J. Biomechunics Vol. 25, No. 10, pp. 1237-1239, Printed in Great Britain
1992. 0
&X-9290/92 1992 Pergamon
S5.00+.00 Press Ltd
TECHNICAL NOTE
STEP LENGTH
AND FREQUENCY EFFECTS ON GROUND FORCES DURING WALKING
REACTION
PHILIPE. MARTIN*and ANTHONY P. MARSH Exercise and Sport Research Institute, Arizona State University, Tempe, AZ 85287-0404, U.S.A. Abstract-It is well established that the speed of walking or running significantly affects ground reaction force (GRF) characteristics. While it is sometimes assumed that the variations in step length (SL) and step frequency (SF) also affect GRF patterns, little documentation of this can be found in the literature. Ten young adults performed overground walking at 1.43 m s-i across a force platform under five SL conditions: preferred SL and SLs that were longer and shorter than the preferred by 5 and 10% of greater trochanter height. The contact time, anteroposterior braking and propulsive force and impulse descriptors, and vertical impulse per step increased systematicany as SL increased. In contrast, vertical peak forces and impulse per meter walked showed little change with SL manipulation. Despite the systematic effect of SL on several GRF descriptors, constraint of SL and SF in gait assessments is not recommended as this would prohibit the evaluation of representative gait kinematics and kinetics. Rather, these results suggest that researchers should report SL and SF data when comparing GRF characteristics between experimental groups or conditions, and should be alert to the association between SL/SF and GRFs when interpreting GRF trends.
INTRODUCTION
METHODS
It is well accepted that the speed of walking or running significantly affects ground reaction force (GRF) characteristics (Andriacchi et al., 1977; Jansen and Jansen, 1978; Munro et al., 1987; Vaughan et al., 1987). Thus, many investigators have chosen to compare experimental groups or conditions at controlled speeds while allowing subjects to freely select step length (SL) and step frequency (SF). It is also possible, however, that systematic differences in SL and SF between groups or conditions could also affect GRF patterns (Soames and Richardson, 1985), which could influence interpretation of GRF findings. As an example, Larish et al. (1988) reported lower peak vertical and anteroposterior GRFs for aged adults than for young adults when walking speed was controlled. One interpretation of this trend provided by Larish et al. is that the older adults select a gait pattern in order to reduce GRF magnitudes and, thus, decrease stress on the musculoskeletal system. They also observed, however, that the older adults had shorter SLs than the young. Thus, the GRF differences between the old and young may simply have been related to the SL differences between the age groups. There is surprisingly little documentation in the research literature on the effect of changes in SL and SF on GRF characteristics, perhaps due to the common practice of allowing subjects to freely choose SL and SF while controlling the speed. Thus, the aim of this study was to examine the effects of SL and SF on selected GRF characteristics during walking. It was hypothesized that increases in SL and decreases in SF at a specific walking speed would result in higher peak forces for both the vertical and anteroposterior components of the GRF.
Ten young adults (5 male, 5 female, R,,=27.6&2.5 y, 8,,, = 169 f 13 cm, 8,, = 624+ 146 N) were used as subjects. Initially each subject’s preferred SL/SF combination at the test speed of 1.43 ms-’ was determined using a motorized treadmill. This speed is within the range of preferred walking speeds of healthy, young adults typically reported in the literature(e.g. Larish et al:, 1988; Laurent and Pailhous, 1986; Murray bt al., 1966: Pearce et al., 1983). Treadmill walkinn was used to determine preferred SL/SF because of the abilit; to control walking speed with precision and the simplicity of determining average SF and SL for multiple step cycles. Comparisons of treadmill and overground locomotion have reflected minor differences in SL and SF (Elliott and Blanksby, 1976; Nelson et al., 1972; Pearce er al., 1983; Taves et al., 1985). Because we expected a systematic SL/SF effect on the GRF and were not specifically interested in the GRF features at the preferred SL/SF combination, minor errors in the estimation of overground preferred SL/SF from the treadmill assessment should have had little or no effect on our ability to test our hypothesis. The treadmill speed was monitored by timing 10 belt revolutions and adjusted as necessary to achieve the desired target speed. Following a minimum of 15 min of treadmill accommodation, each subject’s average preferred SF at 1.43 ms-i was computed from the time required to complete 50 steps. SL was then computed from the deterministic relationship between speed, SL, and SF. Subjects then performed multiple overground walking trials along a 10 m walkway at approximately 1.43 m s- * as the GRF was sampled from an AMTI force platform at 500 Hz. The walking speed was monitored using two photocells spaced 3.20 m apart, the first of which triggered a timer and force platform data collection immediately before the subject contacted the force platform and the second which stopped the timer. Only those trials that were within + 3% of the target speed were considered for analysis. Each subject
Received in finalform12 February 1992.
*Author to whom correspondence should be addressed.
1237
Technical Note
1238
completed five acceptable trials for each of the five randomly ordered SL/SF conditions: his or her preferred SL and SLs that were longer and shorter than the preferred by 5 and 10% of leg length (LL) as represented by greater trochanter height (sample X,,=O.851 kO.069 m). SLs for all trials, including the preferred SL condition, were controlled using tape markers on the floor. The subjects were instructed to walk as normally as possible while using the markers as a guide for heel contact. A potential risk in using this approach is that the subjects were required to target the markers which may have influenced gait kinematics and kinetics. Laurent and Pailhous (1986), however, observed that it was easier for subjects to change SL on the basis of visual marks than to synchronize movement with an auditory signal. In addition, Patla et al. (1989) reported no significant differences in GRF impulse descriptors between preferred SL conditions during walking with and without a visual foot contact cue, lending support to our assumption that targeting effects would be minor. A 60 Hz video camera was used to record a sagittal view of subjects’ walking patterns. This was used later to quantify subjects’ actual SLs as they crossed the force platform to establish the success of SL manipulation. The deviation of the actual SL from the nominal SL was initially measured in meters and subsequently expressed as a percentage of LL on a trial by trial basis. Quantified GRF features included contact time, first and second vertical (V) peak forces, anteroposterior (AP) peak propulsive and braking forces, and V and AP impulses. GRFs and impulses were normalized to the body weight. Impulses were also normalized with respect to the distance walked. A single factor repeated measures ANOVA was used for the statistical analysis of the GRF descriptors.
RESULTS
Video analysis of the actual SLs used by the subjects demonstrated that SL was manipulated effectively in all cases and that there was little difference between the nominal conditions and the actual SLs (Table 1). Initially, GRF characteristics for males and females were compared. Because there were no significant GRF differences between males and females, these data were pooled for subsequent considerations of SL effects. The systematic effects of SL were primarily limited to contact time and the AP GRF descriptors (Table 1, Figs 1 and 2). As expected, these factors increased in magnitude as SL increased. While the first and second vertical peak forces showed little change as SL increased, the vertical impulse per step increased because of the increasing contact time. When normalized per unit of distance traveled, vertical impulse showed a statistically significant SL effect, but there was no systematic nature to the changes in impulse.
DISCUSSION
Clarke et al. (1983) reported increasing peak shank decelerations immediately following foot contact with increasing SL during running, and suggested that the changes in SL at a given speed can affect the amount of shock that must be absorbed by the musculoskeletal system. These results provided circumstantial evidence that increased SL resulted in increased impact forces during running. These observations are consistent with our results for walking, although in-
Table 1. Mean GRF characteristics for five SL/SF combinations during walking at 1.43 m s-i Nominal SL conditions GRF descriptor
- lO%LL
-5%LL
Preferred
+5%LL
+ lO%LL
Actual SL (m)t
0.683 -9.5%LL 0.580 kO.035
0.719 - 5.2%LL 0.611 * 0.039
0.765 0.2%LL 0.630 kO.038
0.802 4.5%LL 0.666 kO.044
0.847 9.8%LL 0.694* f0.044
1.167 *0.070 1.155 +0.041
1.159 kO.065 1.164 kO.051
1.166 + 0.065 1.180 k 0.052
1.187 kO.068 1.176 kO.058
1.185 f 0.074 1.178 *0.060
0.472 * 0.030
0.498 *0.035
0.518 k 0.032
0.552 kO.038
0.579* &0.040
0.691 + 0.020
0.693 kO.019
0.677 kO.015
0.689 kO.018
0.683’ + 0.024
0.194 + 0.030 0.207 *0.031
0.213 f 0.036 0.219 + 0.033
0.224 kO.035 0.238 f 0.032
0.246 kO.037 0.249 f 0.034
0.255* kO.037 0.267* *0.033
0.052 f 0.008 0.077 *0.010
0.061 io.010 0.084 +0.011
0.067 kO.011 0.088 kO.011
0.077 kO.011 0.096 kO.011
0.084* *0.012 0.099’ kO.011
Contact time (s) Vertical force
1st peak (BW) 2nd peak (BW) Impulse (BWsstep-r) (BWsm-‘) Anteriorposterior force Braking peak (BW) Propulsive peak (BW) Impulse (BW s step- I) (BW sm-r)
tActua1 SLs were measured from video records and are presented both as absolute values and as deviations from the preferred SL expressed as a percentage of LL. l Statistically significant SL effect @ < 0.05).
Technical Note
s
1.0
al ;;
0.8
t! Lf 0.6 z .z 0.1 t, > 0.2
---
25
50 % Contact
Pd.SL -1oxLL
75
100
Time
Fig. 1. Ensemble average curves for the vertical component of the ground reaction force for three of the five step length conditions: the preferred step length (Pref. SL) and step lengths either longer or shorter than the preferred by 10% of leg length ( f 10% LL). Step length had no significant effect on the amplitude of the vertical force.
0.3 ,
-0.3
1
0
I
I 25
54 % Contact
75
loo
Time
Fig. 2. Ensemble average curves for the anteroposterior component of the ground reaction force for the preferred step length and step lengths either longer or shorter than the preferred by 10% of leg length. Peak AP braking and propulsive forces and impulses systematically increased as step length increased.
creases in vertical force peaks were also expected with increasing SL. It was not surprising that both V and AP impulse computed per step (BW s step-i) showed a significant effect for SL since impulse is a function of both force and contact time, the latter showing a strong relationship with SL. If one is interested in considering impulse or related momentum changes of the body as a gross indicator of muscular effort during gait, impulse expressed per unit of distance (BW sm-‘) would be more relevant since this would offset the differences between SL conditions in the number of steps required to traverse a given distance. The systematic effect of SL on AP impulse expressed per unit of distance remained because of the strong trend for AP force to increase with increasing SL. V impulse, however, no longer showed the systematic SL effect when impulse was expressed per unit distance since V force changed very little with increasing SL. In our analysis, SL/SF conditions were examined at a single speed of walking so that the effect of speed on GRF characteristics was controlled. In contrast, Soames and Richardson (1985) constrained both SL and SF, but did so in a manner such that multiple walking speeds were produced as they evaluated GRF characteristics. They reported that both SL and SF significantly influenced GRFs, particularly at heel strike, and concluded that constraining only speed is insufficient if GRFs or joint transmission forces for different groups or conditions are being compared. They suggested that cadence should be constrained within unspecified limits
1239
whenever possible. It is unclear whether they were suggesting SF constraint as an alternative to speed constraint or in addition to it. Controlling only SF would make it difficult to interpret GRF trends because it has been shown that SF and speed are highly dependent (Laurent and Pailhous, 1986) and because of speed’s well-established effect on GRF (Andriacchi er al., 1977; Jansen and Jansen, 1978; Munro et al., 1987; Vaughan et al., 1987). A constraint of speed and SF (and, thus, SL); however, would likely prohibit the evaluation of representative gait kinematics and kinetics which is a goal of many analyses. While we favor controlling speed, we do not encourage the constraint of SL and SF in gait assessments. Our results, however, suggest that researchers should also report SL and SF data when comparing GRF characteristics between experimental groups or conditions, and should be alert to the association between SL/SF and GRFs when interpreting GRF trends. REFERENCES Andriacchi, T. P., Ogle, J. A. and Galante, J. 0. (1977) Walking speed as a basis for normal and abnormal gait parameters. J. Biomechanics 10, 261-268. Clarke, T. E., Cooper, L., Clark, D. and Hamill, C. (1983) The effect of varied stride rate and length upon shank deceleration during ground contact in running. Med. Sci. Sports Exert. 15, 170. Elliott, B. C. and Blanksby, B. A. (1976) A cinematographic analysis of overground and treadmill running by males and females. Med. Sci. Sports 8, 84-87. Jansen, E. C. and Jansen, K. F. (1978) Vis-velocitas-via: alteration of foot-to-ground forces during increased speed of gait. In Eiomechanics VI-A (Edited by Asmussen, E. and Jorgensen, K.), pp. 267-271. University Park Press, Baltimore. Larish, D. D., Martin, P. E. and Mungiole, M. (1988) Characteristic patterns of gait in the healthy old. Ann. N. Y. Acad. Sci. 515, 18-32. Laurent, M. and Pailhous, J. (1986) A note on modulation of gait in man: effects of constraining stride length and frequency. Hum. Mvmt Sci. 5, 333-343. Munro, C. F., Miller, D. I. and Fuglevand, A. J. (1987) Ground reaction forces in running: a reexamination. J. Biomechanics 20, 147-155. Murray, M. P., Kory, R. C., Clarkson, B. H. and Sepic, S. B. (1966) Comparison of free and fast speed walking patterns of normal men. Am. J. Phys. Med. 45, 8-24. Nelson, R. C., Dillman, C. J., Lagasse, P. and Bickett, P. (1972) Biomechanics of overground versus treadmill running. Med. Sci. Sports 4, 233-240. Patla, A. E., Robinson, C., Samways, M. and Armstrong, C. J. (1989) Visual control of step length during overground locomotion: task-specific modulation of the locomotor synergy. J. Exp. Psych: Hum. Perception Perform. 15, 603-617.
Pearce, M. E., Cunningham, D. A., Donner, A. P., Rechnitzer, P. A., Fullerton, G. M. and Howard, J. H. (1983) Energy cost of treadmill and floor walking at self-selected paces. Eur. J. appl. Physiol. 52, 115-I 19. Soames, R. W. and Richardson, R. P. S. (1985) Stride length and cadence: their influence on ground reaction forces during gait. In Biomechanics IX-A (Edited by Winter, D. A. et al.), pp. 406410. Human Kinetics, Champaign. Taves, C. L., Charteris, J. and Wall, J. C. (1985) A speedrelated kinematic analysis of overground and treadmill walking. In Biomechanics IX-A (Edited by Winter, D. A. et al.), pp. 423-426. Human Kinetics, Champaign. Vaughan, C. E., du Toit, L. L. and Roffey, M. (1987) Speed of walking and forces acting on the feet. In Biomechanics X-A (Edited by Jonsson, B.), pp. 349-354. Human Kinetics, Champaign.
1992. 0
&X-9290/92 1992 Pergamon
S5.00+.00 Press Ltd
TECHNICAL NOTE
STEP LENGTH
AND FREQUENCY EFFECTS ON GROUND FORCES DURING WALKING
REACTION
PHILIPE. MARTIN*and ANTHONY P. MARSH Exercise and Sport Research Institute, Arizona State University, Tempe, AZ 85287-0404, U.S.A. Abstract-It is well established that the speed of walking or running significantly affects ground reaction force (GRF) characteristics. While it is sometimes assumed that the variations in step length (SL) and step frequency (SF) also affect GRF patterns, little documentation of this can be found in the literature. Ten young adults performed overground walking at 1.43 m s-i across a force platform under five SL conditions: preferred SL and SLs that were longer and shorter than the preferred by 5 and 10% of greater trochanter height. The contact time, anteroposterior braking and propulsive force and impulse descriptors, and vertical impulse per step increased systematicany as SL increased. In contrast, vertical peak forces and impulse per meter walked showed little change with SL manipulation. Despite the systematic effect of SL on several GRF descriptors, constraint of SL and SF in gait assessments is not recommended as this would prohibit the evaluation of representative gait kinematics and kinetics. Rather, these results suggest that researchers should report SL and SF data when comparing GRF characteristics between experimental groups or conditions, and should be alert to the association between SL/SF and GRFs when interpreting GRF trends.
INTRODUCTION
METHODS
It is well accepted that the speed of walking or running significantly affects ground reaction force (GRF) characteristics (Andriacchi et al., 1977; Jansen and Jansen, 1978; Munro et al., 1987; Vaughan et al., 1987). Thus, many investigators have chosen to compare experimental groups or conditions at controlled speeds while allowing subjects to freely select step length (SL) and step frequency (SF). It is also possible, however, that systematic differences in SL and SF between groups or conditions could also affect GRF patterns (Soames and Richardson, 1985), which could influence interpretation of GRF findings. As an example, Larish et al. (1988) reported lower peak vertical and anteroposterior GRFs for aged adults than for young adults when walking speed was controlled. One interpretation of this trend provided by Larish et al. is that the older adults select a gait pattern in order to reduce GRF magnitudes and, thus, decrease stress on the musculoskeletal system. They also observed, however, that the older adults had shorter SLs than the young. Thus, the GRF differences between the old and young may simply have been related to the SL differences between the age groups. There is surprisingly little documentation in the research literature on the effect of changes in SL and SF on GRF characteristics, perhaps due to the common practice of allowing subjects to freely choose SL and SF while controlling the speed. Thus, the aim of this study was to examine the effects of SL and SF on selected GRF characteristics during walking. It was hypothesized that increases in SL and decreases in SF at a specific walking speed would result in higher peak forces for both the vertical and anteroposterior components of the GRF.
Ten young adults (5 male, 5 female, R,,=27.6&2.5 y, 8,,, = 169 f 13 cm, 8,, = 624+ 146 N) were used as subjects. Initially each subject’s preferred SL/SF combination at the test speed of 1.43 ms-’ was determined using a motorized treadmill. This speed is within the range of preferred walking speeds of healthy, young adults typically reported in the literature(e.g. Larish et al:, 1988; Laurent and Pailhous, 1986; Murray bt al., 1966: Pearce et al., 1983). Treadmill walkinn was used to determine preferred SL/SF because of the abilit; to control walking speed with precision and the simplicity of determining average SF and SL for multiple step cycles. Comparisons of treadmill and overground locomotion have reflected minor differences in SL and SF (Elliott and Blanksby, 1976; Nelson et al., 1972; Pearce er al., 1983; Taves et al., 1985). Because we expected a systematic SL/SF effect on the GRF and were not specifically interested in the GRF features at the preferred SL/SF combination, minor errors in the estimation of overground preferred SL/SF from the treadmill assessment should have had little or no effect on our ability to test our hypothesis. The treadmill speed was monitored by timing 10 belt revolutions and adjusted as necessary to achieve the desired target speed. Following a minimum of 15 min of treadmill accommodation, each subject’s average preferred SF at 1.43 ms-i was computed from the time required to complete 50 steps. SL was then computed from the deterministic relationship between speed, SL, and SF. Subjects then performed multiple overground walking trials along a 10 m walkway at approximately 1.43 m s- * as the GRF was sampled from an AMTI force platform at 500 Hz. The walking speed was monitored using two photocells spaced 3.20 m apart, the first of which triggered a timer and force platform data collection immediately before the subject contacted the force platform and the second which stopped the timer. Only those trials that were within + 3% of the target speed were considered for analysis. Each subject
Received in finalform12 February 1992.
*Author to whom correspondence should be addressed.
1237
Technical Note
1238
completed five acceptable trials for each of the five randomly ordered SL/SF conditions: his or her preferred SL and SLs that were longer and shorter than the preferred by 5 and 10% of leg length (LL) as represented by greater trochanter height (sample X,,=O.851 kO.069 m). SLs for all trials, including the preferred SL condition, were controlled using tape markers on the floor. The subjects were instructed to walk as normally as possible while using the markers as a guide for heel contact. A potential risk in using this approach is that the subjects were required to target the markers which may have influenced gait kinematics and kinetics. Laurent and Pailhous (1986), however, observed that it was easier for subjects to change SL on the basis of visual marks than to synchronize movement with an auditory signal. In addition, Patla et al. (1989) reported no significant differences in GRF impulse descriptors between preferred SL conditions during walking with and without a visual foot contact cue, lending support to our assumption that targeting effects would be minor. A 60 Hz video camera was used to record a sagittal view of subjects’ walking patterns. This was used later to quantify subjects’ actual SLs as they crossed the force platform to establish the success of SL manipulation. The deviation of the actual SL from the nominal SL was initially measured in meters and subsequently expressed as a percentage of LL on a trial by trial basis. Quantified GRF features included contact time, first and second vertical (V) peak forces, anteroposterior (AP) peak propulsive and braking forces, and V and AP impulses. GRFs and impulses were normalized to the body weight. Impulses were also normalized with respect to the distance walked. A single factor repeated measures ANOVA was used for the statistical analysis of the GRF descriptors.
RESULTS
Video analysis of the actual SLs used by the subjects demonstrated that SL was manipulated effectively in all cases and that there was little difference between the nominal conditions and the actual SLs (Table 1). Initially, GRF characteristics for males and females were compared. Because there were no significant GRF differences between males and females, these data were pooled for subsequent considerations of SL effects. The systematic effects of SL were primarily limited to contact time and the AP GRF descriptors (Table 1, Figs 1 and 2). As expected, these factors increased in magnitude as SL increased. While the first and second vertical peak forces showed little change as SL increased, the vertical impulse per step increased because of the increasing contact time. When normalized per unit of distance traveled, vertical impulse showed a statistically significant SL effect, but there was no systematic nature to the changes in impulse.
DISCUSSION
Clarke et al. (1983) reported increasing peak shank decelerations immediately following foot contact with increasing SL during running, and suggested that the changes in SL at a given speed can affect the amount of shock that must be absorbed by the musculoskeletal system. These results provided circumstantial evidence that increased SL resulted in increased impact forces during running. These observations are consistent with our results for walking, although in-
Table 1. Mean GRF characteristics for five SL/SF combinations during walking at 1.43 m s-i Nominal SL conditions GRF descriptor
- lO%LL
-5%LL
Preferred
+5%LL
+ lO%LL
Actual SL (m)t
0.683 -9.5%LL 0.580 kO.035
0.719 - 5.2%LL 0.611 * 0.039
0.765 0.2%LL 0.630 kO.038
0.802 4.5%LL 0.666 kO.044
0.847 9.8%LL 0.694* f0.044
1.167 *0.070 1.155 +0.041
1.159 kO.065 1.164 kO.051
1.166 + 0.065 1.180 k 0.052
1.187 kO.068 1.176 kO.058
1.185 f 0.074 1.178 *0.060
0.472 * 0.030
0.498 *0.035
0.518 k 0.032
0.552 kO.038
0.579* &0.040
0.691 + 0.020
0.693 kO.019
0.677 kO.015
0.689 kO.018
0.683’ + 0.024
0.194 + 0.030 0.207 *0.031
0.213 f 0.036 0.219 + 0.033
0.224 kO.035 0.238 f 0.032
0.246 kO.037 0.249 f 0.034
0.255* kO.037 0.267* *0.033
0.052 f 0.008 0.077 *0.010
0.061 io.010 0.084 +0.011
0.067 kO.011 0.088 kO.011
0.077 kO.011 0.096 kO.011
0.084* *0.012 0.099’ kO.011
Contact time (s) Vertical force
1st peak (BW) 2nd peak (BW) Impulse (BWsstep-r) (BWsm-‘) Anteriorposterior force Braking peak (BW) Propulsive peak (BW) Impulse (BW s step- I) (BW sm-r)
tActua1 SLs were measured from video records and are presented both as absolute values and as deviations from the preferred SL expressed as a percentage of LL. l Statistically significant SL effect @ < 0.05).
Technical Note
s
1.0
al ;;
0.8
t! Lf 0.6 z .z 0.1 t, > 0.2
---
25
50 % Contact
Pd.SL -1oxLL
75
100
Time
Fig. 1. Ensemble average curves for the vertical component of the ground reaction force for three of the five step length conditions: the preferred step length (Pref. SL) and step lengths either longer or shorter than the preferred by 10% of leg length ( f 10% LL). Step length had no significant effect on the amplitude of the vertical force.
0.3 ,
-0.3
1
0
I
I 25
54 % Contact
75
loo
Time
Fig. 2. Ensemble average curves for the anteroposterior component of the ground reaction force for the preferred step length and step lengths either longer or shorter than the preferred by 10% of leg length. Peak AP braking and propulsive forces and impulses systematically increased as step length increased.
creases in vertical force peaks were also expected with increasing SL. It was not surprising that both V and AP impulse computed per step (BW s step-i) showed a significant effect for SL since impulse is a function of both force and contact time, the latter showing a strong relationship with SL. If one is interested in considering impulse or related momentum changes of the body as a gross indicator of muscular effort during gait, impulse expressed per unit of distance (BW sm-‘) would be more relevant since this would offset the differences between SL conditions in the number of steps required to traverse a given distance. The systematic effect of SL on AP impulse expressed per unit of distance remained because of the strong trend for AP force to increase with increasing SL. V impulse, however, no longer showed the systematic SL effect when impulse was expressed per unit distance since V force changed very little with increasing SL. In our analysis, SL/SF conditions were examined at a single speed of walking so that the effect of speed on GRF characteristics was controlled. In contrast, Soames and Richardson (1985) constrained both SL and SF, but did so in a manner such that multiple walking speeds were produced as they evaluated GRF characteristics. They reported that both SL and SF significantly influenced GRFs, particularly at heel strike, and concluded that constraining only speed is insufficient if GRFs or joint transmission forces for different groups or conditions are being compared. They suggested that cadence should be constrained within unspecified limits
1239
whenever possible. It is unclear whether they were suggesting SF constraint as an alternative to speed constraint or in addition to it. Controlling only SF would make it difficult to interpret GRF trends because it has been shown that SF and speed are highly dependent (Laurent and Pailhous, 1986) and because of speed’s well-established effect on GRF (Andriacchi er al., 1977; Jansen and Jansen, 1978; Munro et al., 1987; Vaughan et al., 1987). A constraint of speed and SF (and, thus, SL); however, would likely prohibit the evaluation of representative gait kinematics and kinetics which is a goal of many analyses. While we favor controlling speed, we do not encourage the constraint of SL and SF in gait assessments. Our results, however, suggest that researchers should also report SL and SF data when comparing GRF characteristics between experimental groups or conditions, and should be alert to the association between SL/SF and GRFs when interpreting GRF trends. REFERENCES Andriacchi, T. P., Ogle, J. A. and Galante, J. 0. (1977) Walking speed as a basis for normal and abnormal gait parameters. J. Biomechanics 10, 261-268. Clarke, T. E., Cooper, L., Clark, D. and Hamill, C. (1983) The effect of varied stride rate and length upon shank deceleration during ground contact in running. Med. Sci. Sports Exert. 15, 170. Elliott, B. C. and Blanksby, B. A. (1976) A cinematographic analysis of overground and treadmill running by males and females. Med. Sci. Sports 8, 84-87. Jansen, E. C. and Jansen, K. F. (1978) Vis-velocitas-via: alteration of foot-to-ground forces during increased speed of gait. In Eiomechanics VI-A (Edited by Asmussen, E. and Jorgensen, K.), pp. 267-271. University Park Press, Baltimore. Larish, D. D., Martin, P. E. and Mungiole, M. (1988) Characteristic patterns of gait in the healthy old. Ann. N. Y. Acad. Sci. 515, 18-32. Laurent, M. and Pailhous, J. (1986) A note on modulation of gait in man: effects of constraining stride length and frequency. Hum. Mvmt Sci. 5, 333-343. Munro, C. F., Miller, D. I. and Fuglevand, A. J. (1987) Ground reaction forces in running: a reexamination. J. Biomechanics 20, 147-155. Murray, M. P., Kory, R. C., Clarkson, B. H. and Sepic, S. B. (1966) Comparison of free and fast speed walking patterns of normal men. Am. J. Phys. Med. 45, 8-24. Nelson, R. C., Dillman, C. J., Lagasse, P. and Bickett, P. (1972) Biomechanics of overground versus treadmill running. Med. Sci. Sports 4, 233-240. Patla, A. E., Robinson, C., Samways, M. and Armstrong, C. J. (1989) Visual control of step length during overground locomotion: task-specific modulation of the locomotor synergy. J. Exp. Psych: Hum. Perception Perform. 15, 603-617.
Pearce, M. E., Cunningham, D. A., Donner, A. P., Rechnitzer, P. A., Fullerton, G. M. and Howard, J. H. (1983) Energy cost of treadmill and floor walking at self-selected paces. Eur. J. appl. Physiol. 52, 115-I 19. Soames, R. W. and Richardson, R. P. S. (1985) Stride length and cadence: their influence on ground reaction forces during gait. In Biomechanics IX-A (Edited by Winter, D. A. et al.), pp. 406410. Human Kinetics, Champaign. Taves, C. L., Charteris, J. and Wall, J. C. (1985) A speedrelated kinematic analysis of overground and treadmill walking. In Biomechanics IX-A (Edited by Winter, D. A. et al.), pp. 423-426. Human Kinetics, Champaign. Vaughan, C. E., du Toit, L. L. and Roffey, M. (1987) Speed of walking and forces acting on the feet. In Biomechanics X-A (Edited by Jonsson, B.), pp. 349-354. Human Kinetics, Champaign.
