiological Variables of Interest
During Walking at Three Different Depths
Land Thigh-deep Waist-deep
O V&2 (mlminkg) 9.84 10.84 20.16 2.32 17.48 12.47
HR (beatsmin) 78.00 10.00104.00 5.0096.00 15.00
RER 0.85 10.07 0.79 0.05 0.81 10.03
SF (stridesmin) 101.00 16.0096.00 7.0092.00 10.00
O V&2stride (mlminkg) 0.10 10.01 0.21 0.04 0.19 10.04
Note: Values are expressed as the mean standard deviation
Responses to Running and Walking in Water 69
Oxygen Cost Per Stride
Running produced an elevated
O V&2 per stride relative to walking (p 0.001).
Depth also appeared to have an effect. Land-based trials resulted in a significantly
O V&2 per stride in relation to both thigh-deep (p 0.01) and waist-deep (p
0.01) trials. There was no difference between the thigh-deep and waist-deep con-
ditions during either walking or running.
The oxygen cost per stride was not affected by an interaction between the
ambulatory mode and level of immersion (Figure 3). The increase from walking to
running was similar for the three different depth treatments.
Significant effects on HR were observed for ambulatory mode, depth, and the
interaction between the ambulatory mode and depth. Heart rate was elevated dur-
ing running compared to walking (p
0.001) for all depths. Comparing depth
responses revealed that thigh-deep exercise elicited a higher HR response than
0.01) and land-based (p 0.001) exercise. Additionally, waist-
deep HR was higher than land-based HR (p
During walking, HR values for waist-deep and thigh-deep water were similar,
with mean values of 96 5 and 104 5 bpm, respectively. Land-based HR values
were much lower, with a mean value of 78 10 bpm (see Tables 1 and 2). Figure
4 indicates that the HR increase from walking to running is lower for the waist-
Ambulatory ModeRun Walk
Estimated Marginal Means for SF (stridesmin)
Figure 2. Interaction between the ambulatory mode and level of immersion for stride
70 M. B. Pohl and L. R. McNaughton
deep condition than thigh-deep and land-based responses. Waist-deep HR increased
only 35%, compared to 56% and 58% during thigh-deep and land-based trials,
respectively (Tables 1 and 2).
Respiratory Exchange Ratio
The only statistically significant effect on RER was that of the ambulatory mode.
The RER was elevated during running in comparison to walking (p
though not significantly higher, RER was found to be highest during both walking
(Table 1) and running (Table 2) on land. Thigh-deep walking and running dis-
played the lowest RER values.
No significant interaction was found between the ambulatory mode and depth.
Figure 5, however, suggests that thigh-deep immersion produced a relatively larg-
er increase in RER between walking and running. It should be noted that the
standard deviations for land-based and thigh-deep trials were large (Tables 1 and
O V&2 during running compared to walking was true for both water
and land tests, as
O V&2 is proportional to the intensity of exercise. When the in-
Ambulatory ModeRun Walk
Estimated Marginal Means for VO2stride (mlminkg)
Figure 3. Interaction between the ambulatory mode and level of immersion for oxygen
uptake per stride.
Responses to Running and Walking in Water 71
tensity of exercise is increased, the muscles involved have to perform more me-
chanical work, so more oxygen must be both supplied to and taken up by the
Both thigh- and waist-deep levels expressed higher values for walking and
running than the respective land-based values. This could be attributed to the added
resistance imposed by the water on the body during aquatic locomotion, as water
is approximately 800 times more dense than air (Dowzer and Reilly 1998). Thigh-
O V&2 during running, however, also was substantially higher than that of
waist-deep running, which is in keeping with the findings of Gleim and Nicholas
(1989). They postulate that the most likely explanation is that the flight phase
during running is prolonged due to the buoyancy of the human body, which would
allow the subject to float momentarily while the belt runs underneath, thus reduc-
ing the workload. During thigh-deep running, it could be postulated that not enough
of the body is immersed to benefit from the lift forces provided by buoyancy.
Napoletan and Hicks (1995) found that chest-deep running produced a
O V&2 that
was 13.6 mlminkg lower than thigh-deep running at a speed of only 5.5 kmh.
This is a much larger difference in
O V&2 than was found between the thigh- and
waist-deep values in this study. This could be due to buoyancy being increased
further at chest level, serving to decrease the metabolic cost even more.
The enhanced buoyancy would not be of as much significance during walk-
ing, as the ambulatory mode lacks a flight phase, so the subject must match the
speed of the belt more precisely. This is evidenced in the results, as
O V&2 for thigh-
deep walking is only 2.68 mlminkg higher than waist-deep values. Once again,
Ambulatory ModeRun Walk
Estimated Marginal Means for HR (beatsmin)
Figure 4. Interaction between the ambulatory mode and level of immersion for heart rate.
72 M. B. Pohl and L. R. McNaughton
this is similar to the study conducted by Gleim and Nicholas (1989), in which
thigh-deep walking was 2.3 mlminkg higher than waist depth. If buoyancy was
not acting during walking, it might be expected that
O V&2 would be higher during
the waist-deep condition because a greater proportion of the body is immersed and
must overcome the greater resistance of the water. It would, therefore, be incorrect
to state that buoyancy is not a factor during walking. The lowered metabolic re-
sponses at waist depth found during running are simply due to the increased influ-
ence of buoyancy.
A significant finding of this study was that the waist-deep condition resulted in a
lower SF than both the thigh-depth and land-based conditions. The ANOVA, how-
ever, combined the walking and running values for the depth conditions and then
compared them. Closer inspection of the interaction revealed that SF for the waist-
deep condition was significantly influenced by the ambulatory mode. When sub-
jects are walking, SF is similar for land-based, thigh-deep, and waist-deep condi-
tions. During running, however, although the land-based and thigh-deep SF re-
main similar, the SF for the waist-deep condition is more than 20 stridesmin lower.
Thus, it appears that the difference for the waist-deep condition can be attributed
mainly to the discrepancy of SF during running.
The findings of this study are in contrast to those of Hall, Macdonald, Maddison
et al. (1998), who discovered a significant difference in SF between water- and
Figure 5. Interaction between the ambulatory mode and level of immersion for RER.
Ambulatory ModeRun Walk
Estimated Marginal Means for RER
Responses to Running and Walking in Water 73
land-based walking. The authors found that at all walking speeds, SF in water was
27 stridesmin lower than on land. A possible explanation for the contrasting find-
ings is that the study by Hall, Macdonald, Maddison et al. (1998) immersed the
subjects to the level of the xiphoid. It has been postulated previously that an in-
creased level of water immersion may serve to enhance the lift forces due to buoy-
ancy. In turn, this might increase the duration of the gait cycle.
The buoyancy theory also would be of use in explaining the discrepancy found
for SF during waist-deep running. It could be argued that the decreased SF was
due to a prolonged flight phase in the gait cycle, which would serve to increase the
time to perform one stride. Thigh-deep SF during running has to match the land-
based value, because little buoyancy is provided by the water at this depth. Thus,
the lower SF found during waist-deep running provides additional evidence that
the lowered metabolic cost for this condition may be due to buoyancy.
Stride frequency alone, however, cannot be used to explain the lowered
O V&2cost during waist-deep running. It could be plausible that the decreased SF was, in
fact, due to the higher resistance of the water acting on a greater proportion of the
body. The lowered
O V&2 response could be explained by decreased work required
between the body and the ground during the support phase of gait. In order to
determine which factor is more important, the
O V&2 cost per stride also needs to be
taken into account.
Oxygen Cost Per Stride
The oxygen cost per stride appears to be lower in land-based walking and running
compared to the respective values at both different water depths. This elevated
cost can be attributed to water viscosity and the increased resistance to locomo-
tion. Hall, Macdonald, Maddison et al. (1998) reported
O V&2 costs per stride of
6.73 and 10.48 mlmin for land and water walking (4.5kmh), respectively. If the
results from the present study are converted to similar units, comparative values of
6.5 and 12.9 mlmin are displayed for land and water. The lower value for water
walking found by Hall, Macdonald, Maddison et al. (1998) could be due to the
deeper level of water immersion. The resulting increase in buoyancy would lower
the work required during the support phases of the gait cycle.
If the increasing water immersion did serve to decrease the work intensity, it
would be expected that the
O V&2 per stride would be lower in waist-deep water
exercise compared to thigh-deep exercise. The results indicate that this is not the
case, and that the two water depths illustrate similar costs for both walking and
running. The explanation may lie in the interaction between water resistance and
buoyancy. As the level of immersion increases from thigh depth to waist depth,
the effects of buoyancy would become more pronounced. At the same time, the
increased depth also would mean that a greater proportion of the body would have
to overcome the resistance of the water. Therefore, the work-reducing effect of the
buoyancy may be countered by the increased water resistance.
74 M. B. Pohl and L. R. McNaughton
O V&2 cost per stride is found for both thigh-deep and waist-deep
running in water. This seems to imply that the
O V&2 cost per stride cannot explain
O V&2 response in waist-deep running compared to thigh-deep values.
Thus, it appears that the lowered metabolic response during waist-deep running is
simply due to the fact that subjects take fewer strides.
When the subjects walked, HR was considerably higher in the water conditions
when compared to the land-based HR value (78 beatsmin). Walking in thigh-
deep water (104 beatsmin) elicited a slightly higher HR than waist-deep water (96
beatsmin), which was exactly the case with
O V&2. It could be suggested that the
HR for all three conditions simply represented the metabolic cost of the walking.
The results of this study are comparable with those of Gleim and Nicholas (1989),
who found HR values of 81, 105, and 98 beatsmin for land-based, thigh-deep,
and waist-deep conditions, respectively. It appears that while walking at different
levels of water immersion, HR is simply a function of exercise intensity and meta-
Thigh-deep running elicited the highest HR response during running, followed
by waist-deep and then land-based conditions. The significant interaction between
depth and ambulatory mode, however, implies that HR trends are different during
running and walking. The main point is that HR for waist-deep water running has
increased only 35% from the value recorded during walking, which is low com-
pared with the 56% and 58% increases found in the thigh-deep and land-based
There is no clear explanation for this dampened HR response, but one theory
is that the phenomenon is a result of an interaction between the baroreceptor and
Bainbridge reflexes. If the level of immersion is sufficient to increase the hydro-
static pressure on the thoracic cavity, a redistribution of blood centrally can be
expected. Thus, the resulting increase in stroke volume would prompt a decrease
in HR via the baroreceptor reflex. It is possible that during low-moderate intensity
exercise the increased atrial pressure acts to offset the bradycardia. Another pos-
sible explanation is that water immersion may affect the autonomic nervous sys-
tem (Christie, Sheldahl, and Tristani 1990). During exercise, HR is controlled by
both divisions of the autonomic nervous system and is elevated by simultaneously
increasing sympathetic and decreasing sympathetic activity. The initial increase
in heart rate (up to 100 beatsmin) during exercise is due to parasympathetic neu-
ral withdrawal, whereas sympathetic neural outflow should have a greater impact
on HR at higher work rates (Powers and Howley 2001). This would imply that HR
while walking in waist-deep water was mainly controlled by parasympathetic with-
drawal, but running was controlled by sympathetic stimulation. It has been sug-
gested that sympathetic neural outflow is reduced in water (Christie, Sheldahl, and
Tristani 1990), which would imply that HR during running might be lower than
Responses to Running and Walking in Water 75
expected. The values of HR for walking are less affected because this condition
would rely less on sympathetic stimulation. The increase in HR between walking
and running in thigh-deep water was not depressed, which may imply that the
level of immersion was not sufficient to cause a decreased sympathetic response.
Respiratory Exchange Ratio
The only clear effect on RER was that it was significantly elevated during running
in comparison to walking. This effect was evident irrespective of the level of wa-
ter immersion. This can be explained by the fact that low-intensity exercise relies
primarily on fat as fuel, but as the intensity increases, muscles increasingly rely on
carbohydrate sources of energy.
During walking and running, the land-based condition demonstrates the high-
est values of RER, but also the lowest
O V&2 and HR values. This implies that
despite a higher exercise intensity required to move in water, RER is lower than
land responses. It would be expected, however, that the increased cost of moving
in water would result in a higher RER. This was illustrated by Hall, Macdonald,
Maddison et al. (1998), who demonstrated RER values of 0.94 and 0.89 when
walking in water and land, respectively. This could be due to calculations of RER,
assuming that the bodys CO
2 exchange in the lung is proportional to its release
from the cells. Body CO
2 pools are quite large, however, and can be altered by
breathing patterns. Under these conditions, the CO
2 released in the lung may not
actually represent that being produced in the tissues. This would bias the ratio of
2 to O V&2 and invalidate the use of RER to estimate fuel utilization. Another
likely explanation for the confounding RER values is the large standard deviation
values associated with the means. No significant differences were found between
land-based, thigh-deep, and waist-deep conditions.
Application of Findings
Despite the assumed reduced impact forces during water exercise, running in thigh-
deep and waist-deep water still seems to induce a physiological response that could
stimulate a training response. A subject who is utilizing water running for reha-
bilitation purposes probably would begin a program using a relatively high water
depth in order to take full advantage of the reduced impact forces. Because the
vertical ground reaction forces increase as the level of immersion in reduced
(Nakazawa, Yano, and Miyashita 1994), it might be expected that as the injury
improves a progressively lower immersion level would be used to increase load-
ing on the injured joint. The implications of this study are that as the level of
immersion is decreased, the workload must be adjusted accordingly, as exercise is
more intense in more shallow water.
Additionally, water running can be used by injury-free subjects as a training
tool, as the reduced impact may decrease the susceptibility of the lower extremi-
76 M. B. Pohl and L. R. McNaughton
ties to injury and could be used in conjunction with on-land training. To enable the
subject to train at a similar intensity to land-based running, the relationship be-
tween water and land running must be known. The present study, however, used
only one speed to represent walking and running. This may be inaccurate, as the
variation of speed within the ambulatory mode might also influence physiological
responses. This was highlighted by Gleim and Nicholas (1989), who found that at
higher running speeds ( 8 kmh) waist-deep
O V&2 was similar to land-based val-
ues. Further investigation is required into the interaction of depth and ambulatory
mode at more speeds.
Because the mechanism for the lowered HR response during waist-deep run-
ning is currently unclear, further investigation into cardiovascular responses to
water immersion per se is needed. The contribution of baroreceptor and Bainbridge
responses must be determined as a function of both exercise intensity and immer-
sion per se. Additionally, the effect of water immersion on the sympathetic ner-
vous system needs investigation. The level of immersion required to influence this
mechanism also needs to be established.
Finally, the justification for this study was provided by the theory that water
exercise provided a cardiovascular workout without the high-impact forces present
in land exercise. Some subjects, however, commented that they felt greater pres-
sure on the joints (especially the knee) during water walking and running. It should
be noted that although Nakazawa, Yano, and Miyashita (1994) demonstrated de-
creasing vertical ground reaction forces with increasing water immersion, this does
not guarantee decreased internal joint forces in the lower extremity. The increased
resistance of water is an additional external force acting on the lower extremity
joints during water walking and running. Biomechanical investigation into the
kinetics of the gait cycle in water also is an important avenue for future work.
Bishop PA, Frazier S, Smith J, Jacobs, D (1989) Physiological responses to treadmill and
water running. The Physician and Sportsmedicine 17:8794.
Buck K, McNaughton L, Sherman R, Bentley DJ, Batterham AM (2001) Physiological
response to treadmill walking in water at different speeds and temperatures. Sports
Medicine Training and Rehabilitation 10:105122.
Bushman BA, Flynn MG, Andres FF, Lambert CP, Taylor MS, Braun WA (1996) Effect of
4 wk of deep water run training on running performance. Medicine and Science in
Sports and Exercise 29:694699.
Byrne HK, Craig JN, Willmore JH (1996) A comparison of the effects of underwater tread-
mill walking to dry land treadmill walking on oxygen consumption, heart rate, and
cardiac output. The Journal of Aquatic Physical Therapy 4:411.
Christie JL, Sheldahl LM, Tristani FE (1990) Cardiovascular regulation during head-out
water immersion. Journal of Applied Physiology 69:657664.
Coyle EF, Martin DR, Sincamore DR, Joyner MJ, Hagverg JM, Holloszy JO (1984) Time
course of loss of adaptation after stopping prolonged intense endurance training. Jour-
nal of Applied Physiology 57:18571864.
Responses to Running and Walking in Water 77
DAcquisto LJ, DAcquisto DM, Renne D (2001) Metabolic and cardiovascular responses
in older women during shallow-water exercise. Journal of Strength and Conditioning
Davidson K, McNaughton L (2000) Deep water running training and road running training
improve Vo(2)max in untrained women. Journal of Strength and Conditioning Re-
Derion T, Guy HJB, Tsukimoto K, Schaffartzik W, Prediletto R, Poole DC, Knight DR,
Wagner PD (1992) Ventilation-perfusion relationships in the lung during head-out
water immersion. Journal of Applied Physiology 72:6472.
Dowzer CN, Reilly T (1998) Deep-water running. Sports Exercise and Training 4:5661.
Dowzer CN, Reilly T, Cable NT (1998) Effects of deep and shallow water running on
spinal shrinkage. British Journal of Sports Medicine 32:4448.
Dowzer CN, Reilly T, Cable NT, Nevill A (1999) Maximal physiological responses to deep
and shallow water running. Ergonomics 42:275281.
Evans BW, Cureton KJ, Purvis JW (1978) Metabolic and circulatory responses to walking
and jogging in water. Research Quarterly for Exercise and Sport 49:442449.
Farhi LE, Linnarsson D (1977) Cardiopulmonary readjustments during graded immersion
in water at 35C. Respiration Physiology 30:3550.
Frangolias DD, Rhodes EC (1996) Metabolic responses and mechanisms during water im-
mersion running and exercise. Sports Medicine 22:3853.
Gehring MM, Keller BA, Brehm BA (1997). Water running with and without a flotation
vest in competitive and recreational runners. Medicine and Science in Sports and Ex-
Gleim GW, Nicholas JA (1989) Metabolic costs and heart rate responses to treadmill walk-
ing in water at different depths and temperatures. The American Journal of Sports
Hall J, Macdonald IA, Maddison PJ, OHare JP (1998) Cardiorespiratory responses to
underwater treadmill walking in healthy females. European Journal of Applied Physi-
Hall J, Skevington SM, Maddison PJ, Chapman K (1996) A randomized and controlled
trial of hydrotherapy in rheumatoid arthritis. Arthritis Care Research 9:206215.
McArdle WD, Magel JR, Lesmes GR, Pechar GS (1976) Metabolic and cardiovascular
adjustment to work in air and water at 18, 25, and 33C. Journal of Applied Physiol-
Mercer JA, Jensen RL (1998) Heart rates at equivalent submaximal levels of VO
2 do not
differ between deep-water running and treadmill running. Journal of Strength and
Conditioning Research 12:161165.
Michaud TJ, Brennan DK, Wilder RP, Sherman NW (1995) Aquarunning and gains in
cardiorespiratory fitness. Journal of Strength and Conditioning Research 9: 7884.
Morrow MJ, Jensen RL, Peace CR (1996) Physiological adaptations to deep water and land
based running training programmmes. Medicine and Science in Sports and Exercise
Nakazawa K, Yano H, Miyashita M (1994) Ground reaction forces during walking in wa-
ter. Medicine and Science in Aquatic Sports 39:2834.
Napoletan J, Janes J, Hicks R (1991) The effect of underwater treadmill exercise in the
rehabilitation of ACL reconstruction. Medicine and Science in Sports and Exercise 5:
78 M. B. Pohl and L. R. McNaughton
Napoletan JC, Hicks RW (1995) The metabolic effects of underwater treadmill exercise at
two depths. American Physical Therapy Research 3:913.
Powers SK, Howley ET (2001) Exercise Physiology (4th Edition). New York: McGraw-
Hill. p166168, 173174.
Svedenhag J, Seger J (1992) Running on land and in water: comparative exercise physiol-
ogy. Medicine and Science in Sports and Exercise 24:11551160.
Town GP, Bradley SS (1991) Maximal metabolic responses of deep and shallow water
running in trained runners. Medicine and Science in Sports and Exercise 23:238241.
Wilber RL, Moffatt RJ, Scott BE, Lee DT, Cucuzzo NA (1996) Influence of water run
training on the maintenance of aerobic performance. Medicine and Science in Sports
and Exercise 28:10561062.