ORIGINAL ARTICLE

Robert Ropret á Milos Kukolj á Dusan UgarkovicDragan Matavulj á Slobodan Jaric

Effects of arm and leg loading on sprint performance

Accepted: 10 November 1997

Abstract The e�ects of loading on sprint kinematicswere examined in 24 male students. The moment of in-ertia of either the arms or legs was increased by up to50% of their unloaded values and the time for distancesof 0.5±15 m and 15±30 m from a sprint start was mea-sured. An increase in leg loading was associated with agradual decrease in velocity of both sprint phases, whilethe change associated with arm loading was modest andsigni®cant only in the second phase. The decrease insprint velocity was predominantly due to a reduction instride rate, while the stride length remained almost un-changed. It was concluded that leg loading a�ectedsprint velocity more than arm loading, and also that thevelocity was reduced due to a decrease in the stride raterather than in the stride length.

Key words Sprint running á Loading á Velocity á Stridelength á Stride rate

Introduction

External loading of the human body has been widelyapplied in the study of various aspects of running. Forexample, it has been shown that moderate loads at-tached to either the legs or the trunk have no e�ect onstride length or stride frequency while running at givenvelocity (Cavanagh and Kram 1989; Martin and Cava-nagh 1990), although an increased stride frequency hasalso been reported (Cooke et al. 1991). Adding a load toa running subject would be expected to increase thework performed expressed as the oxygen consumption.

Loading of the arms, and in particular, of the trunk hasbeen demonstrated to have a moderate e�ect on the workperformed (Cooke et al. 1991), while loading of the legshas been found to increase oxygen consumption consid-erably (Martin 1985; Rusko and Bosco 1987; Bhambhaniet al. 1989; see also Anderson 1996 for review). Thus,external loading of a running subject has been considereda potential training method (e.g. Bosco et al. 1986; All-emeier et al. 1994) in order to combine high velocity withresistance training (Delecluse et al. 1995) and, therefore,to secure conversion of an enhanced muscle strength tomuscle power (Sleivert et al. 1995).

Since most of the previous studies have involvedrunning on a treadmill, the main aim of this study was toexamine possible e�ects of both the mass and distribu-tion of external loads on the kinematics of sprint run-ning. We recorded sprinting velocity and the stridelength of the subjects loaded with weights which weassumed would increase the moment of inertia of eithertheir arms or legs by up to 50%. Since sprint running hasat least two partly independent phases (c.f. Mero et al.1992), we studied separately the initial accelerationphase and the maximal running velocity.

Methods

Subjects

A group of 24 male students of physical education who had notreported any injuries [mean age 20.1 (SD 0.9) years] served assubjects. Their mean height was 179.6 m (SD 8.4) m, while theirmean mass was 74.5 (SD 9.8) kg. In addition to their regularphysical activities at college, 18 of the subjects were active in var-ious athletic disciplines such as games (n � 10), combat sports(n � 4) and middle-distance running (n � 2). Written informedconsent was obtained from all the subjects.

Loads

Loads were applied to increase the moments of inertia of either thesubjects' arms or legs. During arm loading the subjects held 0, 1, 2,or 3 short lead rods (0.22 kg each) in each hand. During leg

Eur J Appl Physiol (1998) 77: 547±550 Ó Springer-Verlag 1998

R. Ropret á M. Kukolj á D. Ugarkovic á D. Matavulj á S. JaricThe Research Center, Faculty for Physical Culture, Belgrade,Yugoslavia

S. JaricInstitute for Medical Research, Belgrade, Yugoslavia

S. Jaric (&)200 Biomechanics Laboratory, Department of Kinesiology,Penn State University, University Park, PA 16802, USA

loading, load belts of 0, 0.6, 1.2, or 1.8 kg were fastened above theankle joint of each leg. According to Stegemann (1981), these loadswould have increased the moment of inertia of the arms or the legsby up to 50%. The subjects were habituated to the external loadsduring their regular running practice over 4 weeks, with particularemphasis on all-out acceleration and maintenance of the maximalrunning velocity.

Procedure

For maximal sprinting velocity over 30 m, photocells were posi-tioned to time the subjects' passing distances of 0.5, 15, and 30 m toenable measurement of the average velocity of both the initial ac-celeration phase and the phase of maximal running velocity (i.e.0.5±15 m and 15±30 m, respectively). Two video cameras recordedthe position of the subject's foot contact with the ground to cal-culate the average stride length.

The arm-load sprint was followed, after a 2-week interval, bythe leg-load sprint. During any session, the subjects ran twice with

each of the loads and the higher overall velocity was noted forfurther analysis. Therefore, each test required eight running trails.The sequence of loads was randomized and 7-min rest was allowedbetween trails. A period of 10-min standardized warming-up, in-cluding stretching exercises, preceded each test. The subjects wereinstructed to attempt to achieve their all-out running performance.Speci®cally, they were asked to attain their maximal running ve-locity as soon as possible and, thereafter, to maintain it until theypassed the ®nishing line.

Data acquisition and processing

Running velocity was the distance covered divided by time neededas recorded by the photocells. Stride length for each phase was thedistance between the ®rst and last foot contact with the groundwithin the phase divided by the number of strides. Stride frequencywas calculated as running velocity divided by stride length. AnANCOVA was applied to assess the possible e�ects of di�erentloads on sprinting kinematics.

Fig. 1 E�ects of arm (left handgraphs) and leg loading (righthand graphs) on the runningvelocity (A, B), stride length (C,D), and stride rate (E, F) withinthe initial acceleration phase(un®lled squares, full line) andthe phase of maximal runningvelocity, ( ®lled squares, dashedline). Error bars depict standarderrors. Asterisks beside each setof data indicate F2,23 signi®-cance level obtained usingANCOVA where the no load-ing condition served as a co-variate, i.e. *P< 0.05,**P< 0.01. Arrows indicatesigni®cant di�erences betweenthe e�ects of particular loadsusing the Tukey signi®cant dif-ference test, +P< 0.05,++ P< 0.01

548

Results

The e�ects of arm and leg loading on the sprint kine-matics are shown in Fig. 1. The decrease in velocity inthe initial acceleration and the maximal running velocityphase associated with the maximal arm load (comparedwith the no load conditions, averaged among all thesubjects) were 0% and 1.0%, respectively. The sameresults associated with leg loading were 7.8% and12.8%, respectively (Fig. 1A, B). Thus, arm loadingcaused either no, or only a small decrease in runningvelocity, while the same e�ects associated with legloading were prominent.

Although an increase in arm and, particularly, legloading caused running velocity to decrease, stridelength remained una�ected (Fig. 1C,D). Running ve-locity is a product of stride length and stride rate andsince the stride length remained una�ected by loading,the decrease in the sprinting velocity was due to a de-crease in stride rate (Fig. 1E, F). An increase in legloading was associated with a gradual decrease in thestride rate. It is noteworthy that a decrease in the striderate of the maximal running velocity phase due to armloading remained insigni®cant although the sprintingvelocity decreased.

Discussion

Although none of the subjects were active sprinters, theyachieved relatively high average velocities, particularlywithin the distance from 15 to 30 m. This result could beexplained by the speci®c selection of the subjects by aqualifying examination which they took for entry to thephysical education college, as well as by the trainingprogramme taught at the college and by their own ath-letic activity. High-level sprinters have been found nor-mally to reach their maximal running velocity after 30 to50 m from a sprint start (Mero et al. 1992), while thedistance of 30 m we studied corresponded to that moreoften seen in ball games.

The most important result was the marked e�ect ofleg loading on the velocity of both sprinting phases,while only a small and in part an insigni®cant e�ect wasassociated with arm loading. A prominent decrease inrunning velocity associated with an increase in the legloading was expected. It has been reported that even alight load attached to either the ankle joint or foot of arunner causes a relatively high increase in the physio-logical demand (Anderson 1996) as a consequence of theincreased work needed to move the leg (Martin 1985),but also as a consequence of the enhanced gravitationalforces (Rusko and Bosco 1987). In these studies only amoderate speed of running was investigated and thesubjects were able to maintain an even velocity as wouldusually be set by a treadmill. In the present study, thesprint running required maximal e�ort in all trials. As aresult, the subjects' muscle forces could not counteract

the increased leg inertia and, consequently, theirsprinting velocity decreased.

The small and in part insigni®cant e�ects of armloading on sprinting velocity could be explained if thesubjects had compensated for the increased demand forthe swings of their arms by a reduced moment of inertiaby additional elbow ¯exion. Also, they could have re-duced the amplitudes of their arm swing, but such e�ectswere not measured.

Change in sub-maximal running velocity has beendescribed as being accomplished almost exclusively bythe stride rate (Mero et al. 1992). Similarly, our resultsdemonstrated that a decrease in sprinting velocity ob-served during leg loading was predominantly due to adecrease in stride rate. However, it has been found thatthe stride length of externally loaded subjects running atmoderate velocity also remained approximately unaf-fected (Cavanagh and Kram 1989; Martin and Cava-nagh 1990; Cooke et al. 1991). Taken together, theseresults could also be interpreted in such a way that stridelength, rather than stride rate, could be an invariantcharacteristic of running with di�erent external loads.Therefore, the conclusion of Cavanagh and Kram (1989)that ``factors other than limb segment mass or inertia areprimary determinants of stride strength'' based on theirexperiments performed at moderate velocity running,could be generalized to the full range of running veloc-ities.

Our study would suggest that leg loading a�ectssprint velocity more than arm loading, and also that thechanges in velocity are due to a reduction in the striderate rather than the stride length. Since the loading weused ful®lled important criteria for the e�ectiveness oftraining, such as ``movement speci®city'' and ``velocityspeci®city'' (see Delecluse et al. 1995), potential bene®tsof loading applied in sprint training also deserve atten-tion.

Acknowledgements The authors wish to thank D. Davidovic forhelp in data collection and processing. The work was supported, inpart, by a grant from the Serbian Research Foundation (13T23).

References

Allemeier CA, Fry AC, Johnson P, Hikida RS, Hagerman FC,Staron RS (1994) E�ects of sprint cycle training on humanskeletal muscle. J Appl Physiol 77:2385±2390

Anderson T (1996) Biomechanics and running economy. SportsMed 22:76±89

Bhamhani Y, Burnham R, Singh M, Gomes P (1989) Ankle andwrist weights: their e�ect on physiological responses duringtreadmill running. Arch Phys Med Rehabil 70:291±296

Bosco C, Rusko H, Hirvonen J (1986) The e�ect of extra-loadconditioning on muscle performance in athletes. Med Sci SportsExerc 18:415±419

Cavanagh PR, Kram R (1989) Stride length in distance running:velocity, body dimensions, and added mass e�ects. Med SciSports Exerc 21:467±479

Cooke CB, McDonagh MJN, Nevill AM, Davies CTM (1991)E�ects of load on oxygen intake in trained boys and men duringtreadmill running. J Appl Physiol 71:1237±1244

549

Delecluse C, Van Coppenolle H, Willems E, Van Leemputte M,Diels R, Goris M (1995) In¯uence of high-resistance and high-velocity training on sprint performance. Med Sci Sports Exerc27:1203±1209

Martin PE (1985) Mechanical and physiological responses to lowerextremity loading during running. Med Sci Sports Exerc17:427±433

Martin PE, Cavanagh PR (1990) Segment interactions within theswing leg during unloaded and loaded running. J Biomech23:529±536

Mero A, Komi PV, Gregor RJ (1992) Biomechanics of sprintrunning. Sports Med 13:376±392

Rusko H, Bosco C (1987) Metabolic response of endurance athletesto training with added load. Eur J Appl Physiol 56:412±418

Sleivert GG, Backus RD, Wenger HA (1995) The in¯uence of astrength-sprint training sequence on multi-joint power output.Med Sci Sports Exerc 27:1655±1665

Stegemann J (1981) Physiological bases of work. In: Skinner JS (ed)Exercise physiology. Year Book Medical Publishers, Chicago,Ill., pp 258±264

550