- Email: [email protected]
Importance of sagittal kick symmetry for underwater dolphin kick performance
Human Movement Science 33 (2014) 298–311
Contents lists available at ScienceDirect
Human Movement Science journal homepage: www.elsevier.com/locate/humov
Importance of sagittal kick symmetry for underwater dolphin kick performance Ryan R. Atkison ⇑, James P. Dickey, Andrew Dragunas, Volker Nolte School of Kinesiology, The University of Western Ontario, Thames Hall 2125, 1151 Richmond Street, London, Ontario N6A 3K7, Canada
a r t i c l e
i n f o
Article history: Available online 27 November 2013 PscyINFO classification: 3720 Keywords: Sports Swimming Kinematics Performance Upkick Downkick
a b s t r a c t The purpose of this study was to determine how sagittal kick symmetry in the underwater dolphin kick (UDK) between the downkick and upkick phases is related to UDK performance. Fifteen adult male competitive swimmers ranging from provincial to international level were filmed performing three trials each of maximum effort UDK over 15 m using an underwater video camera. Video frames were manually digitized and each subjects’ single fastest trial was evaluated for between-subject comparisons. Kinematic variables were calculated for each individual and Pearson product-moment correlations between the average horizontal centre of mass velocity (Vx) and all kinematic variables were calculated. Horizontal velocity during the downkick, horizontal velocity during the upkick, relative time spent in each phase, maximum chest flexion angle, maximum knee and ankle extension angles, the ratio of flexion/extension for chest, knee and ankle angles, and maximum vertical toe velocity during the upkick phase correlated significantly with Vx (p < 0.05). The ratio of downkick vertical toe velocity/upkick vertical toe velocity was significantly negatively correlated with Vx (p < 0.05). These results indicate the importance of kick symmetry for UDK performance, and indicate that performing the upkick phase well appears to be most important for UDK performance. Ó 2013 Elsevier B.V. All rights reserved.
⇑ Corresponding author. Present address: Canadian Sport Institute Ontario, 12 Concorde Place, Suite 204, Toronto, Ontario M3C 3R8, Canada. Tel.: +1 647 210 7672. E-mail addresses: [email protected] (R.R. Atkison), [email protected] (J.P. Dickey), [email protected] (A. Dragunas), [email protected] (V. Nolte). 0167-9457/$ - see front matter Ó 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.humov.2013.08.013
R.R. Atkison et al. / Human Movement Science 33 (2014) 298–311
299
1. Introduction There is a growing research interest in the underwater dolphin kick (UDK) in competitive swimming due to the apparent advantages that can be gained by successfully performing UDK sequences following starts and turns. The major difference between swimming below versus on the surface of the water is improved propulsive efficiency through reduced resisted drag force (Lyttle, Blanksby, Elliott, & Lloyd, 1998). This is accomplished through the reduction of wave drag, as it is minimized at a depth of 0.6–0.7 m at speeds experienced by elite swimmers following turns and starts (1.9–3.1 m/s; Lyttle et al., 1998; Vennell, Pease, & Wilson, 2006). It is legal for swimmers to remain underwater for 15 m following a start or turn in butterfly, backstroke, and freestyle (FINA, 2009); equal to 30% of the total race distance in a 50 m long course (Olympic size) pool, and 60% of the total race distance in a 25 m short course pool. Since nearly a third of the race distance in long course and two thirds in short course competition can be covered underwater, an effective UDK can benefit swimming race performance. The underwater dolphin kick is a cyclic motion where the toes oscillate in a regular fashion with one spatial maximum (up-peak) and one spatial minimum (down-peak) in the vertical direction occurring over the course of one cycle. It has been suggested that a wave travels caudally from the torso to the toes, increasing in amplitude at each body segment in a whip-like action, whereby momentum is transferred from the larger body segments to the smaller body segments (Gavilán, Arellano, & Sanders, 2006; Sanders, Cappaert, & Devlin, 1995; Ungerechts, 1983). Underwater dolphin kick can be broken down into two phases; the downkick and upkick. In a ventral body position, the downkick is characterized by hip flexion and knee extension and occurs from the up-peak position and ends in the down-peak position; the upkick is characterized by hip extension and knee flexion and occurs from the down-peak position and ends in the up-peak position. Figs. 1 and 2 respectively demonstrate a swimmer over the course of one complete kick cycle and the corresponding toe marker path during that cycle. Previous studies have compared the kinematics of human undulatory propulsion to the kinematics of dolphins/cetaceans (Ungerechts, 1983; Von Loebbecke, Mittal, Fish, & Mark, 2009a, 2009b). Ungerechts (1983) compared the kinematics of the butterfly stroke with dolphins and found that the primary difference between the human swimmers and dolphins was that dolphins were able to perform symmetrical downkick and upkick phases; in contrast, the human swimmers were relatively less effective at the upkick phase, and he concluded that only swimmers who were able to hyperextend their knees would be able to perform the upkick phase effectively. Despite similarities between the butterfly stroke and UDK, butterfly is performed on the surface and involves propulsion with the arms and legs. Von Loebbecke et al. (2009a) and Von Loebbecke et al. (2009b) compared the kinematics of the UDK in humans with odontocete cetaceans (class of mammalian swimmers, including dolphins, toothed whales, porpoises, etc.), finding that humans were less propulsively efficient and slower than cetaceans over the range of kicking frequencies and kicking amplitudes selected by the human swimmers. The differences between humans and cetaceans were attributed to the disadvantageous anatomy and musculature of humans, such as narrow feet and less-flexible joints, which especially limit the performance of the upkick phase. Kicking symmetry in the UDK is defined as the ability to produce equivalent propulsion during the downkick and upkick phases. This is accomplished through similar kinematics between the downkick and upkick phases. Theoretically, symmetry between downkick and upkick phases should result in more consistent CM velocity as there are two equivalently propulsive phases, compared with one propulsive and one resistive phase in an assymetrical kick cycle. Von Loebbecke et al. (2009a) evaluated the propulsive efficiency of human underwater dolphin kicking using Computational Fluid Dynamics (CFD) and found that parameters evaluating the entire kick cycle could not predict propulsive efficiency in UDK, but is connected to overall swimming style. Kinematics between downkick and upkick were compared by Arellano, Pardillo, and Gavilán (2000), but in this study the upkick phase was broken into two phases, which is inconsistent with the rest of the literature. Furthermore, data such as kick displacement and toe velocities were reported for the whole kick cycle, not for the individual downkick and upkick phases. Given their anatomical differences, symmetry between the downkick
300
R.R. Atkison et al. / Human Movement Science 33 (2014) 298–311
Fig. 1. From top to bottom, images of elite swimmer showing the components of a complete kick cycle: up-peak, downkick, down-peak, upkick, and up-peak.
R.R. Atkison et al. / Human Movement Science 33 (2014) 298–311
301
Fig. 2. The toe path of the swimmer shown in Fig. 1 with the corresponding peaks and phases labeled.
and upkick phases is an obvious limitation for human swimmers when compared with cetaceans; however, the relationship between kicking symmetry and performance has not been adequately studied in human swimmers. The purposes of this study were to evaluate the kinematics of downkick and upkick phases and how symmetry between downkick and upkick is related to performance. Symmetry in this experiment was evaluated by comparing joint marker paths, joint angles, horizontal displacement of the centre of mass, horizontal velocity of the centre of mass, and vertical toe velocities during the downkick and upkick phases. 2. Methods 2.1. Participants Fifteen adult male swimmers (mean ± SD; age = 21.5 ± 3.2 years, body length = 2.39 ± 0.13 m, thigh length = 0.42 ± 0.03 m, competitive swimming experience = 11.4 ± 5.6 years) from the university varsity swim team and local swim club volunteered to participate in the study. The swimmers ranged in skill level, from the provincial level to the international level (individual best performance ranged from 445 to 868 FINA points based on the 2010 LC FINA point chart; mean ± SD = 663 ± 134 FINA points). The protocol was fully explained to the participants and they provided written consent to participate in the study, which was approved by the university ethics committee. 2.2. Trial setup All trials were filmed from a Lorex CVC-6991 (Lorex Technology Inc., Ontario, Canada) underwater video camera (frame resolution = 720 480 pixels, frame rate = 30 Hz) and recorded to the hard drive of a digital camcorder (JVC GZMG555, JVC Kenwood Holdings, Japan). The underwater camera was secured 0.5 m below the surface of the water and 7.5 m from the initial impulse wall. The swimmers’ plane of motion was four metres from the camera and perpendicular to the camera’s field of view. A 2 m reference line was submerged parallel to the water surface along the swimmers’ path of motion. A picture of the reference line was taken, and the line was subsequently removed from the pool so that it would not interfere with the swimmers performing their trials. Coloured bricks were positioned on the pool bottom along the path of motion to assist the swimmers in maintaining a straight line and to
302
R.R. Atkison et al. / Human Movement Science 33 (2014) 298–311
indicate the completion of each trial (15 m). The camera position was not disturbed throughout the entire testing session. This study was a two-dimensional analysis of movements in the sagittal plane, with the z-axis defined as the vertical direction and x-axis defined as the horizontal direction. Upward movements toward the surface of the water were defined as positive, horizontal movements along the direction of motion were defined as positive. 2.3. Swim trials To accommodate the schedules of the athletes, three testing dates were offered spaced one week apart, each falling on the same day of the week. Testing began two weeks following the season championships before heavy training resumed, which ensured training fatigue did not play a factor in the results. Before entering the water, the joint locations for each swimmer were marked with black waterproof wax on predetermined landmarks on the right side of their body (Fig. 3). Since this was a two-dimensional study, symmetry was assumed between the left and right side. The landmarks used
Fig. 3. Swimmer demonstrating the position of the joint markers. From top to bottom, the landmarks (and naming convention used in this study) are the most distal phalange of the right hand (fingertip), styloid process of the ulna (wrist), olecranon process (elbow), vertex of the skull (head), greater tubercle of the humerus (shoulder), spinous process of T7 and nipples (chest), spinous process of T12 and umbilicus (upper waist), iliac crest (lower waist), greater trochanter (hip), lateral epicondyle of the femur (knee), lateral malleolus (ankle), and the most distal phalange of the foot (toe).
R.R. Atkison et al. / Human Movement Science 33 (2014) 298–311
303
(and name convention used in this study) are the most distal phalange of the right hand (fingertip), styloid process of the ulna (wrist), olecranon process (elbow), vertex of the skull (head), greater tubercle of the humerus (shoulder), spinous process of T7 and nipples (chest), spinous process of T12 and umbilicus (upper waist), iliac crest (lower waist), greater trochanter (hip), lateral epicondyle of the femur (knee), lateral malleolus (ankle), and the most distal phalange of the foot (toe). These landmarks are based on de Leva’s 2-dimensional segment model which was used in the calculation of the subjects centre of mass (de Leva, 1996), and matched landmarks used in a recent kinematic analysis of underwater monofin swimming (Nicolas & Bideau, 2009) and in an UDK simulation study (Nakashima, 2009). All landmarks were applied by the same individual to ensure consistency between subjects. All swimmers performed the same 30 min pre-testing warm-up which included UDK-specific practice and a familiarization session before proceeding with the swimming trials (Connaboy, Coleman, Moir, & Sanders, 2010). Each swimmer was filmed performing three trials of maximum effort UDK over 15 m with at least three minutes of rest between trials. The swimmers were required to perform the trials in a ventral body position, similarly to previous research (Arellano, Gavilán, & Garcia, 1998; Arellano et al., 2000). Swimmers were instructed to maintain a constant depth (between 0.5 and 1.0 m) to minimize the effects of wave-making resistance and to avoid interference from the pool bottom. The swimmers were monitored on a video screen and were provided feedback about their trajectory after each trial. If a swimmer did not maintain a constant depth between 0.5 and 1.0 m then they were required to repeat the trial. 2.4. Video analysis Videos of all trials were transferred from the digital camcorder hard disk drive to a portable hard disk drive. Videos were cropped and then augmented with a photo of the reference line at the beginning of each video (Pinnacle Studio Version 9, Pinnacle Systems, California, USA). The cropped videos contained three to five kick cycles depending on the distance that the swimmer travelled during each kick. Before further analysis, videos of all trials were reviewed and were discarded if it was clear that an adequate constant depth was not maintained or if there was poor visibility of joint markers in the video. In total, 49 videos were collected and 30 were retained for processing (two for each swimmer). Videos were imported and digitized manually in HUman Movement ANalysis software (HMA Technology, Ontario, Canada). The central kick cycle was analyzed. All digitized joint coordinates were exported to Microsoft Excel 2007 (Microsoft Corporation, Washington, USA) for processing and analysis. 2.5. Data processing All raw data was filtered using a Butterworth filter with cut-off frequencies from 4 to 5 Hz based on a residual analysis of the landmark data (Winter, 1990). Timing was calculated for each trial based on the camera frame rate of 30 Hz. To calculate the centre of mass (CM) in the sagittal plane, the coordinate of the CM for each body segment was calculated and then the CM coordinate for each body segment were weighted and summed based on the relative mass of each body segment to determine the coordinate of the whole body CM. Body segment CM values were based on de Leva’s two-dimensional body segment model (de Leva, 1996). 2.6. Kinematic variables We define one complete dolphin kick cycle starting immediately following the up-peak position, followed by the downkick phase that ends in the down-peak position, followed by the upkick phase that ends in the up-peak position. Kick amplitude (A) was calculated as the absolute value of the average of the vertical displacement covered during the downkick and during the upkick of each kick cycle, as the vertical toe position at the start of the kick cycle did not always match the vertical toe position at the end of the kick cycle. Kick frequency (f) was calculated as the reciprocal of the time taken for the toe to complete one cycle.
304
R.R. Atkison et al. / Human Movement Science 33 (2014) 298–311
Body length (L) was calculated as the maximum distance from the fingertip to the toe-tip within the measured kick cycle. Horizontal kick displacement (dx) was calculated as the distance travelled by the CM during one complete kick cycle. The horizontal velocities of each swimmer’s CM (Vx) and vertical velocities of the toe (VT) were calculated using a central finite difference formula. Durations of each phase as a percentage of kick cycle were also calculated. All reported kinematic variables were determined for the whole kick cycle and also for the downkick and upkick phases independently. Symmetry between the kinematics of the downkick and upkick were evaluated by dividing downkick values by upkick values, where ratios approaching one were considered to indicate symmetry between the downkick and upkick for that variable. All angles were measured using the convention shown in Fig. 4. To facilitate comparison between flexion and extension, 180° was subtracted from each angle value. Positive angle values were considered extension and negative angle values were considered flexion (for the ankle, dorsiflexion was reduced to flexion, and plantar flexion was reduced to extension). The maximum, minimum and range values for each joint angle were calculated for each kick cycle. To evaluate symmetry in joint angles between the downkick and upkick, the ratio of maximum flexion/maximum extension was calculated for each joint.
2.7. Data analysis For each subject, the measured time series data were reduced to single variables to allow for comparison between trials and between subjects. Maxima, minima, means and standard deviations were then calculated for each variable for each trial. The time of each swimmer’s kick cycle were normalized to 100% to compare the temporal symmetry between subjects and to facilitate graphical comparisons between subjects. Because separate trials for each swimmer were not independent of one another, only one trial for each swimmer was selected for further analysis. Since this was a performance-related study, the trial chosen for analysis was the one with the highest horizontal centre of mass velocity. Horizontal centre of mass velocity (Vx) was used as the primary measure of performance. Similar to previous research (Arellano et al., 2000), the x–z path of the toe and the time-course of Vx, was investigated to evaluate the symmetry between the downkick and upkick of the swimmers’ toe path and Vx. The swimmers with the slowest, median and fastest Vx were selected for comparison. All subsequent analysis was performed in SPSS 17 (IBM Corporation, New York, USA) on Microsoft Windows. Results of Shapiro-Wilks statistics (p > 0.05) allowed normality to be assumed. Pearson product-moment correlation analyses were performed between Vx and downkick and upkick kinematic variables to determine how kinematics of each phase relate to performance, and between Vx and symmetry ratios to determine how kinematic symmetry between downkick/upkick relates to performance. Data were considered to be statistically significant at p < 0.05.
Fig. 4. Definitions of angle measurements at the ankle (hA) knee (hK), hip (hH), shoulder (hS), elbow (hE) and wrist (hW). All angles are measured as relative angles. The measurement convention for the angles not indicated (lower waist, hL-W; upper waist, hU–W; and chest, hC) follow the same measurement convention as the hip and shoulder angles.
305
R.R. Atkison et al. / Human Movement Science 33 (2014) 298–311
2.8. Reliability and accuracy To test digitizing reliability, one randomly selected trial was digitized twice and the filtered coordinates for each time interval for the central kick cycle were compared. To test how accurately the digitized data represents actual measurements, thigh length was calculated from the digitized data and compared to real thigh length measurements taken before the tests (Nolte, 1985). The swimmers with the slowest, median and fastest Vx were selected for thigh length accuracy comparisons. 3. Results 3.1. Reliability and accuracy Error as a result of manual digitization can affect the reliability of the data. The largest difference was observed for the head in the x-direction (max difference = 0.043 m) but the head was only used in CM calculations to which it only contributed 6.9% of body mass. Max differences for all other
Table 1 Individual summary of measured and calculated kinematic data, including group means and SD.
Swimmer Swimmer Swimmer Swimmer Swimmer Swimmer Swimmer Swimmer Swimmer Swimmer Swimmer Swimmer Swimmer Swimmer Swimmer Mean SD
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
L (m)
Thigh (m)
Vx (m/s)
Vx max (m/s)
Vx min (m/s)
Vx/L (s
2.19 2.34 2.35 2.39 2.29 2.47 2.50 2.56 2.40 2.56 2.25 2.59 2.27 2.47 2.20 2.39 0.13
0.40 0.42 0.40 0.42 0.43 0.41 0.45 0.44 0.45 0.45 0.43 0.43 0.39 0.48 0.37 0.42 0.03
1.73 1.64 1.39 1.76 1.81 1.64 1.52 1.57 1.67 1.73 1.69 1.84 1.62 1.73 1.30 1.64 0.15
1.81 1.75 1.48 1.92 1.95 1.80 1.62 1.73 1.71 1.86 1.77 1.98 1.69 1.82 1.43 1.75 0.16
1.63 1.50 1.29 1.65 1.71 1.49 1.37 1.46 1.61 1.60 1.59 1.65 1.49 1.60 1.18 1.52 0.15
0.79 0.70 0.59 0.74 0.79 0.67 0.61 0.61 0.70 0.67 0.75 0.71 0.71 0.70 0.59 0.69 0.07
1
)
f (Hz)
A (m)
dx (m)
2.20 2.10 2.31 2.31 2.14 1.92 2.14 2.31 2.05 2.25 2.14 2.20 2.00 1.80 1.72 2.11 0.18
0.57 0.53 0.40 0.47 0.54 0.54 0.53 0.53 0.51 0.60 0.57 0.49 0.57 0.69 0.67 0.55 0.07
0.81 0.82 0.60 0.76 0.91 0.87 0.71 0.68 0.78 0.75 0.79 0.79 0.81 0.92 0.78 0.79 0.08
Fig. 5. Toe paths for one complete kick cycle of three swimmers selected from this study with the fastest, slowest, and median Vx swimmers for comparison. The markers have been presented with each cycle beginning at (0,0) to facilitate comparison.
306
R.R. Atkison et al. / Human Movement Science 33 (2014) 298–311
landmarks were less than 0.01 m, with the exception of the fingertip in the x-direction (max difference = 0.013 m), toe in the z-direction (max difference = 0.018 m) and toe in the x-direction (max difference = 0.018 m). Therefore, an uncertainty of 0.02 m was assumed for all data points used in kinematic calculations. This contributed to uncertainty in body length measurements of ±0.04 m. Maximum uncertainty of time measurements was ±1/30 s. Maximum difference in thigh lengths calculated during the central kick cycle was +0.03 m for the fastest UDK swimmer, 0.04 m for the median UDK swimmer, and +0.04 m for the slowest UDK swimmer when compared with their actual thigh measurements. These values all fall within the uncertainty of ±0.04 m expressed above. Thus, given the small differences between the video and physical measures, we determined that the data accurately represented actual measures.
Fig. 6. Horizontal centre of mass velocity and vertical toe velocity as a function of relative kick cycle duration for the fastest (a), median (b), and slowest (c) swimmer in the study. A vertical dashed line in each figure represents the downpeak of the kick cycle and the transition for downkick to upkick.
307
R.R. Atkison et al. / Human Movement Science 33 (2014) 298–311
3.2. Individual kinematic data For each subject the measured variables were reduced to single values. A summary of measured and calculated kinematic parameters for all subjects is presented in Table 1. The three swimmers with the fastest (1.84 m/s), slowest (1.30 m/s), and median (1.67 m/s) Vx were selected for joint marker comparison. Fig. 5 depicts the x–z toe paths of these three swimmers. The downkick portion for the fast swimmer began immediately, whereas there was a small horizontal movement for the median swimmer and a longer horizontal movement for the slow swimmer. The toe paths of the fast and slow swimmers moved in the negative horizontal direction during the downkick phase. The toe path of the median swimmer is constantly moving horizontally in a positive direction. All swimmers’ toes moved primarily in a horizontal direction at the end of the upkick. Horizontal CM velocities and vertical toe velocities over the course of complete kick cycles are shown for these three swimmers in Fig. 6. During the downkick phase, a peak in the horizontal CM velocity can be seen for each swimmer. These peaks correspond to the minima values of vertical toe velocity. Similarly, during the upkick phase distinct peaks of horizontal CM velocity can be seen for the slowest and median swimmer, and an increase in horizontal CM velocity can be seen for the fastest swimmer. These increases in horizontal CM velocity occur at the same time as, or immediately following the minima/maxima values of vertical toe velocity. The slowest swimmer and median swimmer demonstrated distinct horizontal deceleration periods at the beginning of the downkick before accelerating; whereas the fastest swimmer showed minimal deceleration at the beginning of the downkick. All swimmers decelerated toward the end of the upkick phase. Downkick vertical toe velocity peaked toward the end of the downkick phase, whereas upkick vertical toe velocity peaked early in the upkick phase.
3.3. Correlations between kinematic parameters and horizontal velocity Maxima, minima, means, SD, correlations with Vx and correlations with Vx/BL for each variable are presented in table 2. Kick cycle duration was significantly correlated with Vx, with the trend indicating that faster Vx was related to equal durations of downkick and upkick, and slower Vx was related to Table 2 Variables used to assess propulsive symmetry in the UDK include kick displacement, horizontal CM velocities and vertical toe velocities, and are reported for (a) downkick phase, (b) upkick phase, and (c) ratios of downkick/upkick phases. N, maxima, minima, means, SD, correlation with Vx, and correlation with Vx/BL are reported for each kinematic variable. N
*
Max
Min
Mean
SD
r with Vx
r with Vx//BL
(a) Downkick phase % Downkick Vx-DK (m/s) dx-DK (m) VzTavg-DK (m/s) VzTmax-DK (m/s)
15 15 15 15 15
50% 1.87 0.42 2.73 3.26
38% 1.32 0.24 2.08 4.35
45% 1.67 0.35 2.38 3.88
3% 0.14 0.04 0.23 0.34
0.486* 0.983* 0.625* 0.133 0.013
0.603* 0.761* 0.755* 0.335 0.067
(b) Upkick phase % Upkick Vx-UK (m/s) dx-UK (m) VzTavg-UK (m/s) VzTmax-UK (m/s)
15 15 15 15 15
63% 1.81 0.54 2.53 3.82
50% 1.28 0.36 1.41 2.08
55% 1.62 0.43 1.99 3.29
3% 0.15 0.05 0.29 0.43
0.486* 0.993* 0.338 0.395 0.631*
0.603* 0.831* 0.332 0.335 0.579*
(c) Downkick/upkick ratios % Downkick/% upkick Vx-DK/Vx-UK dx-DK/dx-UK VzTavg-DK/VzTavg-UK VzTmax-DK/VzTmax-UK
15 15 15 15 15
0.1 0.03 0.10 0.21 0.16
0.495* 0.412 0.364 0.453* 0.732*
0.624* 0.524* 0.486* 0.512* 0.698*
1 1.08 0.99 1.65 1.56
Correlation is significant at the p < 0.05 level.
0.6 0.98 0.63 0.95 1
0.82 1.03 0.82 1.22 1.19
308
R.R. Atkison et al. / Human Movement Science 33 (2014) 298–311
more relative time spent in the upkick phase. Horizontal CM velocity during the downkick and during the upkick was correlated significantly with Vx; however, horizontal CM velocity during the upkick phase was correlated more strongly with Vx. Surprisingly, the ratio of horizontal CM velocity during the downkick divided by horizontal CM velocity during the upkick was not significantly correlated with Vx. Horizontal kick displacement during the downkick was significantly correlated with Vx. Neither kick displacement during the upkick nor the ratio of displacement during downkick divided by displacement during upkick was significantly correlated with Vx. This is not surprising since kick displacement does not take duration into account. The ratio of average vertical toe velocity during the downkick divided by average vertical toe velocity during the upkick correlated significantly with Vx. Maximum vertical toe velocity during the upkick and the ratio of maximum vertical toe velocity during the downkick divided by maximum vertical toe velocity during the upkick both correlated significantly with Vx, and showed a stronger correlation than the average values. This indicates that maximum vertical toe velocity was likely more important for UDK performance than average vertical toe velocity. In addition, normalized horizontal CM velocity correlated significantly with the ratio of horizontal downkick velocity divided by horizontal upkick velocity, and the ratio of downkick displacement divided by upkick displacement. Maximum chest flexion angle correlated significantly with Vx (r = 0.607, p < 0.05). Maximum knee and ankle extension angles correlated significantly with Vx (r = 0.841, p < 0.05; r = 0.448, p < 0.05). The ratio of maximum flexion/extension chest, knee and ankle angles correlated significantly with Vx (r = 0.517, p < 0.05; r = 0.877, p < 0.05; r = 0.667, p < 0.05).
4. Discussion 4.1. General findings Our results supported the hypothesis that faster swimmers have a more symmetrical kick than slower swimmers, and that upkick performance is significantly related to horizontal velocity; however, certain variables were more strongly related to UDK performance. Kick symmetry has been previously assessed by measuring the path of the toes (Arellano et al., 2000). The x–z toe-paths in Fig. 5 show little indication of symmetry between the downkick and upkick. The toe-path of the median swimmer showed similarities between the downkick and upkick, but is still not symmetrical. The toe-paths for the downkick are more vertical than during the upkick for all three swimmers. Previous literature suggests that slower swimmers have a negative horizontal component during the downkick, whereas faster swimmers are always moving horizontally positive (Arellano et al., 2000). This is contrary to the results seen in this study, which show negative horizontal components during the downkick for fast and slow UDK swimmers, but not the median UDK swimmer. While these results are specific to these individuals, it seems that observing a swimmers toe-path alone is not indicative of UDK performance. According to the data presented in Fig. 6, peaks in horizontal CM velocity occurred at the same time as, or immediately following peaks in vertical toe velocity. Furthermore, the increase in horizontal velocity was greater for the downkick phase than for the upkick phase, corresponding to faster peak vertical toe velocities during the downkick phase than the upkick phase. This indicates a relationship between magnitude and timing of maximum vertical toe velocity and magnitude and timing of horizontal acceleration of the CM. These results mirror previously published research linking vortex size to magnitude of horizontal acceleration in UDK (Arellano, 1999; Arellano et al., 2000; Matsuuchi et al., 2006; Miwa, Matsuuchi, Shintani, Kamata, & Nomura, 2006; Von Loebbecke, Mittal, Mark, & Hahn, 2009c; Von Loebbecke et al., 2009a), and timing of vortex shedding to timing of horizontal acceleration in UDK (Von Loebbecke et al., 2009c). In swimming, vortices represent the transfer of momentum from water to body and vice versa, resulting in body acceleration (Ungerechts, Persyn, & Colman, 1999). Previous studies have demonstrated that efficient swimmers create a large static vortex at the end of the downward kick and a small vortex at the end of the upward kick, whereas inefficient swimmers create small translating vortices at the end of the downward kick and no vortices at the end of upward kick (Arellano, 1999; Arellano et al., 2000). Based on our findings and previous research
R.R. Atkison et al. / Human Movement Science 33 (2014) 298–311
309
on vortex shedding, there appears to be a relationship between magnitude of peak vertical toe velocity and vortex magnitude, and timing of peak vertical toe velocity and timing of vortex shedding. Because these findings are observational, future research should measure the relationship between vertical toe velocity and vortex shedding in UDK. All swimmers spent more time in the upkick than the downkick phase; however, the faster swimmers tended to spend more similar amounts of time in each kick phase, whereas the slower swimmers tended to spend relatively less time in the downkick and more time in the upkick. This suggests that the relative amount of time spent in each phase may be a valuable indicator of UDK performance. Fig. 6 shows that both vertical toe velocity and horizontal CM velocity peaks occur early in the upkick phase; thus, it seems advantageous to decelerate the feet and change direction quickly to initiate the next downkick phase sooner. Optimization simulations of the UDK predict that swimmers who exhibited greater similarity between maximum joint extension and maximum joint flexion would have higher average horizontal velocities (Nakashima, 2009). We observed evidence of this at the chest and it appears that upper thoracic flexibility is important for UDK performance. It is possible that the upper thoracic spine is used to (a) dampen the body undulations of the lower body segments, and (b) reduce resistive drag by maintaining a small angle of attack with the arms. We observed further evidence of this at the knee, indicating that less flexion and more hyperextension of the knee are related to UDK performance. Our results validate findings from previous studies, which have shown that faster swimmers perform the UDK with less flexion of the knee (Arellano et al., 2000), and that faster butterfly swimmers are able to hyperextend the knees to a greater degree than slower swimmers (Ungerechts, 1983). Similar to the knee angle findings, our results indicate that less maximum dorsiflexion and greater maximum plantar flexion of the ankle are related to UDK performance; this is consistent with the results from a simulation study of plantar flexion angle in swimmers which suggested that, up to a certain point, an increase in plantar flexion angle would lead to an increase in UDK performance (Sugimoto, Nakashima, Ichikawa, & Nomura, 2008). These results indicate a relationship of mechanical symmetry and velocity at the chest, knee and ankle joints. Our results also indicate the importance of flexibility in the upper thoracic spine, the knee joint and ankle joint for UDK performance, in particular the ability to hyperextend the upper thoracic spine, knee and ankle joints. We observed that the CM displacement during the downkick phase was significantly correlated with Vx. There was a small and non-significant correlation between centre of mass displacement during the upkick and horizontal velocity; similarly, there was a small and non-significant correlation between the ratio of displacement during downkick divided by displacement during upkick and horizontal velocity. Because the displacement values do not consider the time spent during each phase, some swimmers with low kicking frequencies could travel long distances during a kick phase despite a low horizontal velocity. It appears that kick displacement is not a reliable measure of UDK performance or of kick symmetry since it does not take into account the time spent during each kick phase. As expected, all measures of horizontal velocity were highly correlated with Vx, but average velocity during the upkick showed the highest correlation with Vx. This provides evidence that faster swimmers are more proficient at the upkick than slower swimmers. Surprisingly, the ratio of horizontal CM velocity during the downkick divided by horizontal CM velocity during the upkick did not correlate significantly with Vx. Our initial hypothesis was that faster swimmers would maintain a more even velocity between downkick and upkick, but our results did not support that statement. The maximum vertical toe velocity during the downkick was not significantly correlated with Vx; however, maximum vertical toe velocity during the upkick was significantly correlated with Vx. In contrast, average vertical toe velocity during the downkick and average vertical toe velocity during the upkick were not significantly correlated with Vx. It appears that peak vertical toe velocities are more important than average vertical toe velocities for UDK propulsion. All swimmers achieved higher magnitudes of maximum vertical toe velocity during the downkick than during the upkick, but those who were more symmetric tended to have faster horizontal CM velocities. Generating similar propulsion during the downkick and upkick by having similar maximum vertical toe velocities appears to be most important for attaining high Vx.
310
R.R. Atkison et al. / Human Movement Science 33 (2014) 298–311
Normalizing the data by body length resulted in two additional significant findings: a significant correlation between Vx/BL and the ratio of average horizontal downkick velocity divided by average horizontal upkick velocity; and a significant correlation between Vx/BL and the ratio of horizontal displacement during the downkick divided by horizontal displacement during the upkick. Taken together, our body length-normalized data indicate a stronger relationship between kick symmetry and UDK performance than the non-normalized data. 4.2. Implications Based on these findings coaches and swimmers should emphasize kick symmetry, specifically by focusing on equal upkick and downkick durations and high toe velocities during both upkick and downkick. After maximum upkick vertical toe velocity is achieved, the swimmer should quickly change kick direction to prevent a long deceleration period. Further studies should explore the mechanism of the upkick so that coaches are better able to instruct swimmers on how to successfully perform it. Determining the relationship between UDK performance and hip and leg flexor and extensor strength would be a logical step forward. 4.3. Limitations and assumptions It was assumed that the swimmers performed each trial with maximum effort. Because the swimmers were in varying training loads (pre-competition or post-competition) it is possible that the trials were not representative of the swimmers best effort and/or technique. It was assumed that variability of skill level of the swimmers was due to technique differences. Because swimmers were from the same team, it was assumed that they all had the same level of UDK-specific training during the season preceding the testing. It was assumed that all motion occurred in the sagittal plane, and that each swimmer had a symmetrical kick on their right and left sides; however, detailed analysis of the images revealed that this was not always the case. Future studies should examine the effect of bilateral asymmetry on UDK performance. 5. Conclusion There appears to be a strong relationship between sagittal kick symmetry and performance in the underwater dolphin kick. It appears that most swimmers are able to perform the downkick successfully, but swimmers who are more effectively able to perform the upkick phase tend to be faster at the UDK. Factors that may contribute to improved upkick performance include increased knee hyperextension and less dorsiflexion at the start of the upkick, less knee flexion at the end of the upkick, limiting the duration of the upkick, and achieving higher maximum toe velocities during the upkick. Appendix. Individual kinematic data Average horizontal velocity ranged from 1.30 to 1.84 m/s. The fastest instantaneous velocity was 1.98 m/s and the slowest instantaneous velocity was 1.18 m/s. Body length-normalized average horizontal velocity had a smaller range, from 0.59 to 0.79 body lengths per second. Kick frequency ranged from 1.72 to 2.31 Hz and kick amplitude ranged from 0.40 to 0.69 m. The non-dimensional Strouhal number ranged from 0.59 to 0.88, where smaller values tend to indicate higher efficiency. Horizontal kick displacement ranged from 0.60 to 0.92 m. Appendix. Correlations between kinematic parameters and horizontal velocity Thigh length was significantly correlated with Vx (r = 0.509, p < 0.05). Body length was not significantly correlated with Vx. Neither frequency nor amplitude was significantly correlated with Vx. The non-dimensional St was significantly correlated with Vx (r = 0.580, p < 0.05). Maximum horizontal velocity and minimum horizontal velocity correlated significantly with Vx (r = 0.971, p < 0.05;
R.R. Atkison et al. / Human Movement Science 33 (2014) 298–311
311
r = 0.979, p < 0.05). Horizontal displacement during one complete kick cycle correlated significantly with Vx (r = 0.547, p < 0.05). Average vertical toe velocity over the entire kick cycle did not correlate significantly with Vx. References Arellano, R. (1999). Vortices and propulsion. In R. Sanders & J. Linsten (Eds.), Applied proceedings of the XVII international symposium on biomechanics in sports: Swimming (pp. 53–65). Perth: Edith Cowan University. Arellano, R., Gavilán, A., & Garcia, F. (1998). A comparison of the underwater undulatory swimming technique in two different body positions. In K. Keskinen, P. Komi & A.P. Hollander (Eds.), Proceedings of the VIII International Symposium on Biomechanics and Medicine in Swimming (pp. 25–28). Finland: University of Jyvaskyla. Arellano, R., Pardillo, S., & Gavilán, A. (2000). Underwater undulatory swimming: kinematic characteristics, vortex generation and application during the start, turn and swimming strokes. Coaches’ Infoservice [On-line]. Available from:. Connaboy, C., Coleman, S., Moir, G., & Sanders, R. (2010). Measures of reliability in the kinematics of maximal undulatory underwater swimming. Medicine and Science in Sports and Exercise, 42(4), 762–770. de Leva, P. (1996). Adjustments to Zatsiorsky-Seluyanov’s segment inertia parameters. Journal of Biomechanics, 29, 1223–1230. FINA (2009). FINA Swimming Rules 2009-2013. Fédération Internationale de Natation - fina.org [On-line]. Available from: . Gavilán, A., Arellano, R., & Sanders, R. (2006). Underwater undulatory swimming: study of frequency, amplitude and phase characteristics of the ‘body wave’. In J. P. Vilas-Boas, F. Alves, & A. Marques (Eds.), Biomechanics and medicine in swimming X (pp. 35–37). Porto: Portuguese Journal of Sport Sciences. Lyttle, A. D., Blanksby, A., Elliott, B. C., & Lloyd, D. G. (1998). The effect of depth and velocity on drag during the streamlined glide. Journal of Swimming Research, 13, 15–22. Matsuuchi, K., Hashizume, T., Nakazawa, Y., Nomura, T., Shintani, H., & Miwa, T. (2006). Flow visualization of unsteady flow field around a monofin using PIV. In J. P. Vilas-Boas, F. Alves, & A. Marques (Eds.), Biomechanics and medicine in swimming X (pp. 60–62). Porto: Portuguese Journal of Sport Sciences. Miwa, T., Matsuuchi, K., Shintani, H., Kamata, E., & Nomura, T. (2006). Unsteady flow measurement of dolphin kicking wake in sagittal plane using 2C-PIV. In J. P. Vilas-Boas, F. Alves, & A. Marques (Eds.), Biomechanics and medicine in swimming X (pp. 64–66). Porto: Portuguese Journal of Sport Sciences. Nakashima, M. (2009). Simulation analysis of the effect of trunk undulation on swimming performance in underwater dolphin kick of human. Journal of Biomechanical Science and Engineering, 4, 94–104. Nicolas, G., & Bideau, B. (2009). A kinematic and dynamic comparison of surface and underwater displacement in high level monofin swimming. Human Movement Science, 28, 480–493. Nolte, V. (1985). Die Effektivität des Ruderschlages [The Effectiveness of the Rowing Stroke]. Berlin: Bartels & Wernitz. Sanders, R., Cappaert, J., & Devlin, R. (1995). Wave characteristics of butterfly swimming. Journal of Biomechanics, 28, 9–16. Sugimoto, S., Nakashima, M., Ichikawa, H., & Nomura, T. (2008). The effects of plantar flexion angle increment on the performance during underwater dolphin kick using simulation analysis. Japan Journal of Physical Education, Health and Sport Sciences, 53, 51–60. Ungerechts, B. (1983). A comparison of the movements of the rear parts of dolphins and butterfly swimmers. In A. P. Hollander, P. A. Huying, & G. de Groot (Eds.), Biomechanics and Medicine in Swimming (pp. 215–221). Champaign, Ill: Human Kinetics. Ungerechts, B., Persyn, U., & Colman, V. (1999). Application of vortex flow formation to self-propulsion in water. In K. Keskinen, P. Komi & A.P. Hollander (Eds.), Proceedings of the VIII International Symposium on Biomechanics and Medicine in Swimming (pp. 95–100), Finland: University of Jyvaskyla. Vennell, R., Pease, D., & Wilson, B. (2006). Wave drag on human swimmers. Journal of Biomechanics, 39, 664–671. Von Loebbecke, A., Mittal, R., Fish, F., & Mark, R. (2009a). Propulsive efficiency of the underwater dolphin kick in humans. Journal of Biomechanical Engineering, 131, 054504. Von Loebbecke, A., Mittal, R., Fish, F., & Mark, R. (2009b). A comparison of the kinematics of the dolphin kick in humans and cetaceans. Human Movement Science, 28(1), 99–112. Elsevier B.V. http://dx.doi.org/10.1016/j.humov.2008.07.005. Von Loebbecke, A., Mittal, R., Mark, R., & Hahn, J. (2009c). A computational method for analysis of underwater dolphin kick hydrodynamics in human swimming. Sports Biomechanics, 8, 60–77. Winter, D. A. (1990). Biomechanics and motor control of human movement. New York: John Wiley & Sons.
Contents lists available at ScienceDirect
Human Movement Science journal homepage: www.elsevier.com/locate/humov
Importance of sagittal kick symmetry for underwater dolphin kick performance Ryan R. Atkison ⇑, James P. Dickey, Andrew Dragunas, Volker Nolte School of Kinesiology, The University of Western Ontario, Thames Hall 2125, 1151 Richmond Street, London, Ontario N6A 3K7, Canada
a r t i c l e
i n f o
Article history: Available online 27 November 2013 PscyINFO classification: 3720 Keywords: Sports Swimming Kinematics Performance Upkick Downkick
a b s t r a c t The purpose of this study was to determine how sagittal kick symmetry in the underwater dolphin kick (UDK) between the downkick and upkick phases is related to UDK performance. Fifteen adult male competitive swimmers ranging from provincial to international level were filmed performing three trials each of maximum effort UDK over 15 m using an underwater video camera. Video frames were manually digitized and each subjects’ single fastest trial was evaluated for between-subject comparisons. Kinematic variables were calculated for each individual and Pearson product-moment correlations between the average horizontal centre of mass velocity (Vx) and all kinematic variables were calculated. Horizontal velocity during the downkick, horizontal velocity during the upkick, relative time spent in each phase, maximum chest flexion angle, maximum knee and ankle extension angles, the ratio of flexion/extension for chest, knee and ankle angles, and maximum vertical toe velocity during the upkick phase correlated significantly with Vx (p < 0.05). The ratio of downkick vertical toe velocity/upkick vertical toe velocity was significantly negatively correlated with Vx (p < 0.05). These results indicate the importance of kick symmetry for UDK performance, and indicate that performing the upkick phase well appears to be most important for UDK performance. Ó 2013 Elsevier B.V. All rights reserved.
⇑ Corresponding author. Present address: Canadian Sport Institute Ontario, 12 Concorde Place, Suite 204, Toronto, Ontario M3C 3R8, Canada. Tel.: +1 647 210 7672. E-mail addresses: [email protected] (R.R. Atkison), [email protected] (J.P. Dickey), [email protected] (A. Dragunas), [email protected] (V. Nolte). 0167-9457/$ - see front matter Ó 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.humov.2013.08.013
R.R. Atkison et al. / Human Movement Science 33 (2014) 298–311
299
1. Introduction There is a growing research interest in the underwater dolphin kick (UDK) in competitive swimming due to the apparent advantages that can be gained by successfully performing UDK sequences following starts and turns. The major difference between swimming below versus on the surface of the water is improved propulsive efficiency through reduced resisted drag force (Lyttle, Blanksby, Elliott, & Lloyd, 1998). This is accomplished through the reduction of wave drag, as it is minimized at a depth of 0.6–0.7 m at speeds experienced by elite swimmers following turns and starts (1.9–3.1 m/s; Lyttle et al., 1998; Vennell, Pease, & Wilson, 2006). It is legal for swimmers to remain underwater for 15 m following a start or turn in butterfly, backstroke, and freestyle (FINA, 2009); equal to 30% of the total race distance in a 50 m long course (Olympic size) pool, and 60% of the total race distance in a 25 m short course pool. Since nearly a third of the race distance in long course and two thirds in short course competition can be covered underwater, an effective UDK can benefit swimming race performance. The underwater dolphin kick is a cyclic motion where the toes oscillate in a regular fashion with one spatial maximum (up-peak) and one spatial minimum (down-peak) in the vertical direction occurring over the course of one cycle. It has been suggested that a wave travels caudally from the torso to the toes, increasing in amplitude at each body segment in a whip-like action, whereby momentum is transferred from the larger body segments to the smaller body segments (Gavilán, Arellano, & Sanders, 2006; Sanders, Cappaert, & Devlin, 1995; Ungerechts, 1983). Underwater dolphin kick can be broken down into two phases; the downkick and upkick. In a ventral body position, the downkick is characterized by hip flexion and knee extension and occurs from the up-peak position and ends in the down-peak position; the upkick is characterized by hip extension and knee flexion and occurs from the down-peak position and ends in the up-peak position. Figs. 1 and 2 respectively demonstrate a swimmer over the course of one complete kick cycle and the corresponding toe marker path during that cycle. Previous studies have compared the kinematics of human undulatory propulsion to the kinematics of dolphins/cetaceans (Ungerechts, 1983; Von Loebbecke, Mittal, Fish, & Mark, 2009a, 2009b). Ungerechts (1983) compared the kinematics of the butterfly stroke with dolphins and found that the primary difference between the human swimmers and dolphins was that dolphins were able to perform symmetrical downkick and upkick phases; in contrast, the human swimmers were relatively less effective at the upkick phase, and he concluded that only swimmers who were able to hyperextend their knees would be able to perform the upkick phase effectively. Despite similarities between the butterfly stroke and UDK, butterfly is performed on the surface and involves propulsion with the arms and legs. Von Loebbecke et al. (2009a) and Von Loebbecke et al. (2009b) compared the kinematics of the UDK in humans with odontocete cetaceans (class of mammalian swimmers, including dolphins, toothed whales, porpoises, etc.), finding that humans were less propulsively efficient and slower than cetaceans over the range of kicking frequencies and kicking amplitudes selected by the human swimmers. The differences between humans and cetaceans were attributed to the disadvantageous anatomy and musculature of humans, such as narrow feet and less-flexible joints, which especially limit the performance of the upkick phase. Kicking symmetry in the UDK is defined as the ability to produce equivalent propulsion during the downkick and upkick phases. This is accomplished through similar kinematics between the downkick and upkick phases. Theoretically, symmetry between downkick and upkick phases should result in more consistent CM velocity as there are two equivalently propulsive phases, compared with one propulsive and one resistive phase in an assymetrical kick cycle. Von Loebbecke et al. (2009a) evaluated the propulsive efficiency of human underwater dolphin kicking using Computational Fluid Dynamics (CFD) and found that parameters evaluating the entire kick cycle could not predict propulsive efficiency in UDK, but is connected to overall swimming style. Kinematics between downkick and upkick were compared by Arellano, Pardillo, and Gavilán (2000), but in this study the upkick phase was broken into two phases, which is inconsistent with the rest of the literature. Furthermore, data such as kick displacement and toe velocities were reported for the whole kick cycle, not for the individual downkick and upkick phases. Given their anatomical differences, symmetry between the downkick
300
R.R. Atkison et al. / Human Movement Science 33 (2014) 298–311
Fig. 1. From top to bottom, images of elite swimmer showing the components of a complete kick cycle: up-peak, downkick, down-peak, upkick, and up-peak.
R.R. Atkison et al. / Human Movement Science 33 (2014) 298–311
301
Fig. 2. The toe path of the swimmer shown in Fig. 1 with the corresponding peaks and phases labeled.
and upkick phases is an obvious limitation for human swimmers when compared with cetaceans; however, the relationship between kicking symmetry and performance has not been adequately studied in human swimmers. The purposes of this study were to evaluate the kinematics of downkick and upkick phases and how symmetry between downkick and upkick is related to performance. Symmetry in this experiment was evaluated by comparing joint marker paths, joint angles, horizontal displacement of the centre of mass, horizontal velocity of the centre of mass, and vertical toe velocities during the downkick and upkick phases. 2. Methods 2.1. Participants Fifteen adult male swimmers (mean ± SD; age = 21.5 ± 3.2 years, body length = 2.39 ± 0.13 m, thigh length = 0.42 ± 0.03 m, competitive swimming experience = 11.4 ± 5.6 years) from the university varsity swim team and local swim club volunteered to participate in the study. The swimmers ranged in skill level, from the provincial level to the international level (individual best performance ranged from 445 to 868 FINA points based on the 2010 LC FINA point chart; mean ± SD = 663 ± 134 FINA points). The protocol was fully explained to the participants and they provided written consent to participate in the study, which was approved by the university ethics committee. 2.2. Trial setup All trials were filmed from a Lorex CVC-6991 (Lorex Technology Inc., Ontario, Canada) underwater video camera (frame resolution = 720 480 pixels, frame rate = 30 Hz) and recorded to the hard drive of a digital camcorder (JVC GZMG555, JVC Kenwood Holdings, Japan). The underwater camera was secured 0.5 m below the surface of the water and 7.5 m from the initial impulse wall. The swimmers’ plane of motion was four metres from the camera and perpendicular to the camera’s field of view. A 2 m reference line was submerged parallel to the water surface along the swimmers’ path of motion. A picture of the reference line was taken, and the line was subsequently removed from the pool so that it would not interfere with the swimmers performing their trials. Coloured bricks were positioned on the pool bottom along the path of motion to assist the swimmers in maintaining a straight line and to
302
R.R. Atkison et al. / Human Movement Science 33 (2014) 298–311
indicate the completion of each trial (15 m). The camera position was not disturbed throughout the entire testing session. This study was a two-dimensional analysis of movements in the sagittal plane, with the z-axis defined as the vertical direction and x-axis defined as the horizontal direction. Upward movements toward the surface of the water were defined as positive, horizontal movements along the direction of motion were defined as positive. 2.3. Swim trials To accommodate the schedules of the athletes, three testing dates were offered spaced one week apart, each falling on the same day of the week. Testing began two weeks following the season championships before heavy training resumed, which ensured training fatigue did not play a factor in the results. Before entering the water, the joint locations for each swimmer were marked with black waterproof wax on predetermined landmarks on the right side of their body (Fig. 3). Since this was a two-dimensional study, symmetry was assumed between the left and right side. The landmarks used
Fig. 3. Swimmer demonstrating the position of the joint markers. From top to bottom, the landmarks (and naming convention used in this study) are the most distal phalange of the right hand (fingertip), styloid process of the ulna (wrist), olecranon process (elbow), vertex of the skull (head), greater tubercle of the humerus (shoulder), spinous process of T7 and nipples (chest), spinous process of T12 and umbilicus (upper waist), iliac crest (lower waist), greater trochanter (hip), lateral epicondyle of the femur (knee), lateral malleolus (ankle), and the most distal phalange of the foot (toe).
R.R. Atkison et al. / Human Movement Science 33 (2014) 298–311
303
(and name convention used in this study) are the most distal phalange of the right hand (fingertip), styloid process of the ulna (wrist), olecranon process (elbow), vertex of the skull (head), greater tubercle of the humerus (shoulder), spinous process of T7 and nipples (chest), spinous process of T12 and umbilicus (upper waist), iliac crest (lower waist), greater trochanter (hip), lateral epicondyle of the femur (knee), lateral malleolus (ankle), and the most distal phalange of the foot (toe). These landmarks are based on de Leva’s 2-dimensional segment model which was used in the calculation of the subjects centre of mass (de Leva, 1996), and matched landmarks used in a recent kinematic analysis of underwater monofin swimming (Nicolas & Bideau, 2009) and in an UDK simulation study (Nakashima, 2009). All landmarks were applied by the same individual to ensure consistency between subjects. All swimmers performed the same 30 min pre-testing warm-up which included UDK-specific practice and a familiarization session before proceeding with the swimming trials (Connaboy, Coleman, Moir, & Sanders, 2010). Each swimmer was filmed performing three trials of maximum effort UDK over 15 m with at least three minutes of rest between trials. The swimmers were required to perform the trials in a ventral body position, similarly to previous research (Arellano, Gavilán, & Garcia, 1998; Arellano et al., 2000). Swimmers were instructed to maintain a constant depth (between 0.5 and 1.0 m) to minimize the effects of wave-making resistance and to avoid interference from the pool bottom. The swimmers were monitored on a video screen and were provided feedback about their trajectory after each trial. If a swimmer did not maintain a constant depth between 0.5 and 1.0 m then they were required to repeat the trial. 2.4. Video analysis Videos of all trials were transferred from the digital camcorder hard disk drive to a portable hard disk drive. Videos were cropped and then augmented with a photo of the reference line at the beginning of each video (Pinnacle Studio Version 9, Pinnacle Systems, California, USA). The cropped videos contained three to five kick cycles depending on the distance that the swimmer travelled during each kick. Before further analysis, videos of all trials were reviewed and were discarded if it was clear that an adequate constant depth was not maintained or if there was poor visibility of joint markers in the video. In total, 49 videos were collected and 30 were retained for processing (two for each swimmer). Videos were imported and digitized manually in HUman Movement ANalysis software (HMA Technology, Ontario, Canada). The central kick cycle was analyzed. All digitized joint coordinates were exported to Microsoft Excel 2007 (Microsoft Corporation, Washington, USA) for processing and analysis. 2.5. Data processing All raw data was filtered using a Butterworth filter with cut-off frequencies from 4 to 5 Hz based on a residual analysis of the landmark data (Winter, 1990). Timing was calculated for each trial based on the camera frame rate of 30 Hz. To calculate the centre of mass (CM) in the sagittal plane, the coordinate of the CM for each body segment was calculated and then the CM coordinate for each body segment were weighted and summed based on the relative mass of each body segment to determine the coordinate of the whole body CM. Body segment CM values were based on de Leva’s two-dimensional body segment model (de Leva, 1996). 2.6. Kinematic variables We define one complete dolphin kick cycle starting immediately following the up-peak position, followed by the downkick phase that ends in the down-peak position, followed by the upkick phase that ends in the up-peak position. Kick amplitude (A) was calculated as the absolute value of the average of the vertical displacement covered during the downkick and during the upkick of each kick cycle, as the vertical toe position at the start of the kick cycle did not always match the vertical toe position at the end of the kick cycle. Kick frequency (f) was calculated as the reciprocal of the time taken for the toe to complete one cycle.
304
R.R. Atkison et al. / Human Movement Science 33 (2014) 298–311
Body length (L) was calculated as the maximum distance from the fingertip to the toe-tip within the measured kick cycle. Horizontal kick displacement (dx) was calculated as the distance travelled by the CM during one complete kick cycle. The horizontal velocities of each swimmer’s CM (Vx) and vertical velocities of the toe (VT) were calculated using a central finite difference formula. Durations of each phase as a percentage of kick cycle were also calculated. All reported kinematic variables were determined for the whole kick cycle and also for the downkick and upkick phases independently. Symmetry between the kinematics of the downkick and upkick were evaluated by dividing downkick values by upkick values, where ratios approaching one were considered to indicate symmetry between the downkick and upkick for that variable. All angles were measured using the convention shown in Fig. 4. To facilitate comparison between flexion and extension, 180° was subtracted from each angle value. Positive angle values were considered extension and negative angle values were considered flexion (for the ankle, dorsiflexion was reduced to flexion, and plantar flexion was reduced to extension). The maximum, minimum and range values for each joint angle were calculated for each kick cycle. To evaluate symmetry in joint angles between the downkick and upkick, the ratio of maximum flexion/maximum extension was calculated for each joint.
2.7. Data analysis For each subject, the measured time series data were reduced to single variables to allow for comparison between trials and between subjects. Maxima, minima, means and standard deviations were then calculated for each variable for each trial. The time of each swimmer’s kick cycle were normalized to 100% to compare the temporal symmetry between subjects and to facilitate graphical comparisons between subjects. Because separate trials for each swimmer were not independent of one another, only one trial for each swimmer was selected for further analysis. Since this was a performance-related study, the trial chosen for analysis was the one with the highest horizontal centre of mass velocity. Horizontal centre of mass velocity (Vx) was used as the primary measure of performance. Similar to previous research (Arellano et al., 2000), the x–z path of the toe and the time-course of Vx, was investigated to evaluate the symmetry between the downkick and upkick of the swimmers’ toe path and Vx. The swimmers with the slowest, median and fastest Vx were selected for comparison. All subsequent analysis was performed in SPSS 17 (IBM Corporation, New York, USA) on Microsoft Windows. Results of Shapiro-Wilks statistics (p > 0.05) allowed normality to be assumed. Pearson product-moment correlation analyses were performed between Vx and downkick and upkick kinematic variables to determine how kinematics of each phase relate to performance, and between Vx and symmetry ratios to determine how kinematic symmetry between downkick/upkick relates to performance. Data were considered to be statistically significant at p < 0.05.
Fig. 4. Definitions of angle measurements at the ankle (hA) knee (hK), hip (hH), shoulder (hS), elbow (hE) and wrist (hW). All angles are measured as relative angles. The measurement convention for the angles not indicated (lower waist, hL-W; upper waist, hU–W; and chest, hC) follow the same measurement convention as the hip and shoulder angles.
305
R.R. Atkison et al. / Human Movement Science 33 (2014) 298–311
2.8. Reliability and accuracy To test digitizing reliability, one randomly selected trial was digitized twice and the filtered coordinates for each time interval for the central kick cycle were compared. To test how accurately the digitized data represents actual measurements, thigh length was calculated from the digitized data and compared to real thigh length measurements taken before the tests (Nolte, 1985). The swimmers with the slowest, median and fastest Vx were selected for thigh length accuracy comparisons. 3. Results 3.1. Reliability and accuracy Error as a result of manual digitization can affect the reliability of the data. The largest difference was observed for the head in the x-direction (max difference = 0.043 m) but the head was only used in CM calculations to which it only contributed 6.9% of body mass. Max differences for all other
Table 1 Individual summary of measured and calculated kinematic data, including group means and SD.
Swimmer Swimmer Swimmer Swimmer Swimmer Swimmer Swimmer Swimmer Swimmer Swimmer Swimmer Swimmer Swimmer Swimmer Swimmer Mean SD
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
L (m)
Thigh (m)
Vx (m/s)
Vx max (m/s)
Vx min (m/s)
Vx/L (s
2.19 2.34 2.35 2.39 2.29 2.47 2.50 2.56 2.40 2.56 2.25 2.59 2.27 2.47 2.20 2.39 0.13
0.40 0.42 0.40 0.42 0.43 0.41 0.45 0.44 0.45 0.45 0.43 0.43 0.39 0.48 0.37 0.42 0.03
1.73 1.64 1.39 1.76 1.81 1.64 1.52 1.57 1.67 1.73 1.69 1.84 1.62 1.73 1.30 1.64 0.15
1.81 1.75 1.48 1.92 1.95 1.80 1.62 1.73 1.71 1.86 1.77 1.98 1.69 1.82 1.43 1.75 0.16
1.63 1.50 1.29 1.65 1.71 1.49 1.37 1.46 1.61 1.60 1.59 1.65 1.49 1.60 1.18 1.52 0.15
0.79 0.70 0.59 0.74 0.79 0.67 0.61 0.61 0.70 0.67 0.75 0.71 0.71 0.70 0.59 0.69 0.07
1
)
f (Hz)
A (m)
dx (m)
2.20 2.10 2.31 2.31 2.14 1.92 2.14 2.31 2.05 2.25 2.14 2.20 2.00 1.80 1.72 2.11 0.18
0.57 0.53 0.40 0.47 0.54 0.54 0.53 0.53 0.51 0.60 0.57 0.49 0.57 0.69 0.67 0.55 0.07
0.81 0.82 0.60 0.76 0.91 0.87 0.71 0.68 0.78 0.75 0.79 0.79 0.81 0.92 0.78 0.79 0.08
Fig. 5. Toe paths for one complete kick cycle of three swimmers selected from this study with the fastest, slowest, and median Vx swimmers for comparison. The markers have been presented with each cycle beginning at (0,0) to facilitate comparison.
306
R.R. Atkison et al. / Human Movement Science 33 (2014) 298–311
landmarks were less than 0.01 m, with the exception of the fingertip in the x-direction (max difference = 0.013 m), toe in the z-direction (max difference = 0.018 m) and toe in the x-direction (max difference = 0.018 m). Therefore, an uncertainty of 0.02 m was assumed for all data points used in kinematic calculations. This contributed to uncertainty in body length measurements of ±0.04 m. Maximum uncertainty of time measurements was ±1/30 s. Maximum difference in thigh lengths calculated during the central kick cycle was +0.03 m for the fastest UDK swimmer, 0.04 m for the median UDK swimmer, and +0.04 m for the slowest UDK swimmer when compared with their actual thigh measurements. These values all fall within the uncertainty of ±0.04 m expressed above. Thus, given the small differences between the video and physical measures, we determined that the data accurately represented actual measures.
Fig. 6. Horizontal centre of mass velocity and vertical toe velocity as a function of relative kick cycle duration for the fastest (a), median (b), and slowest (c) swimmer in the study. A vertical dashed line in each figure represents the downpeak of the kick cycle and the transition for downkick to upkick.
307
R.R. Atkison et al. / Human Movement Science 33 (2014) 298–311
3.2. Individual kinematic data For each subject the measured variables were reduced to single values. A summary of measured and calculated kinematic parameters for all subjects is presented in Table 1. The three swimmers with the fastest (1.84 m/s), slowest (1.30 m/s), and median (1.67 m/s) Vx were selected for joint marker comparison. Fig. 5 depicts the x–z toe paths of these three swimmers. The downkick portion for the fast swimmer began immediately, whereas there was a small horizontal movement for the median swimmer and a longer horizontal movement for the slow swimmer. The toe paths of the fast and slow swimmers moved in the negative horizontal direction during the downkick phase. The toe path of the median swimmer is constantly moving horizontally in a positive direction. All swimmers’ toes moved primarily in a horizontal direction at the end of the upkick. Horizontal CM velocities and vertical toe velocities over the course of complete kick cycles are shown for these three swimmers in Fig. 6. During the downkick phase, a peak in the horizontal CM velocity can be seen for each swimmer. These peaks correspond to the minima values of vertical toe velocity. Similarly, during the upkick phase distinct peaks of horizontal CM velocity can be seen for the slowest and median swimmer, and an increase in horizontal CM velocity can be seen for the fastest swimmer. These increases in horizontal CM velocity occur at the same time as, or immediately following the minima/maxima values of vertical toe velocity. The slowest swimmer and median swimmer demonstrated distinct horizontal deceleration periods at the beginning of the downkick before accelerating; whereas the fastest swimmer showed minimal deceleration at the beginning of the downkick. All swimmers decelerated toward the end of the upkick phase. Downkick vertical toe velocity peaked toward the end of the downkick phase, whereas upkick vertical toe velocity peaked early in the upkick phase.
3.3. Correlations between kinematic parameters and horizontal velocity Maxima, minima, means, SD, correlations with Vx and correlations with Vx/BL for each variable are presented in table 2. Kick cycle duration was significantly correlated with Vx, with the trend indicating that faster Vx was related to equal durations of downkick and upkick, and slower Vx was related to Table 2 Variables used to assess propulsive symmetry in the UDK include kick displacement, horizontal CM velocities and vertical toe velocities, and are reported for (a) downkick phase, (b) upkick phase, and (c) ratios of downkick/upkick phases. N, maxima, minima, means, SD, correlation with Vx, and correlation with Vx/BL are reported for each kinematic variable. N
*
Max
Min
Mean
SD
r with Vx
r with Vx//BL
(a) Downkick phase % Downkick Vx-DK (m/s) dx-DK (m) VzTavg-DK (m/s) VzTmax-DK (m/s)
15 15 15 15 15
50% 1.87 0.42 2.73 3.26
38% 1.32 0.24 2.08 4.35
45% 1.67 0.35 2.38 3.88
3% 0.14 0.04 0.23 0.34
0.486* 0.983* 0.625* 0.133 0.013
0.603* 0.761* 0.755* 0.335 0.067
(b) Upkick phase % Upkick Vx-UK (m/s) dx-UK (m) VzTavg-UK (m/s) VzTmax-UK (m/s)
15 15 15 15 15
63% 1.81 0.54 2.53 3.82
50% 1.28 0.36 1.41 2.08
55% 1.62 0.43 1.99 3.29
3% 0.15 0.05 0.29 0.43
0.486* 0.993* 0.338 0.395 0.631*
0.603* 0.831* 0.332 0.335 0.579*
(c) Downkick/upkick ratios % Downkick/% upkick Vx-DK/Vx-UK dx-DK/dx-UK VzTavg-DK/VzTavg-UK VzTmax-DK/VzTmax-UK
15 15 15 15 15
0.1 0.03 0.10 0.21 0.16
0.495* 0.412 0.364 0.453* 0.732*
0.624* 0.524* 0.486* 0.512* 0.698*
1 1.08 0.99 1.65 1.56
Correlation is significant at the p < 0.05 level.
0.6 0.98 0.63 0.95 1
0.82 1.03 0.82 1.22 1.19
308
R.R. Atkison et al. / Human Movement Science 33 (2014) 298–311
more relative time spent in the upkick phase. Horizontal CM velocity during the downkick and during the upkick was correlated significantly with Vx; however, horizontal CM velocity during the upkick phase was correlated more strongly with Vx. Surprisingly, the ratio of horizontal CM velocity during the downkick divided by horizontal CM velocity during the upkick was not significantly correlated with Vx. Horizontal kick displacement during the downkick was significantly correlated with Vx. Neither kick displacement during the upkick nor the ratio of displacement during downkick divided by displacement during upkick was significantly correlated with Vx. This is not surprising since kick displacement does not take duration into account. The ratio of average vertical toe velocity during the downkick divided by average vertical toe velocity during the upkick correlated significantly with Vx. Maximum vertical toe velocity during the upkick and the ratio of maximum vertical toe velocity during the downkick divided by maximum vertical toe velocity during the upkick both correlated significantly with Vx, and showed a stronger correlation than the average values. This indicates that maximum vertical toe velocity was likely more important for UDK performance than average vertical toe velocity. In addition, normalized horizontal CM velocity correlated significantly with the ratio of horizontal downkick velocity divided by horizontal upkick velocity, and the ratio of downkick displacement divided by upkick displacement. Maximum chest flexion angle correlated significantly with Vx (r = 0.607, p < 0.05). Maximum knee and ankle extension angles correlated significantly with Vx (r = 0.841, p < 0.05; r = 0.448, p < 0.05). The ratio of maximum flexion/extension chest, knee and ankle angles correlated significantly with Vx (r = 0.517, p < 0.05; r = 0.877, p < 0.05; r = 0.667, p < 0.05).
4. Discussion 4.1. General findings Our results supported the hypothesis that faster swimmers have a more symmetrical kick than slower swimmers, and that upkick performance is significantly related to horizontal velocity; however, certain variables were more strongly related to UDK performance. Kick symmetry has been previously assessed by measuring the path of the toes (Arellano et al., 2000). The x–z toe-paths in Fig. 5 show little indication of symmetry between the downkick and upkick. The toe-path of the median swimmer showed similarities between the downkick and upkick, but is still not symmetrical. The toe-paths for the downkick are more vertical than during the upkick for all three swimmers. Previous literature suggests that slower swimmers have a negative horizontal component during the downkick, whereas faster swimmers are always moving horizontally positive (Arellano et al., 2000). This is contrary to the results seen in this study, which show negative horizontal components during the downkick for fast and slow UDK swimmers, but not the median UDK swimmer. While these results are specific to these individuals, it seems that observing a swimmers toe-path alone is not indicative of UDK performance. According to the data presented in Fig. 6, peaks in horizontal CM velocity occurred at the same time as, or immediately following peaks in vertical toe velocity. Furthermore, the increase in horizontal velocity was greater for the downkick phase than for the upkick phase, corresponding to faster peak vertical toe velocities during the downkick phase than the upkick phase. This indicates a relationship between magnitude and timing of maximum vertical toe velocity and magnitude and timing of horizontal acceleration of the CM. These results mirror previously published research linking vortex size to magnitude of horizontal acceleration in UDK (Arellano, 1999; Arellano et al., 2000; Matsuuchi et al., 2006; Miwa, Matsuuchi, Shintani, Kamata, & Nomura, 2006; Von Loebbecke, Mittal, Mark, & Hahn, 2009c; Von Loebbecke et al., 2009a), and timing of vortex shedding to timing of horizontal acceleration in UDK (Von Loebbecke et al., 2009c). In swimming, vortices represent the transfer of momentum from water to body and vice versa, resulting in body acceleration (Ungerechts, Persyn, & Colman, 1999). Previous studies have demonstrated that efficient swimmers create a large static vortex at the end of the downward kick and a small vortex at the end of the upward kick, whereas inefficient swimmers create small translating vortices at the end of the downward kick and no vortices at the end of upward kick (Arellano, 1999; Arellano et al., 2000). Based on our findings and previous research
R.R. Atkison et al. / Human Movement Science 33 (2014) 298–311
309
on vortex shedding, there appears to be a relationship between magnitude of peak vertical toe velocity and vortex magnitude, and timing of peak vertical toe velocity and timing of vortex shedding. Because these findings are observational, future research should measure the relationship between vertical toe velocity and vortex shedding in UDK. All swimmers spent more time in the upkick than the downkick phase; however, the faster swimmers tended to spend more similar amounts of time in each kick phase, whereas the slower swimmers tended to spend relatively less time in the downkick and more time in the upkick. This suggests that the relative amount of time spent in each phase may be a valuable indicator of UDK performance. Fig. 6 shows that both vertical toe velocity and horizontal CM velocity peaks occur early in the upkick phase; thus, it seems advantageous to decelerate the feet and change direction quickly to initiate the next downkick phase sooner. Optimization simulations of the UDK predict that swimmers who exhibited greater similarity between maximum joint extension and maximum joint flexion would have higher average horizontal velocities (Nakashima, 2009). We observed evidence of this at the chest and it appears that upper thoracic flexibility is important for UDK performance. It is possible that the upper thoracic spine is used to (a) dampen the body undulations of the lower body segments, and (b) reduce resistive drag by maintaining a small angle of attack with the arms. We observed further evidence of this at the knee, indicating that less flexion and more hyperextension of the knee are related to UDK performance. Our results validate findings from previous studies, which have shown that faster swimmers perform the UDK with less flexion of the knee (Arellano et al., 2000), and that faster butterfly swimmers are able to hyperextend the knees to a greater degree than slower swimmers (Ungerechts, 1983). Similar to the knee angle findings, our results indicate that less maximum dorsiflexion and greater maximum plantar flexion of the ankle are related to UDK performance; this is consistent with the results from a simulation study of plantar flexion angle in swimmers which suggested that, up to a certain point, an increase in plantar flexion angle would lead to an increase in UDK performance (Sugimoto, Nakashima, Ichikawa, & Nomura, 2008). These results indicate a relationship of mechanical symmetry and velocity at the chest, knee and ankle joints. Our results also indicate the importance of flexibility in the upper thoracic spine, the knee joint and ankle joint for UDK performance, in particular the ability to hyperextend the upper thoracic spine, knee and ankle joints. We observed that the CM displacement during the downkick phase was significantly correlated with Vx. There was a small and non-significant correlation between centre of mass displacement during the upkick and horizontal velocity; similarly, there was a small and non-significant correlation between the ratio of displacement during downkick divided by displacement during upkick and horizontal velocity. Because the displacement values do not consider the time spent during each phase, some swimmers with low kicking frequencies could travel long distances during a kick phase despite a low horizontal velocity. It appears that kick displacement is not a reliable measure of UDK performance or of kick symmetry since it does not take into account the time spent during each kick phase. As expected, all measures of horizontal velocity were highly correlated with Vx, but average velocity during the upkick showed the highest correlation with Vx. This provides evidence that faster swimmers are more proficient at the upkick than slower swimmers. Surprisingly, the ratio of horizontal CM velocity during the downkick divided by horizontal CM velocity during the upkick did not correlate significantly with Vx. Our initial hypothesis was that faster swimmers would maintain a more even velocity between downkick and upkick, but our results did not support that statement. The maximum vertical toe velocity during the downkick was not significantly correlated with Vx; however, maximum vertical toe velocity during the upkick was significantly correlated with Vx. In contrast, average vertical toe velocity during the downkick and average vertical toe velocity during the upkick were not significantly correlated with Vx. It appears that peak vertical toe velocities are more important than average vertical toe velocities for UDK propulsion. All swimmers achieved higher magnitudes of maximum vertical toe velocity during the downkick than during the upkick, but those who were more symmetric tended to have faster horizontal CM velocities. Generating similar propulsion during the downkick and upkick by having similar maximum vertical toe velocities appears to be most important for attaining high Vx.
310
R.R. Atkison et al. / Human Movement Science 33 (2014) 298–311
Normalizing the data by body length resulted in two additional significant findings: a significant correlation between Vx/BL and the ratio of average horizontal downkick velocity divided by average horizontal upkick velocity; and a significant correlation between Vx/BL and the ratio of horizontal displacement during the downkick divided by horizontal displacement during the upkick. Taken together, our body length-normalized data indicate a stronger relationship between kick symmetry and UDK performance than the non-normalized data. 4.2. Implications Based on these findings coaches and swimmers should emphasize kick symmetry, specifically by focusing on equal upkick and downkick durations and high toe velocities during both upkick and downkick. After maximum upkick vertical toe velocity is achieved, the swimmer should quickly change kick direction to prevent a long deceleration period. Further studies should explore the mechanism of the upkick so that coaches are better able to instruct swimmers on how to successfully perform it. Determining the relationship between UDK performance and hip and leg flexor and extensor strength would be a logical step forward. 4.3. Limitations and assumptions It was assumed that the swimmers performed each trial with maximum effort. Because the swimmers were in varying training loads (pre-competition or post-competition) it is possible that the trials were not representative of the swimmers best effort and/or technique. It was assumed that variability of skill level of the swimmers was due to technique differences. Because swimmers were from the same team, it was assumed that they all had the same level of UDK-specific training during the season preceding the testing. It was assumed that all motion occurred in the sagittal plane, and that each swimmer had a symmetrical kick on their right and left sides; however, detailed analysis of the images revealed that this was not always the case. Future studies should examine the effect of bilateral asymmetry on UDK performance. 5. Conclusion There appears to be a strong relationship between sagittal kick symmetry and performance in the underwater dolphin kick. It appears that most swimmers are able to perform the downkick successfully, but swimmers who are more effectively able to perform the upkick phase tend to be faster at the UDK. Factors that may contribute to improved upkick performance include increased knee hyperextension and less dorsiflexion at the start of the upkick, less knee flexion at the end of the upkick, limiting the duration of the upkick, and achieving higher maximum toe velocities during the upkick. Appendix. Individual kinematic data Average horizontal velocity ranged from 1.30 to 1.84 m/s. The fastest instantaneous velocity was 1.98 m/s and the slowest instantaneous velocity was 1.18 m/s. Body length-normalized average horizontal velocity had a smaller range, from 0.59 to 0.79 body lengths per second. Kick frequency ranged from 1.72 to 2.31 Hz and kick amplitude ranged from 0.40 to 0.69 m. The non-dimensional Strouhal number ranged from 0.59 to 0.88, where smaller values tend to indicate higher efficiency. Horizontal kick displacement ranged from 0.60 to 0.92 m. Appendix. Correlations between kinematic parameters and horizontal velocity Thigh length was significantly correlated with Vx (r = 0.509, p < 0.05). Body length was not significantly correlated with Vx. Neither frequency nor amplitude was significantly correlated with Vx. The non-dimensional St was significantly correlated with Vx (r = 0.580, p < 0.05). Maximum horizontal velocity and minimum horizontal velocity correlated significantly with Vx (r = 0.971, p < 0.05;
R.R. Atkison et al. / Human Movement Science 33 (2014) 298–311
311
r = 0.979, p < 0.05). Horizontal displacement during one complete kick cycle correlated significantly with Vx (r = 0.547, p < 0.05). Average vertical toe velocity over the entire kick cycle did not correlate significantly with Vx. References Arellano, R. (1999). Vortices and propulsion. In R. Sanders & J. Linsten (Eds.), Applied proceedings of the XVII international symposium on biomechanics in sports: Swimming (pp. 53–65). Perth: Edith Cowan University. Arellano, R., Gavilán, A., & Garcia, F. (1998). A comparison of the underwater undulatory swimming technique in two different body positions. In K. Keskinen, P. Komi & A.P. Hollander (Eds.), Proceedings of the VIII International Symposium on Biomechanics and Medicine in Swimming (pp. 25–28). Finland: University of Jyvaskyla. Arellano, R., Pardillo, S., & Gavilán, A. (2000). Underwater undulatory swimming: kinematic characteristics, vortex generation and application during the start, turn and swimming strokes. Coaches’ Infoservice [On-line]. Available from: