ORIGINAL ARTICLE

Guillaume Millet á Ste phane Perrey á Robin CandauAlain Belli á Fabio Borrani á Jean-Denis Rouillon

External loading does not change energy cost and mechanicsof rollerski skating

Accepted: 20 March 1998

Abstract The purpose of this study was to examine thee�ects of external loading on the energy cost andmechanics of roller ski skating. A group of 13 highlyskilled male cross-country skiers roller skied at19.0 ( SD 0.1) km á h)1 without additional load andwith loads of 6% and 12% body mass (mb). Oxygenuptake ( _V O2), knee and ankle joint kinematics, roller-skielectromyogram (EMG) of the vastus lateralis and gas-trocnemius lateralis muscles, and roller ski velocity wererecorded during the last 40 s of each 4-min period ofroller skiing. One-way repeated measures ANOVA re-vealed that the _V O2 expressed relative to total mass(mtot), joint kinetics, eccentric-to-concentric ratio of theintegrated EMG, velocity changes within a cycle, andcycle rate did not change signi®cantly with load. Thesubsequent analysis of the e�ect of load on each resis-tance opposing motion suggested that the power tosustain changes in translational kinetic energy, potentialenergy, and overcoming rolling resistance increasedproportionately with the load. The lack of a signi®cantchange in _V O2/mtot with external loading was associatedwith a lack of marked change in external mechanicalpower relative to mtot. The existence of an EMG signalduring the eccentric phase prior to the thrust (concentricphase), as well as the lack of signi®cant delay betweenthe two phases, showed that a stretch-shortening cycle(SSC) occurs in roller ski skating. Taken together, thepresent results would suggest that external loading up to12% mb does not increase storage and release of elasticenergy of lower limb muscles during SSC in roller skiskating.

Key words Ski skating á Energy cost á Externalloading á Stretch-shortening cycle

Introduction

Oxygen uptake ( _V O2) normalized to total transportedmass (mtot) during running has been found to decreasewith external loading (Bourdin et al. 1995; Cooke et al.1991; Davies 1980; Thorstensson 1986) or not to changesigni®cantly (Cureton and Sparling 1980; Davies 1980;Taylor et al. 1980). In experiments where _V O2/mtot wasfound to decrease with vertical load, it has been arguedthat the decrease in energy cost with loaded running wasdue to a modi®cation of running patterns in such a waythat a lower mechanical cost (i.e. mechanical power di-vided by velocity) was achieved (Bourdin et al. 1995).Alternatively, it has been proposed that the addition ofan external load could enhance the utilization of muscleelastic energy which may lead to an increase in economyduring loaded running (Cooke et al. 1991; Thorstensson1986). In fact, when an active muscle is stretched prior toa concentric contraction, the elastic energy stored duringthe eccentric phase can be released during the positivephase. It has been shown that this phenomenon, knownas the stretch-shortening cycle (SSC), is more e�cientthan pure concentric contractions (Bosco et al. 1982;Goubel 1987; Komi and KyroÈ laÈ inen 1996; Thys et al.1972).

In cross-country skiing, Komi and Norman (1987)have demonstrated the occurrence of SSC in the lowerlimb extensor muscles using the diagonal stride tech-nique. One investigation has also been devoted to thestudy of SSC in relation to the skating style of cross-country skiing (Candau et al. 1994). However, themuscle activity during the eccentric phase has not beendirectly measured in the study of Candau et al. (1994).Moreover, it has been reported that the techniquestudied in this investigation (V1 skate) is no longergenerally employed by skiers on ¯at or small gradientterrain (Bilodeau et al. 1996). Thus, it is worth to re-examining the occurrence of SSC in the lower limbsusing the V2-alternate (V2A) technique of ski skating.The ®rst purpose of the present study was to examine the

Eur J Appl Physiol (1998) 78: 276±282 Ó Springer-Verlag 1998

G. Millet (&) á S. Perrey á R. Candau á A. BelliF. Borrani á J-D. RouillonLaboratoire des Sciences du Sport, Unite de Formation et deRecherche en Sciences et Techniques des Activitie s Physiques etSportives, Place St-Jacques, F-25030 BesancË on Cedex, France

e�ects of external loading on the energy cost of roller skiskating (RSS). If SSC could be demonstrated, then thesecondary purpose of this study was to test in RSS thetwo hypotheses proposed to explain the decrease in en-ergy cost with loaded running, i.e. a decrease in me-chanical cost versus an increase in e�ciency.

Methods

Subjects

A group of thirteen highly skilled male cross-country skiers of re-gional to national level [mean age 22.3 (SD 2.5) years, height182.2 (SD 7.5) cm, mass 75.9 (SD 5.6) kg] completed the study.Informed consent was obtained from each subject prior to hisparticipation in the study.

Experimental protocol

The study was performed on a 424-m track with two straightsmeasuring 165 m and 168 m. The maximal gradient of the track,measured at 20-m increments with a water level device, was lessthan 1.4%. All measurements were done on one straight.

Each subject performed three 4-min periods of roller skiing withthe same roller skis (Elpex F1, Swedski, Sweden) under the fol-lowing conditions: unloaded (UL), loaded with 6.0 (SD 0.1)% ofbody mass (L6%) and loaded with 12.0 (SD 0.1)% of body mass(L12%). The order in which the subjects were tested with the dif-ferent loads was randomized. A weight jacket was used to load thesubjects and they were instructed to use their usual skating ski polesand boots. No subject had experience of loaded skiing. Roller-skiing speed was paced by an investigator riding 5 m in front of thesubject on a bicycle equipped with a calibrated electronic speed-ometre. Each period was also timed independently to determine theactual speed. The actual speeds were 19.0 (SD 0.1), 18.9 (SD 0.1)and 18.9 (SD 0.1) km á h)1, for the UL, L6% and L12% condi-tions, respectively. The subjects rested for at least 6 min betweenperiods of skiing.

The V2A technique (also referred to as the 2-skate, Gunde skateand open ®eld skate technique) was used in this study. It uses asymmetrical double pole plant as weight is transferred to one ski(Millet et al. in press). The side of the body to which double polingoccurs is considered to be the strong side, with the opposite sidebeing referred to as the weak side.

During each period of roller skiing, the wind velocity was re-corded from an anemometre (NG type 5393-2, SA Jules Richard,Argenteuil, France). Wind velocities were 2.6 (SD 1.2), 2.5 (SD 1.2)and 2.7 (SD 1.3) m á s)1 for UL, L6% and L12%, respectively, andwere not found to be signi®cantly di�erent. Air temperature was18 (SD 4)°C and ranged from 12 to 26°C.

Experiment procedures

Physiological variables

The _V O2 was determined during the last 30 s of each period ofskiing by collection of expired gases through a non-rebreathingvalve (Hans-Rudolph) into a meteorological balloon carried on thesubject's back as previously described by Ho�man and Cli�ord(1990). The subject began and ended gas collection by moving alever attached to a three-way valve (W.E Collins) secured with aharness over the upper abdomen. Movement of the lever could beachieved within a cycle with minimal disruption to the roller skiing.The collection periods were timed with a stopwatch mounted onthe gas collection system which was mechanically triggered bymovement of the lever. Across the range of speeds of the study,

collection periods varied from 34.9 (SD 2.1) to 36.2 (SD 2.9) s. Thefractions of oxygen and carbon dioxide in the expired gas weredetermined within 3 min of the end of each period by calibratedelectronic oxygen and carbon dioxide analysers (CPX, MedicalGraphics Corporation, St Paul, Minn.). The volume determina-tions were carried out in a balanced Tissot spirometre and cor-rected for the volume removed by the electronic analysers.

A 80-ll blood sample was also obtained from the ®ngertipimmediately after each period of skiing. Blood samples werecollected into heparinized capillary tubes and analysed using anenzymatic method to determine lactate concentration.

Roller ski rolling resistances

The rolling coe�cient (Rr) was determined using the coasting de-celeration method by iteration minimising the sum of the meansquares of the di�erences between the measured and the calculateddeceleration (DEC). The measured DEC was determined fromroller-ski speed during a coasting DEC test where each subjectmaintained an average position (i.e. between tuck and standingpositions). The calculated DEC was determined from the rela-tionship between the inertial force (Fi), aerodynamic force (Fa), therolling resistance (FRr) and the gravitational force (Fg) in thecoasting DEC:

Fi � Fa � FRr � Fg �1�

m � a � 0:5 � q � CDA � v2 � m � g � Rr � cos a� m � g � sin a �2�where m is mass of the skier (in kilograms), a is the calculated DEC(in metres per second squared), q is the air density (in kilograms permetre cubed), CDA is the drag area (in metres squared), v is thevelocity of the skier (in metres per second), g is the gravitationalacceleration (in metres per second squared), Rr is the rolling co-e�cient, a is the slope (radians). A value of 1 m2 was taken forCDA. This value represents a mean CDA between tuck and stan-ding positions for a subject with body height and mass comparableto the mean value found in the present study (Leino et al. 1983).The Rr was found to be 0.013 (SD 0.006.)

Joint kinematics, electromyographic activity and skiing velocity

Knee and ankle joint kinematics of both lower limbs were mea-sured from electrogoniometres (Sfernice 78 CSB, Nice, France).Calibration of the goniometres was carried out in the ®eld prior tothe testing of each subject. Gastrocnemius lateralis muscle (GL)and vastus lateralis muscle (VL) electromyogram (EMG) signalswere recorded from bipolar surface electrodes ®xed longitudinallyover the muscle belly (interelectrodes distance: 1.2 cm) undergoingampli®cation (gain 600, input impedance 2 GW, CMRR 90 dB) and®ltration (passband ®lter 6±600 Hz, Biochip, Grenoble, France).The EMG signals were recti®ed and integrated (iEMG). Bothroller-skis were also equipped with electronic speedometres(adapted from Sigma Sport, Neustadt, Germany). For each ski, thefront wheel circumference allowed the determination of about 20di�erent velocity values within each cycle.

Data-acquisition and averaging procedures

The EMG, goniometre and velocity data of both roller-skis weresimultaneously recorded by a portable data logger (SEIP, Elec-tronique Informatique du Pilat, Jonzieux, France) carried by thesubject in a waist-bag (total mass 0.6 kg). Data collection was be-gun by the subject in the turn preceding the last straight by trig-gering an electronic switch. Data were recorded for 40 s at asampling frequency of 800 Hz. After each period of roller skiing,the data were downloaded to a computer for subsequent analyses.The joint displacement curves were then inspected (with the ex-ception of the ®rst 10 s to avoid cycles that might have been af-fected by the turn) and ten consecutive cycles were selected for

277

analysis. Within each cycle, ®ve phases were determined (Fig. 1):eccentric, concentric, ¯ight, downward glide and upward glide. Thebeginning of all phases was determined by the changes in jointangle, except for the beginning of the downward glide which wasde®ned as the onset of VL EMG activity. For the downward glidephase, it was also veri®ed that the EMG signal slightly preceded the®rst recorded increase in ski velocity of the lower limb being con-sidered. The following variables were determined for each lowerlimb and phase: joint angle at the beginning and end of each phase,joint displacement, angular velocity, time, average EMG andiEMG.

Power outputs in roller skiing

For analysis of the e�ect of load on mechanical power, the di�erentresistances opposing motion must be considered. Subjects rollerskiing on ¯at terrain have to supply power to maintain changes intranslational kinetic energy (Pkt), rotational kinetic energy (Pkr),potential energy (Pp), overcoming air resistance (Pa), and over-coming rolling resistance (Pl). These powers can be calculated asfollows:

Pkt � 0:5 � m � �v21 ÿ v20� � CR �3�Pkr � R�0:5 � m1 � l2 � �x2

1 ÿ x20� � CR� �4�

Pp � m � g � h � CR �5�Pa � 0:5 � q � CDA � v3 �6�Pl � m � g �Rr � v �7�where m is mass (in kilograms), v1 and v0 are the extreme velocitiesduring the cycle (in metres per second), CR is the cycle rate (inhertz), m1 is the mass of segment (in kilograms), l is the length ofsegment (in metres), x1 and x0 are the extreme segment angularvelocities during the cycle (in radian per second), g is the acceler-ation of gravity (in metres per second square), h is the verticaldisplacement (in metres), q is the air density (in kilograms per metrecube), CDA is the frontal area (in the metres squared), v is thevelocity of displacement (in metres per second), Rr is the rollingresistance coe�cient (without unity).

Data analysis

Each study variable was compared between conditions with one-way analysis of variance (ANOVA) with repeated measures. For allstatistical analyses, a P value of 0.05 was accepted as the level ofstatistical signi®cance.

Results

Blood lactate concentration was not di�erent among thedi�erent load conditions (Table 1). The respiratoryquotient was equal to 0.87 in the unloaded condition. Asseen in Table 1, no signi®cant e�ect of load was foundfor _V O2/mtot.

Typical changes in joint angle and associated EMGactivity are given in Fig. 1. The ®ve phases de®ned in themethods were clearly recognizable in all subjects forboth lower limbs. All the skiers had a phase of negativeangular velocity prior to the concentric phase with nopause (or a delay shorter than 0.05 s) between these twophases. The average velocities over the eccentric phasewere )0.75 (SD 0.4) and )0.81 (SD 0.2) rad á s)1 for theknees and the ankles, respectively (UL condition). Si-multaneous with the eccentric action, all the skiersshowed EMG signals with an average value of0.10 (SD 0.04) and 0.09 (SD 0.03) mV for VL and GLmuscles, respectively (UL condition). Values for angularvelocity, phase duration, and iEMG for the eccentricand concentric phases are given in Tables 2 and 3. Theaverage EMG activity was greater during the eccentric

Table 1 Oxygen uptakes expressed per kilogram of body mass (mb) and per kilogram of total mass (mtot), and blood lactate con-centrations. UL Unloaded; L 6%, L 12% loaded with 6% and 12% mb, respectively; _V O2 oxygen uptake

UL L 6% L 12%

Mean SD Mean SD Mean SD

_V O2 (ml á min)1 á kg mÿ1b ) 44.1 5.4 46.1 4.5 47.9 5.0**_V O2 (ml á min)1 á kg mÿ1tot ) 44.1 5.4 43.5 4.3 42.8 4.4Blood lactateconcentration (mmol á l)1)

2.7 1.4 2.9 1.6 2.8 1.4

Comparison among conditions**P < 0.01

Fig. 1 Examples of goniometre/electromyogram (EMG) curves forthe knee/vastus lateralis muscle (upper panel), and the ankle/gastrocnemius lateralis muscle (lower panel), showing the ®ve phases:eccentric (Ecc), concentric (Conc), ¯ight, downward glide (DownGL),upward glide (UpGL)

278

phase than during the concentric phase for both the VL(P < 0.001) and GL (P < 0.05) over all the experi-mental conditions (Fig. 2).

No di�erences were found between UL, L6% andL12% in cycle rate [0.58 (SD 0.06), 0.59 (SD 0.07),0.60 (SD 0.06) Hz, respectively] and in variation ofspeed within a cycle [12.2 (SD 4.2), 12.2 (SD 4.3),11.4 (SD 4.9)% of average speed]. Similarly, joint ki-nematics (Fig. 3), average EMG (Fig. 2) and iEMG(Table 3) were not found to be signi®cantly di�erentamong the di�erent conditions of load. Also, eccentric-to-concentric ratio of iEMG was not altered by theaddition of an external mass for either VL or GL(Table 3).

Discussion

Energy cost of roller ski skating

The highest mean blood lactate concentration was 2.9mmol á l)1, suggesting that energy derived from anaer-obic metabolism was negligible in the present study. As aresult, _V O2 could be used to determine the energy cost ofskiing (Cs). A value of 5 mlO2 á min)1 á kg)1, corre-sponding to _V O2 of a subject presenting all the charac-teristics of a running man, except for running (Lacour

Table 2 Angular velocities for the eccentric (ECC) and concentric (CONC) phases and duration of these two phases for the knee andankle joints. UL unloaded; L 6%, L 12% loaded with 6% and 12% body mass, respectively

UL L 6% L 12%

Mean SD Mean SD Mean SD

Knee joint Angular velocity(rad á s)1)

ECC )0.75 0.4 )0.74 0.4 )0.64 0.3

CONC 1.97 0.5 1.99 0.6 1.84 0.5Time (s) ECC 0.25 0.08 0.23 0.07 0.23 0.08

CONC 0.22 0.03 0.22 0.03 0.23 0.03

Ankle joint Angular velocity(rad á s)1)

ECC )0.81 0.2 )0.87 0.3 )0.82 0.3

CONC 3.60 1.1 3.63 1.0 3.58 1.0Time (s) ECC 0.21 0.04 0.20 0.03 0.20 0.04

CONC 0.17 0.02 0.18 0.01 0.18 0.02

Table 3 Integrated electromyogram (iEMG) for the eccentric (ECC) and concentric (CONC) phases and eccentric-to-concentric ratio ofthe iEMG (Ecc/Conc) for the vastus lateralis (VL) and gastrocnemius lateralis (GL) muscles. UL unloaded; L 6%, L 12% loaded with6% and 12% body mass, respectively

UL L 6% L 12%

Mean SD Mean SD Mean SD

VL iEMG (mV á s) ECC 0.026 0.014 0.025 0.012 0.025 0.013CONC 0.015 0.007 0.017 0.008 0.017 0.008

Ecc/Conc 1.8 0.8 1.6 0.6 1.5 0.6GL iEMG (mV á s) ECC 0.018 0.008 0.018 0.009 0.019 0.008

CONC 0.012 0.006 0.012 0.006 0.013 0.006Ecc/Conc 1.7 0.7 1.7 0.6 1.5 0.5

Fig. 2 Mean values for the average electromyogram (EMG) activityof the vastus lateralis muscle (upper panel) and the gastrocnemiuslateralis muscle (lower panel) for the eccentric and concentric phases.*, **, *** Signi®cant di�erences between the eccentric and theconcentric phases at the P < 0.05, P < 0.01 and P < 0.001 levels,respectively. Brackets represent 1 SD. UL unloaded; L 6%, L 12%loaded with 6% and 12% body mass, respectively

279

1996), was removed from _V O2 to calculate Cs. Inthe present study, Cs was 123.6 mlO2 á kg)1 á km)1

(2.53 J á kg)1 á m)1) in the UL condition. To the best ofour knowledge, no other study has examined the energycost of ski or RSS using the V2A technique. However,Cs in the three principal skating techniques (V1 skate,V2 skate and V2A) has been indirectly determined fromheart rate measurements (Bilodeau et al. 1991) and hasnot been found to be di�erent among the three tech-niques. Thus, it was worth comparing our data withprevious Cs established in ski skating. The value of Cs

was 8%±23% lower in the present study than in previousones that have been conducted in ski or RSS at com-parable speeds (Ho�man and Cli�ord 1990; Ho�manet al. 1990; Saibene et al. 1989). This di�erence mightseem large but it must be considered that mechanicalpower at a given velocity (i.e. mechanical cost) islargely altered by rolling resistance (see Eq. 7). In fact,Cs was found to be 31% higher when Rr was increasedfrom 0.012 to 0.025 in RSS (Millet et al. 1997). Thelow Rr (0.013) of the roller skis used in the presentstudy allows us to conclude that the value of Cs isreasonable.

E�ects of load on energy cost and mechanics of RSS

The main result of the present study was that adding anexternal load up to 12% of mb did not change _V O2 perkilogram mtot in RSS. This is in contrast to the ®ndings

of previous studies of running (Bourdin et al. 1995;Cooke et al. 1991; Davies 1980; Thorstensson 1986). Thedecrease in _V O2/mtot resulting from carrying an addi-tional load during locomotion on land can be due to twofactors:

1. The mechanical power related to mtot does not in-crease proportionately to external mass or

2. The e�ciency is altered.

These two hypotheses have been proposed for running(Bourdin et al. 1995; Thorstensson 1986, respectively).We speculated that studying the e�ects of loading on themechanics of RSS may con®rm one or other of these twohypotheses. In the present study, the mechanical costwas not determined. However, it is worth evaluatinghow load changed the mechanical powers due to eachresistance acting on the skiers (see Eqs. 3±7).

The present results would suggest that all butrotational kinetic and aerodynamic powers increasedproportionately to mtot. In fact, neither CR nor thevariation of speed within a cycle as a function of theaverage speed changed signi®cantly with load. As aconsequence, Pkt was likely to have increased propor-tionately to mtot when adding external load (i.e.DPkt µ Dmtot). The present results showed that externalloading did not change the knee and ankle kinematics.Even if the weight jacket may have slightly changed theposition of the centre of gravity (CG) and the ¯exionand extension range of motion for the trunk, we wouldspeculate that loading did not have a major in¯uence onthe vertical oscillations of CG. Since CR did not changewith load, Pp may have increased proportionately to mtot

(DPp µ Dmtot). Assuming Rr to be independent of mass

(i.e. Rr µ m0) as has previously been shown (Ho�manet al. 1995; Millet et al. in press), Pl must also haveincreased as a function of mtot).

However, since the load was not applied to the limbs,rotational kinetic power would not have increased withexternal load (DPkr µ Dm0

tot). Similarly, the weight jacketwould not have had a major in¯uence on the aerody-namic drag (i.e. CDA in proportion to the external loadraised to power 0) so that it is reasonable to assume thatthe increase of Pa was independent of the increase in mtot

(DPa µ Dm0tot). From this, it may be deduced that the

power needed by the skier to overcome all the resistancespreviously described would have increased slightly lessthan the increase in mtot (DSP µ Dmtot < 1). An increaseof 1% of mb should not induce an increase of 1% oftotal external power, i.e. an increase of 1% of _V O2 if thee�ciency does not change. The slight and non signi®cantdecrease of _V O2 when expressed per kilogram of mtot

observed in this study may have been due only to thisfactor. In other words, the independence of rotationalkinetic energy and aerodynamic drag with additionalload could explain why a given increase in mb (6% and12%) led to a slightly smaller increase in _V O2 (4.7% and8.8%) in the present study.

Fig. 3 Mean values for angle of the knee (upper panel) and ankle(lower panel) as a function of time for the unloaded (UL, ®lledsquares), the loaded with 6% body mass (L6%, open circles) and theloaded with 12% body mass (L12%, ®lled triangles) conditions.Brackets represent 1 SD. For other de®nitions see Fig 1

280

E�ects of load on e�ciency of lower limb muscles

The previous analysis shows that:

1. The lack of signi®cant change in energy cost wasassociated with a lack of signi®cant e�ects of loadingon the mechanics of roller-skiing and

2. The slight and non-signi®cant but consistent de-creases in energy cost with loading could be explainedby the independence of loading on rotational kineticenergy and aerodynamic drag.

As a consequence, it can be suggested that the e�ciencyof the lower limb muscles was not altered by verticalloading. E�ciency has been given as a possible expla-nation for the lower energy cost of loaded running(Cooke et al. 1991; Thorstensson 1986) mainly in rela-tion to a better storage and release of elastic energy inSSC. One could argue that SSC has not been directlydemonstrated in ski skating so that e�ciency of thelower limb muscles cannot be improved on by the use ofelastic energy.

However, the three conditions needed to identify aSSC during locomotion were shown in the present study.First, the existence of an eccentric phase (i.e. stretchingof the propulsive muscles) prior to the thrust was ob-served. For the extensor muscles of the lower limbs, thisis the case when a negative angular velocity is measuredfor the knee and ankle joints. Since it has been shownthat inactive muscles are unable to store elastic energy(Komi 1984), EMG activity also must be observedduring the eccentric phase. In the present study, EMGactivity was clearly identi®ed during the eccentric phase,so the ®rst two conditions for a SSC were met. No delay,or only a very short delay (less than 0.05 s), wasidenti®ed between the eccentric and concentric phases,so the stored elastic energy was unlikely to have beentransformed to heat. Moreover, the average EMG ac-tivity was found to be higher during the eccentric phasethan during the concentric phase. Komi (1984) hasdemonstrated that a major activation of the muscleduring the eccentric phase is typical of SSC.

Thus, despite the existence of SSC in ski skating, itseems that 6%±12% of mb load did not adequatelychange the e�ciency of the lower limb muscles. Thisresult can be reinforced by the fact that the eccentric-to-concentric ratio of the iEMG, which has been shown tobe a good index of e�ectiveness of SSC (Bosco et al.1982; Bourdin et al. 1995), was not modi®ed by load. Inthe present study, the de®nition of the eccentric andconcentric phases was based on the angular displace-ment. Since the ski was in contact with the road duringthe eccentric phase, this phase was probably accuratelyde®ned. The lack of data regarding the contact phasemay make the de®nition of the concentric phase some-what uncertain. However, even if the end of the thrust(i.e. the end of the concentric phase) probably slightlypreceded the maximal knee extension, we speculatedthat the di�erence between these two events did notchange with load.

It has also been hypothezised that variation of stridelength in running may cause a shift in the force-veloci-ty curve and improve e�ciency (Thorstensson 1986).However, there were no variations in CR in the presentstudy (i.e. cycle length at a given velocity), suggestingthat this factor is unlikely to have signi®cantly in¯u-enced the results. Thus, the lack of signi®cant variationsof

1. The eccentric-to-concentric ratio of the iEMG forboth VL and GL, and

2. The cycle length, tend to con®rm that the e�ciencywas not a�ected by external loading. This has previouslybeen suggested by Bourdin et al. (1995), who have foundthat the kinematic pattern of running was adjusted, byincreases in stride frequency and decreases in verticaloscillation, to limit the increase in mechanical power. Itis then possible that subjects of studies where _V O2/mb

have been found to increase proportionately to verticalload (present study; Cureton and Sparling 1980; Davies1980; Taylor et al. 1980) did not change adequately theirmovement pattern to limit the rise in mechanical cost,i.e. increase of _V O2/mb. Thus, the results of the presentstudy would suggest that adaptation of movement pat-terns during loaded locomotion could explain the dis-crepancy between studies devoted to the e�ects of loadon energy cost.

In conclusion, the present results showed that externalloading up to 12% of mb does not change signi®cantlythe energy cost of RSS and has no signi®cant e�ect onjoint kinematics, CR and change of velocity within acycle. The analysis of these mechanical modi®cationssuggest that all mechanical power outputs increasedproportionately with mtot with the exception of rota-tional kinetic power and power to overcome aerody-namic drag. The independence of aerodynamic androtational kinetic powers with external mass can explainthe slight and non signi®cant decrease of _V O2 expressedper kilogram of mtot. This suggests that the e�ciency ofthe muscles of the lower limbs was not altered by loaddespite the occurrence of SSC in RSS.

Acknowledgements The authors are grateful to Drs. Martin Ho�-man and Philip Cli�ord from the Sports Performance and Tech-nology Laboratory of the Medical College of Wisconsin forprovision of the oxygen uptake collection system used in this study.We would also like to acknowledge Isabelle Millet for her help withdata collection.

References

Bilodeau B, Roy B, Boulay MR (1991) A comparaison of threeskating techniques and the diagonal stride on heart rate re-sponses and speed in cross-country skiing. Int J Sports Med12:71±76

Bilodeau B, Rundell KW, Roy B, Boulay MR (1996) Kinematics ofcross-country ski racing. Med Sci Sports Exerc 28:128±138

Bosco C, Ito A, Komi PV, Fekete G, Apor P (1982) Store andrecoil of elastic energy in slow and fast types of human skeletalmuscles. Acta Physiol Scand 114:543±550

281

Bourdin M, Belli A, Arsac LM, Bosco C, Lacour JR (1995) E�ectof vertical loading on energy cost and kinematics of running intrained male subjects. J Appl Physiol 79:1078±1085

Candau R, Belli A, Chatard JC, Carrez G, Lacour JR (1994)Stretch-shortening cycle in the skating technique of cross-country skiing. Sci Mot 22:17±22

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

Cureton KJ, Sparling PB (1980) Distance running performance andmetabolic response to running in men and women with excessweight experimentally equated. Med Sci Sports Exerc 12:288±294

Davies CTM (1980) Metabolic cost of exercise and physical per-formance in children with some observations on externalloading. Eur J Appl Physiol 45:95±102

Goubel F (1987) Muscle mechanics. In: Hebbelinck M, Shepard RJB (eds) Medicine and Sport Science, vol 26. Karger, Basel,pp 24±35

Ho�man MD, Cli�ord PS (1990) Physiological responses to dif-ferent cross-country skiing techniques on level terrain. Med SciSports Exerc 22:841±848

Ho�man MD, Cli�ord PS, Foley PJ, Brice AG (1990) Physiolog-ical responses to di�erent roller skiing techniques. Med SciSports Exerc 22:391±396

Ho�man MD, Cli�ord PS, Watts PB, O'Hagan KP, MittelstadtSW (1995) Delta e�ciency of roller skiing with the doublepole and diagonal stride techniques. Can J Appl Physiol20:465±479

Komi PV (1984) Physiological and biomechanical correlates ofmuscle function: e�ects of muscle structure and stretch-short-ening cycle on force and speed. In: Terjung RL (ed) Exerciseand Sport Science Review vol 12. Collamore Press, Toronto, pp81±121

Komi PV, KyroÈ laÈ inen H (1996) Mechanical e�ciency of stretch-shortening cycle exercise. In: Marconnet P, Saltin B, Komi P,Poortmans J (eds) Human muscular function during dynamicexercise. Medicine and Sport Science, vol 41. Karger, Basel,pp 44±56

Komi PV, Norman RW (1987) Preloading of the thrust phase incross-country skiing. Int J Sports Med 8:48±54

Lacour JR (1996) In¯uence of body dimensions, sex and trainingon the energy cost of running. In: Marconnet P, Saltin B, KomiP, Poortmans J (eds) Human muscular function during dynamicexercise. Medicine and Sport Science, vol 41. Karger, Basel,pp 432±43

Leino MAH, Spring E, Suominen H (1983) Methods for the si-multaneous determination of air resistance to a skier and thecoe�cient of friction of his skis on the snow. Wear 86:101±104

Millet GY, Ho�man MD, Candau RB, Buckwalter JB, Cli�ord PS(1997) High resistance roller skis induce greater upper bodydemands than low resistance roller skis in ski skating. Med SciSports Exerc 29:S8

Millet GY, Ho�man MD, Candau RB, Cli�ord PS Poling forcesduring roller skiing: e�ects of technique and speed. Med SciSports Exerc, in press

Saibene F, Cortili G, Roi G, Colombini A (1989) The energy costof level cross-country skiing and the e�ect on the friction of theski. Eur J Appl Physiol 58:791±795

Taylor C, Heglund NC, McMahon TA, Looney TR (1980) Ener-getic cost of generating muscular force during running: acomparison of large and small animals. J Exp Biol 86:9±18

Thorstensson A (1986) E�ects of moderate external loading on theaerobic demand of submaximal running in men and 10 year-oldboys. Eur J Appl Physiol 55:569±574

Thys H, Farggiana T, Margaria R (1972) Utilization of muscleelasticity in exercise. J Appl Physiol 32:491±494

282