Journal of Strength and Conditioning Research, 2005, 19(1), 177–183
q 2005 National Strength & Conditioning Association
CHANGES IN BAR-PATH KINEMATICS
AFTER POWER-CLEAN TRAINING
JASON B. WINCHESTER,1 TRAVIS M. ERICKSON,1 JOHN B. BLAAK,1
Department of Exercise and Sport Science, Musculoskeletal Research Center, University of Wisconsin-La Crosse,
1
La Crosse, Wisconsin 54601; 2Department of Health, Leisure, and Exercise Science, Neuromuscular Laboratory,
Appalachian State University, Boone, North Carolina 28607.
ABSTRACT. Winchester, J.B., T.M. Erickson, J.B. Blaak, and
J.M. McBride. Changes in bar-path kinematics and kinetics af-
ter power-clean training. J. Strength Cond. Res. 19(1):177–183.
2005.—The purpose of this study was to investigate the direction
and magnitude of kinematic changes in bar path and kinetic
variable changes in the power clean (PC) after 4 weeks of PC
training. Eighteen healthy adult men who had a minimum of 1
year of previous experience in the PC participated as subjects in
this study. The subjects were pretested for their 1 repetition
maximum (1RM) and provided with visual and verbal cues dur-
ing PC training sessions, which took place 3 times per week for
4 weeks. Variables measured during data collection include pre–
and post–peak force, peak power, and several bar-path kine-
matic variables through videography at 50, 70, and 90% of the
subjects’ pre-1RM. Peak force was improved at 50% of 1RM from
936 6 338 N to 1,299 6 384 N, at 70% from 1,216 6 315 N to
1,395 6 331 N, and at 90% from 1,255 6 329 N to 1,426 6 321
N. Peak power was increased at 50% of 1RM from 3,430 6 1,280
W to 4,230 6 1,326 W. All variables with respect to bar-path
kinematics were improved significantly. These results indicate
that both kinematic and kinetic variables improve through
training and feedback. It is possible that persons beginning the
PC exercise or coaches who provide instruction on the PC to
beginning lifters should focus on proper bar path during the
movement. This may result in force and power output to develop
as technique improves. However, further investigation is re-
quired to establish the link between bar-path changes and ki-
netic variable performance improvements.
KEY WORDS. weightlifting, technique, peak power, peak force
INTRODUCTION
ar-path parameters have been outlined and re-
B lated to body position and kinetic performanc-
es in weightlifting movements (12, 16) (Figure
1). Furthermore, a relationship between suc-
cess in weightlifting movements and bar-path
kinematics has been proposed (14). Three important ki-
nematic factors have been highlighted: (a) an initial rear-
ward movement of the bar during the first pull (Figure
1, D 3 2), (b) a catch position no more than 20 cm behind
the most forward bar position (Figure 1, D 3 L), and (c)
an amount of looping (D 3 L) less than the net rearward
horizontal displacement (D 3 T). Maximal bar velocities
during the second pull have been identified as a contrib-
uting factor to performance in weightlifting movements
as well (8). In addition, recent kinetic analysis has high-
lighted the importance of maximal force production dur-
ing the second pull on performance in the power clean
(PC) (14). These important factors are supported by other
kinematic and kinetic analyses that have been performed
over the past 25 years (1, 2, 4–7). However, these factors
have never been tracked longitudinally in a relatively un-
AND KINETICS
AND JEFFREY M. MCBRIDE2
trained athletic population attempting to learn and use
weightlifting movements, such as the PC, to improve per-
formance.
Volumes of non–peer-reviewed material have been
published on PC technique (2–9). However, little compar-
ative data in untrained subjects are available as a refer-
ence point for optimal bar-path progression with training
(13). Significant correlations have been described among
variables in successful attempts in weightlifting compe-
tition (16), including catch position to amount of loop (Fig-
ure 1, D 3 T vs. D 3 L), catch position to second pull
position (D 3 T vs. D 3 2), and catch position to amount
of hipping (D 3 T vs. D 3 V). There has also been an
established link between bar kinematics (such as bar ve-
locity) and kinetic variables (such as ground reaction
force) and subsequently power output (16). This includes
relationships among variables such as peak vertical force
to peak power (PP) and PP to peak first pull velocity.
However, these variables have never been examined lon-
gitudinally in the context of an athlete attempting to im-
prove performance in the PC.
The PC has typically been used as a training tool for
improving athletic performance for 2 reasons. First, its
transference and close kinematic and kinetic relationship
to vertical activities such as jumping is clear (3). Second,
the optimal stimulus for training for the development of
muscle power is to use an activity and load that optimizes
power output (10, 11, 17). In weightlifting movements,
the appropriate load to maximize power output is unclear.
However, it can be inferred from past data that it may be
at a load between 50 and 90% of dynamic 1 repetition
maximum (1RM) strength in a PC movement (6, 14). In
ballistic activities, such as jump squats or bench press
throws, it is a load, on average, of 30% of dynamic max-
imum strength in same movement (10, 11, 15, 17). How-
ever, proper use of lifting techniques, especially in the PC,
is essential to maximize the stimulus applied. Now that
bar-path kinematic parameters for proper weightlifting
technique have been established (12, 16), it is essential
to investigate progression of these variables with PC
training in athletic populations and observe their rela-
tionship to kinetic variables as well.
METHODS
Experimental Approach to the Problem
Two-dimensional (2-D) video and a force plate were used
to provide a kinematic and kinetic assessment of the PC
over a training period of 4 weeks. Several variables that
are important to PC performance were outlined in the
‘‘Introduction.’’ The hypothesis or question to be ad-
177
178 WINCHESTER, ERICKSON, BLAAK ET AL.
FIGURE 1. Description of bar-path kinematic variables used
to assess quantitative changes in bar path before (pre) and af-
ter (post) training (modified from Stone et al. [15]).
dressed was whether visual and verbal feedback on bar
path would relate to kinetic aspects of PC performance
with practice. In addition, a loading range of 50–90% of
1RM was used in training and testing to determine its
influence on bar-path kinematics and kinetics as well. Be-
cause bar path is an easy variable to measure and record,
this method may be an effective tool for practitioners to
improve PC technique in athletes.
Subject Characteristics
Eighteen healthy adult men with a mean age of 22.22 6
2.13 years volunteered to participate in this investigation.
Subjects were NCAA Division III athletes who had some
(minimum 1 year) experience in the PC. To ensure the
safety of all participants, all treatment and testing con-
ditions in this investigation were examined and approved
by the University of Wisconsin–La Crosse Institutional
Review Board (IRB). In compliance with IRB guidelines,
all subjects in this study read and signed an informed
consent before participation. Subjects were allowed to
continue their normal physical activities throughout the
duration of the study but were instructed to have a full
day of rest before the pre- and posttesting sessions.
Study Design
Data collection for this project was divided into 2 sessions
for pre- and posttesting. The sessions for data collection
were separated by a 4-week training cycle in the PC, in
which the subjects were given verbal feedback from a Cer-
tified Strength and Conditioning Specialist and visual
cues through 2-D videography to improve individual tech-
nique in the PC. At the time of the first testing session,
a 1RM for each participant in the study and a video rep-
resentation of the subject’s technique for evaluation of ki-
nematic variables of bar position were obtained. In ad-
dition, peak force (PF) and PP during the second pull
were measured with force plate and video data at 50, 70,
and 90% of subjects’ 1RM. During the posttesting session,
all previously mentioned kinematic and kinetic variables
were again measured.
1RM Maximum Protocol
The subjects’ 1RM for the PC was used to measure con-
centric muscle strength for pre- and posttesting sessions
for this study. Warm-up trials based on a percentage of
the subjects’ estimated 1RM were given before 1RM test-
ing in both the pre- and posttesting sessions. The per-
centage and number of repetitions (reps) for the warm-
up protocol is as follows: 4–6 reps at 30% of 1RM, 3–4
reps at 50% of 1RM, 2–3 reps at 70% of 1RM, and 1 rep
at 90% or 1RM. From this point, subjects were allowed
3–4 maximal trials to establish their actual 1RM. A 5-
minute rest period was given between each warm-up trial
and maximal effort.
Training Protocol
Subjects trained in the PC 3 days per week for 4 weeks.
The training days were broken up by level of intensity
from low to high to moderate on Monday, Wednesday,
and Friday, respectively. For the low-intensity or Monday
training session, subjects performed a warm-up set of PCs
at 50% of 1RM and then performed 5 training sets of 5
reps at 50% of 1RM with verbal and visual feedback of
kinematics provided to the subject during the training
session. Visual feedback was in the form of the subject
watching his bar path on a video screen within 3 minutes
of performing his lift. Verbal feedback was provided as to
how to adjust body movements to obtain an optimal bar
path as outlined by Stone et al. (16). Subjects were al-
lowed 2 minutes of rest between all sets on the low-inten-
sity day. On the high-intensity day, subjects performed 1
warm-up set of 5 reps at 50% of 1RM and 1 set of 3 reps
at 70% of 1RM. Subjects then performed 3 training sets
of 1 rep at 90% of 1RM with 4 minutes of rest allowed
between each set. During the moderate-intensity training
sessions, subjects were allowed 1 warm-up set of 5 reps
at 50% of 1RM before performing the training sets. The
training protocol for the moderate-intensity day was 4
sets of 3 reps at 70% of 1RM. A 3-minute rest period be-
tween each set was allowed during the training session.
During the 4-week training cycle, visual and verbal feed-
back was provided during warm-up and training sets.
However, performance with respect to kinematic vari-
ables, PF, and PP was determined only for actual training
sets.
Measurements
Force data were collected from both feet of the lifter at
600 Hz by using an Advanced Medical Technology, Inc.,
(AMTI) model OR5-5 Biomechanics Platform (Watertown,
MA). Collection of kinematic data at 60 Hz was performed
with 2-D videography by using a JVC model TK-C1380
 BAR PATH 179

D 3 2 5 start position to beginning of second pull; D 3 V 5 second pull position to most forward position; DV1 5 peak velocity during first pull; D 3 L 5 most forward
† PF 5 peak force during the second pull; PP 5 peak power during the second pull; D 3 T 5 start position to catch position; DV2 5 peak velocity during second pull;
Color Video Camera (Yokohama, Japan) and an Event

20.208 6 0.089*
and Video Control Unit (PEAK Performance Technolo-

20.157 6 0.045
D 3 L (m)
gies, Inc., Englewood, CO). A reflective marker was
placed on the end of the right-hand side of the bar for 2-
D videography. Data were converted to digital format
with a 12-bit Analog-to-Digital card and stored on hard
disk for later analysis. A lateral-view bar path was cre-
ated by digitizing the bar end marker and visualizing a
tracing of its path. Peak power was calculated with force-
time curves and bar velocity-time curves (generated from

2.846 6 0.315
2.553 6 0.492
DV1 (m·s21)
video digitizing of bar end marker) by using Motus 3.2
software (PEAK). Reliability data concerning bar-path
variables were assessed by using videography in 7 sepa-
rate subjects. Subjects were tested twice, with a 1-week
intermission between testing. The following variables
were calculated, and no significant difference was ob-
served (p # 0.05) among values between testing sessions:

0.107 6 0.063
0.081 6 0.092
D 3 L, 20.093 6 0.030m, 20.113 6 0.070m; D 3 T,

D 3 V (m)
20.055 6 0.063m, 20.048 6 0.093m; D 3 2, 20.012 6
0.040m, 0.001 6 0.041m; D 3 V, 0.051 6 0.073m, 0.065
6 0.082m).
Statistical Analyses
A repeated measures analysis of variance with contrasts

20.077 6 0.065*
20.007 6 0.076
was used to determine among- and within-group differ-

D 3 2 (m)
ences as well as differences among testing sessions with
respect to percentage change. All statistical calculations
were performed by using SPSS version 11.0 for Windows
(SPSS, Inc., Chicago, IL). For all procedures, statistical
significance was set at p # 0.05.

RESULTS

1.497 6 0.321
4.079 6 0.253
DV2 (m·s21)
50% Testing Load
Table 1 shows that PF was improved at the 50% load from
936 6 338 N to 1,299 6 384 N. Peak power also improved
at the 50% testing load from 3,430 6 1,280 W to 4,230 6
1,326 W. Measurement of kinematic variables saw im-
provements in the catch position (D 3 T) from 20.057 6
0.015 m to 20.204 6 0.083 m, in the position for the sec-
20.204 6 0.083*

ond pull (D 3 2) from 20.007 6 0.076 m to 20.077 6

20.057 6 0.115
D 3 T (m)

0.065 m, and in the amount of loop (D 3 L) from 20.157

6 0.045 m to 20.208 6 0.065 m. Significant improve-
ments were found between the pre- and posttesting ses-
sions. Figure 2 shows a visual representation of the av-
TABLE 1. 50% load. Data presented as mean 6 SD.†

erage change in bar path before (pre) and after (post)

training with all 3 testing loads. Figure 3 shows signifi-
cant changes to bar-path kinematic variables.
4,230 6 1,326*
3,430 6 1,280

70% Testing Load

PP (W)

Table 2 shows that PF was improved at the 70% testing

* Significant differences at p # 0.05.

load from 1,216 6 315 N to 1,395 6 331 N. D 3 T im-

proved from 20.094 6 0.017 m to 20.218 6 0.038 m. Po-
sition for D 3 2 also showed improvement pre- to post-
training from 20.043 6 0.075 m to 20.096 6 0.069 m.
Figure 4 shows significant changes to bar-path kinematic
1,299 6 384*
936 6 38

variables.
PF (N)

90% Testing Load

position to catch.

Table 3 shows that PF was improved at the 90% testing

load from 1,255 6 329 N to 1,426 6 321 N. Peak power
was also increased at 90% from 3,688 6 1,172 W to 4,165
Posttest

6 1,225 W. Kinematic measurements for D 3 2 improved

Pretest

from 20.063 6 0.086 m to 20.085 6 0.024 m, D 3 T

showed significant improvement from 20.068 6 0.152 m
 180

and after (post) training at p # 0.05.

WINCHESTER, ERICKSON, BLAAK
ET AL.

maximum testing loads. Average of all subject data.

and after (post) training at 50, 70, and 90% of 1 repetition

ables with 50% of 1 repetition maximum test load before (pre)

FIGURE 2. Qualitative representation of bar path before (pre)

FIGURE 3. Significant changes (*) to bar-path kinematic vari-

TABLE 2. 70% load. Data presented as mean 6 SD.†
PF (N) PP (W) D 3 T (m) DV2 (m·s21) D 3 2 (m) D 3 V (m) DV1 (m·s21) D 3 L (m)
Pretest 1,216.33 6 315 3,896.83 6 1,035 20.094 6 0.117 3.67 6 0.275 20.043 6 0.075 0.104 6 0.077 2.81 6 0.335 20.155 6 0.051
Posttest 1,395.99 6 331* 4,048.3 6 944 20.218 6 0.038* 3.18 6 0.229 20.096 6 0.069* 0.079 6 0.069 2.38 6 0.423 20.201 6 0.087
* Significant differences at p # 0.05.
† PF 5 peak force during the second pull; PP 5 peak power during the second pull; D 3 T 5 start position to catch position; DV2 5 peak velocity during second pull;
D 3 2 5 start position to beginning of second pull; D 3 V 5 second pull position to most forward position; DV1 5 peak velocity during first pull; D 3 L 5 most forward
position to catch.
 BAR PATH 181

D 3 2 5 start position to beginning of second pull; D 3 V 5 second pull position to most forward position; DV1 5 peak velocity during first pull; D 3 L 5 most forward
† PF 5 peak force during the second pull; PP 5 peak power during the second pull; D 3 T 5 start position to catch position; DV2 5 peak velocity during second pull;
20.136 6 0.066
20.199 6 0.072
D 3 L (m)
2.46 6 0.296
2.38 6 0.311
DV1 (m·s21)

0.089 6 0.071*
0.131 6 0.092
D 3 V (m)

20.085 6 0.024*
20.063 6 0.086
D 3 2 (m)
FIGURE 4. Significant changes (*) to bar-path kinematic vari-
ables with 70% of 1 repetition maximum test load before (pre)

3.18 6 0.227
3.18 6 0.229
DV2 (m·s21)
and after (post) training at p # 0.05.

to 20.195 6 0.039 m. Improvement in position for D 3 V

from 0.131 6 0.092 m to 0.089 6 0.071 m was also noted.
Figure 5 shows significant changes to bar-path kinematic
variables.
20.195 6 0.039*
20.068 6 0.152
D 3 T (m)

DISCUSSION
The primary finding in this study is that during perfor-
mance of the PC exercise, as kinematic variables with
respect to bar path improved, the subjects were able to
generate increased levels of PP and PF, thus improving
TABLE 3. 90% load. Data presented as mean 6 SD.†

bar kinetics. As mentioned earlier, previous research has

established that several factors contribute to a successful
3,709.65 6 1,225*
3,461.47 6 1,172

lift in weightlifting movements such as the PC: (a) hori-

zontal (rearward) displacement of the bar in the first pull
PP (W)

with respect to the starting position (D 3 2); (b) the

amount of looping of the bar in the catch phase (D 3 L);
(c) the ratio of looping to the net rearward displacement
* Significant differences at p # 0.05.

of the bar (ratio of D 3 L to D 3 T); (d) increased peak

bar velocity in the second pull (DV2); and (e) increased
PF production during the second pull phase of the PC
(PF) (5, 16).
1,426.00 6 321*
1,255.83 6 329

Rearward movement of the bar from the first to the

PF (N)

second pull (D 3 2) has been established as an important

indicator of the likelihood of success of the lift in weight-
lifting movements (16). As noted in the ‘‘Results’’ section,
position to catch.

D 3 2 improved for all 3 testing loads in this study. Es-

tablishment of a good starting position where the lifters’
knees are positioned in front of the bar before the first
Posttest
Pretest

pull is an important first step in establishing good bar

kinematics. In addition, keeping a flat-footed stance for
as long as possible while keeping the hips directly over
182 WINCHESTER, ERICKSON, BLAAK ET AL.
FIGURE 5. Significant changes (*) to bar-path kinematic vari-
ables with 90% of 1 repetition maximum test load before (pre)
and after (post) training at p # 0.05.
or a little behind the ankles and instructing the lifter to
extend first at the knee joint will allow for rearward mo-
tion of the bar between the first and the second pull.
Maintaining this body position and bar path will allow
for increased vertical velocities and increased vertical
force production through the second pull as more of the
force generated is directed vertically (9, 16).
The second kinematic variable that is closely correlat-
ed with success in weightlifting movements such as the
PC is the amount of displacement between the most for-
ward position during the second pull and the catch posi-
tion commonly referred to as the amount of looping at the
top of the movement (D 3 L) (16). In this particular study,
D 3 L was significantly changed only at the 50% testing
load. This reflects earlier research that suggests less hor-
izontal displacement of the bar in heavier loads, which
may be due to the lifter allowing the bar to drift forward
slightly or coming up on the balls of the feet too early.
This may cause the athlete to change the direction of ap-
plied force to a more horizontal component, commonly re-
ferred to as ‘‘hipping’’ the bar. Such an action will cause
an exaggerated D 3 L and force the lifter to catch the bar
in front of his starting position, thereby allowing for very
little net horizontal displacement (D 3 T). In previous
research, 68% of lifts that had an exaggerated D 3 L were
failed attempts (9, 16). D 3 L typically increased during
training in the current study; however, the amount of
loop never exceeded the recommended amount of 20 cm
at any of the testing loads.
Another important kinematic variable is the total
amount of horizontal displacement from the beginning of
the lift to the catch position (D 3 T). Previous studies by
Stone et al. (16) found a significant correlation where 76%
of successful attempts in weightlifting activities had a
rearward displacement of the bar and 64% of failed at-
tempts had no horizontal displacement or the lifter
caught the bar in front of the starting position. In this
present study, D 3 T was improved for all testing loads,
which may indicate an improved performance success
rate.
The PC has been used for many years to improve per-
formance in athletic endeavors for multiple reasons,
such as its transference to the event performed and its
ability to provide an optimal stimulus for production of
power within a training athlete. It is important to note
that in this study no conscious effort was made to im-
prove force or power production during the training ses-
sions and feedback. The sole focus of the investigation
was on improving the bar-path kinematics of each lifter.
During the final testing phase of this study, a new 1RM
was not measured and the 50, 70, and 90% loads were
of the original 1RM measured at the beginning of the
investigation. It is therefore interesting to note that at
the same loads PP increased at all testing loads and PF
increased at 2 of the 3 testing loads. This would suggest
that focusing on correct bar kinematics with particular
attention to D 3 2, D 3 L, and D 3 T will not only allow
the lifter to improve technique in the PC and possible
other weightlifting movements, but will also translate to
increased power and force production in the athletes
performing the lifts, even with no increase in training
loads.
PRACTICAL APPLICATIONS
Coaches should use visual and verbal feedback to track
bar path with athletes learning to use the PC for training.
By making adjustments in bar path, coaches will ensure
that the proper body mechanics are occurring, which will
translate into improved force and power output during
the exercise and maximize the stimulus placed on the
body for an optimal training response. The bar path out-
lined by Stone et al. (15) should be used as a reference or
template by which comparisons can be made. In this
study, very significant changes in bar path occurred with
just 4 weeks of training. This would seem to be a realistic
time line for coaches attempting to teach effective PC
technique to athletes.
REFERENCES
1. BARTONIETZ, K.E. Biomechanics of the snatch: Toward a higher
training efficiency. J. Strength Cond. Res. 18(3):24–31. 1996.
2. BAUMANN, W., V. GROSS, K. QUADE, P. GALBIERZ, AND A.
SCHWIRTZ. The snatch technique of world class weightlifters at
the 1985 world championships. Int. J. Sports Biomech. 4:68–
89. 1988.
3. CANAVAN, P.K., G.E. GARRET, AND L.E. ARMSTRONG. Kinematic
and kinetic relationships between an Olympic-style lift and the
vertical jump. J. Strength Cond. Res. 10:127–130. 1996.
4. GARHAMMER, J. Cinematographical and mechanical analysis of
the snatch lift. Int. Olympic Lifter. 2:5–15. 1975.
5. GARHAMMER, J. Biomechanical characteristics of the 1978
world weightlifting champions. In: Biomechanics VII-B. A. Mo-
recki, K. Fidelus, K Kedzior, and A. Wit, eds. Baltimore: Uni-
versity Park Press, 1980. pp. 300–304.
6. GARHAMMER, J. Biomechanical profiles of Olympic weightlif-
ters. Int. J. Sports Biomech. 1:122–130. 1985.

7. GARHAMMER, J. Bar trajectories of world class male and female
weightlifters. Part 2. Int. Olympic Lifter. 10(6):12–13. 1990.
8. GARHAMMER, J. Weightlifting performance and techniques of
men and women. In: International Conference on Weightlifting
and Strength Training. K. Häkkinen, ed. Lahti, Finland: Gum-
merus Printing, 1998. pp. 89–94.
9. ISAKA, T., J. OKADA, AND K. FUNATO. Kinematics analysis of
the barbell during the snatch movement of elite Asian weight-
lifters. Int. J. Sports Biomech. 12:508–516. 1996.
10. KANEKO, M. Training effects of differing loads on the force-
velocity relationship and mechanical power output in human
muscle. Scand. J. Sports Sci. 5(2):50–55. 1983.
11. MCBRIDE, J.M., T. TRIPLETT-MCBRIDE, A. DAVIE, AND P.U.
NEWTON. The effect of heavy versus light load jump squats on
the development of strength, power, and speed. J. Strength
Cond. Res. 16:75–85. 2002.
12. SCHILLING, B.K., M.H. STONE, H.S. O’BRYANT, A.C. FRY, R.H.
COGLIANESE, AND K.C. PIERCE. Snatch technique of collegiate
national level weightlifters. J. Strength Cond. Res. 16:551–555.
2002.
BAR PATH 183
13. SEWELL, L.P., T.G. REEVE, AND R.A. DAY. Effect of concurrent
visual feedback on acquisition of a weightlifting skill. Percept.
Mot. Skills. 67:715–718. 1988.
14. SOUZA, A.L., S.D. SHIMADA, AND A. KOONTZ. Ground reaction
forces during the power clean. J. Strength Cond. Res. 16:423–
427. 2002.
15. STONE, M.H., H.S. O’BRYANT, L. MCCOY, R. COGLIANESE, M.
LEHMKUHL, AND B. SCHILLING. Power and maximum strength
relationships during performance of dynamic and static
weighted jumps. J. Strength Cond. Res. 17:140–147. 2003.
16. STONE, M.H., H.S. O’BRYANT, F.E. WILLIAMS, R.L. JOHNSON,
AND K.C. PIERCE. Analysis of bar paths during the snatch in
elite male weightlifters. Strength Cond. J. 20(4): 56–64. 1998.
17. WILSON, G.J., R.U. NEWTON, A.J. MURPHY, AND B.J. HUM-
PHRIES. The optimal training load for the development of dy-
namic athletic performance. Med. Sci. Sports Exerc. 25:1279–
1286. 1993.
Address correspondence to Jeffrey M. McBride, mcbridejm@
appstate.edu.