Handball is a strenuous contact sport that emphasizes running, jumping, sprinting, throwing, hitting, blocking, and pushing (6). Basic (body mass, body height, and body mass index) and specific (hand) anthropometric characteristics are important to technical and tactical skills (25), but elite performance also demands strength and power in both the upper and the lower limb muscles. An early study by Fleck et al. (5) demonstrated the relationship of handball throwing velocity to upper extremity isokinetic torque. More recently, greater medicine ball throw scores were noted in girls and boys selected for handball teams relative to those who were not selected (14). Throwing performance is thus a key to success (11,21), with maximal isometric strength serving as one laboratory correlate of throwing velocity in both sexes (6,21). Nevertheless, isometric testing is a less than optimal method of evaluating a dynamic act such as handball throwing, and it could underestimate the potential performance of the muscles involved. Others have examined relationships between ball throwing velocity and the strength, power, and bar velocity of the upper extremity as measured in concentric-only bench press exercises (7,8). Concentric-only contractions are again rare during handball, a succession of eccentric-concentric muscle contractions being more typical of normal play. The “pullover” seems the most relevant field test for a handball player.
No one has previously studied the power of the upper and lower limbs in relation to handball throwing; however, there have been a few such studies of swimmers. Swimming depends on the power developed by both the upper and the lower limbs, especially in short distance events. Thus, Wingate Test measurements of power for the upper and the lower limbs were related to speed in a 50-m freestyle swim (9,10). Almuzaini et al. (1) found a strong relationship between the mean power of the lower limbs as determined by the Wingate test and performance in a modified long-jump test. This suggests that Wingate test data could be used to predict anaerobic performance in the field (1,9,10). Likewise, standing long-jump scores (which could be considered a measure of explosive power for the lower limb) have distinguished selected handball players from those who were not selected (14). This last finding underlines the important contribution of the lower limbs to handball performance. In this regard, Fleck et al. (5) noted that handball throwing velocities were 6% smaller for a jump shot than for a set shot. They suggested that during a set shot, the feet were in contact with the floor, allowing the use of the lower extremity muscles to increase ball velocity; plainly, this was difficult if not impossible during a jump shot. The action of javelin throwing bears some similarity to a handball throw, and javelin performance is also related to the peak power developed by both the upper and the lower limbs, as assessed by force-velocity testing (3).
The aim of our study was thus to examine relationships between handball throwing velocity and selected measures of lower and upper limb muscle strength and power. Our hypothesis was that both the upper and lower limbs would make significant contributions to such performance.
Experimental Approach to the Problem
Subjects were given standardized instructions and verbal encouragement in all tests, which were preceded by a standardized warm-up. Our aim was to examine relationships between a move typical of handball play (a radar assessment of 3-step running handball throwing velocity [T3-Step]) and selected measures of upper and lower limb muscle strength and peak power (PP). The PP of the upper and the lower limbs was assessed by appropriate force-velocity tests, using a suitably adapted cycle ergometer; additional data from these tests included maximal force (N) and maximal velocity (rpm). Such assessments are related to javelin performance (3), which bears many similarities with T3-Step. Muscle volume, which is well recognized as a determinant of anaerobic muscle power, was assessed for both upper and lower limbs using an anthropometric method (13). Maximal arm strength was also assessed by 1 repetition maximum (1RM) bench press and pullover exercises, as commonly used in field evaluations of handball players (6-8,16).
The study was reviewed and approved by the Institute's Committee on Research for the Medical Sciences. The coach and parents were informed about the various tests to be performed and the experimental risks. Written informed consent was obtained from players over 20 years of age, and both legal guardian and subject consents were obtained for those under 20 years of age.
All 14 participants were male players in the top National Handball League, and all had trained for 8 or more years (mean ± SD: 8.9 ± 0.8 years), they had also been injury free for 2 or more years before testing. Their physical characteristics are summarized in Table 1. All had experience of upper limb strength training and were familiar with the bench press and pullover techniques. All tests were performed midseason (third week of December), with laboratory visits in the morning. Goal keepers (3 subjects) were excluded from our analyses.
Handball Throwing Test
The maximal handball throw was preceded by a 3-step run. The subject had to keep at least 1 foot in contact with the floor at the moment of throwing. This kind of throw is very similar to that used in javelin competition, and indeed, our coach terms it a javelin handball throw.
The definitive throw was made after a 10-minute standardized warm-up. After applying resin as desired, subjects threw a standard handball (mass 480 g, circumference 0.58 m) as fast as possible toward a standard goal, using a single hand and their personal technique. They took a preparatory run of 3 regular steps before releasing the ball behind a line set 9 m from the goal. T3-Step was measured using a radar Stalker ATS system™ (Radar Sales, Minneapolis, MN, USA) hand held at shoulder level. Preliminary checks of the reliability of the radar system had been made (4) for moving subjects (1-7 m·s−1) and rolling balls (8-22 m·s−1); radar data (vr) over a 3-m distance were compared against velocities measured with photoelectric cells (vpc). The relationship between the 2 estimates (vr = 0.99·vpc + 0.22) was linear and close (r = 0.999) (4). To simulate a typical handball action, players were told to aim for the upper right corner of the goal. The maximal ball velocity was noted for 3 consecutive trials, each separated by at least 15 seconds of recovery. To maximize motivation, players were immediately informed of their performance, and the highest of their 3 values was recorded. The interclass correlation coefficient (ICC) of the 1RMPO is presented in Table 2.
The lower limb force-velocity tests were performed on a standard cycle ergometer (model 894 E, Monark Exercise AB, Vansbro, Sweden). The instantaneous peak velocity at each braking force was used to calculate the corresponding maximal anaerobic power. The maximal velocity (V0) was defined as the highest velocity attained without external loading. Peak power was defined as the power at which additional loading induced a decrease in power output. Parabolic relationships were calculated only if we observed a decline of PP over 2 successive braking forces.
Upper limb tests were made using an appropriately modified version of the same apparatus. The pedals were replaced by hand cranks, and the saddle pillar was removed to avoid injuries. The ergometer was then mounted on a metal support that brought the crankshaft to shoulder level. The unrestrained subjects stood freely in front of the ergometer, with the exception that smaller subjects were allowed to stand on a step.
The relationship between braking force F and velocity V can be expressed by the following equation:
where V0 is the intercept with the velocity axis, that is, the theoretical maximal velocity for a braking force of zero, and F0 is the intercept with the force axis, that is, the theoretical maximal braking force corresponding to a velocity of zero (22). The measured and calculated parameters for both tests included Peak power of the upper (PPUL) and lower (PPLL) limbs, each expressed in Watts, W·kg−1 of total body mass, and W·L−1 of limb muscle volume, and the corresponding maximal forces (F0UL and F0LL) and maximal velocities (V0UL and V0LL).
The force-velocity tests required short all-out sprints (duration about 7 seconds) using a suitable sequence of ergometer braking forces (2). All subjects completed familiarization trials 1 week before the definitive measurements, to avoid learning effects. Definitive data for the upper and lower limbs were collected in randomized order on separate days within the same week. Subjects were verbally encouraged to reach maximal pedaling rate as quickly as possible. The peak velocity was noted, and was used to calculate force-velocity relationships. After a 10-minute standardized warm-up, lower limbs tests began at a braking force equal to 2.5% of the subject's body mass (2). After a 5-minute recovery, the braking was increased to 5, 7.5, 8.5, 9.5, 10.5, and 11.5% of body mass in randomized order. The same sequence was performed again, until an additional load induced a decrease of power at each of 2 repetitions; this value was accepted as the PP. Six to 8 all-out sprints were generally performed in a session. The protocol for the upper limbs was similar, beginning with a braking force equal to 1.5% of the subject's body mass. After a 10-minute warm-up, the braking was increased by 0.5% every bout, until the subject was unable to reach the previous peak of power in 2 successive bouts.
The maximal strength of the upper extremity was assessed by a 1 repetition maximum bench press (1RMBP), an action that involves arm muscles specific to overhand throwing. The protocol was similar to that described previously (12), except that players had to execute a flexion-extension with load (eccentric-concentric muscle contraction) to make a fully concentric contraction. The test was performed in a squatting apparatus; the barbell was attached at both ends, linear bearings on the bars allowing only vertical motion. The bar was initially positioned about 0.2 m above the subject's chest and was supported by the bottom stops of the device. Subjects performed an eccentric followed by a concentric action from the starting position. To ensure consistent positioning of the shoulder and elbow joints throughout, the players held their shoulders in 90°abduction during the concentric contractions. No bouncing or arching of the back was allowed. This technique was familiar to the players, because they had used it in their weekly strength training sessions. The approximate RM value had been assessed during the previous weekly strength training session. Warm-up for the definitive test consisted of 5 repetitions at 40-60% of the pretest RMBP. Thereafter, 4-5 separate single attempts were performed until the subject was unable to extend the arms fully. The highest load with an acceptable extension was noted as the 1RMBP. A 2-minute rest period was allowed between attempts. The ICC for the 1RMBP is presented in Table 2.
This exercise is much like the dumbbell pullover, but intensity is added to the movement by using a barbell. The bar was initially positioned 0.2 m above the subject's chest and was supported by the bottom stops of the device. The subject performed an eccentric and then a concentric contraction from the starting position. The eccentric action took the weight over and behind the individual's head, with the elbow fully extended. At the end of the backward movement, when the upper limbs were approximately parallel to the ground and the elbows were slightly flexed, subjects pushed the barbell to bring it back to the starting position, keeping their abdominal muscles well contracted and their body stable without bouncing or arching of the back. All subjects had used this technique regularly in their weekly strength training sessions. A pretest assessment of RMPO had been made during the previous training session. Warm-up for the definitive test comprised 5 repetitions at loads of 40-60% of the pretest RMPO. Thereafter, 4-5 separate attempts were performed until the subject was unable to extend the arms fully. The highest load with acceptable extension was recorded as the 1RMPO. Two minutes of rest was allowed between trials. The ICC of the 1RMPO is presented in Table 2.
Circumferences and skin-fold thickness at different levels of the thigh and the calf, the arm and the forearm, the length of the lower and upper limb, and the breadth of the humeral and femoral condyles were measured to estimate the muscle volume of the upper and lower limbs, as described previously (13,19,20).
Muscle volumes were estimated as follows:
The total limb volume was estimated as the volume of a cylinder, based on its length (L), corresponding to the distance from the acromion to the minimum wrist circumference for the upper limb, and from the trochanter major to the lateral malleolus for the lower limb, and the mean of 5 limb circumferences (axilla, maximum relaxed biceps, minimum above the elbow, maximum over the relaxed forearm, and minimum above the styloid process for the upper limb, and maximal thigh, midthigh, just below the patella, maximal calf and just above the ankle for the lower limb) according to the following formula:
where ∑C2 is the sum of the squares of the 5 circumferences of the corresponding limb.
Skin folds were assessed using a standard Harpenden caliper (Baty International, Burgess Hill, Sussex, United Kingdom). The fat volume was calculated as follows:
where ∑S is the sum of 3 skin folds for the upper limb (biceps, triceps, and midforearm), or 4 skin folds for the lower limb (front of midthigh, back of midthigh, back of calf, and outside of calf) and n represents the number of skin folds measured on each limb.
Standard equations were used to predict body fat from the biceps, triceps, subscapular, and suprailiac skin-fold readings (26):
where ∑S is the sum of the 4 skin-fold readings (in mm), and a and b are constants dependent on sex and age.
Bone volume was calculated as follows:
where D is the humeral or femoral intercondylar diameter, F is a geometric factor (0.21 for the upper limb or 0.235 for the lower limb), and L is the limb length as measured above.
Variables are expressed as mean ± SD. Pearson product-moment correlations tested relationships between T3-Step and the force-velocity test parameters, upper and lower limb muscle volumes, and 1RM upper limb measurements. p ≤ 0.05 was taken as the limit of significance in all statistical tests. The reliability of the T3-Step, bench press, and pullover measurements was assessed using ICC (18). An ICC over 0.90 is considered as high, 0.80-0.90 is moderate, and values below 0.80 indicate that a physiological field test is inadequate (24).
Intraclass correlation coefficients (0.98-0.99) for T3-Step, 1RMBP, and 1RMPO establish the reliability of these tests (Table 2). Absolute peak power values, as calculated from the force-velocity tests are summarized in Table 3 (respective ranges for upper and lower limbs are 590-1200 and 321-717 W). The corresponding values expressed relative to body mass and to muscle volume were 3.9-6.7 kg−1 and 111.8-184.7 W·L−1 for the upper limbs and 3.9-6.7 W·kg−1, and 71.2-96.8 W·L−1 for the lower limbs.
T3-Step is closely related to the absolute PP of the upper limbs and to the maximal force (F0) calculated from the force-velocity test (r = 0.69, p < 0.01 for both relationships, Table 4 and Figure 1). Moreover, T3-Step is moderately related to the 1RMBP and 1RMPO (r = 0.56, p < 0.05; r = 0.55, p < 0.05, respectively) (Table 4 and Figure 2). T3-Step is also moderately related to absolute PP of the lower limbs and to the maximal force (F0) calculated from the force-velocity test (r = 0.56, p < 0.05; r = 0.62, p < 0.05, respectively, Table 5 and Figure 3).
Our main purpose was to examine relationships between T3-Step and muscle strength and power, examining PP and muscle volumes for both the upper and lower limbs, and also the maximal strength of the arms as seen in the 1RMBP and 1RMPO field tests. The hypothesis that both upper and lower limbs contributed to throwing performance was confirmed. T3-Step was closely related to both PPUL and F0UL (r = 0.69, p < 0.01 for both relationships), and it also showed moderately strong relationships to PPLL and F0LL (r = 0.56, p < 0.05; r = 0.62, p < 0.05, respectively). This seems the first investigation to demonstrate the substantial contribution of the lower limb muscles to the throwing velocity of handball players.
PPUL and PPLL values (Table 3) were relatively weak in comparison with the findings in javelin throwers (3), where PPLL and PPUL averaged 1,208 and 720 W, respectively. On average, our handball players developed about 250 and 300 W less power than the javelin competitors in their upper and lower limb muscles, respectively (3). Given that both subject groups were in the elite category, this difference seems related to the type of sport practiced. Bouhlel et al. (3) noted significant correlations between javelin performance and the PP of both the upper and lower limb muscles, using an identical force-velocity protocol. Their findings seem in agreement with our results, as we found significant correlations between T3-Step and both PPUL and PPLL (Figures 1 and 3). Although V0 was determined during force-velocity testing, our data support the view that in handball PPUL and PPLL are the primary determinants of throwing velocity. When throwing at high velocity, critical factors are the ability to transfer the momentary impulse of power from the lower body to the upper body and then to the ball during the release transmit (17); this ability is closely correlated with competitive performance (23). The upper limb muscles of our subjects showed at least as much hypertrophy as the lower limbs relative to untrained young adults, although the greatest part of the total impulse was derived from the powerful muscles of the lower limbs, and the ability to develop a large force in the lower limbs (FLL0) seems important in this regard. This assumption is supported by the early study of Fleck et al. (5), who noted the greater distance thrown in a set shot, when the feet were in contact with the floor and the lower limbs could be used to increase throwing velocity.
Hawley et al. (10) previously reported a strong correlation between lower body anaerobic power, as measured by the Wingate test, and performance in a 50-m sprint swim (r = 0.76; n = 22). In our study, PPLL was also significantly correlated with T3-Step, although the correlation was only moderate (r = 0.69; p < 0.01). Nevertheless, one would not expect a precise prediction of handball throwing ability from measurements of PPLL during ergometer cranking. Both upper and lower limb cranking are essentially cyclic movements, whereas throwing reflects the power output obtained from a single whole-body movement. Moreover, Hawley et al. (10) found a correlation coefficient of 0.63 between the PP of the upper limbs as evaluated by the Wingate test and the speed of a 50-m swimming sprint. In their study, the ratio of upper limb PP to lower limb PP was 45%. These results seem in accordance with our present study; we found a correlation of 0.69 between PPUL and T3-Step (Table 4 and Figure 1), and in our data the ratio PPUL/PPLL was 52%. In addition, Hawley et al. (10) reported that the 50-m swim speed was more closely correlated with lower limb (r = 0.76) than with upper limb peak power (r = 0.63). In contrast, we found the closer relationship was with PPUL (Tables 4 and 5). This could reflect a difference in the respective contributions of the limbs to these 2 types of event, with the lower limbs making a more decisive contribution to swimming performance. In contrast, the upper limbs appear to have the dominant influence on T3-Step in handball players.
When our results were expressed relatively to body mass and limb muscle volume (i.e., W·kg−1 and W·L−1) (Tables 4 and 5), the relationships to throwing performance became statistically nonsignificant (r = 0.50 and r = 0.34 for the upper limbs; r = 0.27 and r = 0.41 for the lower limbs respectively). This suggests the important contributions of overall body mass and local muscle volume to throwing velocity, implying in turn that body mass and particularly the local limb volume should be considered if we wish to enhance throwing velocity. Others, also, have noted that differences in strength and power between elite and amateur handball players disappear if results are expressed relatively to body mass or fat-free mass (7).
Several recent studies of elite male handball players (7,8,15,16) investigated the relationships of throwing velocity to bar velocity and bar power during bench press or half squat. Gorostiaga et al. (7) reported a close relationship between 3-step running velocity and the bar velocity at 30% of 1RMBP (r = 0.72, p < 0.01), with a moderate relationship to power at 100% of body mass in the half squat exercise (r = 0.62, p < 0.05). A close relationship between standing throwing velocity and 1RMBP (r = 0.80, p < 0.001) was also reported (8). Moreover, throwing velocities showed moderate relationships with the bench press bar velocity and the power achieved at 38, 52 and 52, 67% of body mass, respectively (16). Nevertheless, it is difficult to compare these results with our findings, because of differences in methodology, including the type of ergometer that was used. The studies cited used a rotary encoder linked to the end of the bar to record bar displacement, average velocity and average power of the bar. Moreover, all of these parameters were only assessed during a concentric bench press exercise. In our investigation, we measured the 1RMBP and 1RMPO as dependent variables. In addition, we adopted a simultaneous eccentric-concentric upper limb muscle contraction. To our knowledge, the relationships of T3-Step to 1RMBP and 1RMPO (Figure 2) as measured with a simultaneous eccentric-concentric contraction have not been described previously. Moreover, the pullover exercise, although rarely measured in handball studies, is rather specific to the action of handball throwing. Our results suggest that the simple measurements of 1RMBP and 1RMPO could be useful tools for the handball coach, because these 2 exercises are often used in resistance strength training programs. Regular use of bench press and pullover exercises could form an important component of a resistance training program designed to increase the throwing velocity of handball players. This suggestion merits testing by further prospective research.
One limitation of our study was a relatively small sample size, which may have influenced the magnitude of the correlations that we observed, although the ICCs show that our subjects yielded very consistent performances. Moreover, the upper limb force-velocity tests were performed while the subjects were unrestrained; although this posture is typical of handball throwing, at high braking forces it may have allowed a recruitment of accessory muscles that varied from one subject to another. Furthermore, the mass of upper limb muscle that contributes to handball throwing is small, and it may not correlate closely with the range of muscles solicited during upper limb-cranking exercise. Upper limb cranking is an activity with several degrees of freedom, and indeed the active muscle volume is likely to be substantially greater than that estimated by our anthropometric approach; in particular, our estimates do not include the shoulder muscles. This may explain why the correlation between V0UL and V0LL was weak and not statistically significant. The parallel between the lower limb muscles used in throwing and those used during the lower limb ergometry test is also necessarily somewhat limited.
Within the limitations of our study, the present findings highlight a substantial relationship between the peak power of both upper and lower limb muscles and maximal handball throwing velocity. Thus, knowledge of the maximal forces (F0 and 1RM) developed by the upper and lower limbs could be of practical interest for handball coaches. Force-velocity data require sophisticated laboratory testing, but may find application in regulating conditioning and rehabilitation programs for elite players. Because the lower limbs contribute to throwing velocity, coaches should focus strength and power training programs on both the upper and the lower limbs.
The authors would like to thank Dr. Haj Sassi, R. (Head of the research unit: “Pratiques sportives scolaire et universitaire et performance” ISSEP du Kef) for lending the radar system. Also, we thank the “Ministère de l'enseignement supérieur, de la Recherche Scientifique et de la Technologie, Tunisia” for financial support.
1. Almuzaini, KS and Fleck, SJ. Modification of the standing long jump test enhances ability to predict anaerobic performance. J Strength Cond Res
22: 1265-1272, 2008.
2. Arsac, LM, Belli A, and Lacour, JR. Muscle function during brief maximal exercise: Accurate measurements on a friction-loaded cycle ergometer. Eur J Appl Physiol Occup Physiol
74: 100-106, 1996.
3. Bouhlel, E, Chelly, MS, Tabka, Z, and Shephard, R. Relationships between maximal anaerobic power of the arms and legs and javelin performance. J Sports Med Phys Fitness
47: 141-146, 2007.
4. Chelly SM and Denis, C. Leg power and hopping stiffness: Relationship with sprint running performance. Med Sci Sports Exerc
33: 326-333, 2001.
5. Fleck, SJ, Smith, SL, Craib, MW, Denahan, T, Snow, RE, and Mitchell, MR. Upper extremity isokinetic torque and throwing velocity in team handball. J Appl Sport Sci Res
6: 120-124, 1992.
6. Gorostiaga, EM, Granados, C, Ibañez, J, González-Badillo, JJ, and Izquierdo, M. Effects of an entire season on physical fitness changes in elite male handball players. Med Sci Sports Exerc
38: 357-366, 2006.
7. Gorostiaga, EM, Granados, C, Ibáñez, J, and Izquierdo, M. Differences in physical fitness and throwing velocity among elite and amateur male handball players. Int J Sports Med
26: 225-232, 2005.
8. Granados, C, Izquierdo, M, Ibañez, J, Bonnabau, H, and Gorostiaga, EM. Differences in physical fitness and throwing velocity among elite and amateur female handball players Int J Sports Med
28: 860-867, 2007.
9. Hawley JA and Williams, MM. Relationship between upper body anaerobic power and freestyle swimming performance. Int J Sports Med
12: 1-5, 1991.
10. Hawley, JA, Williams, MM, Vickovic, MM, and Handcock, PJ. Muscle power predicts freestyle swimming performance. Br J Sports Med
26: 151-155, 1992.
11. Hoff, J and Almasbakk, B. The effects of maximum strength training on throwing velocity and muscle strength in female team-handball players. J Strength Cond Res
9: 255-258, 1995.
12. Izquierdo, M, Häkkinen, K, Gonzalez-Badillo, JJ, Ibáñez, J, and Gorostiaga, EM. Effects of long-term training specificity on maximal strength and power of the upper and lower extremities in athletes from different sports. Eur J Appl Physiol
87: 264-271, 2002.
13. Jones, PR and Pearson, J. Anthropometric determination of leg fat and muscle plus bone volumes in young male and female adults. J Physiol
204: 63-66, 1969.
14. Lidor, R, Falk, B, Arnon, M, Cohen, Y, Segal, G, and Lander, Y. Measurement of talent in team handball: The questionable use of motor and physical tests. J Strength Cond Res
19: 318-325, 2005.
15. Marques, MC and González-Badillo, JJ. In-season resistance training and detraining in professional team handball players. J Strength Cond Res
20: 563-571, 2006.
16. Marques, MC, van den Tillaar, R, Vescovi, JD, and González-Badillo, JJ. Relationship between throwing velocity, muscle power, and bar velocity during bench press in elite handball players. IJSPP
2: 414-422, 2007.
17. Morriss, C and Bartlett, R. Biomechanical factors critical for performance in the men's javelin throw. Sports Med
21: 438-446, 1996.
18. Rankin, G and Stokes, M. Reliability of assessment tools in rehabilitation: An illustration of appropriate statistical analyses. Clin Rehabil
12: 187-199, 1998.
19. Shephard, RJ, Bouhlel, E, Vandewalle, H, and Monod, H. Muscle mass as a factor limiting physical work. J Appl Physiol
64: 1472-1479, 1988.
20. Shephard, RJ, Vandewalle, H, Bouhlel, E, and Monod, H. Sex differences of physical working capacity in normoxia and hypoxia. Ergonomics
31: 1177-1192, 1988.
21. van den Tillaar, R and Ettema, G. Effect of body size and gender in overarm throwing performance. Eur J Appl Physiol
91: 413-418, 2004.
22. Vandewalle, H, Peres, G, Heller, J, Panel, J, and Monod, H. Force-velocity relationship and maximal power on a cycle ergometer. Correlation with the height of a vertical jump. Eur J Appl Physiol Occup Physiol
56: 650-656, 1987.
23. Viitasalo, J, Mononen, H, and Norvapalo, K. Release parameters at the foul line and the official result in javelin throwing. Sports Biomech
2: 15-34, 2003.
24. Vincent, WJ. Statistics in Kinesiology
. Champaign, IL: Human Kinetics, 1995.
25. Visnapuu, M and Jürimäe, T. Handgrip strength and hand dimensions in young handball and basketball players. J Strength Cond Res
21: 923-929, 2007.
26. Womersley, J and Durnin, JV. An experimental study on variability of measurements of skinfold thickness on young adults. Hum Biol
45: 281-292, 1973.