122 Training and Testing
In Fencing, Does Intensive Practice Equally Improve
the Speed Performance of the Touche when it is Performed Alone
and in Combination with the Lunge?
E. Yiou, M. C. Do
1
Laboratoire de Physiologie du Mouvement, INSERM U483, UniversitØ Paris-Sud, Orsay, France
Yiou E, Do MC. In Fencing, Does Intensive Practice Equally Im-
prove the Speed Performance of the Touche when it is Per-
formed Alone and in Combination with the Lunge? Int J Sports
Med 2000; 21: 122 ± 126
Accepted after revision: September 20, 1999
nnnn Global movements are generally composed of several sim-
ple movements. In this study we tested the hypothesis that the
effects of practice on the performance of a simple movement
cannot be highlighted when it is performed alone but only
when it is part of a global movement. For this purpose we ex-
amined the performance of the touche movement in fencing, in
terms of maximal speed of the foil, when it is performed alone
and in combination with the lunge. The touche can be assimilat-
ed to a pointing task towards a target while the lunge can be
assimilated to a forward stepping movement. Four international
class fencers and five novices were tested. The results show that
1) the performance in the isolated touche condition was compar-
able between the two test populations, 2) the performance was
higher in the expert population compared to the novice popula-
tion in the sequential touche + lunge condition, and 3) the veloc-
ity of the centre of mass at the time of the peak velocity of the
touche was higher in the expert population than in the novice
population in the sequential touche + lunge condition. In line
with Bernstein©s concepts on the effects of practice on motor
control it is suggested that the experts exploit to advantage
the forces developed during the lunge whereas the novices do
not.
n Key words: Fencing, speed performance, sequential move-
ments.
Introduction
It is commonplace to say that intensive practice of a gesture or
a sport leads to enhanced performance, e.g. improved speed.
Nevertheless, several studies in the field of sport reported that
Int J Sports Med 2000; 21: 122 ± 126
 Georg Thieme Verlag Stuttgart ´ New York
ISSN 0172-4622
the performance of high level athletes could not be differenti-
Downloaded by: Florida State University Libraries. Copyrighted material.
ated from that of novices in simple motor tasks.
Zehr et al. [12] reported that the peak velocity and the move-
ment time of a ballistic elbow extension performed at maximal
speed in a sitting posture by competitive karate athletes or no-
vices were not different, although this movement is frequently
used in this sport. In fencing Harmenberg et al. [9] showed that
in terms of reaction time and movement time of a lunge move-
ment performed from the initial garde posture, world class
ØpØe fencers could not be differentiated from beginners. In
contrast, when the subjects were asked to initiate the lunge
movement while they were moving backwards using a fencing
movement called ªretreatº, the movement time of the lunge
was shorter in the expert population compared to the novice
population. In French boxing BØraud et al. [2] reported that
the peak velocity of a kicking movement performed in an up-
right posture at maximal speed was the same in an interna-
tional class boxing expert population and in a novice popula-
tion.
These results suggest that the effects of practice on the per-
formance of a simple movement cannot be highlighted when
it is performed alone but only when it is part of a global move-
ment.
To test this hypothesis, we chose a fencing movement com-
posed of the touche followed by the lunge. The touche can be
assimilated to a classical pointing task, except that in the pres-
ent study it is performed with a foil. The lunge can be assimila-
ted to a forward stepping movement.
In the field of fencing sport it is known that the efficient execu-
tion of the sequence of these two simple basic fencing move-
ments (the touche and the lunge) is the result of a complex mo-
tor co-ordination between upper and lower limbs that requires
at least one year of practice. In this study we compared the
maximal velocity of the touche when it was performed alone
to the maximal velocity of the touche when it was a component
of a global movement including the touche and the lunge.
Based on the above data, we hypothesize that the maximal
velocity of a touche is the same between the novice and the
fencing expert population when the touche is performed alone
but is higher in the expert population relative to the novice
population in the sequential touche + lunge condition.
Fencing Practice and Improvement of Speed Performance
Methods
Subjects
Four international class fencers (expert population) and five
novices in fencing (control population) took part in the experi-
ment. The novices are regularly involved in other athletic prac-
tices (dance, tennis, body-building or karate). All test subjects
were right-handed male adults. Anthropometric characteris-
tics of the test subjects are listed in Table 1.
Tab. 1 Anthropometric characteristics of the subjects. Average  one
standard deviation
Experts Novices
Number of subjects 4 5
Age (years) 25 ( 0.8) 29.4 ( 10.3)
Height (cm) 175 ( 4.5) 171 ( 8.8)
Weight (kg) 74.9 ( 7.9) 69 ( 10)
Leg length (cm) 91 ( 6.9) 88.2 ( 7.1)
Years of fencing 13 ( 5) 0 of the limb involved in the voluntary movement, and the pos-
Test procedure
The subjects had to perform three experimental series: 10
trials of touche (isolated touche condition), 10 trials of lunge
(isolated lunge condition), and a test-series of 40 trials touche
+ lunge (sequential touche + lunge condition) where they were
instructed to initiate the touche before voluntary lunge execu-
tion, i.e. before foot-off (as the fencing rules impose). All trials
where the touche was initiated after foot-off were discarded.
The experiment was conducted with a simplified reaction-
time paradigm, i.e. an oral signal instructed the subject to
stand motionless, then, after a variable delay (1 ± 2 s), a visual
execution signal (green light) located in the centre of the target
was given. The subject was instructed to perform both the
touche and the lunge at their maximal speeds. The target was
a rigid cardboard square of 10 ” 10 cm mounted on a vertical
bar articulated at its base so that the movement of the foil
was not stopped at the time of hit. In other words, the subjects
could go through the target. The target was placed at the level
of the right hand, in line with the prolongation of the foil. It
could be reached with a complete extension of the elbow. The
distance between the target and the foil was approximately
20 cm. No precision was requested. All trials were performed
from an initial posture that is similar to the specific garde pos-
ture in fencing. The knees were flexed (at an angle of approxi-
mately 140 8), the feet were perpendicular, the right foot (for-
ward foot) was in front of the left foot (backward foot) and
was directed towards the target. The distance between the
heels in the garde posture (L1) and in the lunge posture (L2)
was adjusted to the length of the subject©s leg (L), L1 = L/3 and
L2 = 2L/3. The subject was asked to distribute his body weight
equally between the two legs. The trunk was rotated in such a
way that the angle between the shoulder axis and the moving
upper limb was approximately 130 8. The angle of the moving
limb©s elbow was fixed at approximately 120 8. The forearm
and the foil were parallel to the floor.
Int J Sports Med 2000; 21 123
Recordings
The variables recorded were the displacement of the centre of
foot pressure and the acceleration of the body centre of mass,
which were obtained with a large force platform [6]; the accel-
eration of the foil, obtained with a miniature accelerometer
(ENTRAN,  5 g) fixed at the hand-guard of the foil and orien-
ted following the Antero-Posterior direction; and the surface
electromyographic activity (EMG) of the deltoïdus anterior
muscle, recorded with bipolar electrodes and amplified with
an amplifier chain of 3 ± 1000 Hz bandwidth. These variables
were sampled with an A/D converter (CED, 1401 +) at the sam-
pling frequency of 1000 Hz and stored on a PC hard disk for off-
line analysis.
Data processing
It has been known since Belenkii©s pioneer work [1] that the
Downloaded by: Florida State University Libraries. Copyrighted material.
execution of a voluntary movement is preceded and accompa-
nied by postural kinetic adjustments. Thus a voluntary move-
ment is composed of a postural support movement and a focal
movement. The focal movement corresponds to the movement
tural support movement corresponds to the movement of the
rest of the body. In our study the focal movement correspond-
ed to the movement of the upper limb involved in the touche
and was assimilated to the movement of the foil, and the pos-
tural support movement corresponded to the movement of the
rest of the body. The velocity of the focal movement (VF) was
obtained by simple integration of the accelerometer signal
fixed on the hand-guard of the foil and thus corresponded to
the velocity of the foil. The velocity of the postural support
movement (VP) was obtained by simple integration of the ac-
celeration of the body centre of mass. The peak velocity of the
foil (or maximal VF value) is termed VFmax which is the speed
performance of the touche. VPmax is the value of VP at the time
of the peak of foil velocity. In other words VPmax corresponded
to the postural support velocity value when the peak of foil
velocity was reached. As the touche movement is directed
mainly following the Antero-Posterior (A/P) axis, only the A/P
kinetics was considered in this study.
Time of foot-off corresponded to the instant when the trace of
the displacement of the foot pressure reached a plateau after
an initial displacement towards the backward foot (xP < 0) (cf.
Fig. 1). Time of foot-contact corresponded to the instant when
the trace of the displacement of the foot pressure shifted to-
wards the forward foot (xP > 0) (cf. Fig. 1).
Results
Fig. 1 shows the recordings of a single touche, a single lunge
and a sequence of touche + lunge. The time course of the vari-
ables was globally the same in the two populations except in
the sequential touche + lunge condition.
Isolated touche condition
As in classical pointing tasks with upper limb elevation [4, 5],
the voluntary touche execution was preceded by small Antici-
patory Postural Adjustments, i.e. the onset of the voluntary
touche movement, dated with the onset of the EMG activity of
the deltoidus anterior muscle, was preceded by a shift of the
124 Int J Sports Med 2000; 21
centre of foot pressure towards the backward foot (xP < 0) and
a forward acceleration of the body centre of mass (XªG > 0). The
duration of these Anticipatory Postural Adjustments, i.e. time
elapsed between the onset of variation of the acceleration
trace of the body centre of mass and the onset of the voluntary
touche movement, was on average 25  10 ms (average of all
subjects).
In both populations the acceleration of the body centre of mass
and the displacement of the centre of foot pressure reached a
peak value approximately 150 ms after the onset of the ac-
celeration of the body centre of mass. The comparison of the
peak value of the acceleration of the body centre of mass
(0.9  0.3 m/s2 vs 1.1  0.3 m/s2) and the peak value of the back-
ward displacement of foot pressure (± 15  5 cm vs. ± 17  4 cm)
did not show a difference between the novice and the expert
population (P > 0.05; ANOVA test). After target hit the body
centre of mass was accelerated backward (XºG < 0) and the dis-
placement of the centre of foot pressure was directed forward,
i.e. towards the forward foot (xP > 0). The postural support
velocity trace (VP trace) remained close to the baseline during
the touche movement.
Isolated lunge condition
As in the gait initiation process [6 ± 8,10,11], the lunge elicited
large anticipatory postural adjustments, i.e. prior to the volun-
tary lunge execution, dated with foot-off, the centre of foot
pressure moved towards the backward foot (xP < 0), and the
acceleration of the body centre of mass was directed forward
(XªG > 0). The duration of these Anticipatory Postural Adjust-
ments, corresponding to the time between the onset of varia-
tion of the trace of the acceleration of the body centre of mass
and foot-off, was on average 198  39 ms in the novice popula-
tion (all subjects together) and 195  46 ms in the expert pop-
ulation. At foot-off the amplitude of the backward displace-
Yiou E, Do MC
Fig. 1 EMG activity of the deltoïdus
anterior (DA) and biomechanical traces of
the body centre of mass and the centre of
foot pressure in the three experimental
conditions.
The traces of the same novice subject
were reported in the three experimental
conditions. In the sequential touche + lunge
condition the trace of one expert was
reported.
XªG , xP: acceleration of the body centre
of mass and displacement of the centre of
foot pressure (F: forward, B: backward),
respectively; VP, VF: velocity of the postural
support and velocity of the foil, respectively.
Vertical solid line: onset of variations of
XªG trace, FO: foot-off, HT: instant the foil
hits the target, FC: foot-contact.
Downloaded by: Florida State University Libraries. Copyrighted material.
ment of the centre of mass (± 30  5 cm vs. ± 31  4 cm) and
the forward acceleration of the body centre of mass
(3.1  0.5 m/s2 vs. 3.3  0.4 m/s2 ) did not show a difference be-
tween the novice and the expert population. The VP trace in-
creased progressively and reached a maximal value a short
time after foot-contact. The comparison of the maximal value
of VP did not show a difference between the novice and the ex-
pert population (1.3  0.3 m/s vs. 1.4  0.2  m/s). After foot-
contact the acceleration of the body centre of mass decreased
quickly towards negative values, and the centre of foot pres-
sure shifted quickly towards positive values, i.e. towards the
forward foot (xP > 0). During the lunge the deltoïdus anterior
muscle remained silent.
Sequential touche + lunge condition
In the novice population visual examination of the traces of the
acceleration of the centre of mass (XªG) and of the displace-
ment of the centre of foot pressure (xP) (see Fig. 1 touche +
lunge [novices]) showed that before target hit these traces
were similar to the traces of the isolated touche condition.
After target hit these traces were similar to the traces of the
isolated lunge condition. In addition, the latency between the
onset of the voluntary touche and the voluntary lunge move-
ments was equal to 330  80 ms. Time of target hit preceded
foot-off by 175  40 ms. In other words, the novices performed
the global touche + lunge movements as if they were perform-
ing the touche and lunge one after the other.
In contrast, in the expert population the traces of the accelera-
tion of the centre of mass and of the displacement of the centre
of foot pressure were similar to the traces of the isolated lunge
condition during the global touche + lunge movement (see
Fig. 1 touche + lunge [experts]). The latency between the onset
of the voluntary touche and the voluntary lunge movements
was equal to 70  40 ms and was shorter than in the novice
Fencing Practice and Improvement of Speed Performance
population (P < 0.01). Contrary to the novice population foot-
off preceded time of target hit. The latency between foot-off
and target hit was equal to 110  35 ms. The deltoïdus anterior
muscle was activated a short time prior to foot-off, i.e. during
the development of the APA of the lunge. In other words, the
experts performed the touche during the development of the
lunge.
Comparison of the performance intra- and inter-population
(Table 2)
a) Intra-population comparison
Table 2 Comparison of the performance intra and inter-population.
Measurements in m/s values  one standard deviation
Isolated touche condition Sequential touche + lunge
condition
VFmax VPmax VFmax VPmax
Experts 2.66  0.29 0.05  0.09 2.90  0.30** 0.41  0.20**
Novices 2.47  0.36 0.08  0.12 2.52  0.29 0.04  0.10
NS NS S S
VFmax: maximal foil velocity; VPmax: postural support velocity at the time the
maximal velocity of the foil is reached. S means statistical difference between
experts and novices (P < 0.01), NS means no statistical difference (P > 0.05; AN-
OVA test). ** means statistical difference in VFmax or VPmax values between the
isolated touche condition and the sequential touche + lunge condition
(P < 0.01).
Novices: The mean value of the peak of foil velocity (VFmax) was
comparable between the isolated touche condition (2.47  0.36
m/s all subjects together) and the sequential touche + lunge
condition (2.52  0.29 m/s). The mean value of VPmax (value of
the postural support velocity when the peak of foil velocity is
reached) was also comparable between the isolated touche
condition (0.08  0.12 m/s) and the sequential condition
(0.04  0.10 m/s).
Experts: The mean value of VFmax was statistically higher in the
sequential touche + lunge condition than in the isolated touche
condition (2.90  0.30  m/s vs. 2.66  0.29 m/s; P < 0.01, ANO-
VA test). The mean value of VPmax was statistically higher in
the sequential condition than in the isolated touche condition
(0.41  0.20 m/s vs. 0.05  0.09 m/s).
b) Inter-population comparison
Isolated touche condition: The mean value of VFmax (2.66 
0.29 m/s vs. 2.47  0.36 m/s) and VPmax (0.05  0.09 m/s vs.
0.08  0.12 m/s) was comparable between the two populations.
Sequential touche + lunge condition: The mean value of VFmax
was higher in the expert population than in the novice popula-
tion (2.90  0.30 m/s vs. 2.52  0.29 m/s; P < 0.01). The mean
value of VPmax was also higher in the expert population
(0.41  0.20 m/s vs. 0.04  0.10 m/s; P < 0.01).
Int J Sports Med 2000; 21 125
Discussion
The intra-population comparison showed that the speed
performance of the touche, in terms of maximal foil velocity
(VFmax), was comparable between the isolated touche condition
and the sequential touche + lunge condition in the novice
population. In contrast, the speed performance of the touche
was higher in the sequential touche + lunge condition than in
the isolated touche condition in the expert population
(2.66  0.29  m/s vs. 2.90  0.30  m/s). The inter-population
comparison showed that the speed performance of the touche
was comparable in the two populations in the isolated touche
condition. In contrast, the speed performance of the touche
was higher in the expert population than in the novice popula-
tion (2.90  0.30  m/s vs. 2.52  0.29 m/s) in the sequential
touche + lunge condition.
Although the touche is a highly practised movement in fencing,
Downloaded by: Florida State University Libraries. Copyrighted material.
the speed performance of the touche was comparable between
the two populations in the isolated touche condition. This find-
ing is in accordance with previous studies in the field of sport
[2, 9,12] which did not report any significant difference be-
tween experts and novices when performing simple motor
tasks. Therefore it seems that one could generalize the original
finding that the level of expertise cannot be determinated with
the speed performance of isolated simple movements.
In contrast, the higher speed performance (VFmax) (in fencing
experts compared to novices) when the touche is an integral
part of the touche + lunge sequential movement suggests that
the effects of practice on the speed performance can only be
highlighted when the simple test movement is part of a global
movement.
According to Bernstein©s concepts on the acquisition of motor
skills [3], one could have expected such a result. Indeed Bern-
stein has proposed that the organisation of movements be-
comes more economical with practice, in the sense that pas-
sive (reactive, inertial) forces become exploited to the fullest
in such a way as to enhance the efficiency of active (muscular)
forces leading to a higher speed performance. In the fencing
touche + lunge movement the higher efficiency of movement
organisation in the expert population was observed in the tim-
ing of the release of these two voluntary movements. The
touche was initiated during the development of the APA of the
lunge which generated an initial velocity (VPmax) that added to
the increase of the velocity of the touche. In contrast, the novi-
ces, by performing the touche before the development of the
APA of the lunge, could not benefit from this initial velocity
which was very weak compared to the experts© population.
In conclusion, the main finding of the present study is that ex-
perts in fencing extend their limb armed with the foil with
maximal velocity towards the target during a forward motion
of the postural support faster than novices. This faster postural
support velocity leads to an enhanced velocity of the foil. In
terms of practical application, this result suggests that it
would be more useful to train the novice fencers to do the glo-
bal movement rather than to separately repeat the simple
movements which make up the global movement.
126 Int J Sports Med 2000; 21 Yiou E, Do MC

References Corresponding Author:

1
Belenkii YY, Gurfinkel V, Paltsev YI. Element of control of volun- Éric Yiou
tary movements. Biofizika 1967; 12: 135 ± 141 Laboratoire de Physiologie du Mouvement
2
BØraud P. Analyse expØrimentale des coordinates posturo-cinØ- INSERM U483
tiques liØes à la dØlivrance d©un coup de pied. Cas de la boxe fran- UniversitØ Paris-Sud
cËaise savate. University Aix-Marseille II: PhD thesis, 1995: 134 F-91405 Orsay
3
Bernstein N. The coordination and regulation of movements. New France
York: Pergamon Press, 1967
4
Bouisset S, Zattara M. A sequence of postural movements pre- Phone: + 33-169-157231
cedes voluntary movements. Neurosci Lett 1981; 22: 263 ± 270 Fax: + 33-169-855219
5
Bouisset S, Zattara M. Biomechanical study of the programming E-mail: eric.yiou@lpm.u-psud.fr
of anticipatory postural adjustments associated with voluntary
movement. J Biomech 1987; 20: 735 ± 742
6
Breni›re Y, Do MC, Sanchez J. A biomechanical study of the gait
initiation process. J Biophys Med Nucl 1981; 5: 197 ± 205
7
Breni›re Y, Do MC, Bouisset S. Are dynamic phenomena prior to
stepping essential for walking? J Mot Behav 1987; 19: 62 ± 76
8
Carslöö S. The initiation of walking. Acta Anat 1966; 65: 1 ± 9

Downloaded by: Florida State University Libraries. Copyrighted material.

9
Harmenberg J, Ceci R, Barvestad P, Hjerpe K, Nyström J. Compari-
son of different tests of fencing performance. Int J Sports Med
1991; 12: 573 ± 576
10
Herman R, Cook T, Cozzens B, Freedman W. Control of postural re-
actions in man: the initiation of gait. In: Stein RB, Pearson KG,
Smith RS, Redford JB (eds). Control of posture and locomotion.
New York: Plenum, 1973: 363 ± 388
11
Mann RA, Hagy JL, White V, Liddell D. The initiation of gait. J Bone
Joint Surg 1979; 61: 232 ± 239
12
Zehr EP, Sale DG, Dowling JJ. Ballistic movement performance in
karate athletes. Med Sci Sports Exerc 1997; 29: 1366 ± 1373