Journal of Sports Sciences , 1999, 17 , 449 466

Upper extremity augmentation of lower extremity kinetics during countermovement vertical jumps
M ICHAEL E. FELTN ER,* DAN IEL J. F RASCH ETTI and RO BERT J. CRISP
Departm ent of Sports M edicine and Physical E ducatio n, Pepperdine University, M alibu, CA 90263, U SA

Accepted 26 July 1998

Twenty-W ve volleyball players (14 m ales, 11 fem ales) were videotaped (60 Hz) performing countermovement vertical jumps with and without an arm swing. Ground reaction force and video-based coordinate data were collected simultaneously. The resultant joint force and torque at the hip, knee, ankle and shoulder for two trials per subject per condition were computed and norm alized. Average kinem atic, resultant joint force and torque data were compared using repeated-measures analysis of variance. Larger values were recorded for the vertical velocity of the centre of mass at take-oV in the jumps with (m ean 2.75, s = 0.3 m s - 1 ) versus without (m ean 2.44, - 1 s = 0.23 m s ) an arm swing. The jumps with no arm swing produced larger torques at the hip during the W rst third of the propulsive phase (from zero to m aximum vertical velocity of the centre of m ass). During the W nal two-thirds of the propulsive phase, the arm swing augmented hip extensor torques by slowing the rate of trunk extension and placing the hip extensor muscles in slower concentric conditions that favoured the generation of larger forces and resultant joint torques. During the W rst two-thirds of the propulsive phase, knee extensor torque increased by 28% in the jumps with an arm swing, but m aintained a relatively constant m agnitude in the jumps with no arm swing.
K eywords : hip, joint torque, jumping, kinematics, kinetics, knee.

Introduction
Vertical jum ping is a vital skill com ponent in m any sports and recreational activities. W hen perform ing a m axim um heigh t vertical jum p, m ost athletes use a preparatory counterm ovem ent that results in a coordinated X exion of the hips, knees and ankles and a subsequent rapid extension of these sam e articulations before take-o V . In m ost jum ps, a rapid arm swing occurs sim ultaneously w ith the leg m otions. Several investigators have demonstrated that a counterm ovem ent increases the height of a ver tical jum p (Enoka, 1988; Khalid et al., 1989; H arm an et al. , 1990), as it increases the `pre-load on the lower extrem ity m usculature (Enoka, 1988; Anderson and Pandy, 1993; Zajac, 1993) and enables these m uscles to utilize the stretch shortening dynam ics of m uscular contraction (C avagn a et al. , 1968; Asm ussen and Bonde-Peterson, 1974; Kom i and Bosco, 1978; E noka, 1988; Anderson and Pandy, 1993; Z ajac, 1993). Additionally, the counter* Author to whom all correspondence should be addressed. e-mail: m feltner@pepperdine.edu 0264 0414/99

m ovem ent increases the time that the body has a positive upward acceleration (H arm an et al. , 1990; Zajac, 1993). Research has also indicated that an arm swing increases jum p height during counterm ovem ent vertical jum ps (Luhtanen and K om i, 1978; Payne et al. , 1968; Khalid et al. , 1989; O ddson, 1989; Shetty and Etnyre, 1989; Harm an et al. , 1990). H owever, the m echanism s that enable the arm s to increase jum p height are not well understood. A determ inistic m odel of vertical jum ping (Hay and Reid, 1988) is presented in Fig. 1. Jum p height, the height of the centre of m ass of the body at its peak, is a function of the height of the centre of m ass of the body at take-o V and X ight height, the vertical displacem ent of the centre of m ass of the body w hile airborne. D uring a vertical jum p, an arm sw ing that results in a body position of extreme shoulder X exion and elbow extension at take-o V (i.e. arm s extended and raised above the head) will increase the take-o V height of the centre of m ass of the body, resulting in an increased jum p height. However, the m otions of the arm s m ay also a V ect the m agnitude of the vertical com ponent of the ground reaction force and enhance the propulsive and

1999 E & FN Spon

450

Feltner et al.

F igure 1

Deterministic model of the vertical jump adapted from Hay and Reid (1988).

net im pulses exerted on the jum per. In turn, the larger net im pulse would result in increased vertical velocity of the centre of m ass of the body at take-o V and augm ented X ight height (Fig. 1). Payne et al. (1968) suggested that the arm swing enhances the m agnitude of the ver tical com ponent of the ground reaction force im m ediately before take-o V . D uring ver tical jum ping, M iller (1976) found that the net force exerted on both the body and trunk exhib ited a `double-peaked or quar tic trend consisting of two m axim a surrounding a local m inim um . She suggested that the arm m otion reduced the m agnitude of the m axim a values, but increased the m agnitude of the local m inim um , and had an overall e V ect of increasing jum p height. W hen the m usculature of the lower extremity was in an `advantageous position to exert vertical ground reaction force (Harm an et al. , 1990, p. 832), the upward acceleration of the arm s (associated with the arm sw ing) was seen to create a dow nward force on the body at the shoulders that slowed the rate of shortening of the quadriceps and gluteal m uscles. According to the force velocity relationship for m uscular contraction (H ill, 1938; Perrine and Edgerton, 1978), slower concentric actions of the leg m uscles would result in enhanced m uscle tension and presum ably larger vertical ground reaction forces. H owever, no experim ental data were provided to validate this m echanism . T he aim of this study was to exam ine the kinem atics and kinetics of the body during counterm ovem ent vertical jum ps perform ed both with and without an arm

swing. SpeciW cally, the study sought to determ ine if an arm sw ing can augm ent the ability of the lower extrem ity m usculature to generate tension, as m easured by the resultant joint torques at the hip, knee and ankle.

M ethods
Participants Twenty- W ve m em bers (14 m ales, 11 fem ales) of the 1995 Pepperdine University m en s and wom en s volleyball team s (m ales: height, m ean 193.8, s = 6.1 cm ; body m ass, m ean 88.2, s = 6.6 kg; age, m ean 20.5, s = 1.7 years; wom en: height, m ean 173.7, s = 8.8 cm ; body m ass, m ean 69.3, s = 7.1 kg; age, m ean 18.5, s = 0.7 years) provided voluntary w ritten inform ed consent and ser ved as participants. Each participant was videotaped perform ing W ve trials of counterm ovem ent vertical jum ps from a force platform both with and without a bilateral arm sw ing. Before perform ing the jum ps, each participant was given detailed instructions and allowed a brief period of practice. All jum ps were initiated from a stationary uprigh t posture and, during the jum ps w ith no arm swing, the hands rem ained on the subject s iliac crests. T he order of the jum ps was random ized and 1 2 m in rest was allowed between trials. The W rst two trials per condition that were perform ed correctly were used for the subsequent analyses.

Augmentation of lower extremity kinetics

Proc edures Before videotaping, several anthropom etric m easures were obtained to com pute m om ent of inertia data (Hinrichs, 1985). Retrore X ective m arkers were placed on nine body landm arks: centre of the head and the righ t shoulder, elbow, w rist, hip, knee, ankle, heel and toe. Assum ing bilateral sym m etry, these m arkers de W ned a six-se gm ent m odel of the body: head and trunk, upper arm , forearm and hand, thigh, shank, and foot. T he distance between the knee and ankle m arkers was also m easured to scale the video im ages. Ow ing to physical lim itations within the laboratory, the distance between the cam era and the participants was lim ited to 1.35 m and two video cam eras were required to record the m otions of the par ticipants. T he cam eras were positioned on the right side of the participants with their optical axes oriented perpendicular to the sagittal plane of the athletes. Cam era 1 was located at a height of 0.72 m and cam era 2 at a height of 2.47 m . T he cam eras were genlocked to synchronize their instants of exp osure and their respective W elds of view overlapp ed by app roxim ately 30 cm . T hus the two cam eras (shutter speeds 1/1000 s; sam pling rate 60 H z) provided a W eld of view (approxim ately 1.5 3.5 m ) large enough to record the m otions of the participants. Coordinates (in pixels) representing the location of the body landm arks were obtained using an autom ated digitizing process (Peak Perform ance Technologies, Englewood, CO ) at instants (`output fram es ) separated by 1/60 s. To m inim ize perspective errors associated with the short distance between the cam era and the participants, prelim inary scale factors for the two cam eras (SC 1P and SC 2 P, respectively) were determ ined using a m etre-stick placed in the approxim ate plane of m otion of the athlete. At the instant of each output fram e, the scale factor for cam era one (SC 1 ) was com puted as follows: SC 1 = l SK (m )/ l SK (pixels) (1)

451
of m otion of the other body landm arks. All coordinate data were exam ined for discontinuities as the landm arks m oved between the W elds of view of the two cam eras. N o discontinuities were noted. At the instant of each output fram e, the coordinates (in m etres) of each landm ark were com puted as the product of its digitized coordinates (in pixels) and the app ropriate scale factor (SC 1 or SC 2 ). All coordinates were expressed in term s of an or thogonal laborator ybased inertial reference fram e R 0 . T he axes of R 0 (X 0 , Y 0 and Z 0 ) were de W ned by unit vectors i 0 , j 0 and k 0 , respectively. Vector k 0 pointed vertically upward; j 0 was horizontal and pointed anteriorly; i 0 was horizontal and pointed to the participant s right side. Sim ultaneous with the recording of the videotape inform ation, ground reaction force and centre of pressure data were collected from a Kistler force plate (M odel 9281B) at a sam pling rate of 1000 H z using Bioware software (K istler Instrum ent Com pany, Am herst, N Y). The correspondence between the video and force plate data was determ ined using a video synchronization unit (Peak Perform ance Technologies) that sim ultaneously displayed a m arker in the video im age and transm itted an analog signal to the Bioware software. In all trials, the instant of take-o V was determ ined as the instant the vertical com ponent of the ground reaction force dropped below 4 N. A time of 10.00 s was assign ed to the instant of take-o V . The varied sam pling rates for the landm ark coordinates (60 Hz) and the ground reaction force and centre of pressure data (1000 H z) necessitated com putation of average ground reaction force and centre of pressure values at the instant of each output fram e. T his was accom plished using the equation: ( t) = Q 1 tf
t + tf /2

t - tf /2

Q (i ) dt

(3)

where l SK is the distance between the knee and the ankle m arkers. The scale factor for cam era 2 (SC 2 ) was then com puted using SC 1 and the ratio between SC 2P and SC 1P : SC 2 = SC 1 (SC 2P /SC 1P ) (2)

As the shank is in the plane of m ovem ent, com puting cam era scale factors at the instant of each output fram e eliminated the e V ects of perspective error in the calculation of the scaled coordinates for all landm arks except the elbow and wrist. The e V ects of persp ective errors on the scaled coordinates of the elbow and wrist were m inim ized by ensuring that the participants sw ung their arm s in a sagittal plane adjacent and parallel to the plane

( t) where t is the time of the videotape output fram e, Q is the value of the respective ground reaction force or centre of pressure com ponent at the instant of each output fram e, Q ( i ) is the instantaneous value of the respective ground reaction force or centre of pressure com ponent determ ined from the force plate, and t f is the time between each output fram e (1/60 s). At the instant of each output fram e, the respective ground reaction force and centre of pressure com ponents were expressed in term s of reference fram e R 0 . Data smooth ing T he time-dependent coordinates of each body landm ark were sm oothed using a quintic spline sm oothing routine (Wood and Jennings, 1979; Vaughan , 1980) to reduce sm all random errors that m ay have occurred

452
during digitizing. To determ ine the sm oothing factor for each trial, the ver tical acceleration of the body s centre of m ass ( G ) at the instant of each output fram e [ a zG-FP( i )] was com puted using the follow ing equation: a zG-FP ( i ) = ( F Z ( i ) - mg )/m (4)

Feltner et al.
w here a s1/s2 is the acceleration of segm ent 1 relative to segm ent 2, and a s1 and a s2 are the absolute accelerations of segm ents 1 and 2, respectively. A ngu lar kinematics T he abso lute angular orientation ( h) of the extremity segm ents were com puted using the general equation:

w here F Z ( i ) is the value of the ver tical com ponent of the ground reaction force at the instant of each output fram e, m is the m ass of the subject and g is gravitational acceleration. U sing an iterative procedure, the sm ooth6 ing factor was increased in 2.0 10 - increm ents and the vertical acceleration of the body s centre of m ass based upon the digitized landm ark coordinates ( a zG-COR ) was com puted (Dapena, 1978; see follow ing section). For each value of the sm oothing factor, the average residual (r ) between a zG-CO R and a zG-FP for the trial was com puted: r = 1
n
zG-CO R

h = cos - 1 ( r PR OX IM AL /DISTAL j 0 )

(7)

w here r P ROX IM A L/D ISTAL is a unit vector pointing from the distal to proxim al endpoint of the segm ent and the sym bol ? indicates the dot product operation. The absolute angular orientation of the trunk ( hTR ) was:

hTR = cos - 1 ( r SH LD /H IP j 0 )

(8)

S |a n i=1

( i ) - a zG-FP( i ) |

(5)

w here n is the num ber of output fram es in a trial. The sm oothing factor value that m inim ized r was selected for each trial. For the jum ps w ith an arm swing, the m ean 6 sm oothing factor was 9.2 ( s = 5.5) 10 - per fram e in the Y 0 and Z 0 directions and the average r -valu e 2 was 0.91 ( s = 0.26) m s - . For the jum ps without an arm swing, the m ean sm oothing factor was 7.6 6 ( s = 7.4) 10 - per fram e in the Y 0 and Z 0 directions and - 2 the average r -valu e was 0.83 ( s = 0.30) m s . The sm oothed two-dim ensional landm ark data expressed in term s of reference fram e R 0 were used for all subsequent com putations. Linear kinematics T he linear velocity ( v ) and acceleration ( a ) of each landm ark were com puted as the W rst and second derivatives, respectively, of quintic spline functions W tted to the time-dependent com ponents of the landm ark coordinates. T he displacem ent ( s ), velocity and acceleration of segm ental centres of m ass were com puted using the equation presented by D apena (1978) and centre of m ass location data reported by Clauser et al. (1969) and adjusted according to H inrichs (1990). T he displacem ent, velocity and acceleration of the body s centre of m ass ( s G , v G and a G , respectively) were also com puted using the procedures detailed by D apena (1978). Equations sim ilar to those used to com pute s G , v G and a G were also used to com pute the displacement, velocity and acceleration of the following segm ents: arm s, legs, head and trunk, and head, trunk and arm s. The relative accelerations of the segm ents were com puted using the general equation: a s1/s2 = a s1 - a s2 (6)

w here r S H LD /HIP is a unit vector pointing from the hip to the shoulder. T he absolute angular velocity ( v ) and acceleration ( a) of each segm ent were com puted as the W rst and second derivatives, respectively, of quintic spline functions W tted (zero sm oothing) to the tim edependent h-valu es. Joint angles at the ankle ( hANK ), knee ( hKN EE ) and hip ( hH IP ) were com puted using the follow ing equations:

hANK = p + hFT - hSK hK NEE = hTH - hSK hHIP = p - hTR + hTH

(9) (10) (11)

w here hFT , hSK , hTH and hTR are the angular orientations of the foot, shank, thigh and trunk, respectively. The angular velocity at each joint was com puted as the W rst derivatives of quintic spline functions W tted (zero sm oothing) to the time-dependent joint angle ( h) values. Signs and reference values for the segm ent and joint angles are displayed in F ig. 2. Joint kinetics Inverse dynam ics (Andrews, 1974, 1982; Feltner and D apena, 1986) were used to com pute the proxim al resultant joint force and resultant joint torque exerted at the ankle, knee, hip and shoulder. For the lower extrem ity, the foot was m odelled as being subjected to a proxim al resultant joint torque ( T ANK ) and three forces: the ground reaction force applied at the centre of pressure, weight at its centre of m ass, and a proxim al resultant joint force ( F AN K ). T he shank and thigh were each assu m ed to be subjected to a proxim al resultant joint torque ( T KN EE or T H IP, respectively), a distal resultant joint torque, and three forces: weight at its centre of m ass, a proxim al resultant joint force ( F K NEE or F HIP ,

Augmentation of lower extremity kinetics

453
the distal force applied to the segm ent (if applicable). T he local angular m om entum of each segm ent about a transverse axis through its centre of m ass was com puted using a m odiW cation of the procedures presented by D apena (1978). T he net torque ( S T ) on each segm ent about its centre of m ass was calculated as the W rst derivative of its local angular m om entum . T he proxim al resultant joint torque (RJ T PROX IM AL ) exer ted on a segm ent was then com puted using the general equation: RJ T PROX IM AL = S T - RJ T DISTAL T F PROX IM AL - T F DISTAL (13)

where RJ T DISTAL is the distal resultant joint torque app lied to the segm ent (if applicable), and T F PROX IM AL and T F DISTAL are the torques created about the segm ent s centre of m ass by the proxim al and distal resultant joint forces applied to the segm ent (if app licable). Positive torque values result in extension of the trunk, hip and knee, and plantar X exion at the ankle.

Data analysis To aid the interpretation of the kinetic and kinem atic data, the jum p was divided into four periods (A D ) using W ve instants (see Fig. 3): (1) t FM , the time of W rst m ovem ent was subjectively determ ined as the instant the m agnitude of the vertical com ponent of the ground reaction force decreased approxim ately 5 N below the participant s body weight; (2) t N V , the instant of m axim um negative vertical velocity of the centre of m ass; (3) t LP , the instant of m inim um ver tical displacem ent of the centre of m ass of the body; (4) t PV , the instant of m axim um positive ver tical velocity of the centre of m ass of the body; and (5) t TO , the instant of take-o V . At t NV and t PV , the vertical acceleration of the body s centre of m ass ( a zG ) equals zero. T herefore, the follow ing general equation was used to com pute t NV and t PV : t IN T = t ( i ) +

F igure 2 Signs and reference values for the angular kinematic data.

respectively) and a distal resultant joint force. For the upper extrem ity, the forearm and hand segm ent was assu m ed to be subjected to weight acting at its centre of m ass and a proxim al resultant joint force and resultant joint torque. The upper arm was m odelled as being acted upon by a proxim al resultant joint torque ( T SH LD), a distal resultant joint torque and three forces: weight acting at its centre of m ass, a proxim al resultant joint force ( F S HLD) and a distal resultant joint force. T he m om ent of inertia values about the transverse axis of each segm ent reported by C handler et al. (1975) were used and personalized for each par ticipant (H inrichs, 1985). The proxim al resultant joint force (RJ F PROX IM AL ) exerted on a segm ent was com puted using the general equation: RJ F PROX IM AL = m ( a CM - g ) - F DISTAL (12)

- a zG ( i ) ( a zG ( i + 1) - a zG ( i ))

tf

(14)

where t INT is the interpolated time value ( t NV or t PV ), t is time, t f is the time between each output fram e, and i and i + 1 indicate the output fram es im m ediately before and after the instant when a zG equals zero. At the instant of t LP , the vertical velocity of the centre of m ass of the body ( v zG ) equals zero. T herefore, the follow ing general equation was used to com pute t LP : t LP = t ( i ) +

where a CM and g are the accelerations of the segm ent s centre of m ass and gravity, respectively, and F DISTAL is

- v zG ( i ) ( v zG ( i + 1) - v zG ( i ))

tf

(15)

454

Feltner et al.

F igure 3

Sketch depicting the periods used to analyse the vertical jumps.

To exam ine fur ther the jum ps with and w ithout an arm swing, the propulsive phase (period C ) was divided into three sub-per iods (C 1 , C 2 and C 3 ) of equal time by de W ning the instants t 1 and t 2 , where t 1 = t LP + 1/3 ( t PV - t LP ) and t 2 = t LP + 2/3 ( t PV - t LP ) (see F ig. 3). For each period A D and each sub-period C 1 C 3 , the average values of joint torques, forces and angular velocities were com puted using an equation of the general form : = Q 1 t PER IOD
tB

tA

Q (i ) d t

(16)

w here t PER IOD is the duration of the period; t A and t B are the times that de W ne the start and end of the period, is the average value of the data param eter respectively; Q during the period; and Q ( i ) indicates the instantaneous values of the respective data param eters. All force values were norm alized by dividing by the participant s weight and are reported in body weight units. All torque values were norm alized by dividing by the product of the participant s m ass and height squared (m ass height 2 ). O ne-way repeated-m easures analyses of varian ce (K eppel, 1982) were used to com pare the discrete biom echanical data for the jum ps with versus without an arm sw ing. O wing to the large num ber of statistical com parisons ( n = 117) and in an e V ort to avoid Type I errors, the exp erim ent-wide error rate was established a priori at P 0.05 using a Bonferroni correction

(K eppel, 1982). This resulted in a per com parison a of P 0.0005. H owever, the Bonferroni correction will result in an increased Type II error rate. Exam ination of the statistical results indicated that in 16 cases the probab ility asso ciated w ith the per com pariso n F -ratio was greater than P 0.0005, but less than P 0.05. T hese cases have been indicated in Tables 1 6. As the aim of the study was to identify factors that clearly discrim inated between the jum ps with and without an arm swing, the m ore stringent experim ent-w ide error rate of P 0.05 was used to determ ine statistical signiW cance during data analysis. To present ensem ble average curves for the arm swing and no-ar m -sw ing groups, the times during each jum p [ t ( i )] were norm alized and exp ressed as a function of the length of the propulsive phase (period C): % t( i ) =

t( i ) - t LP tPV - tLP

100

(17)

w here % t ( i ) is the norm alized tim e at the instant of output fram e i. Q uintic spline functions (zero sm oothing) were then W tted to the time-norm alized-dependent kinetic and kinematic data. T he spline functions were used to com pute interpolated values at 5% increments from an instant after the time of W rst m ovem ent ( - 150% ) until an instant after take-o V (130% ). The norm alized values for all subjects in the arm -sw ing and no-ar m swing groups were then averaged at each % t ( i ) value.

Augmentation of lower extremity kinetics

Table 1 Linear kinem atic data for body and segment centres of mass and temporal data (mean s)

455

Arm swing m in. s zG (%) a a s zG at take-o V (%) - 1 v zG at take-o V (m s ) - 1 m ax. - v z G (m s ) - 2 m ax. a zG (m s ) - 2 m ax. a zAR M (m s ) m ax. a zHT (m s - 2 ) - 2 m ax. a zHTA (m s ) time v zG equals zero ( t LP ) (s) time t 1 (s)b c time t 2 (s) time of max. positive v z G (t PV ) (s) time of take-oV (t TO ) (s) 40.9 71.1 2.75 - 1.16 14.4 82.4 18.0 18.3 9.68 9.77 9.87 9.96 10.00

No arm swing 38.4 67.8 2.44 - 1.26 12.4 18.6 15.9 15.9 9.67 9.76 9.86 9.96 10.00

3.3 1.7 0.28 0.32 2.8 16.8 3.4 3.8 0.04 0.03 0.02 0.01 0.00

3.9 d 1.8 d 0.23 d 0.31 e 2.2 d 4.6 d 3.0 d 3.0 d 0.04 0.03 0.02 0.01 0.00

s zG = vertical linear displacement of the centre of m ass of the body (G ); v zG = vertical linear velocity of the centre of m ass of the body; a zG = vertical linear acceleration of the centre of m ass of the body; a zA R M = vertical linear acceleration of the centre of mass of the arm segm ent (ARM ); a zH T = vertical linear acceleration of the centre of m ass of the head and trunk segm ent (HT ); a zH TA = vertical linear acceleration of the centre of m ass of the head, trunk and arm segm ent (HTA). a Percentage of standing height; b t 1 = t LP + 1/3 ( tP V - t LP ); c t 2 = t LP + 2/3 ( tP V - t LP ). d Experiment-wide P 0.05 (per comparison P 0.0005). e Per comparison P 0.05.

Results and discussion

T he jum ps w ith an arm swing had a larger vertical velocity of the body centre of m ass ( v zG ) at take-o V and the body s centre of m ass was located at a higher relative position above the ground (Table 1). As jum p height is determ ined entirely by the height of the body s centre of m ass and its vertical velocity at the instant of take-o V (Fig. 1), these two factors resulted in higher jum p heights for the arm -sw ing versus no-arm -sw ing jum ps. T he arm swing raised the body s centre of m ass by 3% of standing height at take-o V relative to the jum ps with no arm sw ing (an average increase of 6.1 cm for the height of the centre of m ass at take-o V ) (Table 1). T he vertical velocity of the body s centre of m ass at take-o V was 12.7% larger in the arm -sw ing versus no-ar m -sw ing jum ps (Table 1) and would result in an increase of 8.2 cm in the vertical displacem ent of the body s centre of m ass between take-o V and the peak of the jum p. T hus the arm sw ing added app roxim ately 14.3 cm to the peak height of the body s centre of m ass during a vertical jum p; 43% of the increase in jum p height was due to the arm s being in a raised position at take-o V and 57% of the increase was due to e V ects asso ciated with the arm m otion that occurred before take-o V . For an average athlete (height = 186 cm), peak height of the centre of m ass of the body would have been 171 cm in the jum ps with an arm swing and 156 cm in the jum ps w ithout an arm swing. T hus, arm m otion increased peak jum p

height by approxim ately 9% . T his is in accordance with the W ndings of Luhtanen and Kom i (1978), Shetty and Etnyre (1989) and H arm an et al. (1990), who repor ted increases in take-o V velocity of 10 11% in counterm ovem ent vertical jum ps using an arm swing, and those of Payne et al. (1968) and Khalid et al. (1989), who reported a 5 10% increase in jum p height in counterm ovem ent vertical jum ps perform ed using an arm swing. H owever, the 21% increase in take-o V velocity reported by Oddson (1989) was not substantiated. The jum ps w ith an arm swing exhibited larger m axim al ver tical accelerations for the centre of m ass of the body, arm s, head and trunk, and head, trunk and arm s segm ents (Table 1; Figs 4 and 5). In the arm -sw ing jum ps, the arm swing resulted in a large positive (upward) vertical acceleration of the arm s relative to the trunk through out m ost of periods C 1 and C 2 ( a zARM /HT in Fig. 4). For the arm s to acquire a positive vertical acceleration, the trunk m ust m ake positive vertical forces on the arm s at the shoulder ( F SHLD). By reaction, the arm s m ake negative (dow nward) vertical forces on the trunk at the shoulder ( F TR(SHLD)) during periods C 1 and C 2 . D uring the arm -sw ing jum ps, the vertical acceleration of the head and trunk segm ent ( a zH T ) exhibited a quartic trend: it reached local m axim a near the instant of m inim um vertical displacem ent of the centre of m ass ( t LP ) and t 2 , and a local m inim um at t 1 w hen the acceleration of the arm s relative to the head and trunk ( a zAR M /H T ) was

456

Feltner et al.

F igure 4 Ensemble averages of the vertical acceleration of the centre of mass of the body ( a z G), arm s (a zAR M ) and arms relative to the head and trunk (a zAR M/HT ) for the arm-swing and no-arm-swing jumps. Time is expressed as the percentage of the propulsive phase (period C ). The m aximum negative velocity of the centre of mass of the body during the countermovement occurred near t = - 42% and take-o V occurred at t = 113% in both types of jump. The four solid vertical lines indicate the instants of t LP , t 1 , t 2 and t PV , respectively, and the dashed vertical line indicates the instant of take-oV ( t TO ). Periods B, C 1 , C 2, C 3 and D are labelled below the graph. The sequence above the graph indicates the m otions of a representative athlete performing an arm-swing jump.

at its m axim al value (Figs 4 and 5). In the jum ps without an arm sw ing, the vertical acceleration of the head and trunk ( a zH T ) and the head, trunk and arm ( a zHTA) segm ents reached their m axim um m agnitudes near the instant of m inim um vertical displacem ent of the centre of m ass and decreased through take-o V . In the jum ps w ith an arm swing, despite the decrease in the m agnitude of the vertical acceleration of the head and trunk near t 1 , the vertical acceleration of the head, trunk and arm segm ent reached its m axim um value near t 1 and exhib ited larger m agnitudes relative to the jum ps w ithout an arm swing until near the m idpoint of period C 3 . As the head, trunk and arm segm ent accounts for over 70% of the m ass of the body, larger m agnitudes for its acceleration in the jum ps w ith an arm sw ing resulted in larger m agnitudes for the vertical acceleration of the centre of m ass ( a zG ) (Fig. 4). Both types of jum p used a preparator y counterm ovem ent. D uring vertical jum ps, the vertical velocity of the centre of m ass of the body at take-o V is due entirely to the net vertical im pulse exerted upon the

jum per between the instant of m inim um vertical displacem ent of the centre of m ass and take-o V . For this reason, only this period was exam ined in detail when com paring the arm -sw ing and no-ar m -sw ing jum ps. As a jum per m ust have zero vertical velocity at the instant of minim um displacement, augm entation of jum p height from a counterm ovem ent m ust result from changes in the `state of the m usculature at t LP or to changes in the neurom echanical properties of the m usculature asso ciated with the counterm ovem ent that persist after t LP . T hus, exam ination of the period between t LP and the instant of take-o V would identify the m echanical factors that result in the vertical velocity of the centre of m ass of the body at take-o V , regard less of w hether these factors are associated with the counterm ovem ent. T he net vertical im pulse exerted on the jum per between the instant of m inim um vertical displacem ent of the centre of m ass and take-o V is due to the net force m ade on the jum per and the time that the net force acts. As indicated by the tem poral data (Table 1), the time between t LP and the instant of take-o V was nearly

Augmentation of lower extremity kinetics

457

F igure 5 Ensemble averages of the vertical acceleration of the centre of m ass of the head and trunk segment (a zHT ) and of the head, trunk and arm segment (a zHTA ) for the arm -swing and no-arm-swing jumps. Time is expressed as the percentage of the propulsive phase (period C). The four solid vertical lines indicate the instants of t LP , t 1 , t 2 and t PV , respectively, and the dashed vertical line indicates the instant of take-oV (t TO ). Periods B, C 1 , C 2 , C 3 and D are labelled below the graph. The sequence above the graph indicates the motions of a representative athlete perform ing a no-arm-swing jump.

identical in the arm -sw ing and no-ar m -sw ing jum ps (arm -sw ing jum ps: 0.32, s = 0.04 s; no-ar m -sw ing jum ps: 0.33, s = 0.04 s). H arm an et al. (1990) also reported sim ilar values for this period in the arm -sw ing and no-ar m -sw ing jum ps. Thus, diV erences in the net im pulse exerted on the jum per in the arm -sw ing versus no-ar m -sw ing jum ps m ust be a result of changes in the vertical com ponent of the ground reaction force ( F Z in Table 2). As the ground reaction force results from the actions of the jum per, sequential exam ination of the kinetic and kinematic data between the instant of m inim um vertical displacem ent and take-o V should elucidate the causal m echanism s asso ciated with the arm action that resulted in increased jum p height. Exam ination of the kinem atic and kinetic data for the lower extrem ity indicated no signiW cant diV erences in the kinematic and kinetic data asso ciated with the ankle. T herefore, the following descriptions will not focus upon the ankle. Failure to discuss the ankle in the following sections does not im ply a lack of im portance for the m usculature of the ankle in vertical jum ping. Rather, the W ndings of the current investigation revealed

that the kinetics and kinem atics of the ankle joint were not altered by the presence or absence of an arm sw ing. Period C 1 At the start of the propulsive phase, the no-ar m -sw ing jum pers were in positions of increased knee and hip X exion, but less forward trunk inclination, relative to the arm -sw ing jum pers (Table 3 and sequences in F igs 4 and 5). Both groups were rotating the trunk counterclockwise at the instant of m inim um vertical displacem ent of the centre of m ass, but the arm -sw ing jum pers rotated the trunk at a faster rate ( v TR in F ig. 6 and Table 4). T he thigh was relatively m otionless at the instant of m inim um vertical displacement in both groups; thus, trunk rotation resulted in faster rates of hip extension for the arm -sw ing jum pers ( v TH and v HIP in Fig. 6 and Table 4). At the start of the propulsive phase, the knees were continuing to X ex at a low rate for both jum p styles ( v KN EE in Fig. 7 and Table 5) and the arm s were beginning their upward acceleration relative to the trunk in the arm -sw ing jum ps ( a zAR M /HT in Fig. 4). Lastly, the

458
Table 2 Instantaneous and average values (BWs) for the vertical component of the ground reaction force ( F Z ) during period C. Instantaneous values are reported at the times of the four events ( t LP , t1 , t 2 and t PV ) that de W ne the subperiods of period C and average values are reported - 1 for the intervals C 1 , C 2 and C 3 . Instantaneous values (m s ) of the vertical velocity of the body centre of m ass (v zG ) are also reported for the instants of t LP , t 1 , t 2 and t PV (m ean s)

Feltner et al.

Arm swing
FZ Instantan eous t LP t1 t2 t PV Avera ge Period C 1 Period C 2 Period C 3
a

No arm swing
v zG FZ v zG

2.0 1.9 2.3 1.3

0.3 b 0.2 0.2 a 0.2

0.00 1.07 2.20 2.93

0.00 0.16 0.18 a 0.25 a

2.1 1.9 1.9 1.3

0.3 b 0.2 0.1 a 0.2

0.00 1.07 2.03 2.66

0.00 0.14 0.16 a 0.25 a

1.9 0.2 a 2.1 0.2 2.1 0.1 a

b

2.0 0.2 a 1.9 0.2 1.9 0.1 a

Experiment-wide P 0.05 (per comparison P 0.0005). Per comparison P 0.05.

Table 3 Angular displacements ( ) of lower extremities and trunk (see Fig. 2 for signs and referen ce values) at take-o V ( t TO ) and the low point ( t LP ) of the jump (mean s)

Arm swing
H ip t TO t LP K nee t TO t LP A nkle t TO t LP Trun k t TO t LP
a

No arm swing

07 - 105 16

- 9 6a - 116 12 a

- 63 - 90 10

- 9 4a - 98 13 a

33 6 - 27 6

34 6 - 27 5

85 4 35 13

81 5 a 30 9 a

Experiment-wide P 0.05 (per comparison P 0.0005).

resultant joint torques at the hip ( T HIP ) and knee ( T K NEE ) had larger extension values in the no-ar m -sw ing versus arm -sw ing jum ps at the instant of m inim um vertical displacem ent (Fig. 8 and Tables 4 and 5). T hroughout period C 1 , the no-ar m -sw ing jum pers increased the rate of counterclockw ise trunk rotation, clockw ise thigh rotation and the asso ciated extension of the hip (Fig. 6 and Table 4). Consequently, the m usculature at the hip went from near isom etric conditions at the instant of m inim um vertical displacement of the centre of m ass to concentric conditions at t 1 . Since the

force-generating capacity of skeletal m uscle decreases as the rate of concentric action increases (C avagn a et al. , 1968; Asm ussen and Bonde-Peterson, 1974; K om i and Bosco, 1978), the extension torque at the hip decreased in the no-ar m -sw ing jum ps throughout period C 1 . For the arm -sw ing jum ps, the trunk rotated rapidly counterclockw ise after the instant of m inim um vertical displacem ent of the centre of m ass, but it reached its m axim al rate of extension before t 1 and was decreasing its rate of counterclockwise rotation before the end of period C 1 . As the thigh rotated clockwise at a relatively slow rate during period C 1 in the arm -sw ing jum ps, the decreasing rate of counterclockw ise trunk rotation resulted in a decreasing rate of hip extension at t 1 (Fig. 6). C onsequently, the extension torque at the hip decreased slightly throughout period C 1 and reached a local m inim um value near t 1 (Fig. 8 and Table 4). At t 1 , the m agnitudes of the extension torques at the hip were nearly identical in the jum ps w ith and without an arm swing. H owever, the average hip extension torque during C 1 was greater in the jum ps without than w ith an arm swing, owing to the slower average rate of hip extension in the no-ar m -sw ing jum ps during C 1 (Table 4). W hy did the trunk decrease its rate of counterclockw ise rotation in the jum ps with an arm swing during period C 1 ? As the arm s accelerate upward relative to the trunk, they result in a larger downward vertical force on the trunk ( F TR(SHLD); the reaction to F SH LD in Table 6). In turn, this force creates a clockwise torque about the centre of m ass of the trunk ( T TR(SHLD)) that tries to decrease the rate of counterclockwise trunk rotation. N ear the instant of m inim um vertical displacem ent of

Augmentation of lower extremity kinetics

459

F igure 6 Ensemble averages of the angular velocity of the trunk (vTR ), thigh (vTH ) and hip (vHIP ) for the arm-swing and noarm-swing jumps. Extension values are positive for the hip; counterclockwise rotation values are positive for the trunk and thigh. Time is expressed as the percentage of the propulsive phase (period C ). The four solid vertical lines indicate the instants of t LP , t1 , t 2 and t PV , respectively, and the dashed vertical line indicates the instant of take-oV ( t TO ). Periods B, C 1 , C 2 , C 3 and D are labelled below the graph.

Table 4 Instantaneous and average angular velocities (rad s ) of the hip (vHIP ), trunk ( vTR ) and thigh ( vTH ), and the instantaneous and average normalized joint torques (s - 2 10) at the hip (T HIP ) during period C. Instantaneous values are reported at the times of the four events that de W ne the subperiods of period C (t LP , t 1 , t 2 and t PV ) and average values are reported for the intervals C 1 , C 2 and C 3 (mean s )

- 1

Arm swing

No arm swing

vTR
Instantan eous t LP 2.34 t1 3.30 t2 1.99 t PV 2.65 Average Period C 1 Period C 2 Period C 3
a

vTH

vHIP

T HIP

vTR
0.58 a 0.93 a 0.61 a 0.45

vTH
0.23 b 0.40 a 0.47 b 0.65 a

vHIP

T HIP

1.17 a 1.06 a 0.56 a 0.59

- 0.02 0.42 b 2.36 - 0.96 0.47 a 4.26 - 2.91 0.49 b 4.90 - 7.66 0.82 a 10.31

1.06 a 16.1 2.6 a 1.12 14.8 2.0 0.77 a 14.5 1.7 a 1.01 a 1.2 1.6

0.75 2.72 3.32 2.76

0.12 - 1.59 - 3.07 - 6.53

0.63 4.31 6.39 9.29

0.44 a 17.6 3.3 a 0.77 14.9 1.5 0.77 a 12.0 1.3 a 0.80 a 1.6 1.6

3.23 1.01 a 2.50 0.75 a 2.27 0.42 a

- 0.51 0.42 a - 1.73 0.47 a - 5.25 0.49 a

3.72 0.94 a 4.21 0.82 a 7.52 0.66 b

15.4 2.0 a 14.9 1.8 a 9.3 1.2 a

1.82 0.79 a 3.18 0.77 a 3.09 0.38 a

- 0.72 0.30 a - 2.34 0.42 a - 4.63 0.37 a

2.55 0.59 a 16.6 2.2 a 5.52 0.77 a 13.4 1.4 a 7.71 0.52 b 8.4 1.1 a

Experiment-wide P 0.05 (per comparison P 0.0005). b Per comparison P 0.05.

the centre of m ass, the dow nward vertical force exer ted by the arm s on the trunk is sm all and creates a negligible torque on the trunk (see Figs 9 and 10a). However, near t 1 , the centre of m ass of the arm s is close to its low point, the m agnitude of the vertical com ponent of the force on the trunk is increased, the m agnitude of its horizontal com ponent is near to zero, and the force creates a large

clockwise torque about the centre of m ass of the trunk (Figs 9 and 10b). T hus, during the latter stages of period C 1 , the arm sw ing results in a decrease in the rate of counterclockwise trunk rotation. In turn, the decreased rate of counterclockwise trunk rotation slows the rate of hip extension. T he decreasing rate of hip extension prevents the hip extensor m uscles from

460

Feltner et al.

F igure 7 Ensemble averages of the angular velocity of the thigh (vTH ), shank (vSK ) and knee ( vKNEE ) for the arm-swing and noarm -swing jumps. Extension values are positive for the knee; counterclockwise rotation values are positive for the thigh and shank. Time is expressed as the percentage of the propulsive phase (period C). The four solid vertical lines indicate the instants of t LP , t 1 , t 2 and t PV , respectively, and the dashed vertical line indicates the instant of take-o V (t TO ). Periods B, C 1, C 2 , C 3 and D are labelled below the graph. Table 5 Instantaneous and average angular velocities (rad s ) of the knee (vKNEE ), thigh (vTH ) and shank ( vSK ), and the instantaneous and average normalized joint torques (s - 2 10) at the knee (T KNEE ) during period C. Instantaneous values are reported at the times of the four events that de W ne the subperiods of period C (t LP , t 1 , t 2 and t PV ) and average values are reported for the intervals C 1 , C 2 and C 3 (mean s)
- 1

Arm swing

No arm swing
T KNEE

vTH
Instan tane ou s t LP - 0.02 t1 - 0.96 t2 - 2.91 t PV - 7.66 Averag e Period C 1 Period C 2 Period C 3
a

vS K

vK NEE

vTH

vSK

vKNEE

T KNE E

0.42 b 0.47 a 0.49 b 0.82 a

- 1.10 0.51 a - 1.08 0.79 a 14.4 4.4 a - 0.24 0.45 b 0.72 0.80 a 16.3 3.0 b a 1.41 0.56 4.33 0.87 18.5 1.8 6.13 0.87 13.77 1.71 a 5.1 1.9 a -

0.12 1.59 3.07 6.53

0.23 b - 0.58 0.35 a - 0.70 0.51 a 15.8 3.9 a 0.40 a - 0.10 0.44 b 1.47 0.73 a 15.6 2.0 b a 0.47 b 1.31 0.42 4.38 0.73 16.0 1.7 0.65 a 6.13 0.91 12.65 1.29 a 6.7 2.1 a

- 0.51 0.42 a - 0.73 0.49 a - 1.73 0.47 a 0.44 0.45 - 5.25 0.49 a 3.84 0.61 a

- 0.23 0.79 a 15.3 3.4 b - 0.72 0.30 a - 0.37 0.40 a 2.16 0.77 a 17.6 2.5 a - 2.34 0.42 a 0.44 0.37 9.09 0.98 a 14.7 1.4 a - 4.63 0.37 a 3.58 0.52 a
b

0.35 0.66 a 2.78 0.68 a 8.20 0.77 a

16.0 2.7 b 15.6 1.9 a 13.8 1.4 a

Experiment-wide P 0.05 (per comparison P 0.0005). Per comparison P 0.05.

experiencing rapid concentric conditions and results in only a sm all decline in the hip extensor torque during period C 1 . M iller (1976) reported that the angular acceleration of the trunk exhibited two periods of counterclockwise or extension angular acceleration (coinciding app roxim ately with periods C 1 and C 3 in the current study) and an interm ediate period of clockw ise or X exion angular

acceleration (coinciding app roxim ately with period C 2 in the current study) during arm -sw ing jum ps. T his pattern of trunk angular acceleration would be consistent w ith the angular velocity pattern exhib ited for the trunk in the current study. M iller also found that the periods of counterclockwise trunk angular acceleration were asso ciated with biceps fem oris electromyog raphic activity.

Augmentation of lower extremity kinetics

461

F igure 8 Ensemble averages of the torque at the knee (T KN EE ) and hip (T HIP ) for the arm-swing and no-arm-swing jumps. Extension torques have positive values. Tim e is expressed as the percentage of the propulsive phase (period C ). The four solid vertical lines indicate the instants of t LP , t 1, t 2 and t PV , respectively, and the dashed vertical line indicates the instant of take-o V ( t TO ). Periods B, C 1 , C 2 , C 3 and D are labelled below the graph.

F igure 9 Ensemble averages of the torque created on the trunk by the arms (T TR (SHLD)) for the arm-swing and no-arm-swing jumps. Tim e is expressed as the percentage of the propulsive phase (period C). The four solid vertical lines indicate the instants of t LP , t 1 , t 2 and t PV , respectively, and the dashed vertical line indicates the instant of take-o V ( t TO ). Periods B, C 1 , C 2 , C 3 and D are labelled below the graph.

The knee extensor m uscles created larger torques at the instant of m inim um vertical displacement of the centre of m ass in the no-ar m -sw ing versus arm -sw ing jum ps (Table 5). It is unclear why the knee extensors

were creating larger torques at the instant of m inim um vertical displacement in the jum ps without an arm swing. H owever, it m ay be related to an increased stretch placed on the extensor group ow ing to the

462

Feltner et al.
that continued to operate throughout period C 2 (see below). At the end of period C 1 , ver tical velocity of the centre of m ass of the body was identical for both the arm -sw ing and no-ar m -sw ing jum ps (Table 2). However, the vertical accelerations of the centre of m ass of the body ( a zG ), the head and trunk segm ent ( a zHT ), and the head, trunk and arm segm ent ( a zH TA ), all reached their m axim al values near the instant of m inim um vertical displacem ent in the jum ps without an arm swing and decreased in m agnitude throughout period C 1 (Figs 4 and 5). Conversely, the ver tical accelerations of the centres of m ass of both the body and of the head, trunk and arm segm ent increased during period C 1 in the jum ps w ith an arm swing despite a large decline in the acceleration of the head and trunk segm ent owing to the upward acceleration of the arm s (Figs 4 and 5). T he patterns of these three accelerations in period C 1 and throughout the jum p were consistent w ith the results of M iller (1976). Period C 2 In the jum ps w ithout an arm swing, the hip extended at increasingly faster rates during period C 2 (Fig. 6) and the rapid concentric conditions resulted in a continued decline in the m agnitude of the hip extensor torque during this time (Fig. 8 and Table 4). In the jum ps w ith an arm swing, the rate of hip extension initially declined during period C 2 , but exhib ited a net increase relative to its value at t 1 by the instant of t 2 (Fig. 6 and Table 4). T he changes in the rate of hip extension during the arm -sw ing jum ps were prim arily a result of the decreased rate of counterclockwise trunk rotation during period C 2 (Fig. 6 and Table 4). As the arm sw ing was responsible for decreasing the rate of counterclockwise trunk rotation, it resulted in slower average concentric conditions for the hip extensor m usculature and was responsible for the increased hip extensor torque in the arm -sw ing versus no-ar m -sw ing jum ps during C 2 (Fig. 8 and Table 4). N ear t 2 and as the ver tical acceleration of the arm s relative to the trunk was decreasing in the arm -sw ing jum ps (Fig. 4), the m agnitude of the torque exerted on the trunk ( T TR (SHLD)) began to decrease (Figs 9 and 10c). T his halted the decline in the rate of counterclockw ise trunk rotation (Fig. 6) and, together w ith the hip extensor torque, resulted in an increase in the rate of hip extension in the arm -sw ing jum ps by t 2 . In both the jum ps w ith and without an arm swing, the knee extended at increasingly faster rates during period C 2 and was extending at approxim ately the sam e rate in both jum p styles at t 2 (Fig. 7 and Table 5). T he average extension angular velocity at the knee during period C 2 was signiW cantly less in the jum ps with an arm sw ing, and was associated w ith larger average knee extensor

F igure 10 The positions of a representative athlete and F TR (SHLD) at instants near (A) t LP , (B) t 1, (C) t 2 and (D) t PV . The triangle indicates the location of G.

increased X exion of the knee at the instant of m inim um vertical displacem ent (Table 3) or to increased levels of activation of knee extensor m otor units. D uring the preceding counterm ovem ent, the no-ar m -sw ing jum pers had larger m axim al negative values for the vertical velocity of the centre of m ass ( v zG ) (no-ar m -sw ing jum ps: m ean m axim um - v zG = - 1.26, s = 0.31 m s - 1 ; arm -sw ing jum ps: m ean m axim um - v zG = - 1.16, s = 0.32 m s - 1 ), which m ay have required larger extensor torques at the knee to stop the downward m otion of the body s centre of m ass (Fig. 8, Table 1 and Table 5 at the instant of m inim um vertical displacem ent). T hroughout period C 1 , the knee extensor torque m aintained a relatively constant m agnitude in the noarm -sw ing jum ps and increased in m agnitude in the arm -sw ing jum ps (Fig. 8 and Table 5) despite an increasing rate of knee extension during this period (Fig. 7 and Table 5). This m ay re X ect volitional recruitm ent strategies used by the athletes or m uscle tension bene W ts associated with the previous stretch of the quadriceps during the counterm ovem ent (C avagn a et al. , 1968; Asm ussen and Bonde-Peterson, 1974; K om i and Bosco, 1978). It m ay also be associated w ith other neurophysio logical and m echanical m echanism s

Augmentation of lower extremity kinetics

torques in the arm -sw ing jum ps than jum ps without an arm sw ing during C 2 (Fig. 8 and Table 5). D espite an increasing rate of extension at the knee (im plying faster concentric conditions) during period C 2 in both the arm -sw ing and no-ar m -sw ing jum ps, the norm alized knee extensor torque ( T KN EE ) increased in m agnitude for both jum p styles [no-arm -sw ing jum ps: D T K NEE = 0.4 s - 2 10 (2.5% ); arm -sw ing jum ps: D T K NEE = 2.2 s - 2 10 (13.5% ); Fig. 8 and Table 5]. The increase in extensor torque at the knee m ay be due to an increased m om ent arm for the quadriceps m uscles. At t 1 , the knee was at app roxim ately 90 of X exion in both jum ps (arm swing: 91 ; no arm swing: 96 ) and it extended approxim ately 15 during period C 2 (arm swing: 12 ; no arm sw ing: 16 ). The m om ent arm for the quadriceps increases by 13 17% from 90 to 75 of extension (Sm idt, 1973; Spoor and van Leeuwen, 1992) and could account for an increase in extensor torque. In the no-ar m -sw ing jum ps, the apparent increase in the m om ent arm for the quadriceps was o V set by declining m uscular forces through out period C 2 and the extensor torque rem ained relatively constant during this interval. In the arm -sw ing jum ps, the jum pers apparently used both an increase in m om ent arm and neurophysio logical m echanism s that increased m uscular tension to increase knee extensor torque. The increases in knee extensor torque during both types of jum p during period C 2 m ay be due to the action of the two-joint knee extensor (rectus fem oris). As the hip was extending during C 2 , the rate of shortening of the rectus fem oris would be less than that experienced by the vastii m uscles of the quadriceps. The slower concentric conditions for the rectus fem oris m ay be large enough to o V set the declines in force ow ing to the faster concentric actions of the rem aining quadriceps m uscles during this period. Both G regoire et al. (1984) and Bobbert and van Ingen Schenau (1988) reported that the rectus fem oris exhib ited high levels of electromyograph ic activity during this portion of no-ar m swing style jum ps. Changes in volitional levels of m uscular activation and varying recruitment levels or discharge rates of the knee extensors during C 2 m ay also account for increased m uscular tension in the quadriceps. D uring the counterm ovem ent, the no-ar m -sw ing jum pers had a larger m axim al negative value for the vertical velocity of the centre of m ass of the body and they created larger extensor torques at the knee to stop the dow nward m otion of the body s centre of m ass (Fig. 8 and Table 5 at the instant of m inim um vertical displacement). As a result and to stop the downward m otion of the centre of m ass of the body during the counterm ovem ent, jum ps without an arm swing m ay require the recruitment of m ore extensor m otor units or higher discharge rates of the active extensor m otor units relative to jum ps with

463
an arm swing. If the higher level of activation of the quadriceps during the counterm ovem ent results in short-term fatigu e of the extensor m otor units during the no-ar m swing jum ps, the size principle (H ennem an, 1979) would dictate that fast-tw itch, fatigu able m otor units capable of producing large tensions would be de-recruited initially. D ecreased activation of the fasttwitch, fatigu able m otor units m ay result in a decreased ability of the quadriceps to generate tension during the propulsive phases of the no-arm -sw ing jum ps. In turn, this m ay account for the lack of increase in the knee extensor torque, as the quadriceps m om ent arm was increasing during period C 2 in the no-ar m -sw ing jum ps. Conversely for the jum ps w ith an arm swing, the sm aller knee extensor torques required during the counterm ovem ent m ay have allowed the jum pers to avoid recruitment of the fast-tw itch, fatigu able m otor units. As a result, activation of the fast-tw itch, fatigu able m otor units could be delayed until the periods during the propulsive phase of the jum p w hen m axim al knee extensor torque could be produced, resulting in the largest contribution to the vertical velocity of the centre of m ass of the body. Thus, increased m uscular forces produced by the quadriceps (despite increasing concentric conditions), coupled w ith an increasing m om ent arm for the quadriceps, could account for the increase in the knee extensor torque during periods C 1 and C 2 of the arm -sw ing jum ps. H erzog et al. (1991) reported that m axim al quadriceps force is produced at knee angles of approxim ately 78 , and Jensen et al. (1991) repor ted that m axim al knee extensor torque following a preload occurs at app roxim ately 75 of knee X exion. Collectively, this m ay suggest that, during the arm -sw ing jum ps, recruitment of high -threshold, fast-tw itch, fatigu able m otor units is delayed until after the initiation of upward m ovem ent of the centre of m ass of the body. Additionally, as joint torques are indicative of the net m uscular activity at an articulation (Andrew s, 1974, 1982), the changes in the hip and knee extensor torques m ay be due to co-contraction of the agonist and antagonist m uscle groups at the hip and knee. In both styles of jum ps, the athletes need to produce hip and knee extensor torques during the propulsive phase of the jum p. The hip extensor torque would be produced prim arily by the gluteus m axim us (one-jo int m uscle) and the ham strings (two-joint m uscle) and the knee extensor torque by the quadriceps. However, the twojoint ham strings also create a X exion torque at the knee that would decrease the m agnitude of the knee extensor torque produced by the quadriceps. In the jum ps without an arm swing, the gluteus m axim us, ham strings and quadriceps are active near their m axim um levels at t 1 . H owever, as the athletes need to produce an extensor torque at the knee to continue the upward acceleration of the body during the

464

Feltner et al.

Table 6 Instantaneous and average force values (BWs) at the hip (F HIP ), knee (F KNEE ), ankle ( F ANK ) and shoulder ( F SHLD ) during period C. Instantaneous values are reported at the times of the four events that de W ne the subperiods of period C (t LP , t 1 , t 2 and t PV ) and average values are reported for the intervals C 1 , C 2 and C 3 (mean s )

Arm swing
F HIP Instan tane ou s t LP - 1.5 t1 - 1.5 t2 - 1.6 t PV - 0.5 Averag e Period C 1 Period C 2 Period C 3
a

No arm swing
F ANK F SHLD F HIP F KNEE F AN K F SH LD

F K NEE

0.2 b 0.2 b 0.1 a 0.2 b

- 1.8 0.3 b - 1.8 0.2 - 2.1 0.2 a - 0.8 0.2

- 2.0 0.3 b - 1.9 0.2 - 2.2 0.2 a - 1.1 0.2

0.2 0.8 0.1 - 0.2

0.3 0.2 a 0.2 a 0.1 a

- 1.6 0.2 b - 1.4 0.1 b - 1.3 0.1 a - 0.6 0.2 b

- 1.9 0.3 b - 1.8 0.1 - 1.8 0.1 a - 0.9 0.2

- 2.0 0.3 b - 1.9 0.2 - 1.9 0.1 a - 1.2 0.2

0.2 0.3 0.2 0.0

0.0 0.1 a 0.0 a 0.0 a

- 1.5 0.2 b - 1.6 0.1 a - 1.3 0.1 a

- 1.8 0.2 - 1.9 0.2 a - 1.8 0.1 a

- 1.9 0.2 - 2.1 0.2 a - 2.0 0.1 a

0.5 0.2 0.5 0.2 a - 0.1 0.1 a

a

- 1.6 0.1 b - 1.4 0.1 a - 1.2 0.1 a

- 1.8 0.2 - 1.7 0.2 a - 1.6 0.1 a

- 2.0 0.2 - 1.8 0.2 a - 1.8 0.1 a

0.3 0.0 0.3 0.0 a 0.1 0.0 a

a

Experiment-wide P 0.05 (per comparison P 0.0005). b Per comparison P 0.05.

propulsive phase of the jum p, they m ay selectively inactivate the ham string m uscles. T he inactivation of the ham strings coupled w ith the rapid concentric conditions at the hip would result in a rapid decline in hip extensor torque observed during periods C 1 and C 2 in the no-ar m -sw ing jum ps. Sim ultaneously, the decrease in the X exion torque at the knee asso ciated w ith the ham string activity and the increasing m om ent arm of the quadriceps m ay o V set reductions in torque ow ing to the concentric conditions experienced by the quadriceps, and enable the knee extension torque to rem ain roughly constant in the no-ar m -sw ing jum ps during periods C 1 and C 2 . T he above interpretation is substantiated by the W ndings of G regoire et al. (1984) and Bobbert and van Ingen Schenau (1988), who reported electromyographic activity for the gluteus m axim us, sem itendinosus, biceps fem oris, rectus fem oris and vastus m edialis through out the propulsive phase of no-ar m -sw ing style jum ps. Gregoire et al. (1984) and Bobber t and van Ingen Schenau (1988) found that the electromyographic activity of the ham strings decreased after the initial 33% of the propulsive period and speculated that the action of the ham strings m uscles was to exert a restraining in X uence on knee extension. In the jum ps with an arm swing, the gluteus m axim us, ham strings and quadriceps are also active at t 1 . As in the jum ps without an arm swing, the arm -sw ing jum pers also wish to produce large knee extensor torques during the propulsive phase to continue the upward acceleration of the body during the jum p. H owever, as the arm swing jum pers deactivate the ham strings, they are sim ultaneously placing the hip joint in slower concentric conditions by using the arm swing to decrease the rate of hip extension. D eactivation of the ham strings coupled

w ith the slowing concentric conditions at the hip m ay result in the m odest decline in hip extensor torque obser ved during periods C 1 and C 2 in the arm -sw ing jum ps. Sim ultaneously, the elimination or reduction of the X exion torque at the knee associated w ith the ham strings activity, the increasing m om ent arm of the quadriceps and the possible recruitment of highthreshold, fast-tw itch, fatigu able m otor units o V set the concentric conditions experienced by the quadriceps and enable the knee extension torque to increase in m agnitude by app roxim ately 28.4% in the arm -sw ing jum ps during periods C 1 and C 2 . As a result of the arm sw ing and the larger hip and knee extensor torques in the arm -sw ing versus no-ar m swing jum ps, larger downward vertical forces were applied to the thigh, shank, foot and ground by its adjacent respective proxim al segm ent (see F HIP, F K NEE and F ANK in Table 6 and F Z in Table 2). As a result, the vertical acceleration of the head and trunk ( a zHT ) and body ( a zG ) continued to increase in the arm -sw ing jum ps during period C 2 . T he larger vertical accelerations for the centre of m ass of the body in the arm swing versus no-ar m -sw ing jum ps resulted in a larger change in the vertical velocity of the centre of m ass of the body ( v zG ) during period C 2 (arm swing: 1.13 m s - 1 ; no arm sw ing: 0.96 m s - 1 ) and accounted for the larger value of this variable at t 2 (Table 2). Period C 3 D uring period C 3 in the arm -sw ing jum ps, the rates of rotation of the trunk, thigh and hip increased and resulted in signi W cantly larger values for the angular velocity of the hip ( v H IP) and thigh ( v TH ) at the instant of m axim um positive vertical velocity of the centre of

Augmentation of lower extremity kinetics

m ass relative to t 2 (Fig. 6 and Table 4). In the no-ar m swing jum ps, the angular velocity of the hip exhibited a sm aller increase during C 3 , as reductions in the rate of trunk extension o V set increasing rates of clockw ise thigh rotation (Fig. 6 and Table 4). T he arm -sw ing jum ps also had larger extension angular velocities at the knee relative to the no-ar m -sw ing jum ps during period C 3 . As the angular velocity of the shank was sim ilar in both jum p styles, the diV erences in knee extension angular velocity were due principally to diV erent rates of clockwise thigh rotation in the arm -sw ing versus no-ar m swing jum ps (Table 5 and Fig. 7). In the arm -sw ing and no-ar m -sw ing jum ps, the rapid extension angular velocities at the hip and knee during period C 3 were associated with large decreases in the extension torques at the hip and knee (Fig. 8). Between t 2 and the instant of m axim um vertical velocity of the centre of m ass of the body in both types of jum p, the hip extensor torques declined by 92% and 87% respectively, and the knee extensor torques declined by 72% and 58% respectively (Tables 4 and 5). T he decreases in the extensor torques at the hip and knee were prim arily a result of the rapid concentric actions of the m uscles at this time. Before take-o V in both jum p styles, it is necessar y for the athlete to decrease the knee extensor torque to prevent the knee from being in a hyperextended and possible injurious position. A decline in the hip extensor torque before take-o V m ay also be required to position the trunk and thigh in near vertical orientations at take-o V (Table 2). Thus, the athletes m ay have volitionally deactivated the hip and knee extensor m usculature before the instant of m axim um positive vertical velocity to avoid injury and m axim ize the height of the centre of m ass of the body at take-o V . The greater m agnitudes of extensor torque in the arm -sw ing jum ps both at and im m ediately after t 2 were responsible for the larger average torque values at the hip and knee in the arm -sw ing versus no-ar m -sw ing jum ps during period C 3 (Tables 4 and 5). In turn, the larger torques at the hip and knee resulted in larger average downward vertical forces app lied at the joints of the lower extremity (Table 6) and on the ground (Table 2) during period C 3 in the arm -sw ing jum ps. D uring period C 3 , the vertical velocity at the centre of m ass ( v zG ) increased by 0.73 m s - 1 in the arm -sw ing jum ps and by 0.63 m s - 1 in the no-arm -sw ing jum ps (Table 2) and this velocity was 0.27 m s - 1 greater at the instant of m axim um positive vertical velocity of the centre of m ass of the body in the jum ps with an arm swing. Period D Between the instant of m axim um positive vertical velocity of the centre of m ass and take-o V , the torques at the hip and knee continued their rapid decline and had

465
sm all X exion values by the instant of take-o V (Fig. 8). T he m agnitudes of the vertical forces app lied to the ground by the athletes during period D were less than their body weight and the vertical acceleration of the centre of m ass ( a zG ) was negative for both the arm -sw ing and no-arm -sw ing jum ps (Fig. 4). C onsequently, the vertical velocity of the centre of m ass ( v zG ) decreased by app roxim ately 0.2 m s - 1 during period D for both styles of jum ps. At take-o V , the vertical velocity was 2.75 m s - 1 for the arm -sw ing jum ps and 2.44 m s - 1 for the no-ar m -sw ing jum ps. Additionally, the trunk was in a m ore upright position and the hips and knees were in positions of increased extension at take-o V in the arm -sw ing versus no-ar m -sw ing jum ps (Table 3 and sequences at the top of Figs 4 and 5).

Summ ar y
T he W ndings of this study indicate that, during counterm ovem ent vertical jum ps, the arm sw ing augm ents the ability of the hip and knee m usculature to create extensor torques during the propulsive period of the jum p. By slowing the rate of counterclockwise trunk rotation as the propulsive phase of the jum p is initiated, the arm swing placed the hip extensor m uscles in physiological conditions that favou red the generation of large m uscular tensions and torques. The arm swing also allowed the athletes to increase the extensor torque at the knee by 28% between the instant of m inim um vertical displacem ent of the centre of m ass and t 2 and when the knee extensor m uscles were in concentric conditions. T he m echanical, neural or physiological factors responsible for the increase in knee extensor torque between these two instants is unclear but m ay be due to som e or all of the following factors: an increasing m om ent arm for the quadriceps m uscles during period C 2 ; the e V ects of two-jo int hip and knee m uscles; diV erent physio logical conditions experienced by the knee m uscles; altered levels of m uscle activation; and diV erent recruitm ent strategies for the hip and knee extensor m uscles.

Acknowledgem ents
The authors are indebted to M . Dunphy, N. M athies and the 1994 95 m en s and women s Pepperdine University volleyball teams for their time and patience. We also acknowledge the contributions of T. Stewart, P. M andella, M . Cox, K. M illikan, M . Davis, M. Anderson and K. Jonasson to the data collection and reduction. Lastly, we thank D.M . Koceja for his insights regarding the neurophysiological aspects of the manuscript and M . LeBlanc for her review of the completed m anuscript. The subroutines necessary to draw the jumpers in Figs 4, 5 and 10 were adapted from subroutines provided by Jes s Dapena, Department of Kinesiology, Indiana U niversity, Bloomington, IN 47405, USA.

466 R eferences
Anderson, F.C. and Pandy, M .G. (1993). Storage and utilization of elastic strain energy during jumping. Jour na l of B iom ech anics , 26 , 1413 1427. Andrews, J.G. (1974). Biomechanical analysis of human m otion. In K inesiology IV , pp. 32 42. Washington, DC: AAHPERD. Andrews, J.G. (1982). On the relationship between resultant joint torques and muscular activity. M edicine and Science in Sports and Exercise , 14 , 361 367. Asmussen, E. and Bonde-Peterson, F. (1974). Storage of elastic energy in skeletal m uscles in m an. Acta Physiolog ica Scandinavica , 91 , 385 392. Bobbert, M .F. and van Ingen Schenau, G.J. (1988). C oordination in vertical jumping. Jour nal of B iom echanics , 21 , 249 262. C avagna, G.A., Dusman, B. and M argaria, R. (1968). Positive work done by the previously stretched muscle. Jour nal of A pplied Physiolog y , 24 , 21 32. C handler, R.F., C lauser, C.E ., M cConville, J.T., Reynolds, H.M . and Young, J.M . (1975). Investigation of the inertial properties of the human body. AM RL Technical Report 74-137 . Dayton, OH : Wright Patterson Air Force Base. C lauser, C.E ., M cConville, J.T. and Young, J.W. (1969). Weight, volume and centre of mass of segments of the human body. A M R L Tech nica l Report 69-70 . Dayton, OH: W right Patterson Air Force Base. Dapena, J. (1978). A m ethod to determine the angular m omentum of a human body about three orthogonal axes passing through its centre of gravity. Jour na l of B iom ech anics , 11 , 251 256. Enoka, R.M . (1988). N eurom echanica l B asis of K inesiology . C hampaign, IL: Human Kinetics. Feltner, M .E. and Dapena, J. (1986). Dynamics of the shoulder and elbow joints of the throwing arm during a baseball pitch. Inter nationa l Jour nal of Sport B iom echanics , 2 , 235 259. G regoire, L., Veeger, H.E., Huijing, P.A. and van Ingen Schenau, G.J. (1984). Role of m ono- and biarticular m uscles in explosive movements. Inter nationa l Jour na l of Sports M edicine , 5 , 301 305. Harm an, E.A., Rosenstein, M .T., Frykman, P.N. and Rosenstein, R.M . (1990). The e V ects of arms and countermovement on vertical jumping. M edicine and Science in Sports and Exercise , 22 , 825 833. Hay, J.G. and Reid, J.G. (1988). Anatom y , M ech anics and H um an M otion . Englewood C liV s, NJ: Prentice-Hall. Hennem an, E. (1979). Functional organization of motorneuron pools: The size principle. In Integ ration in the N er vous System (edited by H. Asanuma and V.J. Wilson), pp. 13 25. Tokyo: Igaku-Shoin. Herzog, W., Hasler, E. and Abrahamse, S.K. (1991). A comparison of knee extensor strength curves obtained theoretically and experimentally. M edicine and Science in Sports and Exercise , 23 , 108 114. Hill, A.V. (1938). The heat of shortening and the dynamic

Feltner et al.
constants of muscle. Proceedings of the R oyal Society of L ondon , B , 126 , 136 195. Hinrichs, R.N. (1985). Regression equations to predict segmental moments of inertia from anthropom etric m easurem ents: An extension of the data. Jour nal of B iom ech anics , 18 , 621 624. Hinrichs, R.N. (1990). Adjustments to the segment centre of mass proportions of Clauser et al . (1969). Jour nal of B iom ech anics , 23 , 949 951. Jensen, R.C ., Warren, B., Laursen, C . and M orrissey, M .C. (1991). Static pre-load e V ect and knee extensor isokinetic concentric and eccentric perfor mance. M edicine and Science in Sports and E xercise , 23 , 10 14. Keppel, G. (1982). Design and Analysis . Englewood Cli V s, NJ: Prentice-Hall. Khalid, W., Amin, M . and Bober, T. (1989). The in X uence of upper extremities m ovement on take-oV in vertical jump. In B iom echanics in Sports V (edited by L. Tsarouchas, J. Terauds, B. Gowitzke and L. Holt), pp. 375 379. Athens: Hellenic Sports Research Institute. Komi, P.V. and Bosco, C . (1978). Utilization of stored elastic energy in leg extensor muscles by m en and wom en. M edicine and Science in Sports and Exercise , 10 , 261 265. Luhtanen, P. and Komi, P.V. (1978). Segmental contribution to forces in vertical jump. E uropea n Jour na l of Applied Physiolog y , 38 , 181 188. M iller, D.I. (1976). A biomechanical analysis of the contribution of the trunk to standing vertical jump takeoV s. In Physical E ducatio n , Sport and the Sciences (edited by J. Broekho V ), pp. 355 374. Eugene, OR: M icroform Publications. Oddson, L. (1989). What factors determ ine vertical jumping height? In B iom echanics in Sports V (edited by L. Tsarouchas, J. Terauds, B. Gowitzke and L. Holt), pp. 393 401. Athens: Hellenic Sports Research Institute. Payne, A.H., Slater, W.J. and Telford, T. (1968). The use of a force platform in the study of athletic activities. E rgonom ics , 11 , 123 143. Perrine, J.J. and Edgerton, V.R. (1978). M uscle force velocity and power velocity relationships under isokinetic loading. M edicine and Science in Sports , 10 , 159 166. Shetty, A.B. and Etnyre, B.R. (1989). Contribution of arm m ovement to the force components of a m aximum vertical jump. Jour na l of O rthopaedic and Sports Physical Therapy , 11 , 198 201. Smidt, G.L. (1973). Biomechanical analysis of knee X exion and extension. Jour na l of B iom echanics , 6 , 79 92. Spoor, C .W. and van Leeuwen, J.L. (1992). Knee m uscle m oment arms from MRI and tendon travel. Jour nal of B iom ech anics , 25 , 201 206. Vaughan, C.L. (1980). An optimization approach to closed loop problems in biomechanics. Unpublished doctoral dissertation, University of Iowa, Iowa C ity, IA. Wood, G.A. and Jennings, L.S. (1979). On the use of spline functions for data smoothing. Jour na l of B iom ech anics , 12 , 477 479. Zajac, F.E. (1993). M uscle coordination of movement: A perspective. Jour na l of B iom echanics , 26 , 109 124.