Scand J Med Sci Sports 2012: ••: ••–•• © 2012 John Wiley & Sons A/S

doi: 10.1111/j.1600-0838.2012.01503.x

Upper-body kinematics in team-handball throw, tennis serve,

and volleyball spike
H. Wagner1, J. Pfusterschmied1, M. Tilp2, J. Landlinger1, S. P. von Duvillard3, E. Müller1
1
Department of Sport Science and Kinesiology, University of Salzburg, Rif, Salzburg, Austria, 2Institute of Sport Science, University of
Graz, Graz, Styria, Austria, 3Department of Physical Performance, Norwegian School of Sport Sciences, Oslo, Norway
Corresponding author: Herbert Wagner, PhD, Department of Sport Science and Kinesiology, University of Salzburg, Rifer Schlossallee
49, A-5400 Hallein/Rif, Austria. Tel: +43 662 8044 4887, Fax: +43 662 8044 615, E-mail: [email protected]
Accepted for publication 15 June 2012

Overarm movements are essential skills in many different target. Significant differences (P < 0.05) between the three
sport games; however, the adaptations to different sports overarm movements were found in 17 of 24 variables. The
are not well understood. The aim of the study was to order of the proximal-to-distal sequencing was equal in
analyze upper-body kinematics in the team-handball the three analyzed overarm movements. Equal order of
throw, tennis serve, and volleyball spike, and to calculate the proximal-to-distal sequencing and similar angles in
differences in the proximal-to-distal sequencing and joint the acceleration phase suggest there is a general motor
movements. Three-dimensional kinematic data were pattern in overarm movements. However, overarm move-
analyzed via the Vicon motion capturing system. The ments appear to be modifiable in situations such as for
subjects (elite players) were instructed to perform a throwing or hitting a ball with or without a racket, and
team-handball jump throw, tennis serve, and volleyball due to differences at takeoff (with one or two legs).
spike with a maximal ball velocity and to hit a specific

Overarm movements are essential skills in different sport In previous studies, it was found that a transfer of
games. In team handball and water polo, players use momentum from proximal to distal is important to maxi-
different throwing techniques to score goals; in volley- mize performance in javelin (Whiting et al., 1991), base-
ball, beach volleyball, and fistball, players hit the ball to ball (Hong et al., 2001), team-handball throw (Van den
score points; in baseball, softball, and cricket, the Tillar & Ettema, 2009), and tennis serve (Marshall &
pitcher/baller will throw the ball at different speeds Elliott, 2000), whereas Wagner et al. (2012) found sig-
and/or locations to confuse the hitter; in American foot- nificant differences in the proximal-to-distal sequence of
ball, the quarterback throws the ball over long distances the maximal joint movements (angular velocities) among
to his receivers to gain yards; and in tennis, badminton, different skill levels in team-handball throwing. A posi-
and squash, serves and/or smashes are used to induce tive influence of pelvis, trunk flexion and rotation, shoul-
pressure on the opposing player to score points. These der internal rotation, as well as elbow extension to the ball
different overarm techniques for throwing or hitting velocity was shown in the baseball pitch (Fleisig et al.,
(with or without a racket) differ either because of the 1999; Hong et al., 2001; Stodden et al., 2001), team-
rules of each sport, ball size, and weight or uncontrolled handball throw (Van den Tillar & Ettema, 2004, 2007;
attack strategies of the opposing defensive players. Wagner et al., 2010), volleyball spike (Coleman et al.,
However, these overarm movements may also be similar, 1993), and tennis serve (Elliott et al., 1995; Marshall &
especially in the upper-body kinematics between throw- Elliott, 2000). However, the shoulder internal rotation
ing and hitting a ball (with or without a racket). If there angular velocity was identified as the major contributor to
are similarities in these different overarm movements, it the ball velocity and racket speed (Escamilla & Andrews,
may be possible to identify general motor patterns of 2009). Kinematic comparisons between overarm move-
overarm movements that can be adapted to the different ments in different sports were made between baseball
sports. The answer to these questions could add impor- pitching and football passing (Fleisig et al., 1996), as well
tant information to understanding of the transfer of as between team-handball throwing and volleyball
movements in the motor learning process and to provide spiking (Bergun et al., 2009). It was found that the differ-
a reference point that is reasonable to practice different ent movements were similar but not identical (Fleisig
overarm movements especially for adolescent or prepu- et al., 1996; Bergun et al., 2009). However, in the study of
bescent athletes. Bergun et al. (2009), kinematics were determined using

1
Wagner et al.
two cameras, measuring with 50 fps which is insufficient takeoff from the left leg (right-handed player) or the right leg
to analyze such dynamic movements to calculate angles (left-handed player). The jump throw was selected because this
throwing technique is the most frequently applied throwing tech-
or angular velocities. Although studies exist that analyze nique (about 75%) during the game (Wagner et al., 2008). To
kinematics and performance in team-handball throw (Van confirm that the jump throw was a vertical jump throw, the hori-
den Tillar & Ettema, 2004, 2007; Wagner et al., 2010), zontal difference between takeoff and landing may not exceed
tennis serve (Elliott et al., 1995; Marshall & Elliott, more than 2 m. The subjects were instructed to throw the ball
2000), and volleyball spike (Coleman et al., 1993), (International Handball Federation Size 3) at a target from an 8-m
distance, and to hit the center of a 1 ¥ 1 m2 at about eye level
studies comparing these movements in detail (angles, (1.75 m), with maximum ball velocity. The subsequent evaluation
angular velocities, and their timing) under similar condi- was used only for those throws that hit the target. Each subject had
tions (game specific movements of elite players on offi- to continue to throw until five valid jump throws were achieved.
cial courts) are missing. Team handball, tennis, and Because the distance between the center of the ball and the hand
volleyball were selected because they represent typical increases abruptly at the ball release (Van den Tillar & Ettema,
2007; Wagner et al., 2010), to determine the moment of ball
overarm movements of throwing or hitting a ball (with or release, the distance between the center of the ball and the hand
without a racket) and where ball velocity is the main (head of the second metatarsal) was calculated.
performance determining variable and because these
sports are played (mostly professionally) all over the
world and are part of traditional Olympic Games. Tennis serve
The aims of the study were: (a) to analyze upper-body After a general warm-up, the subjects could perform as many
three-dimensional (3D) kinematics (trunk flexion practice strokes and services as needed to familiarize themselves
with the testing requirement. To enable optimal conditions, the
and rotation, shoulder flexion and internal rotation, as
subjects used their own rackets. During testing, the subjects were
well as elbow flexion) and pelvis rotation in the team- instructed to serve the ball (flat serve) from left to right into the left
handball jump throw, tennis serve, and volleyball spike side of the service field, with maximum ball velocity. Only those
of elite players; and (b) to compare the differences in the serves that hit the service field were used for the analysis. Each
proximal-to-distal sequence of the maximal joint move- subject had to continue to serve until five flat serves were achieved.
Ball contact was defined as the point where the first ball–racket
ments (angular velocities) as well as maximal angles and
contact occurred. This point was identified with a Basler digital
angular velocities (their timing) between the different high-speed camera (100 fps) and verified with racket head coor-
overarm movements. We hypothesized there would be dinate data (Landlinger et al., 2010). The racket head horizontal
differences in maximal angles, angular velocities, and acceleration decreased abruptly at ball–racket contact.
their timing, but similar proximal-to-distal sequencing
between the team-handball throw, tennis serve, and the
Volleyball spike
volleyball spike.
Following a general and volleyball specific warm-up of 20 min,
subjects performed the required volleyball spikes. To maintain
Materials and methods constant testing conditions, the ball was suspended on a rope from
the ceiling via an apparatus that held and released the ball when
Subjects
spiked and to allow for similar repeated movement conditions to
Ten male elite team-handball players [mean ⫾ standard deviation accommodate players of different jumping height during testing
(SD) for age: 23 ⫾ 3 years; height: 1.87 ⫾ 0.06 m; weight: (Wagner et al., 2009). Every subject performed up to three spikes
85 ⫾ 10 kg; training experience: 12 ⫾ 3 years; 3 wings, 6 back- before testing. During testing, subjects were instructed to jump as
court players, 1 pivot, 9 right- and 1 left-handed players], 10 tennis high as possible and to hit the ball as fast and hard as possible into
players (mean ⫾ SD for age: 20 ⫾ 4 years; height: 1.88 ⫾ 0.07 m; a 3 ¥ 3 m corridor on the volleyball court. The subsequent assess-
weight: 77 ⫾ 10 kg; training experience: 14 ⫾ 3 years, 9 right- ment was used only for those spikes that hit the corridor. Subjects
and 1 left-handed players), and 10 volleyball players (mean ⫾ SD had to continue to spike the ball until five hits per subject were
for age: 24 ⫾ 4 years; height: 1.91 ⫾ 0.06 m; weight: 82 ⫾ 8 kg; achieved. Ball impact was defined as the point where the hand’s
training experience: 9 ⫾ 4 years; 5 middle blocks, 4 major attack- (head of the second metatarsal) horizontal acceleration of the
ers, 1 setter, 9 right- and 1 left-handed players) participated in the hitting arm decreased abruptly (cf. ball impact in tennis serve).
study. All subjects were physically healthy, in good physical
condition, and reported no injuries during the time of the study.
The study was approved by the local ethics committee and all Kinematic analysis
subjects agreed and signed an informed consent. The subjects were
The experimental setup consisted of eight cameras of Vicon MX13
recruited from the Austrian Team Handball National Team (n = 6),
motion capture system (Vicon Peak, Oxford, UK). Team-handball
2nd and 3rd Austrian Handball League (n = 4), male tennis player
jump throw and volleyball spike were captured with 250 fps
with an Association of Tennis Professionals ranking better than
(Wagner et al., 2009, 2010, 2011), while in the tennis serve the
347 (n = 7), high-performance male youth tennis players with a
measuring frequency was increased to 400 fps. In the tennis serve,
top 10 national ranking at the time of the study (n = 3), and
a Basler digital high-speed camera (100 fps) was synchronized
Austrian Volleyball National Team (n = 2), 1st and 2nd Austrian
with the Vicon MX13 cameras. Therefore, the measuring fre-
Volleyball League (n = 8).
quency of Vicon MX13 cameras had to be increased fourfold (400
fps) to accurately measure such a dynamic movement. To compare
the three movements, the measuring frequency of the team-
Team-handball throw handball jump throw and volleyball spike was increased using
After a general and a team-handball specific warm-up of 20 min, a spline function in MatLab R14a. For kinematic analysis, 39
the subjects performed vertical team-handball jump throws with reflective markers of 14 mm diameter were affixed to specific

2
 Kinematic differences in overarm movements
anatomical landmarks (Plug-In Gait Marker Set, Vicon Peak) for (pelvis and trunk rotation, trunk flexion, shoulder flexion, shoulder
every participant. Three-dimensional trajectories of the 39 markers internal rotation, and elbow extension); (b) timing of maximal
were analyzed utilizing Nexus software (Nexus 1.3, Vicon, angular velocity; (c) maximal angle (pelvis and trunk rotation,
Oxford, UK) and filtered with a Woltring filter (Woltring, 1986). trunk hyperextension, shoulder hyperextension, shoulder external
To calculate the joint positions, a 3D model (Plug-In Gait Model, rotation, and elbow flexion); and (d) timing of the maximal angles.
Vicon Peak) was used (Davis et al., 1991). The model was iden- Multivariate linear models [multiple analysis of variance
tical to that used in the team-handball jump throw (Wagner et al., (MANOVA)] were calculated to determine if all variables within a
2010, 2011), tennis ground stroke (Landlinger et al., 2010), and group of variables differ significantly between the team-handball
volleyball spike jump (Wagner et al., 2009). The global coordinate throw, tennis serve, and volleyball spike. If the multivariate analy-
system was defined and dependent on the movement direction. sis resulted in a significant difference (P < 0.05), one-way analysis
The global x-axis was defined in the direction of the throw, the of variance (ANOVA) was calculated for all variables within this
z-axis vertical and the y-axis perpendicular to these axes, whereas group. To determine the proximal-to-distal sequence, depending
the xy-plane was identical with the court. The orientation of the on the different movements (team-handball throw, tennis serve,
pelvis and trunk segments was identified by calculating three and volleyball spike), we used a repeated measures two-way
orthogonal axes (x-axis anteroposterior, pointed anteriorly; y-axis ANOVA with timing (pelvis rotation, trunk rotation, trunk flexion,
mediolateral, pointed laterally; and z-axis longitudinal). The ori- elbow extension, shoulder rotation, and shoulder flexion) and
entations of the humerus, radius, and hand segments were deter- movement (team-handball throw, tennis serve, volleyball spike) as
mined by the longitudinal z-axis (from the proximal to the distal main factors where the term movement is defined as the between-
joint center, pointed distally), the mediolateral y-axis (from the subjects factor. For the MANOVAs and the two-way ANOVA, we
distal joint center to the distal joint marker, pointed laterally), and used the Bonferroni post hoc test. For all statistic analysis, signifi-
the perpendicular anteroposterior x-axis (pointed anteriorly). Joint cance was set at P < 0.05 and effect size (h2) was defined as small
angles were calculated by the relative orientation of the proximal for h2 > 0.01, medium for h2 > 0.09, and large for h2 > 0.25
and distal segments (Figs 1 and 2). The joint flexion angles (shoul- (Cohen, 1988).
der and elbow flexion) were the angles determining the longitudi-
nal axes of the proximal and distal segments. The shoulder
internal-external rotation angle was defined as the rotation of the Results
humerus along the longitudinal axis of the humerus, where the
rotation of the humerus was determined by the movement of
Mean ⫾ SD values of maximal angular velocities,
the radius relative to the humerus. Trunk (pelvis) rotation angle timing of maximal angular velocities, maximal angles
was defined as the rotation between the anteroposterior axis of the and timing of maximal angles in the team-handball
trunk (pelvis) and the x-axis of the global coordinate system. The throw, tennis serve and volleyball spike, the statistical
direction of the rotation was defined as forward (counterclockwise results of the MANOVAs and one-way ANOVAs (global
for a right-handed player from an overhead view) and backward
rotation (clockwise for a right-handed player from an overhead
significance, effect size, power), as well as the results of
view), whereas we termed a forward rotation only as rotation. The the post hoc tests are depicted in Table 1. Significant
trunk flexion angle was calculated between the projected antero- differences between team-handball throw, tennis serve,
posterior trunk axis and the x-axis of the global coordinate system. and volleyball spike with a large effect (P < 0.001,
Angular velocities and ball velocity were calculated using the h2 > 0.70, 1-b = 1.00) were found for all groups of vari-
5-point differential method (Van den Tillar & Ettema, 2003).
ables used in the analysis. For the single variables, one-
way ANOVAs yielded a global significant difference
Phase classification between the different movements for 17 of 24 kinematic
To compare different measurements, all throws or hits were time variables (Table 1).
normalized to the ball release or ball contact point. The measure- The results show the pelvis and trunk rotated more
ments were conducted from 0.40 s before to 0.10 s after ball (40–50°) backward in the tennis serve compared with the
release or ball contact to calculate all relevant variables (Coleman team-handball throw and volleyball spike in the cocking
et al., 1993; Stodden et al., 2001; Escamilla & Andrews, 2009;
phase (Fig. 1). This difference was also found in the trunk
Wagner et al., 2009, 2010). Cocking phase, acceleration phase,
and follow-through phase were defined as described in Wagner hyperextension between a tennis serve and the other
et al. (2010). movements, whereas the trunk was more hyperextended
(15–20°) in the volleyball spike compared with the team-
handball throw (Fig. 2). For maximal angular velocities,
Proximal-to-distal sequence and timing pelvis rotation, trunk flexion, and trunk rotation were
Proximal-to-distal sequence was determined using the time of higher (50–250°/s) in the tennis serve compared with
occurrence of the maximal joint angular velocities (Marshall & the team-handball throw and volleyball spike. Maximal
Elliott, 2000; Van den Tillar & Ettema, 2009; Wagner et al., 2010).
Timing variables were measured relative to ball release or ball shoulder hyperextension angle (10–20°) and shoulder
contact. A negative value corresponds to a point of time before and flexion angular velocity (150–550°/s) were greater in the
a positive value matches with a time after the ball release or ball team-handball throw compared with the tennis serve and
contact. volleyball spike. No significant differences between the
analyzed movements were found in the maximal shoul-
Statistical analysis der external rotation angle, shoulder internal rotation
angular velocity, and elbow extension angular velocity.
All statistical analyses were performed using SPSS version 15.0.
(SPSS Inc., Chicago, IL, USA). Means and SDs of the variables Analyzing the proximal-to-distal sequencing, it was
were calculated for descriptive statistics. Four groups of variables found that the maximal angular velocities occurred from
were used for statistical analysis: (a) maximal angular velocity proximal to distal beginning with the pelvis rotation,

3
Wagner et al.

Fig. 1. Mean (⫾ standard error) pelvis, trunk, and shoulder rotation angle in the team-handball throw, tennis serve, and volleyball
spike. FT, follow-through.

4
 Kinematic differences in overarm movements

Fig. 2. Mean (⫾ standard error) trunk, shoulder, and elbow flexion angle in the team-handball throw, tennis serve, and volleyball spike.
FT, follow-through.

5
Wagner et al.
Table 1. Maximal angles and angular velocities (and their timing relative to ball release or ball contact) of the pelvis, trunk, shoulder, and elbow (rotation
and/or flexion) in the team-handball throw, tennis serve, and volleyball spike, and associated significant differences

Team handball Tennis Volleyball Global significance h2 Power

Maximal angular velocity (°/s) < 0.001 0.84 1.00

Pelvis rotation 450 ⫾ 120 510 ⫾ 110 370 ⫾ 90 < 0.05c 0.23 0.67
Trunk flexion 480 ⫾ 60 910 ⫾ 130 490 ⫾ 110 < 0.001ac 0.81 1.00
Trunk rotation 740 ⫾ 70 880 ⫾ 140 640 ⫾ 160 < 0.01ac 0.41 0.97
Shoulder flexion 1100 ⫾ 240 590 ⫾ 140 920 ⫾ 130 < 0.001ac 0.61 1.00
Shoulder internal rotation 4700 ⫾ 850 5580 ⫾ 2350 4520 ⫾ 1020 0.29 0.09 0.26
Elbow extension 1570 ⫾ 230 1670 ⫾ 380 1600 ⫾ 520 0.84 0.01 0.08
Timing of maximal angular velocity (s) < 0.001 0.72 1.00
Pelvis rotation -0.118 ⫾ 0.018 -0.081 ⫾ 0.028 -0.113 ⫾ 0.032 < 0.05ac 0.29 0.81
Trunk flexion -0.047 ⫾ 0.013 -0.042 ⫾ 0.011 -0.045 ⫾ 0.010 0.60 0.04 0.13
Trunk rotation -0.089 ⫾ 0.017 -0.045 ⫾ 0.012 -0.063 ⫾ 0.013 < 0.001abc 0.63 1.00
Shoulder flexion 0.050 ⫾ 0.008 0.063 ⫾ 0.050 0.050 ⫾ 0.009 0.15 0.13 0.38
Shoulder internal rotation 0.003 ⫾ 0.003 -0.005 ⫾ 0.011 0.005 ⫾ 0.003 < 0.01ac 0.32 0.87
Elbow extension -0.020 ⫾ 0.018 -0.030 ⫾ 0.005 -0.016 ⫾ 0.008 < 0.05c 0.20 0.59
Maximal angle (°) < 0.001 0.91 1.00
Pelvis rotation (–) -43 ⫾ 9 -99 ⫾ 17 -48 ⫾ 8 < 0.001ac 0.83 1.00
Trunk hyperextension 10 ⫾ 4 39 ⫾ 5 27 ⫾ 7 < 0.001abc 0.84 1.00
Trunk rotation (–) -66 ⫾ 9 -102 ⫾ 18 -64 ⫾ 16 < 0.001ac 0.61 1.00
Shoulder hyperextension 35 ⫾ 7 21 ⫾ 11 17 ⫾ 10 < 0.01ab 0.42 0.97
Shoulder external rotation 58 ⫾ 12 61 ⫾ 19 55 ⫾ 25 0.77 0.02 0.09
Elbow flexion 100 ⫾ 17 112 ⫾ 8 117 ⫾ 9 < 0.05b 0.28 0.78
Timing of maximal angle (s) < 0.001 0.70 1.00
Pelvis rotation -0.310 ⫾ 0.064 -0.333 ⫾ 0.047 -0.300 ⫾ 0.055 0.40 0.07 0.20
Trunk hyperextension -0.199 ⫾ 0.020 -0.099 ⫾ 0.011 -0.119 ⫾ 0.020 < 0.05a 0.25 0.71
Trunk rotation -0.234 ⫾ 0.020 -0.315 ⫾ 0.057 -0.186 ⫾ 0.019 < 0.001abc 0.70 1.00
Shoulder hyperextension -0.211 ⫾ 0.075 -0.238 ⫾ 0.077 -0.130 ⫾ 0.050 < 0.001bc 0.48 0.99
Shoulder extension rotation -0.029 ⫾ 0.016 -0.010 ⫾ 0.014 -0.016 ⫾ 0.005 < 0.01a 0.31 0.84
Elbow flexion -0.156 ⫾ 0.108 -0.171 ⫾ 0.040 -0.141 ⫾ 0.041 0.65 0.03 0.11

a
Between team-handball throw and tennis serve.
b
Between team-handball throw and volleyball spike.
c
Between tennis serve and volleyball spike.

followed by the trunk rotation, trunk flexion, elbow Jumping vs. standing
extension, shoulder internal rotation, and shoulder It is well known that the team-handball jump throw,
flexion in all three movements (Fig. 3). Repeated meas- tennis serve, and volleyball spike differ in lower body
ures two-way ANOVAs yielded a significant effect movements. The team-handball jump throw involves
for timing (F(23,5) = 270, P < 0.001, h2 = 0.98, 1-b = executing a vertical jump off one leg at takeoff after the
1.00), movement (F(27,2) = 6.4, P < 0.01, h2 = 0.32, run-up, whereas the main part of the throwing movement
1-b = 0.87), and the timing ¥ movement interaction is executed during the flight phase. This is also true in the
(F(46,10) = 8.5, P < 0.001, h2 = 0.65, 1-b = 1.00). Time volleyball spike, except that the takeoff is done with both
of occurrence of the maximal angular velocities differed legs. In the tennis serve, the first part of the cocking
and was highly significant (P < 0.001) to each other and phase is done during standing; in the second part of the
significant differences between the different movements cocking phase, tennis players executed a jump-off with
in the timing were found in the pelvis rotation, trunk both legs (0.27 ⫾ 0.02 s before ball contact). During
rotation, shoulder internal rotation, and elbow extension cocking phase, the pelvis and trunk were rotated back-
(Table 1). ward in all three analyzed movements (Fig. 1), whereas
the backward rotation angle of the pelvis (50–60°) and
Discussion trunk (30–40°) rotation were greater in the tennis serve
compared with the team-handball throw and volleyball
The aim of the study was to analyze not only differences spike. We observed that the floor contact in the tennis
but also similarities in the upper-body kinematics serve enabled a greater rotation over the front leg fol-
between the team-handball throw, tennis serve, and vol- lowed by a longer time for acceleration of the trunk and
leyball spike. For a detailed discussion of the results, the pelvis rotation. This allowed for generating greater
discussion was separated into different parts based on maximal angular velocity of the pelvis (10–40%) and
technical and tactical components as well as different trunk (15–45%) rotation in the tennis serve compared
lower body movements that could influence possible dif- with the team-handball throw and volleyball spike. In
ferences in the upper-body kinematics. different team-handball throwing techniques (Wagner

6
 Kinematic differences in overarm movements

Fig. 3. Mean timing of maximal angular velocities in the team-handball throw, tennis serve, and volleyball spike. Significant
differences between two joints: *P < .05, **P < .01, ***P < .001.

et al., 2011), it was shown that the maximal pelvis and proximal-to-distal sequencing was specific because the
trunk rotation backward angle during the cocking phase maximal elbow extension angular velocity occurred
and maximal pelvis and trunk rotation angular velocity before the maximal shoulder internal rotation angular
during acceleration phase differ significantly between velocity. We ascertain that the elbow of the throwing or
standing (with and without run-up) and jumping (takeoff hitting arm was extended earlier to reduce the moment
on one or both legs) throws. In the jump throws, team- arm for the shoulder internal rotation (Hong et al., 2001;
handball players had to rotate the trunk and pelvis via Van den Tillar & Ettema, 2004; Wagner et al., 2012), and
opposed leg movements during the flight. Pelvis and elbow extension angular velocity was reduced to prevent
trunk rotation angle and angular velocity are similar in excessive extension of the elbow and therefore reduce
the tennis serve and in a team-handball standing throw the possibility of muscle and joint injuries (Wierzbicka
(with and without run-up) but are quite different to a et al., 1986; Wagner et al., 2012). The highest elbow
team-handball jump throw or volleyball spike. We extension angle and therefore the smallest moment arm
observed that the floor contact during cocking has a for the shoulder internal rotation was measured near ball
strong influence on pelvis and trunk rotation in overarm release or contact (Fig. 2) when the shoulder internal
movements. rotation angular velocity reached its maximum (Table 1).
The order of occurrence was identical between the
different analyzed overarm movements; only the
Proximal-to-distal sequencing timing was significantly different. As shown in Fig. 3
We hypothesized that there would exist a similar and Table 1, maximal pelvis and trunk rotation angular
proximal-to-distal sequencing in the different overarm velocity occurred later in the tennis serve compared with
movements. As shown in Fig. 3, maximal angular veloci- the team-handball throw and volleyball spike. We
ties occurred in a specific proximal-to-distal order start- suggest that the longer time of acceleration of the pelvis
ing with pelvis rotation followed by trunk rotation, trunk and trunk rotation enabled a higher but delayed maximal
flexion, elbow extension, shoulder internal rotation, and angular velocity in the tennis serve compared with the
shoulder flexion in all analyzed overarm movements. team-handball throw and volleyball spike (Fig. 4).
This order of occurrence is typical for overarm move-
ments and was also found in similar studies (Whiting
et al., 1991; Fleisig et al., 1999; Marshall & Elliott, The influence of ball and racket
2000; Hong et al., 2001; Van den Tillar & Ettema, The maximal elbow extension and shoulder internal
2009; Landlinger et al., 2010; Wagner et al., 2012). The rotation angular velocity occurred earlier in the tennis

7
Wagner et al.

Fig. 4. Example of pelvis rotation, trunk rotation, elbow extension, and shoulder internal rotation angular velocity in the team-handball
throw (a) and tennis serve (b). CP, cocking phase; FT, follow-through.

serve compared with the team-handball throw and vol- The influence of arm swing at takeoff in the
leyball spike (Table 1 and Fig. 3). We maintain that the volleyball spike
tennis racket influences those significant differences in In the volleyball spike jump, the left and right shoulder
timing. In the last sequence of the tennis serve, the flexion angular velocities were maximized during the
racket has to be accelerated. The shoulder internal rota- upward phase to potentially increase jump height
tion angular velocity was reduced and momentum was (Escamilla & Andrews, 2009; Wagner et al., 2009). The
transferred from the upper and lower arms to the racket differences in the shoulder internal rotation and shoulder
(Elliott et al., 1995; Marshall & Elliott, 2000). Another flexion angle in the cocking phase of the volleyball spike
aspect is the weight of the tennis racket and handball. In compared with the tennis serve and team-handball throw
the tennis serve and team-handball throw, the throwing (Figs 1 and 2) are probably due to this shoulder flexion at
arm must accelerate the additional weight of the tennis takeoff that leads to a delay in the maximal shoulder
racket or handball whereas in the volleyball spike there hyperextension angle in the volleyball spike jump.
is no additional weight that has to be accelerated. We However, volleyball players were able to adapt to these
found significant differences in the maximal angular differences that lead to a similar shoulder internal rota-
velocities of pelvis rotation, trunk rotation, and trunk tion and shoulder flexion angle in the acceleration phase.
flexion between the tennis serve and team-handball
throw as well as volleyball spike, but no significant dif-
ferences in the maximal shoulder internal rotation The influence of impact from an opposing player in
and/or elbow extension angular velocity. We suggest team-handball throwing
that the transfer of momentum from the trunk to the In the tennis serve, the initiation of the movement is
throwing arm compensates for the different angular decided by the player and the movement cannot be dis-
momentum due to the tennis racket or handball. The turbed by the opposing player. In the volleyball spike,
difference in the angular momentum due to the tennis initiation of the movement is often influenced by the
racket may also explain the reduced and delayed setter and the movement of opposing players albeit not
maximal angular velocity of shoulder flexion in the disrupting the actual movement. In team handball,
tennis serve compared with the team-handball throw however, the throwing player is often tackled by the
and volleyball spike. opposing defensive player to prevent the throw. Wagner

8
 Kinematic differences in overarm movements
et al. (2010) have reported significant differences in the lutely greatest in all overarm movements compared with
trunk flexion between elite and low-level players in the other angular velocities. As shown in Fig. 1, shoulder
team-handball jump throw. The authors suggested that internal rotation angle is similar in the team-handball
elite players have adapted their throwing movement to throw, tennis server, and volleyball spike. The results of
prevent the impact of the opposing defensive player to this study are in agreement with kinematic studies in
prevent injuries. Gutierrez Davila et al. (2006) analyzed other overarm movements like football passing and base-
the effect of opposition on the team-handball jump throw ball throwing (Fleisig et al., 1996). We assert that elite
in elite players and found no significant differences in team handball, tennis, and volleyball players are able to
the upper-body kinematics between throwing with or reach a high ball velocity in all different overarm move-
without impact of an opposing defensive player. We ments in team handball, baseball, softball, cricket and
report that the elite team-handball players in this study water polo throwing, volleyball, beach volleyball and
may have already habitually adopted this trunk move- fistball hitting, as well as tennis, badminton, and squash
ment and have exhibited the movement also during serving or smashing. Studies that analyze this transfer of
testing. Therefore, the maximal trunk hyperextension movements (the same athletes in different overarm
was lower in the team-handball throw (10 ⫾ 4°) com- movements) are warranted.
pared with the tennis serve (39 ⫾ 5°) and volleyball
spike (27 ⫾ 7°). We suggest that the higher maximal Perspectives
trunk hyperextension allows more time for acceleration
during trunk flexion, and therefore a higher (70–100%) The team-handball throw, tennis serve, and volleyball
maximal trunk flexion angular velocity in the tennis spike show an equal order in the proximal-to-distal
serve compared with the team-handball throw and vol- sequence of the maximal angular velocities as well as
leyball spike. Similar results were found when compar- similar angles in the pelvis, trunk, and throwing or
ing football passing (possible tackling from an opposing hitting arm movements. The overarm movements differ
player) with baseball pitching (Fleisig et al., 1996). in the range of motion and therefore in the time of
acceleration of the pelvis and trunk rotation depending
on the ground contact, in trunk flexion depending on
Similarities between the different overarm movements possible impacts of the opposing player in competition
We found adaptations in the three different overarm (in team handball), and in shoulder flexion and rotation
movements as a result of jumping vs. standing, the influ- depending on the arm movements at takeoff (in the vol-
ence of a ball and/or racket of the arm swing at takeoff in leyball spike). We conclude that the overarm movements
the volleyball spike, as well as the impact from an oppos- are similar but not identical, because there are specific
ing player in team-handball throwing. However, the adaptations based on technical and tactical components
results of our study also show that there is a general of different games as well as different lower body move-
motor pattern in overarm movements of throwing or ments. However, training with other overarm movements
hitting a ball (with or without a racket). These overarm may be beneficial, especially for adolescent and prepu-
movements are characterized by the pelvis and trunk bescent athletes.
backward rotation as well as shoulder hyperextension
and elbow flexion at the beginning of the cocking phase, Key words: 3D kinematics, angular velocity, similarity
followed by the trunk hyperextension at the end of the of movements, specific differences.
cocking phase (Figs 1 and 2). Acceleration started with
the pelvis rotation, followed by the trunk rotation and Acknowledgements
trunk flexion (Figs 1–4). During this phase, the shoulder
was rotated externally. Ball (in the team-handball throw- The authors would like to thank Miriam Klous from the Depart-
ing), racket (in the tennis serve), or hand (in the volley- ment of Kinesiology, Motor Control Laboratory, Pennsylvania
State University; Elif Öz from the Department of Sport Science,
ball spike) were then accelerated by the shoulder internal Gaza University; and Michael Buchecker from the Department of
rotation added by the elbow extension, whereas the Sport Science and Kinesiology, University of Salzburg, for assist-
shoulder internal rotation angular velocity was abso- ance during data collection.

References
Bergun M, Mensure A, Tuncay C, NJ: Lawrence Erlbaum Associates, Davis RB, Õunpuu S, Tyburski D, Gage
Aydin O, Cigdem B. 3D kinematic 1988. JR. A gait analysis data collection and
analysis of overarm movements for Coleman SGS, Benham AS, Northcott reduction technique. Hum Mov Sci
different sports. Kines 2009: 41: SR. A three-dimensional 1991: 10: 575–587.
105–111. cinematographical analysis of the Elliott B, Marshall R, Noffal GJ.
Cohen J. Statistical power analysis for the volleyball spike. J Sport Sci 1993: 11: Contributions of upper limb segment
behavioral sciences. 2nd edn. Hillsdale, 295–302. rotation during the power serve in

9
Wagner et al.
tennis. J Appl Biomech 1995: 11: Marshall RN, Elliott BC. Long-axis Wagner H, Kainrath S, Müller E.
433–442. rotation: the missing link in Coordinative and tactical parameters of
Escamilla RF, Andrews JR. Shoulder proximal-to-distal segmental team-handball throw. The correlation of
muscle recruitment patterns and related sequencing. J Sports Sci 2000: level of performance, throwing quality
biomechanics during upper extremity 18: 247–254. and selected technique-tactical
sports. Sport Med 2009: 39: 569–590. Stodden DF, Fleisig GS, McLean SP, parameters. Leistungssport 2008: 38:
Fleisig SG, Barrentine SW, Zheng N, Lyman SL, Andrews JR. Relationship 35–41. (In German).
Escamilla RF, Andrews JR. Kinematic of pelvis and upper torso kinematics to Wagner H, Pfusterschmied J, von
and kinetic comparison of baseball pitched baseball velocity. J Appl Duvillard S, Müller E. Skill-dependent
pitching among various levels of Biomech 2001: 17: 164–172. proximal-to-distal sequence in
development. J Biomech 1999: 32: Van den Tillar R, Ettema G. A team-handball throwing. J Sports Sci
1371–1375. three-dimensional analysis of overarm 2012: 30: 21–29.
Fleisig SG, Escamilla RF, Andrews JR, throwing in experienced handball Wagner H, Pfusterschmied J, von
Tomoyuki M, Satterwhite Y, Barrentine players. J Appl Biomech 2007: 23: Duvillard SP, Müller E. Performance
SW. Kinematic and kinetic comparison 12–19. and kinematics of various throwing
between baseball pitching and football Van den Tillar R, Ettema G. Is there a techniques in team-handball. J Sport
passing. J Appl Biomech 1996: 12: proximal-to-distal sequence in overarm Sci 2011: 10: 73–80.
207–224. throwing in team-handball? J Sports Sci Wagner H, Tilp M, von Duvillard SP,
Gutierrez Davila M, Garcia PL, Parraga 2009: 27: 949–955. Müller E. Kinematic analysis of
Montilla J, Rojas Ruiz FJ. Effect of Van den Tillar R, Ettema G. Instructions volleyball spike jump. Int J Sport Med
opposition on the handball jump emphasizing velocity, accuracy, or both 2009: 30: 760–767.
shot. J Hum Mov Stud 2006: 51: in performance and kinematics of Whiting WC, Gregor RJ, Halushka M.
257–275. overarm throwing by experienced Body segment and release parameter
Hong DA, Cheung TK, Roberts EM. A team-handball players. Percept Mot contributions to new-rules javelin
three-dimensional, six-segment chain Skills 2003: 97: 731–742. throwing. Int J Sport Biomech 1991:
analysis of forceful overarm throwing. Van den Tillar R, Ettema G. A force- 7: 111–124.
J Electromyogr Kinesiol 2001: 11: velocity relationship and coordination Wierzbicka MM, Wiegner AW, Shahani
95–112. patterns in overarm throwing. J Sport BT. Role of agonist and antagonist
Landlinger J, Lindinger SJ, Stöggl T, Sci Med 2004: 3: 211–219. muscles in fast arm movements in man.
Wagner H, Müller E. Kinematic Wagner H, Buchecker M, von Duvillard Exp Brain Res 1986: 63: 331–340.
differences of elite and SP, Müller E. Kinematic description of Woltring HJ. A FORTRAN package for
high-performance tennis players in the elite vs. low level players in generalized cross-validatory spline
cross court and down the line forehand. team-handball jump throw. J Sport Sci smoothing and differentiation. Adv Eng
Sport Biomech 2010: 9: 280–295. Med 2010: 9: 15–23. Softw 1986: 8: 104–113.

10