Ef fects of a high-intensity swim test on kinematic
parameters in high-level athletes
Yannick A. Aujouannet, Marco Bonifazi, Frédérique Hintzy, Nicolas Vuillerme,
and Annie H. Rouard
Abstract: The present study aimed to investigate the effects of a high-intensity swim test among top-level swimmers
on (i) the spatial and temporal parameters of both the stroke and the 3-D fingertip pattern and (ii) the mechanical,
muscular, and physiological parameters. Ten male international swimmers performeda4×50mswimatmaximal in
tensity. Isometric arm flexion force with the elbow at 90° (F90°), EMG signals of right musculus biceps brachii and
triceps brachii and blood lactate concentrations were recorded before and after the swim test. Kinematic stroke (stroke
length, rate, and velocity) and spatiotemporal parameters of the fingertip trajectory were measured by two underwater
cameras during the first and last 50 m swims. After the swim test, F90° and mean power frequencies of the EMG de
creased significantly when blood lactate concentration increased significantly, attesting the reaching of fatigue. From
the first to the last 50 m, stroke rate, stroke velocity, and temporal parameters of the fingertip trajectory exhibited sig
nificant increases although stroke length and spatial fingertip trajectory remained unchanged. General and individual
adaptations were observed among the top-level swimmers studied. The present findings could be useful for coaches in
evaluating fatigue effects on the technical parameters of swimming.
Key words: sport, fatigue, biomechanic, lactate, force.
Résumé : Les objectifs de cette étude étaient d’étudier les effets d’un test intensif de nage sur (i) les paramètres spa-
tiaux du cycle et de la trajectoire en 3-D du majeur et (ii) les paramètres mécaniques, musculaire et physiologique. Dix
nageurs internationaux ont effectué un4×50màvitesse maximale. La force de flexion isométrique maximale avec le
coude à 90° (F90°), le signal EMG du biceps et du triceps brachii et la concentration sanguine de lactate ont été enre-
gistrés avant et après le test de nage. Les paramètres cinématiques du cycle (amplitude, fréquence, vitesse) et les para-
mètres spatio-temporels de la trajectoire du majeur ont été mesurés à l’aide de deux caméras sous marine durant le
premier et le dernier 50 m. Après le test de nage, F90° et les paramètres fréquentiels de l’EMG diminuent significati-
vement alors que la concentration sanguine de lactate augmente, attestant la présence de la fatigue. Du premier au der-
nier 50 m, la fréquence de cycle, la vitesse de nage et les paramètres temporels de la trajectoire du majeur présentent
une augmentation significative alors que l’amplitude de cycle et la trajectoire spatiale du majeur restent inchangées.
Des adaptations générales et individuelles sont observées entre les nageurs de haut niveau étudiés. Ces résultats peu
vent être utiles pour les entraîneurs désirant évaluer les effets de la fatigue sur les paramètres techniques de la nage.
Mots clés : sport, fatigue, biomécanique, lactate, force.
Aujouannet et al.
Front crawl swimming consists of right and left arm strokes
and a varying number of kicks per arm cycle (e.g., Maglischo
1993). A number of studies in swimming have analysed the
relationships between swimming velocity (SV), stroke length
(SL), and stroke rate (SR) (e.g., Craig and Pendergast 1979;
Craig et al. 1985; East 1970; Keskinen and Komi 1988a,
1993; Wilke 1992). Specific studies concerned the evolution
of these parameters during maximal exercises even in com
petition or under test conditions. Craig et al. (1985) and
Wilke (1992) showed a decrease in SV during the latter
stages of competitive events. These authors indicated that
the impairment of performance for faster swimmers corre
sponded to a decrease in SR with a constant SL, whereas
slower swimmers experienced a decrease in both SR and SL.
The decrease in SR could be due to either a decrease in
force production (e.g., Toussaint and Beek 1992) or to a fail
ure of neural activation (e.g., Keskinen and Komi 1993); the
decrease in SL could be linked to blood lactate accumulation
(e.g., Keskinen and Komi 1993; Weiss et al. 1988).
SV resulted from the propulsive and resistive forces. The
propulsive forces in front crawl are mainly generated by the
arm movements (e.g., Hollander et al. 1987), particularly by
the forearms and hands (e.g., Berger et al. 1995). Miyashita
(1975) highlighted the strong relationship between isometric
Appl. Physiol. Nutr. Metab. 31: 150–158 (2006) doi:10.1139/H05-012 © 2006 NRC Canada
Received 29 March 2004. Accepted 22 June 2005. Published
on the NRC Research Press Web site at http://apnm.nrc.ca on
7 March 2006.
F. Hintzy, N. Vuillerme, and A.H.
Rouard. Laboratoire de Modélisation des Activités Sportives,
Bâtiment Beaufortain, Université de Savoie, 73376 Le
Bourget Du Lac, France.
M. Bonifazi. Dipartimento di Fisiologia, Universita’ degli
Studi di Siena, Via Aldo Moro, 53100 Siena, Italy.
Corresponding author (e-mail: firstname.lastname@example.org).
arm pull force and swimming velocity. Propulsive forces
were strongly linked to kinematic hand parameters as ob
served in the different models of hand force calculations
(e.g., Schleihauf 1979; Berger et al. 1995). Maglischo et al.
(1986) have also shown the importance of the 3-D hand pat
tern in 50 and 100 m freestyle performance. Other authors
have suggested that SV could partly be explained by hori
zontal or vertical hand displacements during the arm stroke
(e.g., Deschodt et al. 1996; Deschodt 1999). However, stud
ies on the evolution of hand patterns in high intensity swim
exercises are limited. Deschodt (1999) reported a significant
decrease in the displacement of the wrist in the sagittal plane
followinga6×50mfreestyle swim at maximal velocity.
During maximal swim trials, a decrease in hand velocity was
observed during the insweep phase in 400 m freestyle (e.g.,
Monteil et al. 1994) and during the upsweep phase in a
200 m butterfly (e.g., Martins-Silva et al. 1997). Monteil et
al. (1994) concluded that the decrease in hand velocity was
due to the inability to maintain high force intensity through
out the test. The present study therefore aimed to investigate
the effects of a high-intensity swim test among top-level
swimmers on (i) the mechanical, muscular, and physiologi
cal parameters and (ii) the spatial and temporal parameters
of both the stroke and the 3-D fingertip pattern. A principal
component analysis (PCA) allowed the determination of
general and individual adaptations to high-intensity swim
exercises among top-level swimmers.
Materials and methods
Ten male international-calibre swimmers volunteered to
participate in this study. Five were medallists and the others
were finalists at the European Championship (Berlin, Ger-
many, 2002). Their mean physical characteristics were as
follows: age, 22.5 ± 2.3 y; height, 1.87 ± 0.07 m; mass,
79.00 ± 6.53 kg.
The test took place in a 25 m swimming pool. After a
standardized warm up consisting of a 1200 m swim, subjects
performed an isometric maximal voluntary contraction (MVC)
of the biceps brachii (Bi) and triceps brachii (Tri) muscles,
which they sustained for 5 s. After the MVC tests, subjects
were asked to perform a maximal right shoulder flexion test
at an arm–trunk angle of 90° with the elbow joint fully
). The arm–trunk angle was chosen to re
produce the middle phase of the stroke (e.g., Fomitchenko
1999; Strasse et al. 1999). The subject was lying prone on a
swim bench and a hand paddle attached to a strain gauge al
lowed the subject to generate force with the hand in prone
position (Fig. 1). This system was developed and validated
by the Istituto di Scienza dello Sport (CONI, Roma, Italy,
2002). Swimmers were familiarized with the apparatus in
their daily training. The lower and upper bodies were straight
and the subjects’ faces were positioned facing the bench.
The swimmer was fixed to the bench to avoid forward dis
placement during the pull. They were instructed to increase
force gradually up to maximum and to maintain this level
for 3 s before relaxing. The contralateral arm rested free
along the body. During this test, strong encouragements were
provided to the subjects.
After this initial isometric test, each subject performed
four 50 m repetitions (4 × 50 m) in freestyle at maximal
effort separated by 10 s rest periods. This exercise corre
sponded to a broken 200 m and was designed to reproduce
as closely as possible the effort of a 200 m freestyle compe
tition (e.g., Pelayo et al. 1996) without the psychological
constraints of the race (e.g., Alberty et al. 2003). Immedi
ately after the completion of the 4 × 50 m, the F90° test was
Blood lactate concentration ([BLa]) was obtained from
capillary blood samples (5
L) in the ear lobe at the end of
the warm up ([BLa
freestyle swim ([BLa
]). This parameter was chosen as a
physiological indicator of the contribution of anaerobic
glycolysis during exercise (e.g., Di Prampero et al. 1978).
The time was measured for each of the four sprints to
quantify the variation of performance related to exercise.
A strain gauge force transducer (Haarbye, Copenhagen,
Denmark) was used to measure the force exerted by the sub-
jects during the F90° test in pre and post conditions (Fig. 1).
The electrical activity of the right Bi and Tri was mea
sured by an EMG device (ME 3000 P8, Mega Electronics
Ltd., Kuopio, Finland) fixed to the lower part of the swim
mer’s back. These muscles were chosen according to their
main function in the central part of the underwater stroke
(e.g., Clarys 1983). The skin was shaved and rubbed with
an alcohol solution and two silver – silver chloride surface
electrodes with preamplifiers (Medicotest blue sensor type
M-00-S, 27 mm diameter) were placed in a bipolar configu
ration (2 cm interelectrode distance) in line with each mus
cle’s fibre orientation. Electrodes were placed in the midpoint
of the contracted muscle belly, as suggested by Clarys and
Cabri (1993). An adhesive film (Tegaderm, 3M, St. Paul,
Minn.) avoided contact with water (e.g., Rouard and Clarys
1995). A third reference electrode was attached to the body
in an area not in proximity to the studied muscles. The EMG
signal was stored on-line with a sampling frequency of
1000 Hz using a data acquisition card (flash memory 32
MB) processed through a computer with high- and low-pass
filters of 8 and 500 Hz, respectively. The gain was set at
1000 with a common mode rejection ratio of 92 dB.
Two digital video cameras (Panasonic WV-CP454E) were
used to record frontal and sagittal views of the underwater
arm stroke during the 1st and 4th sprints. The sagittal cam
era filmed the right side of the swimmer. Each camera was
enclosed in a waterproof box fixed at a depth of 0.60 m. The
© 2006 NRC Canada
Aujouannet et al. 151
Fig. 1. Testing apparatus for the evaluation of maximal isometric
optical axes between the two cameras were set at 90°, ac-
cording to Schleihauf’s software (Kinematic Analysis, San
Francisco, Calif.). The sagittal camera covered a field of 8 m
between the 37th and 45th m of the 50 m swim. This adjust-
ment allowed us to obtain more than one cycle for each
swimmer, in relation to the 2 m/stroke reported by Craig and
Pendergast (1979). Exposure time was set at 1/250 s owing
to the poor lighting conditions. The 25 Hz frame rate was
sufficient with respect to the average stroke frequency of
1.05 Hz (e.g., Craig and Pendergast 1979). The cameras
were switched on simultaneously. At the end of each experi-
ment, a 1.50 m calibration ruler was filmed in the middle
part of the two fields.
Blood lactate concentration was analysed with a Lactate
Pro-analyser (Akray KDK, Kyoto, Japan) according to the
The EMG data obtained during the F90°
tests were processed with MegaWin software (Mega Elec
tronics). EMG signals were treated overa3sstabilized
portion of the maximal force production. The mean power
frequency under pre (MPF
) and post (MPF
tions was calculated for both muscles (Bi-MPF
, and Tri-MPF
). The EMG signals
recorded were full-wave rectified and integrated. The inte
grated EMG (iEMG) under pre (iEMG
) and post
) conditions was calculated for the Bi (Bi-iEMG
) and Tri (Tri-iEMG
cles. To normalize the signals, all of the iEMG data were
reported to the iEMG value of the MVC, expressed as a
percentage of the MVC.
According to the Kinematic Analysis software validated
by Monteil et al. (1996), the right hip joint and the right
fingertip were semi-manually digitized frame-by-frame be
tween two successive right-hand entries. The 3-D hip and
hand trajectories were smoothed with a polynomial function
(3rd degree). For this testing condition, the reliability of the
software was tested and an average error of 3.07% ± 0.6%
was observed on the 3 axes. For a transportable acquisition
system, this result was acceptable in regard to the range val
ues of the studied parameters.
A right-handed Cartesian base reference frame O
established with the origin O fixed at the fingertip entry, as
illustrated in Fig. 2. The horizontal x axis was noted as posi
tive in the forward direction. The transverse y axis was di
rected perpendicular to the right side of the pool and the z
axis was vertically upward. The right fingertip entry was
taken as the temporal and spatial reference (0, 0, 0, 0) to de
termine different points of the 3 dimensions of the underwater
Five spatiotemporal stroke parameters, (i) stroke length
(SL; measured in m), (ii) stroke rate (SR; strokes@min
(iii) stroke velocity (SV; m@s
), (iv) underwater stroke dura
tion (UD; s), and (v) recovery duration (RD; s), were calcu
lated as follows:
 SL = X
is the x coordinate at the first right-hand entry and
is the x coordinate of the hip at the subsequent right-hand
 SR =
is the number of images between the two succes-
sive right-hand entries at a 25 Hz sampling rate of video
 UD = t
is the temporal coordinate of the fingertip exit
is the temporal coordinate of the right fingertip
 RD = t
is the temporal coordinate of the next right
The right fingertip trajectory relative to the reference frame
in the frontal and sagittal planes was described by 8
characteristic points, as previously seen in other studies (e.g.,
Schleihauf 1974; Deschodt et al. 1999; Maglischo 2003)
In the anteroposterior x axis, the coordinate of fingertip
entry (En) and exit (Ex), the maximal coordinates in the
forward (F) and backward direction (B), and the catch
point (C). C was determined when the arm and hand were
facing back and corresponded to the beginning of the pro
pulsive phase (e.g., Maglischo 2003).
In the transverse y axis, the maximal coordinates in the
outward direction (O) and inward direction (I).
In the vertical z axis, the maximal depth (D) of the finger
All of these points (En, C, F, O, D, I, B, and Ex) were
characterized by both temporal and spatial coordinates. To
study the influence of high-intensity exercise, we calculated
© 2006 NRC Canada
152 Appl. Physiol. Nutr. Metab. Vol. 31, 2006
Fig. 2. The reference O
with the origin O at the right fingertip
entry. The right fingertip and right hip digitized points were rep
the differences for the spatial and temporal parameters of
trajectories between the 1st and 4th sprints.
Mean and standard deviations were calculated for each
parameter. A coefficient of variation (CV = standard
deviation/mean) was determined to evaluate the homogene-
ity of the population. The effects of a high-intensity swim
test were evaluated by comparison of pre and post values for
blood lactate concentration, F90°, MPF, and iEMG. For
kinematic parameters, the comparison was realised between
data obtained during the first and last sprints. All compari-
sons were realised using a non-parametric test (Wilcoxon
test with p < 0.05).
For the significantly different kinematic parameters, the
variations between the 1st and 4th sprints were computed
into a principal component analysis (PCA), which was used
in a descriptive way. The results were represented by two
graphics: a correlation diagram and an individual representa
tion. The correlation diagram best represented the relation
ships between the studied parameters. Each axis of the
diagram was defined by a combination of the different pa
rameters, especially by those projected nearest the periphery
of the diagram. The correlation diagram enabled the location
of the coordinates for each subject. This method allowed us
to identify the individual responses with respect to parame
ters defining each axis of the correlation diagram. To evaluate
the influence of [BLa
] on kinematic changes, the subjects
were classified in descending order of [BLa
]; subject 1
had the highest and subject 10 had the lowest.
The results suggest significant effects of high-intensity
swim exercise on time performance during a 50 m freestyle,
[BLa], F90°, EMG, and kinematic parameters of stroke and
fingertip trajectory. Each of these parameters is presented in
Time and [BLa] increased significantly between the first
and last 50 m sprint (p < 0.01) and a significant decrease in
F90° values (p < 0.05) was observed. There was a significant
decrease between Bi-MPF
(p < 0.05). Regardless of the
muscles measured, iEMG data presented great individual
variations and did not exhibit significant differences between
pre and post conditions (p > 0.05).
For all parameters, with the exception of the 50 m time
(first and last 50 m values), the population appeared some
what heterogeneous (Table 1).
High-intensity swim test on whole-stroke and
underwater trajectory parameters
The whole-stroke parameters (SL, SR, SV, UD, and RD)
for the first and last 50 m sprints are presented in Table 2.
SR and SV decreased significantly between the first and
last sprints (p < 0.05 and p < 0.01, respectively). No signifi
cant changes were observed for SL, UD, or RD (p > 0.05).
The population was heterogeneous especially for SV and SR
The spatial and temporal hand parameters for the first and
the last 50 m sprints are presented in Fig. 4.
The most important displacements of the hand were for F
on the anteroposterior x axis and for D on the vertical z axis
(Fig. 4A). There were no significant differences (p > 0.05)
between the first and last 50 m in any of the spatial parame
ters (F, B, Ex, D, O, and I) when the time parameters for O,
I, and C increased significantly in the last 50 m sprint (p <
0.05). With the exception of D, high standard deviation val-
ues were observed for the spatial and temporal parameters of
fingertip trajectory, indicating great heterogeneity within the
population for these parameters.
Overall, these results indicated that the high-intensity swim
test had effects on temporal parameters for both whole-stroke
and underwater trajectory, whereas no significant difference
was observed for spatial whole-stroke or trajectory parameters.
These findings indicated general adaptations for the whole
group without any information on individual adaptations.
The relationships between the parameters significantly
modified by the high-intensity test and the individual adapta
tions were examined by PCA. The five parameters of the
PCA corresponded to the differences between the 1st and
4th sprints for parameters significantly different between the
two conditions (
SV). The variance was
principally explained by two factors (87.9%). On the corre
lation diagram (Fig. 5), axis 1 was mainly defined by the
variation of temporal parameters of the underwater trajectory
C), which were close to the periphery of the dia
gram. Axis 2 was principally defined by the stroke parameters,
The correlation diagram showed that variations among
temporal parameters (
C) were highly correlated
(0.921, p < 0.001) between
p < 0.01) be
O, and between
I (0.755, p < 0.01).
In other words, swimmers who presented the higher temporal
increase for C on the anteroposterior x axis also presented
the higher temporal increase for O and I on the transverse y
axis. It was also evident that variations of stroke parameters
SR) were strongly correlated (0.762, p < 0.01). The
decrease in SV observed during the 4th sprint resulted from
the decline of SR. Variations of stroke parameters (
were not correlated to those of temporal parameters (
C) (p > 0.05), i.e., the decrease in SV and SR could not
© 2006 NRC Canada
Aujouannet et al. 153
Fig. 3. Typical example of characteristic points of the fingertip
trajectory. On the anteroposterior x axis: En, entry into water; C,
catch point; F, maximal forward coordinate; B, maximal back
ward coordinate; and Ex, exit from the water. On the transverse
y axis: O, outward; and I, inward. On the vertical z axis: D,
be explained by the temporal increase of fingertip trajectory
The characteristics of the subjects, classified according to
] value, were determined from their coordinates
with respect to the two axes (Fig. 6).
Axis 1 revealed two groups of swimmers. The first group
(including subjects 1, 2, and 3 characterized by an average
] value of 16.27 ± 0.47 mmol@ L
) was in opposition
to all other subjects. Their position on the left indicated
great variation of the temporal parameters for I, O, and C at
the end of the test. In contrast, the second group (formed by
subjects 4, 5, 6, 7, 8, 9, and 10 characterized by an average
] value of 11.57 ± 0.81 mmol@L
) was characterized
by low variation of the temporal parameters (I, O, and C).
In regard to axis 2, one group of subjects (1, 4, 5, 6, and
10, with high
SV) was in opposition to subject 9
SV). The central position of the other swim
mers (2, 3, 7, and 8) indicated that they did not exhibit typi
cal variations for SR and SV. Consequently, swimmers who
presented the highest [BLa
] values did not have the great
est decrease in SV. These results indicated that [BLa
associated with the variation in temporal parameters (
C) but not with the variation in stroke parameters
The aim of this study was to investigate the effects of a
high-intensity swim test on both (i) mechanical, muscular,
and physiological parameters and (ii) spatial and temporal
parameters of both the stroke and the 3-D fingertip pattern.
Finally, a PCA was computed to further investigate the indi-
vidual adaptations and the relationships between the parame-
ters modified by the high-intensity test.
After the high-intensity swim test, several indicators sug-
gested the reaching of fatigue by all of the swimmers. First,
the decrease in maximal isometric force could be interpreted
as a manifestation of fatigue as referenced by many authors
(e.g., Bigland-Ritchie et al. 1986; Vollestad et al. 1988) who
defined fatigue as a loss of maximal force-generating capac
ity. Second, the overall compression of the power spectrum
to lower frequencies illustrated by the decrease in MPF data
values for the studied muscles suggested the presence of
fatigue as observed by previous investigations of isometric
(e.g., Bigland-Ritchie et al. 1981, 1983; Hagberg 1981;
Mannion and Dolan 1996; Merletti et al. 1992) or dynamic
contractions (e.g., Grassino et al. 1979; Komi and Tesch
1979). The [BLa
] values at the end of the set of4×50m
sprints were similar to those recorded after an elite 200 m
freestyle race (e.g., Bonifazi et al. 1993), indicating the high
intensity of this exercise. These results confirmed that this
protocol closely reproduced the effort of a 200 m freestyle
competitive event, as previously suggested by Pelayo et al.
(1996) and Alberty et al. (2003). Thus, the reduction in mus
cle pH after the last 50 m (as suggested by the [BLa] mea
sures) could be one factor responsible for impaired muscle
function resulting in fatigue (Tesch et al. 1978).
Finally, the significant decrease in 50 m time performance
might indicate the reaching of fatigue according to Enoka
and Stuart (1992), who noted an acute impairment of
performance with fatigue. Overall, the significant changes
© 2006 NRC Canada
154 Appl. Physiol. Nutr. Metab. Vol. 31, 2006
50 m time (s) [BLa] (mmol@L
) F90° (N) Bi-MPF (Hz) Bi-iEMG (%) Tri-MPF (Hz) Tri-iEMG (%)
1st 4th [BLa
] Pre Post Pre Post Pre Post Pre Post Pre Post
Mean 28.68 33.50**
13.0** 122.1 105.2* 63.4 50.3* 43.2 42.8 84.4 83.0* 52.9 51.2
SD 1.35 2.61 0.1 2.8 22.1 18.1 10.2 12.7 25.2 34.9 9.9 9.4 34.2 49.2
CV (%) 4.71 7.80 12.6 21.5 18.1 17.3 16.0 25.3 58.3 81.6 11.7 11.4 64.6 95.9
Note: *, significant difference between pre and post conditions at p < 0.05; **, significant difference between pre and post conditions at p < 0.01. Bi, biceps brachii; Tri, triceps brachii; MPF, mean
Table 1. Mean, standard deviation (SD), and coefficient of variation (CV) of the 50 m time, blood lactate concentrations ([BLa
] and [BLa
]), and isometric forces (F90°)
obtained in pre and post conditions.
SL (m) SR (strokes@min
) UD (s) RD (s)
1st 4th 1st 4th 1st 4th 1st 4th 1st 4th
1.06 1.14 0.72 0.80
SD 0.29 0.22 5.60 4.17 0.19 0.09 0.19 0.26 0.13 0.27
CV (%) 12.83 10.10 16.23 13.25 14.73 8.33 17.92 22.81 18.06 33.75
Note: *, significant difference between the 1st and 4th 50 m sprints at p < 0.05; **, significant difference between the 1st and 4th 50 m sprints at p < 0.01. SL, stroke length; SR, stroke rate;
SV, swimming velocity; UD, underwater stroke duration; RD, recovery duration.
Table 2. Mean, standard deviation (SD), and coefficient of variation (CV) of the whole-stroke parameters during the 1st and 4th 50 m tests.
in mechanical, electromyographic, physiological, and
performance indicators between pre and post conditions
strongly suggested the presence of fatigue at the end of the
high-intensity swim test used in this study.
During the 1st 50 m, the SL results were comparable to
those of previous studies involving 200 m freestyle during
official competitions with lower SR and SV values (e.g.,
Craig et al. 1985; Pai et al. 1984). The lower values found
here could be explained by the recording EMG system,
which increased body drag and limited the performance of
the subjects. In addition, the experimentation took place dur
ing a high-volume endurance training cycle, which was char
acterized by lower stroke rates and velocities in accordance
with Maglischo (2003). The spatial and temporal parameters
of the hand trajectory during the 1st 50 m agreed with previ
ous studies (e.g., Maglischo et al. 1986; Deschodt et al.
1996; Deschodt 1999; Maglischo 2003).
Under fatigue, spatial stroke (SL) and trajectory parame
ters (F, B, Ex, D, O, I) were not significantly different.
These results differed from previous studies, which found a
decrease in SL with fatigue in well-trained swimmers (e.g.,
Deschodt 1999; Keskinen and Komi 1988b, 1993; Weiss et
al. 1988). For the spatial parameter of fingertip trajectory,
Deschodt (1999) observed a decrease in F (maximal forward
coordinate) and D (maximal depth) of the hand duringa6×
50 m swimming set at maximal velocity and performed by
well-trained swimmers. In regard to earlier studies with
lower-skilled swimmers (e.g., Deschodt 1999; Keskinen and
Komi 1988b; Weiss et al. 1988), the maintenance of the spa
tial parameters observed in the present study could be due to
the very high performance level of our subjects (5 of them
were international medallists). The maintenance of the hand
trajectory suggested a robust spatial pattern in these elite
swimmers that is not easily changed even by the impair
ments imposed by fatigue (e.g., Rodacki et al. 2001).
In contrast to the spatial parameters, the temporal parame
ters were significantly altered by fatigue with significant
decrease in SR and increase of I, O, and C. A reduction of
SR has been reported by Craig et al. (1985) during a 200 m
freestyle competition. As suggested by Toussaint and Beek
(1992), the decrease in SR could result from decrease in
force production or by a failure in neural activation as sug
gested by Keskinen and Komi (1993). The 2 hypotheses
© 2006 NRC Canada
Aujouannet et al. 155
FB DO IEx
A. Spatial coordinates
FB D O I CEx
B. Temporal coordinates
Fig. 4. (A) Spatial parameters of the underwater fingertip trajectory (expressed in cm) for the 1st (white bar) and 4th (grey bar) 50 m
sprints. F, maximal forward coordinate; B, maximal backward coordinate; Ex, exit from the water; D, maximal depth; O, outward; and
I, inward. (B) Temporal parameters of F, B, Ex, D, O, I, and C (catch point), expressed in s. Asterisk indicates significant difference
between the 1st and 4th sprints at p < 0.05.
Fig. 5. Correlation diagram of the PCA for the variations of the
temporal parameters (
O, maximal outward coordinate;
mal inward coordinate; and
C, catch point) and stroke parameters
SR, stroke rate; and
SV, stroke velocity).
Fig. 6. Representation of subjects of the PCA. Axis 1 represents
the subjects grouped on the basis of variation of temporal param
C) (dashed line). Axis 2 represents the sub
jects grouped on the basis of variation of stroke parameters (
SR) (solid line).
were supported by the shift of the MPF to low frequencies
and the decrease in force production during the isometric
test in post conditions.
The PCA results indicated that the decrease in SV was
strongly correlated to the decrease in SR, confirming previ
ously reported data (e.g., Cappaert et al. 1995; Keskinen and
Komi 1988a, 1988b, 1993). This underlined the fact that SR
is the most determinant factor of SV in high-level swimmers
under fatigue (e.g., Deschodt 1999). The significant decrease
in SR was not related to temporal changes of a particular
phase of the stroke (such as increased time of UD, RD or I,
O, and C), but rather to the overall temporal pattern.
The temporal variations of the fingertip trajectory (I, O,
C) were observed during the first half of the stroke. The high
correlation between these parameters confirmed the strong
relationships between the 3-D components of the hand
movement as observed by previous studies (e.g., Counsilman
1969; Brown and Counsilman 1971). This present result
highlights that the longer the time the swimmer placed his
hand in a propulsive position, the longer the time the subject
took to complete the insweep phase. The increased time for
C (catch point) indicates that fatigued swimmers spent more
time placing the hand in a propulsive way. This confirmed
Goldfuss and Nelson’s (1971) finding, who studied tethered
swimming to the point of exhaustion and who observed a
consistent increase of time during the entry phase (i.e., the
subject took a rest with the upper arm extended before be-
ginning the pull). The time increases for O and I indicated a
longer duration of the insweep phase for a similar spatial
pattern. As a result, the hand velocity decreased during this
phase. This result was in the line with Monteil et al. (1994),
who found a decrease in hand velocity during the insweep
phase in the latter stages of a 400 m freestyle. Cappaert et
al. (1995) showed that this phase was the most propulsive
part of the stroke and Rouard et al. (1997) observed the
greatest muscle recruitment during it. Decreased hand veloc
ity with fatigue could be due to the inability to maintain me
chanical and muscular constraints, as suggested by Monteil
et al. (1994). Our findings in a maximal shoulder flexion test
in post conditions (decrease in F90°
and Bi- and Tri-
values) could confirm this hypothesis.
In regard to these general adaptations to fatigue, the PCA
results could point to some individual adaptations, particu
larly in reference to the influence of blood lactate concentra
tion on stroke and fingertip trajectory parameters. From axis
1, the PCA result demonstrated that swimmers with the
] values (subjects 1, 2, and 3) presented the
greatest temporal variations for I, O, and C; other subjects
(4, 5, 6, 7, 8, 9, and 10) were not characterized by such im
portant variations. Furthermore, this result suggested that
subjects with highest [BLa
] values (subjects 1, 2, and 3)
experienced the greatest decrease in hand velocity under fa
tigue, particularly during the insweep phase. From axis 2, no
relationship was observed between two different indicators
SV and [BLa
]. Indeed, swimmers with the
] did not present the strongest decrease in
SV. This result can be illustrated by subjects 1, 2, and 3,
who had fairly similar [BLa
] values (16.8, 16.1, and
, respectively). Although subject 1 presented a
strong decrease in SV (from 1.60 to 1.07 m@s
), subjects 2
and 3 were characterized by a slight decrease in SV (from
1.17 to 1.12 m@ s
and from 1.27 to 1.22 m@ s
Another example of individual adaptation concerned sub
jects 3 and 4,l who presented similar [BLa
] values (15.9
and 15.1 mmol@L
, respectively) with different variations of
stroke and temporal parameters. Subject 3 presented high
variations of temporal parameters and low variations of
stroke parameters, whereas subject 4 exhibited the opposite
adaptations. For Keskinen and Komi (1988a), large individ
ual variations in [BLa
] and SV parameters reflected great
variability in both swimming efficiency and stroke parame
ters. These results underline that fatigue in swimming leads
to individual adaptations for stroke and trajectory parameters
among top-level swimmers.
This study aimed to investigate the effects of a high-
intensity swim test on isometric force production and associ
ated muscular recruitments and on spatiotemporal parame
ters of the stroke and 3-D fingertip pattern. At the end of the
final 50 m sprint, several indicators (F90°, MPF, [BLa], 50 m
time performance) revealed that our subjects reached fatigue.
For 10 top-level swimmers, fatigue was characterized by
spatial stability and temporal increase for both stroke and
parameters of fingertip trajectory. The maintenance of the
spatial stroke and trajectory parameters could be due to the
very high performance level of our subjects and suggested a
robust spatial fingertip pattern that is not easily changed
even by the impairments imposed by fatigue. The increase of
temporal parameters under fatigue was specially marked by
an important increase of time for the beginning of the cycle.
Under fatigue, general adaptations suggested that SR be-
comes the most determinant factor of SV for top-level swim-
mers. Individual adaptations showed that [BLa
to influence the variations of temporal trajectory parameters,
but did not appear to decrease SV. These results suggested
that the usual monitoring of fatigue in swimming, based
upon blood lactate concentration and SL, is questionable.
The present study could be useful for coaches to evaluate the
general and individual effects of fatigue on technical param
eters of freestyle stroke.
Alberty, M., Sidney, M., Dekerle, J., Potdevin, F., Hespel, J.M.,
and Pelayo, P. 2003. Effects of an exhaustive exercise on upper
limb coordination and intracyclic velocity variations in front
crawl stroke. In Biomechanics and medicine in swimming IX.
Edited by J.C. Chatard. Publications de l’Université de Saint
Etienne, St. Etienne, France. pp. 87–92.
Berger, M.A.M., De Groot, G., and Hollander, A.P. 1995. Hydro
dynamic drag and lift forces on human hand/arm models. J.
Biomech. 28: 125–133.
Bigland-Ritchie, B., Donovan, E.F., and Roussos, C.S. 1981. Con
duction velocity and EMG power spectrum changes in fatigue of
sustained maximal efforts. J. Appl. Physiol. 51(5): 1300–1305.
Bigland-Ritchie, B., Johansson, R.S., Lippold, O.C.J., and Woods,
J.J. 1983. Contractile speed and EMG changes during fatigue of
sustained maximal voluntary contractions. J. Neurophysiol. 50:
© 2006 NRC Canada
156 Appl. Physiol. Nutr. Metab. Vol. 31, 2006
Bigland-Ritchie, B., Furbush, F., and Woods, J.J. 1986. Fatigue of
intermittent submaximal voluntary contractions: central and pe
ripheral factors. J. Appl. Physiol. 61(2): 421–429.
Bonifazi, M., Martelli, G., Marugo, L., Sardella, F., and Carli, G.
1993. Blood lactate accumulation in top level swimmers follow
ing competition. J. Sports Med. Phys. Fitness, 33: 13–18.
Brown, R.M., and Counsilman, J.E. 1971. The role of lift in pro
pelling the swimmer. In Selected topics in biomechanics. Edited
by J.M. Cooper. Chicago Athletic Institute, Chicago, Ill.
Cappaert, J.M., Pease, D.L., and Troup, J.P. 1995. Three-dimensional
analysis of the men’s 100m freestyle during the 1992 Olympic
Games. J. Appl. Biomech. 11: 103–112.
Clarys, J.P. 1983. A review of EMG in swimming. In Swimming
science IV. Edited by A.P. Hollander, P.A. Huijing, G. De Groot,
and G.J. Van Ingen Schenau. Human Kinetics, Champaign, Ill.
Clarys, J.P., and Cabri, J. 1993. Electromyography and the study of
sports movements: a review. J. Sports Sci. 11(5): 379–448.
Counsilman, J.E. 1969. The role of sculling movements in the arm
pull. Swim. World, 10(12): 6–7, 43.
Craig, A.B., and Pendergast, D.R. 1979. Relationships of stroke
rate, distance per stroke, and velocity in competitive swimming.
Med. Sci. Sports, 11(3): 278–283.
Craig, A.B., Skehan, P.L., Pawelczyk, J.A., and Boomer, W.L.
1985. Velocity, stroke rate, and distance per stroke during elite
swimming competition. Med. Sci. Sports Exerc. 17(6): 625–634.
Deschodt, V. 1999. Modifications de la trajectoire aquatique du
poignet liées à l’apparition de la fatigue lors d’exercices
intermittents en natation. Rev. Sci. Motr. 37: 19–25.
Deschodt, V., Rouard, A.H., and Monteil, K.M. 1996. Relation-
ships between the three coordinates of the upper limb joints
with swimming velocity. In Biomechanics and medicine in
swimming VII. Edited by J.P. Troup, A.P. Hollander, D. Strasse,
S.W. Trappe, J.M. Cappaert, and T.A. Trappe. E & FN Spon,
London,. UK. pp. 52–58.
Di Prampero, P.E., Meyer, M., Cerretelli, P., and Pijper, J. 1978.
Energetics of anaerobic glycolysis in dog gastrocnemius.
Pflugers Arch. 377(1): 1–8.
East, D.J. 1970. Swimming: an analysis of stroke frequency, stroke
length, and performance. N. Z. J. Health Phys. Educ. Recr. 3:
Enoka, R.M., and Stuart, D.G. 1992. Neurobiology of muscle fa
tigue. J. Appl. Physiol. 72: 1631–1648.
Fomitchenko, T. 1999. Relationship between sprint swimming speed
and power capacity in different groups of swimmers. In
Swimming science VIII. Edited by K.L. Keskinen, P.V. Komi,
and A.P. Hollander. University of Jyv
, Finland. pp. 203–
Goldfuss, A.J., and Nelson, R.C. 1971. A temporal and force anal
ysis of the crawl arm stroke during tethered swimming. In
Biomechanics in swimming. Edited by L. Lewillie and J.P. Clarys.
Free University, Brussels, Belgium. pp. 129–142.
Grassino, A., Gross, D., Macklem, P.T., Roussos, C., and
Zagelbaum, G. 1979. Inspiratory muscle fatigue as a factor lim
iting exercise. Bull. Eur. Physiopathol. Respir. 15(1): 105–115.
Hagberg, M. 1981. Electromyographic signs of shoulder muscular
fatigue in two elevated arm positions. Am. J. Phys. Med. 60(3):
Hollander, A.P., de Groot, G., van Ingen Schenau, G.J., Kahman, R.,
and Toussaint, H.M. 1987. Contribution of the legs to propul
sion in front crawl swimming. In Swimming science V. Edited
by B.E. Ungerechts, K. Wilke, and K. Reischle. University Park
Press, Baltimore, Md. pp. 39–43.
Keskinen, K.L., and Komi, P.V. 1988a. Interaction between
aerobic/anaerobic loading and biomechanical performance in
freestyle swimming. In Biomechanics XI-B. Vol. 7-B. Edited by
G. De Groot, A.P. Hollander, P.A. Huijing, and G.J. Van Ingen
Schenau. Free University Press, Amsterdam, Netherlands.
Keskinen, K.L., and Komi, P.V. 1988b. The stroking characteristics
in four different exercises in free style swimming. In Biomech
anics and medicine in swimming. Vol. 9. Edited by A.P. Hol
lander, P.A. Huijing, and G. De Groot. Free University Press,
Amsterdam, Netherlands. pp. 219–226.
Keskinen, K.L., and Komi, P.V. 1993. Stroking characteristics of
front crawl swimming during exercise. J. Appl. Biomech. 9:
Komi, P.V., and Tesch, P. 1979. EMG frequency spectrum, muscle
structure, and fatigue during dynamic contractions in man. Eur.
J. Appl. Physiol. Occup. Physiol. 42(1): 41–50.
Maglischo, C.W. 1993. Swimming even faster. Human Kinetics
Publishers, Champaign, Ill.
Maglischo, E.W. 2003. Swimming fastest. Human Kinetics Pub
lishers, Champaign, Ill.
Maglischo, C.W., Maglischo, E.W., Higgins, J., Hinricks, R., Luedtke,
D., Schleihauf, R.E., and Thayer, A. 1986. A biomechanical analysis
of the 1984 US Olympic swimming team: the distance freestylers. J.
Swim. Res. 2(3): 12–16.
Mannion, A.F., and Dolan, P. 1996. Relationship between myoelectric
and mechanical manifestations of fatigue in the quadriceps femoris
muscle group. Eur. J. Appl. Physiol. Occup. Physiol. 74(5): 411–9.
Martins-Silva, A., Alves, F., and Gomes-Pereira, J. 1997. Fatigue
related technical changes in 3-D directional components of hand
velocity in butterfly swimming during a 200 m maximum trial: a
pilot study. In Proceedings of the XII FINA World Congress on
Sports Medicine, April 1997, G
teborg, Sweden. Edited by B.O.
Eriksson and L. Gullstrand. pp. 463–467.
Merletti, R., Knaflitz, M., and DeLuca, C.J. 1992. Electrically evoked
myoelectric signals. Crit. Rev. Biomed. Eng. 19(4): 293–340.
Miyashita, M. 1975. Arm action in the crawl stroke. In Swimming
II. Edited by L. Lewillie and J.P. Clarys. University Park Press,
Baltimore, Md. pp. 167–173.
Monteil, K.M., Rouard, A.H., and Troup, J.P. 1994. Etude des
paramètres cinétiques du nageur de crawl au cours d’un exercice
maximal dans un “flume”. Rev. Sci. Tech. Act. Phys. Sport. 33:
Monteil, K.M., Chèze, L., Masset, J.B., and Rouard, A.H. 1996.
Three dimensional analysis of human gait: comparison between
the motion analysis and the kinematic analysis systems. In Inter
national symposium on 3-D analysis of human movement. Edited
J.P. Blanchi and P. Allard. Presse Universitaire, Grenoble,
France. pp. 364–365.
Pai, Y.C., Hay, J.G., and Wilson, B.D. 1984. Stroking technique of
elite swimmers. J. Sports Sci. 2(3): 225–239.
Pelayo, P., Mujika, I., Sidney, M., and Chatard, J.C. 1996. Blood
lactate recovery measurements, training, and performance
during a 23-week period of competitive training. Eur. J. Appl.
Physiol. 74: 107–113.
Rodacki, A.L.F., Fowler, N.E., and Bennett, S.J. 2001. Multi-segment
coordination: fatigue effects. Med. Sci. Sports Exerc. 33(7):
Rouard, A.H., and Clarys, J.P. 1995. Cocontraction in the elbow
and shoulder muscles during rapid cyclic movements in an
aquatic environment. J. Electromyogr. Kinesiol. 5(3): 177–183.
Rouard, A.H., Billat, R.P., Deschodt, V., and Clarys, J.P. 1997.
Muscular activations during repetitions of sculling movements
© 2006 NRC Canada
Aujouannet et al. 157
up to exhaustion in swimming. Arch. Physiol. Biochem. 105(7):
Schleihauf, R.E. 1974. A biomechanical analysis of freestyle.
Swim. Tech. 11(3): 88–96.
Schleihauf, R.E. 1979. A hydrodynamic analysis of swimming pro
pulsion. In Swimming science III. Edited by J. Terauds and E.W.
Bedingfield. University Park Press, Baltimore, Md. pp. 70–109.
Strasse, D., Wild, M., and Hahn, A. 1999. A comparison of maxi
mal voluntary force during unilateral and bilateral arm extension
in swimmers. In Biomechanics and medicine in swimming VIII.
Edited by D. Strasse, M. Wild, and A. Hahn. Free University
Press, Amsterdam, Netherlands. pp. 197–201.
Tesch, P., Sjodin, B., Thorstensson, A., and Karlsson, J. 1978.
Muscle fatigue and its relation to lactate accumulation and LDH
activity in man. Acta Physiol. Scand. 103(4): 413–420.
Toussaint, H.M., and Beek, P.J. 1992. Biomechanics of competitive
front crawl swimming. Sports Med. 13(1): 8–24.
Vollestad, N.K., Sejersted, O.M., Bahr, R., Woods, J.J., and
Bigland-Ritchie, B. 1988. Motor drive and metabolic responses
during repeated submaximal contractions in humans. J. Appl.
Physiol. 64(4): 1421–1427.
Weiss, M., Reischle, K., Bouws, N., Simon, G., and Weicker, H.
1988. Relationships of blood lactate to stroke rate and distance
per stroke in top female swimmers. In Swimming science V.
Vol. 18. Edited by B.E. Ungerechts, K. Wilke, and K. Reischle.
Human Kinetics Publishers, Champaign, Ill. pp. 295–303.
Wilke, K. 1992. Analysis of sprint swimming: the 50 m freestyle.
In Biomechanics and medicine in swimming VI. Edited by D.
McLaren, T. Reilly, and A. Lees. pp. 33–46.
© 2006 NRC Canada
158 Appl. Physiol. Nutr. Metab. Vol. 31, 2006