Journal Logo

Applied Sciences: Physical Fitness and Performance

Leg power in young women: relationship to body composition, strength, and function


Author Information
Medicine & Science in Sports & Exercise: October 1996 - Volume 28 - Issue 10 - p 1321-1326
  • Free


Power is the product of force and the velocity of movement. The term “speed-strength” has also been used as a synonym for power(33). Sporting events such as sprinting, jumping, throwing, or kicking demand high velocity movements combined with high force generation, requiring muscles to generate high absolute power(33). Some daily activities such as lifting heavy objects, climbing stairs, rising from a chair, and doing heavy housework also require individuals to be above a minimum threshold of muscle power output(4,15).

There is controversy in the literature as to the maximum external resistance against which maximum muscle power can be generated(2,18,25,31,33). As external resistance increases, power output may be reduced owing to the progressive slowing of contraction velocity. Typically, progressive resistance training programs have used external resistances of 80% or more of an individual's one repetition maximum (1-RM) to optimally increase muscle strength and size(22). However, it is not clear whether improvements in muscle power are achieved at similar training resistances. Determining the appropriate external resistance for maximum power may be useful in establishing the appropriate training stimulus to improve muscle power. Therefore, the purpose of this study was to determine the external resistance(as a percentage of maximum dynamic strength (% 1-RM)) at which maximum power is generated in sedentary young women during leg extension exercise.

Since previous laboratory tests of leg power have also been related to differences in body composition (fat free and muscle mass)(1,12,21) and other field tests of muscle power(3-6,9,20,29), measures of body composition and other tests of muscle power were conducted in this study to assess their relationship with maximum power obtained during dynamic leg extension. Sedentary women were studied exclusively since almost all previous data on power output have been generated using male subjects.


Subjects. Nineteen healthy women (age range 21-29 yr) participated in this study. The subjects were not participating in any regular exercise program. Each subject filled out a medical history questionnaire, an activity questionnaire, and signed an informed consent form. This protocol was approved by the Boston University and the Tufts University-New England Medical Center Human Investigation Review Committees.

Experimental design. The subjects were active but not engaged in any regular exercise program and were instructed not to exercise on any of the testing days. They reported to the laboratory on three separate days. The subjects fasted on the night before day 1. After their body composition was determined by hydrostatic weighing as described below on day 1, they underwent the following tests: leg extensor power, habitual gait velocity, maximal gait velocity, double leg press 1-RM, and double leg press power. A minimum of 10 min of rest was allowed between each of these tests. Day 2 testing was performed at least 1 wk after day 1. On day 2 the subjects performed the following tests: leg extensor power rig, habitual gait velocity, maximal gait velocity, double leg press 1-RM, and double leg press power. Within 5 d of the second day of testing, the subjects also performed the Wingate anaerobic power test, the vertical jump test, and the 40-yard dash. This order of testing was chosen to minimize fatigue between tests and because the subjects had to be transported to another location for the Wingate, vertical jump, and 40-yard dash measurements.

Double leg press power test (KP). This power test was conducted on a computer interfaced pneumatic double leg press machine specifically modified for this study (Keiser Sports Health Equipment Inc., Fresno, CA). The individual's 1-RM was determined by incrementally increasing the force against which the subject pushed until she was unable to perform the lift through her full range of motion (24). The highest force that was generated before failure occurred was determined as her 1-RM. After a 5-min rest period, the force was set at 34% of the subject's 1-RM and a double leg press was performed as fast as possible through the full range of motion(ROM). The test was repeated with the force increased by 5% increments. The power test was performed once at each force setting.

To determine range of motion, the subject fully extended her legs under no load. During each leg press, achievement of full ROM was visually verified by the investigator. Each subject held on to the handles and her back was supported by a seat. A 1-min rest was given between each leg press. For each lift performed, the computer interface calculated work and average power by sampling the pressure at the cylinder 400 times·s-1 during the movement and recorded the distance traveled by the piston. The algorithm looks at the outgoing stroke (leg extension), which is defined as the distance between the low and high force output points between each single repetition. Work was calculated by multiplying the distance moved by the pressure at the end of this calculation period and a constant that converts this value to Joules/256. The velocity of movement was calculated using ultrasonic transducers attached to the air cylinders. To avoid discrepancies and noise encountered at the very beginning and end of the movement, the computer interface calculated the mean power generated from 5% to 95% of the piston movement. This system is unique because mean power output can be measured during a standard muscle strengthening exercise, the double leg press. This test was performed twice with a 1-wk interval between tests. The coefficient of variation was 4% (intraclass correlation coefficient: 0.967), which was determined from the maximal power output value, irrespective of the force setting, achieved during the incremental power test from each week.

Leg extensor power rig (LP). The leg extensor power was determined using a leg extensor power rig (University of Nottingham Medical School, Nottingham, UK), which has been described previously(5). Briefly, the power rig has a seat and a footplate connected to a flywheel through a lever (5). The subject sat with a slight knee bend to determine the seat setting; the seat was then locked in place. The subject was then asked to push the foot plate as hard and as quickly as possible with her hands across her chest. Leg extensor power was determined as the highest of five trials performed 30 s apart. This test was performed twice with a week in between. The coefficient of variation for this test in our laboratory was 13% (intraclass correlation coefficient: 0.916), which was determined by taking the maximal power output value from each week.

Habitual gait and maximal gait velocity. Gait velocity was determined using an Ultratimer (DCPB Electronics, Glasgow, Scotland). The subject walked at her normal walking pace for approximately 10 m to assess habitual gait velocity. She was also asked to walk as quickly as possible without running to determine maximal gait velocity. Each of these tests was repeated five times. The average of the five normal speed trials was considered the habitual gait velocity, and the maximal value obtained from the five maximal trials was considered the maximal gait velocity. Values are reported as m·s-1. The precision for measuring gait velocity was±1%. The coefficient of variation for habitual gait velocity is 8%(intraclass correlation coefficient: 0.977) and was determined from the ten trials from both weeks. Maximal gait velocity coefficient of variation is 6%(intraclass correlation coefficient: 0.596) and was determined from the fastest time from both weeks.

Wingate anaerobic power test. The Wingate test required each subject to perform 30 s of supramaximal exercise on a bicycle ergometer(3). She pedaled as fast as possible under zero resistance; the resistance was then quickly set (0.075 × body weight(kg)), and she was encouraged to exercise as hard as she could for 30 s. The cycle ergometer (Monark, Stockholm, Sweden) was interfaced to a personal computer using reflective tape placed on the flywheel and a light sensor. The mean power and absolute power for each 5-s interval was calculated using software developed specifically for this test (Sports Medicine Industries, St. Cloud, MN). Subjects performed this test twice separated by a 15-min recovery period. The coefficient of variation is 7% (intraclass correlation coefficient: 0.939), which was determined from the maximal power output from both trials.

Vertical jump test. This test was performed using a Vertec (M-F Athletic Company, Cranston, RI) stadiometer to determine the vertical jump height (1). As the subject stood under the Vertec with both arms extended up, her standing reach was determined. She then assumed a bent-knee, arms-back position and was instructed to jump as high as possible, without taking any steps, and hit the plastic sticks(17). The height of the highest plastic stick she hit was determined. The difference between the highest value and the standing reach value was her vertical jump. The highest value out of three trials was taken. The coefficient of variation for this field test is 5% (intraclass correlation coefficient: 0.947), determined from the values obtained from three jumps.

40-Yard dash. This test was performed on Astroturf at Boston University's Nickerson Field. Subjects started in a three-point stance at the goal line and ran as fast as possible until they passed the 40-yard line. Every run was timed to the nearest 0.01 s using a handheld stopwatch. Each subject was given three attempts with a 1-min break, and the fastest time was chosen (21). The coefficient of variation for this field test is 2% (intraclass correlation coefficient: 0.994) and was determined from the values obtained from three trials.

Body composition. Body composition was determined using hydrostatic weighing. A Sauter scale (Model K120, Denshore Scale, Holbrook, MA) interfaced with a/d converter and computer (Hewlett Packard HP3468A and HP71B) was used to measure and record the hydrostatic weight. Density was determined using the equation of Brozek (10). Residual volume was estimated from age, height, and gender (30). To determine hydrostatic weight, each subject sat in a bent forward position with as much air as possible exhaled from her lungs. Five trials were run, and the values were averaged to determine the underwater weight. Fat free mass(FFM) was then determined using the Siri equation (27). The coefficient of variation for this test in our laboratory is 2%.

Habitual physical activity. Energy expenditure during physical activity was estimated by the subjects' responses to the Harvard Alumni Health Questionnaire. Calculation of weekly energy expenditure during physical activity (kcal·wk-1) was determined as previously reported(26).

Statistical analysis. Data are presented as the mean ± SEM. All data were analyzed using paired t-tests or analysis of variance when appropriate. When F statistics indicated significance, mean differences were located using appropriate contrasts. In addition, Pearson's linear correlation coefficient was used to express univariate relationships between measures of interest. A multiple regression model to predict maximum power during the double leg press was constructed using forward and backward stepwise multiple regression. Statistical significance was set at P < 0.05. All data were analyzed using Statview 4.1 and Super ANOVA (Abacus Concepts, Berkeley, CA) on an Apple Macintosh personal computer (Apple Computers, Cupertino, CA).


Subjects. Descriptive characteristics and mean data from the power tests and functional measurements are presented inTables 1 and 2. The subjects reported expending 1521 ± 178 kcal·wk-1 in physical activity.

Test-retest reproducibility. There was no significant difference between the first and second trials on the LR (P < 0.221), mean power output (P < 0.903), and maximum power outputs (P< 0.086) on the Wingate anaerobic power test. However, there was, as expected, a significantly higher 1-RM during the second KP test (test 1 = 619.87 ± 25.28, test 2 = 653.80 ± 28.81, P < 0.014). For all tests the highest value for each subject was used, except for habitual gait velocity in which the mean value was used.

Double leg press power output.Fig. 1 illustrates power output versus the percentage of the 1-RM at which it was generated using KP. The highest power output was recorded during KP at 68% of the 1-RM (404 W ± 22, 9.0 ± 0.3 W·kg of FFM). Power output was not statistically different between repetitions performed from 56% to 78% of the 1-RM. However, power output was significantly lower at 34%(P < 0.001), 40% (P < 0.001), 45% (P < 0.001), 50% (P < 0.040), 84% (P < 0.001), and 89%(P < 0.001) compared with 68% of the 1-RM.

Comparison of KP and LR. A paired comparison revealed a trend for the LR test to be higher than the KP test (P < 0.055). Rank-ordered correlation showed an association between KP and LR when expressed per kg LBM (Rho = 0.565, P < 0.016) as shown inFig. 2.

KP and subject characteristics. As a secondary objective, maximum power during the KP was compared with functional tests and other subject characteristics. There was a significant relationship between FFM and KP power(R2 = 0.543, P < 0.001). There was also a significant relationship between the 1-RM and KP power (R2 = 0.584, P< 0.001). Standing height was also related to KP power (R2 = 0.336,P < 0.009). The univariate predictors of KP power (FFM, 1-RM, and height) were entered into a forward stepwise multiple regression model. The variables remaining in the final model were the 1-RM and height (R2 = 0.718, P < 0.001). Backward stepwise regression produced the same model.

KP and functional performance tests. Because of the strong association between FFM and several measures of power, the data for power were normalized per kilogram of FFM to examine the residual relationship between power and functional performance after adjusting for variance in FFM. Mean(R2 = 0.194, P < 0.06) power output was nearly, and maximal (R2 = 0.30, P < 0.015) power output during the Wingate test was associated with KP power. There was also relationship between the vertical jump and KP power (R2 = 0.54, P < 0.001). However, there were no relationships between 40-yard dash time (R2 = 0.02, P < 0.57), habitual gait velocity (R2 = 0.002,P < 0.87), or maximal gait velocity (R2 = 0.14,P < 0.12) and KP power.


The major focus of this study was to examine power output during the double leg press while resisting a load set at varying percentages of maximum strength (1-RM) to determine where in this range maximum power is generated. We observed that maximum power output occurred at 56-78% of the 1-RM. This is higher than the range for peak power that has been reported for training explosive power (25). Previous studies have recommended training for muscle power development at 30% of maximum force. Berger(8) reported that vertical jump performance was significantly improved when subjects performed squat exercises with 50-60% of the 10 RM load (approximately 30% of the 1-RM) compared with squat exercises using the 10-RM load (approximately 70% of the 1-RM), jump training, or static isometric training. In addition, Kaneko et al. (18) reported that the greatest improvements in peak power occurred when their subjects trained the elbow flexors at 30% of peak isometric force compared with training at 60% and 100% of peak force. Studies using isokinetic training(11,14,19) have demonstrated that training at high angular velocities can induce specific increases in power production at high angular velocities, suggesting that muscle power improvements may be optimized during low force high velocity training. However, Behm and Sale(7) observed similar increases in peak torque of the ankle dorsiflexors in both ankles when individuals trained one ankle isometrically and the other at low resistance but with high velocity ballistic dorsiflexion exercise. The study by Behm and Sale (7) was unique because the isometric and concentric training were performed at high rates of force development.

The question arises as to what the optimum external resistance that should be used to specifically train muscle power at maximum voluntary velocity is. It is possible that power development could be maximized by training at forces between 56% and 78% of the 1-RM and that training against lower forces may not allow the individual to generate maximum power. However, some caution needs to be used in extrapolating the results of this study of young untrained women to more athletic or physically active populations. Untrained individuals may achieve their peak power at a higher percentage of their maximum strength owing to differences in muscle recruitment and other neural factors(23). In addition, the effects of strength training interventions on power generation at increasing percentages of the 1-RM are unknown. Delecluse et al. (13) reported significant improvements in initial acceleration and maximum speed during a 100-m sprint in male physical education students who had performed high velocity“plyometric” training compared with a group of sprinters who had performed resistance training at a high velocity at their 10-RM, validating the need for high velocity low resistance training to maximize sprint performance. Wilson et al. (32) have confirmed that squat training with loads equaivalent to 30% of isometric force significantly improves performance in several selected performance measures compared with individuals performing conventional strength training at 80-90% of maximum but at a slow velocity and individuals performing “nonloaded” plyometric training.

A second focus of this study was to compare peak power during the KP with a previously validated test of leg extensor power (LR) (5). The LR has been used as a measure of lower limb power and has been previously related to functional tasks in young and old individuals(4,6). Given that the two tests involve similar muscle groups during leg extension, we thought that mean power output during both tests would be closely related. Although the absolute power values from the two tests were significantly different, power·kg-1 FFM in the KP test was related to power·kg-1 FFM in the LR test when a rank-ordered correlation analysis was used. Several factors may explain the poor absolute agreement between these seemingly related measures. First, the LR measures the mean power output through the full range of motion. Because of sampling noise at the very beginning and end of each stroke using the double leg press power test, the computer interface calculates the mean power for each repetition from 5% to 95% of the movement. Another potential difference between the two devices is the force application against which the individual extends her leg. The LR test requires the subject to overcome the initial inertia of the flywheel to which it is attached, whereas the KP test uses a pneumatic piston and has no initial inertia to overcome. Finally, the LR tests the individual's ability to generate power internally against a relatively low resistance and the double leg press power test is performed against a high external resistance, which may limit the velocity able to be generated.

In a univariate analysis, KP power was related to the 1-RM, FFM, and stature. However, using a multivariate-model, we found that only the 1-RM and stature were independent predictors of double KP power. The strong association between the 1-RM and the double leg press power indicates that a large component of the variance in double leg press power can be explained by differences in maximum strength. However, the 1-RM was not related to power on the LR because velocity was the determining factor for power generation on the LR. Previous studies (5,16) have shown a similar relationship between mean maximal power and measures of maximal strength, both of which are influenced by muscle mass, motor unit activation, and muscle energy metabolism. Stature was also a significant predictor of double leg press power, but this could simply be a result of the very strong association between height and body mass (R2=0.437, P < 0.002), which is related to muscle mass.

The relationship between the vertical jump height and KP power was strong because power development during the vertical jump depends on the quantity and efficiency of force production at the hip, knee, and ankle joints(1). Ashley and Weiss (1) observed that force and power development during squat weight training exercises were highly correlated with vertical jump height in young women. In addition, other laboratory measures of muscle power have been shown to be associated with vertical jump performance (5,21). This study extends these findings to associate vertical jumping ability with maximal power development during leg extension. In addition, it further supports the utility of the vertical jump test as a suitable field measure for evaluating leg power.

KP power was also related to mean and maximum power output on the Wingate test. Maximum (5-s) power during the Wingate test was more closely related to double leg press power, probably because of the exclusion of the lower mean power generated at the beginning and the end of the 30-s Wingate test. These results agree with previous studies (3,28) that have shown agreement between the Wingate test and functional measures of anaerobic power such as sprinting and jumping ability.

A final result of the present study was the lack of a relationship between the 40-yard dash and the KP test. Differences in sprinting efficiency, along with lack of proper technique necessary for success in the 40-yard dash, may have made this test a poor predictor of leg power in comparison with more standardized and controlled laboratory tests. Because the subjects were sedentary and not trained as sprinters, the time to run 40 yards may not be representative of leg extensor power.

In summary, peak power during the KP occurs between 56% and 78% of the 1-RM and occurs at higher relative forces than has been described using testing devices in which the external load is not set and maximal force must be generated internally by the subject. Of practical significance is that individuals using our results can perform and prescribe training at maximum power as a percentage of maximum dynamic strength (56-78% of the 1-RM). Maximum power during the KP was related to the vertical jump height and maximum strength (1-RM), suggesting that strength is a key component of power development and that the vertical jump is a suitable field test for evaluating explosive leg power. These data also suggest that training interventions that incorporate activities in this range of power development need to be evaluated with regard to their effects on enhancing or increasing muscle power production. Future studies should address the functional importance of muscle power development in athletic and nonathletic populations of different ages.

Figure 1-Plot of power output during double leg press vs the percentage of the 1-RM. *Indicates significantly different from 67% of the 1-RM (:
P < 0.001); **Indicates significantly different from 67% of the 1-RM ( P < 0.040).
Figure 2-Plot of KP power vs LP power normalized per kg fat free mass (FMM) (Rho = 0.565, :
P < 0.016).


1. Ashley, C. D. and L. W. Weiss. Vertical jump performance and selected physiological characteristics of women. J. Strength Condit. Res. 8:5-11, 1994.
2. Atha, J. Strengthening muscle. Exerc. Sport Sci. Rev. 9:1-73, 1981.
3. Bar-Or, O. The Wingate Anaerobic Test: An update on methodology, reliability, and validity. Sports Med. 4:381-394, 1987.
4. Bassey, E. J., M. A. Fiatarone, and E. F. O'Neill. Leg extensor power and functional performance in very old men and women.Clin. Sci. 82:321-327, 1992.
5. Bassey, E. J. and A. H. Short. A new method for measuring power output in a single leg extension: feasibility, reliability and validity.Eur. J. Appl. Physiol. 60:385-390, 1990.
6. Bassey, E. J., G. Tay, and F. West. A comparison between power output in a single leg extension and in weight-bearing activities of brief duration such as stair-climbing in man. J. Physiol. 427:12P, 1990.
7. Behm, D. G. and D. G. Sale. Intended rather than actual movement velocity determines velocity-specific training response. J. Appl. Physiol. 74:359-368, 1993.
8. Berger, R. A. Effects of dynamic and static training on vertical jumping ability. Res. Quart. 34:419-424, 1963.
9. Bosco, C., P. V. Komi, J. Tihanyi, G. Fekete, and P. Apor. Mechanical power test and fiber composition of human leg extensor muscles. Eur. J. Appl. Physiol. 51:129-135, 1983.
10. Brozek, J., F. Grande, J. T. Anderson, and A. Keys. Densitometric analysis of body composition: revision of some quantitative assumptions. N. Y. Acad. Sci. 110 Part 1:113-140, 1963.
11. Coyle, E. F., D. C. Feiring, T. C. Rotkis, et al. Specificity of power improvements through slow and fast isokinetic training.J. Appl. Physiol. 51:1437-1442, 1981.
12. Davies, C. T. M. Human power output in exercise of short duration in relation to body size and composition. Ergonomics 14:245-256, 1971.
13. Delecluse, C., H. V. Coppenolle, E. Willems, M. V. Leemputte, R. Diels, and M. Goris. Influence of high resistance and high velocity training on sprint performance. Med. Sci. Sports Exerc. 27:1203-1209, 1995.
14. Ewing, J. L., D. R. Wolfe, M. A. Rogers, M. L. Amundson, and G. A. Stull. Effects of velocity of isokinetic training on strength, power, and quadriceps muscle fiber characteristics. Eur. J. Appl. Physiol. 61:159-162, 1990.
15. Fleming, B. E., D. R. Wilson, and D. R. Pendergast. A portable easily performed muscle power test and its association with falls by elderly persons. Arch. Phys. Med. Rehabil. 72:886-889, 1991.
16. Garhammer, J. Power production by Olympic weightlifters. Med. Sci. Sports 12:54-60, 1980.
17. Harman, E. The importance of testing power output.NSCA J. 13:72-73, 1991.
18. Kaneko, M., T. Fufuchimoto, H. Toji, and K. Suei. Training effects of different loads on the force-velocity relationship and mechanical power output in human muscle. Scand. J. Sports Sci. 5:50-55, 1983.
19. Lesmes, G. R., D. L. Costill, E. F. Coyle, and W. J. Fink. Muscle strength and power changes during maximal isokinetic training.Med. Sci. Sports Exerc. 10:266-269, 1978.
20. Margaria, R., P. Aghemo, and E. Rovelli. Measurement of muscular power (anaerobic) in man. J. Appl. Physiol. 21:1662-1664, 1966.
21. Mayhew, J. L., M. G. Bemben, M. G. Rohrs, and D. A. Bemben. Specificity among anaerobic power tests in college female athletes.J. Strength Condit. Res. 8:43-47, 1994.
22. McDonaugh, M. and C. Davies. Adaptive response of mamalian skeletal muscle to exercise with high loads. Eur. J. Appl. Physiol. 52:139-155, 1984.
23. Moritani, T. and A. Devries. Neural factors versus hypertrophy in the time course of muscle strength gain. Am. J. Phys. Med. 58:115-130, 1979.
24. Nelson, M. E., M. A. Fiatarone, C. M. Morganti, I. Trice, R. A. Greenberg, and W. J. Evans. Effects of high intensity strength training on multiple risk factors for osteoporotic fractures. JAMA 272:1909-1914, 1994.
25. Newton, R. U. and W. J. Kraemer. Developing explosive muscle power: implications for a mixed method training strategy. J. Strength Condit. Res. 20-31:1994.
26. Paffenbarger, R. S., A. L. Wing, and R. T. Hyde. Physical activity as an index of heart attack risk in college alumni.Am. J. Epidemiol. 108:161-175, 1978.
27. Siri, W. The Gross Composition of the Body. New York: Academic Press, 1956.
28. Tharp, G. D., R. K. Newhouse, L. Uffelman, W. Thorland, and G. O. Johnson. Comparison of sprint and run times with performance on the Wingate Anaerobic Test. Res. Q. Exerc. Sport 56:73-76, 1985.
29. Viitisalo, J. T. Evaluation of explosive strength for young and adult athletes. Res. Quart. 59:9-13, 1988.
30. Wilmore, J. H. A simplified method for determination of residual lung volume. J. Appl. Physiol. 27:96-100, 1969.
31. Wilson, G., R. Newton, A. Murphy, and B. Humphries. The optimal training load for the development of dynamic athletic performance.Med. Sci. Sports Exerc. 25:1279-1286, 1993.
32. Wilson, G. J., R. U. Newton, A. J. Murphy, and B. J. Humphries. The optimal training load for the development of dynamic athletic performance. Med. Sci. Sports Exerc. 25:1279-1286, 1993.
33. Young, W. B. and G. E. Bilby. The effect of voluntary effort to influence speed of contraction on strength, muscular power, and hypertrophy development. J. Strength Condit. Res. 7:172-178, 1993.


©1996The American College of Sports Medicine