Human Movement Science xxx (2013) xxx–xxx

Contents lists available at ScienceDirect

Human Movement Science

journal homepage: www.elsevier .com/locate/humov

Biomechanical and energetic determinantsof technique selection in classical cross-countryskiing

0167-9457/$ - see front matter � 2013 Elsevier B.V. All rights reserved.http://dx.doi.org/10.1016/j.humov.2013.07.010

⇑ Corresponding author. Address: CeRiSM, Center of Research in Mountain Sport and Health, University of Verona, Videl Ben 5/B, 38068 Rovereto, TN, Italy.

E-mail address: barbara.pellegrini@univr.it (B. Pellegrini).

Please cite this article in press as: Pellegrini, B., et al. Biomechanical and energetic determinants of tecselection in classical cross-country skiing. Human Movement Science (2013), http://dx.doi.org/1j.humov.2013.07.010

Barbara Pellegrini a,b,⇑, Chiara Zoppirolli a,b, Lorenzo Bortolan a,b,Hans-Christer Holmberg c, Paola Zamparo b, Federico Schena a,b

a CeRiSM, Center of Research in Mountain Sport and Health, University of Verona, Rovereto, Italyb Department of Neurological, Neuropsychological, Morphological and Movement Sciences, University of Verona, Verona, Italyc Swedish Winter Sports Research Centre, Department of Health Sciences, Mid Sweden University, Östersund, Sweden

a r t i c l e i n f o a b s t r a c t

Article history:Available online xxxx

PsycINFO classification:23002330

Keywords:Human locomotionPoling force

Classical cross-country skiing can be performed using three maintechniques: diagonal stride (DS), double poling (DP), and doublepoling with kick (DK). Similar to other forms of human and animalgait, it is currently unclear whether technique selection occurs tominimize metabolic cost or to keep some mechanical factors belowa given threshold. The aim of this study was to find the determi-nants of technique selection. Ten male athletes roller skied on atreadmill at different slopes (from 0� to 7� at 10 km/h) and speeds(from 6 to 18 km/h at 2�). The technique preferred by skiers wasgathered for every proposed condition. Biomechanical parametersand metabolic cost were then measured for each condition andtechnique. Skiers preferred DP for skiing on the flat and they tran-sitioned to DK and then to DS with increasing slope steepness,when increasing speed all skiers preferred DP. Data suggested thatselections mainly occur to remain below a threshold of polingforce. Second, critically low values of leg thrust time may limitthe use of leg-based techniques at high speeds. A small role hasbeen identified for the metabolic cost of locomotion, which deter-mined the selection of DP for flat skiing.

� 2013 Elsevier B.V. All rights reserved.

a Matteo

hnique0.1016/

2 B. Pellegrini et al. / Human Movement Science xxx (2013) xxx–xxx

1. Introduction

Cross-country skiing is a form of locomotion performed with the use of passive locomotory tools,skis and poles, in which propulsion is generated by both upper and lower body musculature and whichallows travel on snow with a reduced cost of transport compared to walking and running (Minetti,2004; Saibene & Minetti, 2003).

Classical cross-country skiing can be performed by using three main techniques: diagonal stride,double poling, and double poling with kick (Bilodeau, Rundell, Roy, & Boulay, 1996; Smith, 1992).Diagonal stride (DS) is performed by exerting force through the skis and poles while the arms and legsare moved in a coordinated pattern that resembles that of walking or running, since the push of onearm is performed along with the push of the contralateral leg. The leg’s propulsive action entails a stopin the motion of the ski and a rapid downward and backward movement of the leg that can be de-scribed as a backward kick. Double poling (DP) is performed with symmetrical and synchronousmovements of both poles, the propulsive action of which is enhanced by a considerable trunk flexionwhile the legs’ involvement is minimal (Holmberg, Lindinger, Stoggl, Bjorklund, & Muller, 2006). Dou-ble poling with kick (DK) is performed through a poling action similar to that described for DP but, inthis case, propulsion is obtained also by a left or a right leg kick. The techniques are therefore charac-terized by different contributions from the lower and upper limbs and by different coordinativepatterns.

Cross-country skiers can switch between techniques according to changes in speed and track slope;the different techniques can thus be considered to be a gear system (Nilsson, Tveit, & Eikrehagen,2004) and the selection of the appropriate technique could have important implications for locomo-tory efficiency and performance. It is generally known that DS is used mainly on moderate to steepuphill slopes, that DP is selected for skiing on flat tracks (and it is the preferred technique at highspeeds), and that DK is mainly used for skiing on low to moderate slopes. However, the parametersthat determine the selection of these techniques are still unknown.

A widespread assumption in human and animal locomotion is that gaits are selected to minimizemetabolic energy cost (Alexander, 1989; Margaria, 1938; Mercier et al., 1994; Rubenson, Heliams,Lloyd, & Fournier, 2004). Margaria (1938), who first investigated the walk to run transition in humans,reported that walking is metabolically more expensive than running at the speed at which humanschange gait, whereas below this speed the opposite is true. A similar metabolic advantage has beenobserved in horses as they change from walk to trot and gallop (Hoyt & Taylor, 1981). Also hoppinganimals, such as the kangaroo, seem to select metabolically optimal gaits in relation to different loco-motion speeds, the transition in this case being from pentapedal walking to bipedal hopping (Dawson& Taylor, 1973). These findings suggest that gait selection occurs in a number of species in order tominimize the metabolic cost of locomotion. However, some more recent studies on horses (Farley &Taylor, 1991) and humans (Hreljac, 1993; Minetti, Ardigo, & Saibene, 1994; Rotstein, Inbar, Berginsky,& Meckel, 2005) have found that the actual transition occurs at a speed that is lower than theenergetically optimal transition speed. It is moreover questionable that both animals and humanscould have a precise perception of the metabolic cost they are paying (Farley & Taylor, 1991) and alsothe time needed ‘‘to make a decision’’ about the preferred gait mode is much less than the time neededto reach a metabolic steady state. Several researchers have thus hypothesized that gait transitionsare the result of a subjective feeling and that local factors, such as muscular overexertion, couldprevail over central factors (i.e., metabolic related variables). According to this hypothesis, comfortcriteria would be predominant over energy savings criteria (Daniels & Newell, 2003; Prilutsky &Gregor, 2001; Thorstensson & Roberthson, 1987). Many studies have investigated the role of biome-chanical parameters in determining the transition between gaits in humans and animals, some ofthem finding that transition occurs in horses when the critical value of musculoskeletal forces isreached (Biewener & Taylor, 1986; Biewener, Thomason, Goodship, & Lanyon, 1983; Farley & Taylor,1991; Taylor, 1985).

Even if it is reasonable to hypothesize that, in analogy with other vertebrates, the transitionbetween gaits in humans could be triggered by kinetic factors, no agreement has so far been reachedon the existence of a critical force level or other mechanical factors at the transition from walking to

Please cite this article in press as: Pellegrini, B., et al. Biomechanical and energetic determinants of techniqueselection in classical cross-country skiing. Human Movement Science (2013), http://dx.doi.org/10.1016/j.humov.2013.07.010

B. Pellegrini et al. / Human Movement Science xxx (2013) xxx–xxx 3

running. An investigation of gait transition in humans (Raynor, Yi, Abernethy, & Jong, 2002), whichconsidered as possible determinants numerous kinematical, kinetic, and energetic variables concludedthat the time to first peak force and the loading rate are the most likely determinants of gait transition.It was moreover hypothesized that high angular velocities and rapid muscle activation for hip and an-kle speed at increasing walking speed would drive the transition to running (Hreljac, 1995; Hreljac,Arata, Ferber, Mercer, & Row, 2001; Malcolm, Segers, van Caekenberghe, & de Clercq, 2009; Minettiet al., 1994; Segers, Lenoir, Aerts, & de Clercq, 2007).

Up to now, only one study has attempted to analyze the determinants of technique selection inclassical cross-country skiing (Cignetti, Schena, Zanone, & Rouard, 2009). In that study, the switch be-tween techniques was investigated by studying inter-limb coordination from a dynamical systemsperspective. Within this theory, the preferred patterns of coordination in human movements are char-acterized by stability in the dynamical system and thus the transitions between coordination patternsoccur as a result of gradual changes in stability. Cignetti et al. (2009) demonstrated that in cross-coun-try skiing the transition between techniques is associated with a loss of coordination stability and theythus hypothesized that transition is driven by changes in the stability of the coordination pattern. Thisapproach, however, provides no answer to the prevailing question of why the movement patternchanges.

The aim of this study was to determine the factors that determine the selection of techniques inclassical cross-country skiing. We hypothesized that the minimization of metabolic cost may not bethe prevalent determinant of technique selection and that mechanical factors could play a significantrole. To achieve this aim we: (1) recorded the preference of the skiers for a given technique (among DS,DP and DK), for a variety of slope and speed conditions and (2) measured the biomechanical and ener-getic responses of these ‘‘gait modes’’ in the same slope and speed conditions. From the comparisonbetween the subjective selection of the preferred techniques and the objective evaluation of theparameters recorded at each condition, we (3) examined each of these parameters to assess whetherthey could be determinants of technique selection.

2. Methods

2.1. Subjects

Ten male cross-country skiers were recruited for this study. Inclusion criteria were: to have learnedto ski when they were young, to have competed at national level for at least 5 years, to be active at thetime of the investigations (at least at amateur level) and to have experience in the use of roller skis.These criteria aimed to ensure that all the participants were highly skilled in each of the investigatedtechniques, were familiar with the transitions between them and were able to sustain the proposedworkload.

The stature and body mass of the subjects (mean ± standard deviation) were respectively1.78 ± 0.05 m and 72.0 ± 4.7 kg and the age was 30.7 ± 5.3 years. Their VO2max, measured during anincremental test performed with the diagonal stride technique on a treadmill using an ergo-spiromet-ric system (Quark, Cosmed, Rome, Italy), was 62.7 ± 7.3 ml�kg�1 min�1. Before participating, writteninformed consent was obtained from each participant in accordance with the approval of the Institu-tional Review Board.

2.2. Overall design and experimental setup

All tests were performed using roller skis on a motorized treadmill with a belt surface that was2.5 m wide and 3.5 m long (RL3500E, Rodby, Sweden). All athletes used the same pair of roller skis(Nord, Ski Skett, Sandrigo, Italy) and the poles were available at multiple lengths of 2.5 cm so that eachathlete could use their preferred pole length. The experimental setup is schematically represented inFig. 1. The coefficient of rolling resistance on the treadmill surface was determined from measure-ments of the force required to tow a skier standing still on the roller skis used during the test, as donein a previous study (Pellegrini, Bortolan, & Schena, 2011); this was found to be 0.024.

Please cite this article in press as: Pellegrini, B., et al. Biomechanical and energetic determinants of techniqueselection in classical cross-country skiing. Human Movement Science (2013), http://dx.doi.org/10.1016/j.humov.2013.07.010

markers for roller ski motion detection

optoelectonic cameras

ergospirometric measurement system

poling force

safety harness

Fig. 1. Schematic illustration of the experimental setup.

4 B. Pellegrini et al. / Human Movement Science xxx (2013) xxx–xxx

The investigation was divided into two parts; the first part was aimed at determining of the pre-ferred skiing technique and was conducted in a single session. The second part was devoted to themeasurement of biomechanical and metabolic parameters and was divided into four sessions. All test-ing sessions were preceded by an 8 min low intensity warm up (2� and 8 km/h) performed using allthree techniques considered in the study. All sessions were performed at the same time of the dayon non-consecutive days and were completed within three weeks.

2.3. Determination of the preferred skiing technique protocol

In the first part of the study, the preferred skiing techniques in different experimental conditionswere investigated. For this test we selected eight different slopes at a fixed speed, (from 0� to 7�, with1� increments at 10 km/h) and six different speeds at a fixed slope (from 6 km/h to 18 km/h with2 km/h increments at a slope of 2�), giving a total of 14 combinations of slope and speed. The exper-iments were conducted by asking the subject to ski using each of these combinations at least threetimes in random order, to compensate for any possible hysteresis effect (Diedrich & Warren, 1998).Each combination of speed and slope was maintained for at least 60 s and the subjects were allowedto try all techniques before declaring which one they preferred in those conditions. The precise ques-tion that was addressed to the subjects was: ‘‘Which technique would you spontaneously use if youhad to keep skiing at this speed and slope for about 10 min?’’. If the skiers were not able to expressa unique preference they were allowed to report two of them; in that case, a weight of 0.5 was givento each technique. When the answer given after the three repetitions of the same condition was al-ways the same, a unit score was assigned to that technique; otherwise, a weight taking into accountthe ‘‘frequency’’ of the preference declared was considered (e.g. the weight was 0.333 if the techniquewas selected once over three repetitions). The percentage of the subjects that selected each techniquefor any given speed and slope condition was then calculated.

2.4. Metabolic and biomechanical measurements protocol

In the second part of the study, metabolic and biomechanical parameters were measured for eachcombination of slope and speed considered during the first experiment. These experimental

Please cite this article in press as: Pellegrini, B., et al. Biomechanical and energetic determinants of techniqueselection in classical cross-country skiing. Human Movement Science (2013), http://dx.doi.org/10.1016/j.humov.2013.07.010

B. Pellegrini et al. / Human Movement Science xxx (2013) xxx–xxx 5

conditions were randomly divided into 4 testing sessions and the skiers were required to performthem using all techniques in random order; the skiers were then required to also use techniques theywould not have spontaneously used.

Resting oxygen uptake (V’O2) was measured at the beginning of each testing session for 3 minwhile the subject was standing still (Saibene, Cortili, Roi, & Colombini, 1989). A capillary blood sample(20 lL) was taken from the earlobe for at basal measurement of blood lactate concentration. Each ski-ing stage lasted for 4 min to allow for steady state levels of oxygen uptake. Respiratory gas exchangewas measured continuously and kinematic measurements were performed during the second minuteof each stage. At the end of each stage, capillary blood (20 lL) was sampled from the earlobe for bloodlactate concentration. The skiers were allowed to rest between successive stages for at least 3 min anduntil their heart rate returned below a value of 100 bpm, as measured by a heart rate monitor (PolarS610, Polar Electro OY, Kempele, Finland). During resting time, they were allowed to drink ‘‘ad libi-tum’’. When a subject was not able to sustain the required condition for 4 min, the measures were dis-carded. Only conditions completed by all athletes were retained for subsequent analysis.

Oxygen uptake (V’O2), carbon dioxide production (V’CO2) and respiratory exchange ratio(R = V’CO2/V’O2) were continuously measured using a breath-by-breath gas exchange measurementsystem (Quark, b2, Rome, Italy). The subjects were asked to wear a facemask that covered nose andmouth (dead space: 70 mL) equipped with a low-resistance, bidirectional digital turbine (28 mmdiameter) and with a sampling tube for the analysis of gas concentrations. Both turbine and gas sen-sors were calibrated according to the manufacturer’s recommendations before each testing session.Blood lactate concentration was determined using a capillary blood sample (20 lL) taken from theearlobe using a Biosen C-line lactacid meter (EKFdiagnostic GmbH, Magdeburg, Germany).

The net cost of locomotion (C = E’/v) was calculated based on the aerobic and anaerobic contribu-tions to total energy expenditure (E’) and on the speed of progression (v) and was expressed in J m�1 -kg�1. The aerobic contribution was determined by calculating the net (above resting values) averageV’O2 in the last 40 s of each stage and by considering the energy equivalent associated with the mea-sured respiratory exchange ratio (Peronnet & Massicotte, 1991). The anaerobic contribution was cal-culated with an energy equivalent of 3 mL kg�1 per mM of net lactate production (di Prampero &Ferretti, 1999).

Pole force was measured by a lightweight single-axial load cell (Deltatech, Sogliano al Rubicone,Italy) mounted inside standard poles (Diamond Storm 10 Max; OneWay, Helsinki, Finland) underthe handgrip. Analog signals from the force transducer were sampled at 100 Hz by means of a dataacquisition board (NI DAQ-PAD-6016, 16 bit; National Instruments, Austin, Texas, USA). Force trans-ducers were dynamically calibrated few minutes before each test with a load cell used as a reference(546QD; DSEurope, Milan, Italy), as described in a previous study (Pellegrini et al., 2011).

The motion of the poles and of the roller skis was recorded at 100 Hz by means of an optoelectronicmotion capture system (six cameras MCU240, ProReflex; Qualisys, Gothenburg, Sweden). Two reflec-tive markers were axially attached to each pole; the higher marker was placed 40 cm from the top ofthe pole and the lower marker was placed 60 cm above the tip of the pole; the line passing throughthese markers was taken as the pole direction. Two markers were placed on the upper part of the shaftof each roller ski, one 2 cm ahead of the binding and the other 2 cm behind the anterior wheel.

Data collection was triggered by a digital signal in order to ensure synchronization between theforce signals and the kinematic signals. Force and kinematic data were collected for both sides. Therewas a 40 s recording period for each condition in order to have at least 20 poling cycles for each stage.

For all techniques, the skiing cycle was defined as beginning at the start of the right pole groundcontact and ending at the subsequent right pole ground contact (Holmberg, Lindinger, Stoggl, Eitzlm-air, & Muller, 2005; Lindinger, Gopfert, Stoggl, Muller, & Holmberg, 2009; Pellegrini et al., 2011) and itsduration, cycle time (CT), was calculated. The pole contact and the pole take-off were identified in theforce data as the first point above and the first point below a force threshold of 10 N, respectively. Thisthreshold has been verified to be the minimum force value above the level of typical noise recordedduring the recovery phase (Pellegrini et al., 2011). Each cycle contains then the poling action of bothpoles for all techniques. For DS a skiing cycle contains the action of both legs, for DK a skiing cycle con-tains the action of the left leg or of the right leg. The duration of the poling action, poling time (PT),was calculated as the time between pole ground contact and pole take-off. Since the propulsive action

Please cite this article in press as: Pellegrini, B., et al. Biomechanical and energetic determinants of techniqueselection in classical cross-country skiing. Human Movement Science (2013), http://dx.doi.org/10.1016/j.humov.2013.07.010

6 B. Pellegrini et al. / Human Movement Science xxx (2013) xxx–xxx

of the leg through the ski can occur only when the ski stops with respect to the ground (Nilsson et al.,2004), in this study leg thrust time (LTT) was assumed as the time period during which the ski isstanding still (Bellizzi, King, Cushman, & Weyand, 1998). The speed of the roller skis along the direc-tion of progression was calculated and the ski stops were identified as the time in which ski speed re-mains below a threshold of 0.5 km/h. No ski stops were identified for DP techniques as no kick actionis performed in this technique. The average poling force over the entire cycle (PF) was calculated bydividing the integral of the force–time curves by CT. Data processing was performed using Matlab7.0 (MathWorks Inc., Natick, MA, USA) and Excel 2003 (Microsoft Corporation, Redmond, Washington,USA).

2.5. Statistical analysis

Descriptive statistics were expressed as means ± SD for all parameters over the group. All param-eters were analysed by means of a two-way (Technique � Speed or Technique � Slope) ANOVA for re-peated measures to check for the statistical significance of main effects and interactions. A post hocBonferroni correction test was applied to analyse differences in the observed parameters betweeneach pair of techniques. When the polynomial contrast of ANOVA was found to be significant, datawere fitted by means of linear or polynomial regression and R2 was calculated. When significant inter-actions were found, the rate of changes was compared between techniques by checking for differencesbetween the slopes’ coefficients of the linear or polynomial regression by means a one-way ANOVA forrepeated measures. The level of significance was set an a-value of <.05. All statistics were conductedby using SPSS 15.0 for Windows.

3. Results

3.1. Determination of preferred skiing techniques

The results of the survey on the preferred technique for different conditions of slope and speed arereported as the percentage of the subjects that selected each technique; these data are reported inFig. 2. When skiing at 0� and 1� slopes at 10 km/h (Fig. 2A) 100% of the subjects preferred the DP tech-nique. A transition to the DK technique was observed with increasing slope, with DK becoming themost preferred technique at slopes between 2� and 3�. With further increases in slope, a transitionfrom DK to DS occurred (at about 5�) and 100% of the subjects preferred DS at slopes steeper than 6�.

When skiing at 2�, DP was the preferred technique for the majority of the subjects at every testedspeed (Fig. 2B). DS was preferred by a small percentage of the subjects (25%) at the slowest speed(6 km/h), and none indicated DS as the preferred at faster speeds, DK was the technique preferredby 35–40% of subjects between 8 km/h and 12 km/h. At faster speeds, all subjects progressively pre-ferred DP over DK. DP was preferred by all skiers at speeds faster than 16 km/h.

3.2. Metabolic and biomechanical measurements

The conditions sustained by all skiers were: DS from 6 km/h to 14 km/h at 2� and from 0� to 6� at10 km/h; DP from 6 km/h to 16 km/h at 2� and from 0� to 4� at 10 km/h; DK from 6 km/h to 16 km/h at2� and from 0� to 4� at 10 km/h.

The cost of locomotion of skiing was found to be significantly dependent on the slope (p < .001),ranging from 2 Jkg�1 m�1 to 8 Jkg�1 m�1 (Fig. 3A), but to be independent of the speed (p = .074), withC = 4–5 Jkg�1 m�1 at all speeds (Fig. 3B). No effect of the technique was observed neither for incre-ments of slope (p = .298) nor increments of speed (p = .522). A significant Slope � Technique interac-tion was seen (p < .001) and polynomial contrast indicated the existence of a linear relationshipbetween each technique and the slope (R2 = .992, p < .001 for DS, R2 = .995, p < .001 for DK andR2 = .994, p < .001 for DP). The rate of increment of C across slopes was found to be different amongtechniques, DP having the higher rate of increment (C/slope = 1.00 ± 0.25 Jkg�1 m�1/�; p = .0018 vs.DS, p = .0024 vs. DK), and it being intermediate for DK (C/slope = 0.80 ± 0.19 J/(kg�1 m�1/�); p = .037

Please cite this article in press as: Pellegrini, B., et al. Biomechanical and energetic determinants of techniqueselection in classical cross-country skiing. Human Movement Science (2013), http://dx.doi.org/10.1016/j.humov.2013.07.010

0%

20%

40%

60%

80%

100%

0 1 2 3 4 5 6 7Slope [°]

0%

20%

40%

60%

80%

100%

6 8 10 12 14 16 18Speed [km/h]

Num

ber o

f sub

ject

s [%

]

A

B

Num

ber o

f sub

ject

s [%

]

DP DS

DK

DP DS

DK

Fig. 2. Representation for each technique of the percentage of the subjects that declared that technique was the preferred, fordifferent slope at a fixed speed of 10 km/h (panel A) and speed (panel B) conditions. The symbols refer to different techniques:DP technique: dotted area and solid line, DK technique: dashed area and long dashed line, DS technique: high density dottedarea and short dashed line.

B. Pellegrini et al. / Human Movement Science xxx (2013) xxx–xxx 7

vs. DS) and the lowest for DS (C/slope = 0.69 ± 0.13 Jkg�1 m�1/�). Post-hoc analysis demonstrated a sig-nificant difference in C between DS and DP at 0� (p = .005) and 1� (p = .016) and between DP and DK at1� (p = .003), with DP having a lower metabolic cost with respect to DS and DK. No significant speedeffect (p = .522) and no Speed � Technique was found (p = .147) for this parameter.

Poling time was found to be constant across the investigated slope range (p = .552) (Fig. 4A) and tohave a significant speed effect (p < .001) (Fig. 4B). A significant technique effect was found with incre-ments of slope (p < .001) and increment of speed (p < .001), with PT for DS being lower than PT for DP(p < .001) and DK (p = .005 and p < .001). Poling time was found to decrease with speed, with a secondorder polynomial function for all techniques (R2 = .994 for DS, R2 = .994 for DK and R2 = .997 for DP).

Leg thrust time was found to range from 0.10 to 0.35 s depending on technique, slope and speedconditions (Fig. 4A, B). A significant technique effect was found for both slope (p = .017) and speed(p = .008) increases, with LTT showing higher values for DK than for DS at 8 km/h, 10 km/h and12 km/h at 2� and at 10 km/h at 4�. A significant effect of slope (p < .001) and speed (p < .001) was ob-served. Thrust time increased linearly with slope for both DS (R2 = .984) and DK (R2 = .971). A signif-icant slope � technique effect was observed (p = .011) with LTT for DK increasing more rapidly thanLTT for DS (p = .0387). At increasing speed, LTT decreased with a second order polynomial functionfor DS (R2 = .998) and DK (R2 = .994). The rates of decrease were similar, and indeed no significantspeed x technique interaction was found (p = .175). The value of LTT at 14 km/h was about 36% ofthe value observed at 6 km/h.

Please cite this article in press as: Pellegrini, B., et al. Biomechanical and energetic determinants of techniqueselection in classical cross-country skiing. Human Movement Science (2013), http://dx.doi.org/10.1016/j.humov.2013.07.010

0

1

2

3

4

5

6

7

8

9

0 1 2 3 4 5 6 7Slope [°]

[ C [

J·kg

-1·m

-1 ]

#,§#

0

1

2

3

4

5

6

7

8

9

4 6 8 10 12 14 16 18Speed [km/h]

[ C [

J·kg

-1·m

-1 ]

DS

DKDP

A

B

DS

DKDP

Fig. 3. Cost of locomotion as a function of slope at a fixed speed of 10 km/h (panel A) and speed at a fixed slope of 2� (panel B)(mean and standard deviation). DS: filled rhombus and short dashed line, DP: empty triangle and solid line, DK: filled squareand long dashed line. Significant differences between techniques (p < .05) are indicated with the symbol # for DS vs. DP, § for DPvs. DK and £ for DS vs. DK.

8 B. Pellegrini et al. / Human Movement Science xxx (2013) xxx–xxx

The force exerted through the poles (PF) was significantly different among techniques for all slopes(Fig. 5A) (p < .001) and speeds (Fig. 5B) (p < .001); PF was larger for DP while for DK it was intermedi-ate between DP and DS. The values of PF showed a significant effect of the slope (p < .001) and a sig-nificant slope x technique interaction (p < .001). The increase of PF with the slope was found to belinear for all techniques, (R2 = .999, p < .001 for DP, R2 = .993, p < .001 for DK and R2 = .954, p < .001for DS). The slope of the linear regressions is significantly higher for DP (0.130 N/(kg �)) with respectto DK (0.058 N/(kg �)) and DS (0.026 N/(kg �)) (p < .001) and the slope for DK is significantly larger thanfor DS (p = .0011). No significant speed effect was seen on PF for DP and DK; on the contrary, PF for DSshowed a significant increase with the speed (p = .014) but the increment was not described by a lin-ear regression. Significant linear and curvilinear regressions that were found for PF and LTT are shownin Figs. 6 and 7 respectively and used as basis for the discussion.

4. Discussion

The aim of this study was to single out the factors that determine the selection of techniques inclassical cross-country skiing.

As far as technique selection is concerned, we observed that DP is the technique all athletes preferto use on the flat at a speed of 10 km/h. A transition between DP and DK occurred with increasingslope (between 2� and 3�) and a further transition from DK to DS can be observed at about 5�. DS isselected by all skiers at slopes steeper than 6�. On a ‘‘gentle slope’’ (2�) DS is used only at very low

Please cite this article in press as: Pellegrini, B., et al. Biomechanical and energetic determinants of techniqueselection in classical cross-country skiing. Human Movement Science (2013), http://dx.doi.org/10.1016/j.humov.2013.07.010

0.0

0.2

0.4

0.6

0.8

1.0

Tim

e fo

r pro

puls

ion

[s]

0 1 2 3 4 5 6 7Slope [°]

LTT

PT

#, £ #, £ #,£ £

£

0.0

0.2

0.4

0.6

0.8

1.0

4 6 8 10 12 14 16 18Speed [km/h]

Tim

e fo

r pro

puls

ion

[s]

LTT

PT

#, £#, £

#, £

#, £

££ £

A

B

DS

DKDP

DS

DKDP

Fig. 4. Time for propulsion is plotted as a function of slope at a fixed speed of 10 km/h (panel A) and speed at a fixed slope of 2�(panel B) Each graph is delimitated by a dotted line, the lower part of each panel, indicated with LTT reports data referring to legthrust time, the upper part, indicated with PT reports data referring to poling time. DS: filled rhombus and short dashed line,DP: empty triangle and solid line; DK: filled square and long dashed line. Significant differences between techniques (p < .05)are indicated with the symbol # for DS vs. DP and £ for DS vs. DK.

B. Pellegrini et al. / Human Movement Science xxx (2013) xxx–xxx 9

speeds. At faster speeds some of the athletes preferred DK or DP. DP was preferred by all skiers atspeeds faster than 16 km/h.

Although, to our knowledge, no studies on the selection of cross-country skiing techniques are re-ported in the literature, an indirect comparison can be attempted using the techniques and conditionsobserved in other investigations. In a study performed on snow during a race by Bilodeau and col-leagues (1996) they reported that all skiers used DS on a section of the track that was 7� uphill (atspeeds ranging from 9 km/h 14 km/h) whereas on a flat section of the same race all skiers used DPand reached speeds of about 22 km/h. To our knowledge, in the only study in which the protocol re-quired the subjects to use the technique they preferred (Cignetti et al., 2009), DS has been used by allskiers above a slope of 6–7�, DP was almost never used above a slope of 4� and DK was mostly used ona slope of 4�. Our results for the preferred technique at different slopes are thus in good agreementwith the selection of techniques reported in these studies.

The main differences between the three techniques as far as the measured parameters were: i) thatthe force exerted trough the poles is different between the techniques: highest for DP, intermediate forDK and lowest for DS; and ii) that the duration of the propulsive phase through the poles is lower forDS than for DK and DP. Despite the difference in the skiing pattern and in the distribution of propul-sive forces, we found iii) only slight differences in the cost of transport between the techniques (DPwas found to be more economical than DS on flat and than DS and DK at 1�). Our data show, however,

Please cite this article in press as: Pellegrini, B., et al. Biomechanical and energetic determinants of techniqueselection in classical cross-country skiing. Human Movement Science (2013), http://dx.doi.org/10.1016/j.humov.2013.07.010

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

4 6 8 10 12 14 16 18Speed [km/h]

PF [N

/kg]

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

0 1 2 3 4 5 6 7Slope [°]

PF [N

/kg]

#,§,£ #,§,£ #,§,£ #,§,£ #,§,£

#,§,£ #,§,£ #,§,£ #,§,£ #,§,£

A

B

DS

DKDP

DS

DKDP

Fig. 5. Poling force (PF) is plotted as a function of slope (at a fixed speed of 10 km/h) (panel A), and speed at a fixed slope of 2�(panel B). DS: filled rhombus and short dashed line DP: empty triangle and solid line, DK: filled square and long dashed line,(Mean and standard deviation). Significant differences between techniques (p < .05) are indicated with the symbol # for DS vs.DP, § for DP vs. DK and £ for DS vs. DK.

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

0 1 2 3 4 5 6 7

PF

[N/k

g]

DS

DKDP

DP DK DS

Slope [°]

Fig. 6. Diagram illustrating the hypothetical scheme of transitions under the hypothesis of a limit threshold in poling forceexertion. The linear increase of poling force as a function of the slope as determined by data fitting is represented for eachtechnique. DP: solid line, DK: long dashed line, DS: short dashed line. The hypothetical limit of poling force is represented by thehorizontal dotted line (corresponding to 5.5 N/kg). The values correspond to the techniques selected at different slope accordingto the hypothesis of limit of force are indicated by a grey wide line and technique transitions according to the hypothesis areindicated with vertical dash-dotted lines.

10 B. Pellegrini et al. / Human Movement Science xxx (2013) xxx–xxx

Please cite this article in press as: Pellegrini, B., et al. Biomechanical and energetic determinants of techniqueselection in classical cross-country skiing. Human Movement Science (2013), http://dx.doi.org/10.1016/j.humov.2013.07.010

DK

4 6 8 10 12 14 16 18Speed [km/h]

0 1 2 3 4 5 6 7Slope [ °]

DSDK

A

B

0.0

0.1

0.2

0.3

0.4

0.5

LTT

[s]

optimal

not optimal but possible

not possible

DSDK

0.0

0.1

0.2

0.3

0.4

0.5

LTT

[s]

optimal

not optimal but possible

not possible

DP DK DS

DS DP

Fig. 7. Diagram illustrating the hypothetical scheme of transitions under the hypothesis of values of possible and optimalleg thrust time. The linear trend of LTT as a function of slope (panel A) and curvilinear trend of LTT as a function of speed(panel B) are represented for each technique; DK: long dashed line, DS: short dashed line. Hypothetical limit of optimal LTTis represented by the upper dotted horizontal line and the limit of possible LTT is represented by the lower dottedhorizontal line in each panel. The values corresponding to techniques selected according to the hypothesis of the limit ofoptimal LTT are indicated with grey wide lines and technique transitions according to the hypothesis are indicated withvertical dash-dotted lines.

B. Pellegrini et al. / Human Movement Science xxx (2013) xxx–xxx 11

that the cost of locomotion increases with the slope with a higher rate of increment for the DP tech-nique than for DS and DK.

To identify the possible determinants of transition between techniques, we adopted the four crite-ria introduced by Hreljac (1995). The first criterion indicates that the ‘‘potential variable’’ must exhibitan abrupt change in its value at the transition point. The second criterion indicates that the value ofthe variable after transition must return to a level that is characteristic of the previous gait. The thirdcriterion indicates that the variable must be capable of producing an effect on proprioceptive sensors.The fourth criterion indicates that the value of the variable at transition must be maintained evenwhen the locomotory conditions change.

The variable among the physiological and biomechanical parameters we have investigated thatshows the greatest variation among techniques is the force exerted through the poles. We thushypothesize that this parameter has a role in triggering the transition from one technique to another.The transition that occurs when slope increases can be tentatively explained with the hypothesis thatthere is a limit in the force a skier would like to exert. As shown by Fig. 6, if we hypothetically set thelimit of force exertion at 0.55 N/kg, a skier would utilize DP on moderate slopes (up to 2�) and then,when this force threshold is exceeded due to an increase of slope (between 2� to 3�), he wouldswitch to DK; when the force threshold is exceeded again (above 5�) he would then switch to DS. This

Please cite this article in press as: Pellegrini, B., et al. Biomechanical and energetic determinants of techniqueselection in classical cross-country skiing. Human Movement Science (2013), http://dx.doi.org/10.1016/j.humov.2013.07.010

12 B. Pellegrini et al. / Human Movement Science xxx (2013) xxx–xxx

hypothesis is compatible with what we actually observed: a switch from DP to DK indeed occurs in-deed between 2� and 3�, while the transition between DK and DS occurs between 5� and 6�. If we applythe hypothesis of a threshold limit to the changes in speed, we find that this does not explain the pref-erence for DP over DK and DS at high speeds. However, since the poling force observed when skiingwith the different techniques does not change with speed and never exceeds the value of the arbi-trarily defined threshold here (0.55 N/kg), this does not rule out the hypothesis that the poling forcethreshold could act as trigger: at increasing speeds the trigger could be present, but not activated.

The poling force limit we have hypothesized meets some of the criterion set by Hreljac (1995) for avariable that acts as a trigger; pole force indeed exhibits an abrupt change when switching betweentechniques and it returns to the values of the previous gait. As can be seen in Fig. 6A, switching fromDP at 2� to DK at 3� leads the value of PF to switch to a value that is close to the value of PF at 1�;switching from DK at 5� to DS at 6� would lead the PF values to decrease again at a value similar tothat assumed during the previous gait. Poling force meets, moreover, the criterion of producing an ef-fect on proprioceptive sensors, as the level of muscular tension and force exertion can be detected byspecific mechanoreceptors in tendons and muscles (Carson, Riek, & Shahbazpour, 2002).

We have thus observed that the transitions between techniques when skiing on an increasing slopecan be explained by the limit of force. However, it was not observed that this also occurs with increas-ing speed, though we can see that this rule is not contradicted by what we found for increments ofspeed, where the hypothesized force threshold was never exceeded in any technique. Neither wasthe existence of a trigger of force confirmed nor rejected for the different speed conditions we tested.

The hypothesis that gait transition could occur on the basis of a force trigger is controversial in hu-man locomotion (Fewster & Smith, 1996; Hreljac, 1993), but it is largely confirmed in both bipedal andquadrupedal animals that chose to switch from trotting to galloping when musculoskeletal forcesreach a critical level during trotting (Biewener & Taylor, 1986; Farley & Taylor, 1991), and that tran-sitions occur to reduce stress applied to the musculoskeletal system. In a study conducted on goats itwas found that these animals switch from a trot to a gallop at a given value of peak bone strain that issimilar to that found for dogs and horses, despite the great differences in the speed of the transition(Biewener & Taylor, 1986). On the basis of this evidence it has been hypothesized that transitions oc-cur to reduce the stress applied to the musculoskeletal system. Similarly, other researchers have sug-gested that human gait changes occur to prevent overexertion in the relatively small dorsiflexormuscle group which may be working close to its maximum when walking at speeds near the transi-tion to running (Hreljac, 1995). Moreover, Bartlett and Kram (2008) demonstrated the role of muscleexertion in the transition from walking to running by showing that altering the demands on specificmuscles can change the preferred transition speed. In this study we are investigating for the first time,transition between gaits based on arm muscle propulsion and therefore based on propulsive actionperformed using small sized muscles with a low oxidative capacity (Åstrand, Rodahl, Dahl, & Strømme,2003). We could thus hypothesize that a force limit is likely to be reached.

The observation that poling force did not play a significant role in the selection of techniques indifferent speed conditions does not explained why the skiers prefer to use DP at the fastest testedspeeds. Thus, other triggers need to be introduced to ‘‘our model’’ to explain the technique selection.One parameter that, among the others investigated in our study, could be a possible determinant ofthe transition between techniques is the time of application of the force exerted by the legs: the leg’sthrust time (LTT). In our study, leg thrust time was found to decrease with increasing speed and toincrease with increasing slope for both DS and DK; moreover, at speeds of 14 km/h for DS and16 km/h for DK (which correspond to the highest speeds our skiers were able to sustain continuouslyfor 4 min) the leg thrust time reached a minimum value of 0.11 s, e.g. similar to that found in sprintrunning. It therefore seems that, as with running, leg thrust time could be a parameter that limits themaximal attainable speed in DS and DK. It is known from studies of sprint running that foot contacttime decreases with increasing speed, but never falls below a limit of about 0.11 s (Weyand, Sandell,Prime, & Bundle, 2010).

Experimental evidence suggests that a time period of 0.10 s is too short to allow the forces pro-duced by the limb muscles to reach their maximum and that the time required to generate and trans-mit the peak isometric force is more than twice the contact time (Harridge et al., 1996). On the basis ofthese considerations we can hypothesize that there is a thrust time below which the thrust action is

Please cite this article in press as: Pellegrini, B., et al. Biomechanical and energetic determinants of techniqueselection in classical cross-country skiing. Human Movement Science (2013), http://dx.doi.org/10.1016/j.humov.2013.07.010

B. Pellegrini et al. / Human Movement Science xxx (2013) xxx–xxx 13

still possible, but which is not as effective as that exerted with longer thrust times. If we arbitrarilyhypothesize that this value is about 0.22 s, i.e. twice the minimum thrust time limit (0.11 s) and ifwe then note these two thresholds on the graph of leg thrust times (Fig. 7, dotted lines), we cansee that the selection of DS and DK as preferred techniques occurs at approximately the range ofspeeds and slopes for which leg thrust action has been hypothesized to be optimal.

The rationale of preferring a gait that allows a longer time for force exertion is supported by theobservation that a reduction in this time leads to an increase in the cost of muscle force generation,as found for running (Kram & Taylor, 1990) and confirmed for cross-country skiing (Bellizzi et al.,1998). The ability to generate force indeed depends on muscle length, as well as the speed at whichit shortens. In particular, high contraction speeds are associated with low force production (Harridgeet al., 1996; Hill, 1938). Moreover, the existence of a stretch shortening cycle mechanism in the muscleinvolved in the kick movement of DS has been recognized (Komi & Norman, 1987) and this mechanismwould be optimized by proper motion timing.

The trigger criteria fixed by Hreljac (1995) are mainly fulfilled by leg thrust time. We can indeedhypothesize that the duration of force exertion (or the rate of force development) can be perceivedby the skier (Gruber & Gollhofer, 2004). Moreover, switching from DS or DK to DP involves an abruptchange in this parameter since in DP no leg thrust action occurs. Transitions from DK to DS at increas-ing slope and from DS to DK at increasing speed, however, cause only slight changes in leg thrust time.Moreover, the criterion of the independence of the value of the trigger parameter is not completelymet, since transitions for speed increment occurs in correspondence of values of thrust time thatare larger than those found in correspondence with transitions that occurs with slope increment.

Taken together, the hypothesis of a limit in the poling force and in the time for leg thrust canrespectively explain why DP is abandoned in favour of DS on steeper slopes and why DS and DKare abandoned for DP at high speeds, only leaving the reason for the preference for DP for flat skiingunexplained.

This preference could be justified by the criterion of economy of locomotion, as we found that forskiing on flat or shallow slopes, the cost of locomotion is different between techniques and that DP isthe more economical technique. The cost of locomotion increases at a higher rate for DP than for DSand DK, so that no differences between techniques are detectable at slopes steeper than 3�. Moreover,we have observed the existence of crossing points between the lines that describe the increments ofcost in the different techniques; these can be called, in agreement with other studies that have inves-tigated gait transitions, ‘‘energetic optimal transition slopes’’ (Hoyt, Wickler, Dutto, Catterfeld, & John-sen, 2006). According to our data we can thus estimate that the cost of skiing in DS would becomelower than the cost for DP and DK on slopes steeper than 4�.

The higher rate of increment in energy cost with slope for DP (only arm propulsion) compared tothe techniques in which propulsion is also produced by the legs (DS and DK) can be attributed to thedifferent economy in arm and leg work in skiing. Bellizzi and co-workers (Bellizzi et al., 1998) ob-served that, arms-only skiing requires 40% less energy than legs-only skiing on the flat. They proposedthat this is due to the fact that body weight is supported ‘‘passively’’ by maintaining straight legs,while the arms generate propulsive forces that are low compared to those produced by the legs,but which have a longer time of application. When skiing uphill, the need to increase the force exertedthrough the poles for DP greatly increases (in DP the poling force value at 4� is 3 times larger than thatfor the flat) while for DK and DS the poling force at this slope reaches smaller values (2 and 1.5 timeslarger than that exerted on the flat, respectively). On the other hand, it is possible to hypothesize thatleg force requirements underwent little change with increasing slope because the additional propul-sive force required is a small fraction of the forces generated to support the weight of the body (Bel-lizzi et al., 1998). The arms then incur a greater additional cost when an athlete is skiing uphill and,consequently, arm skiing is no more economical than combined arm-and-leg skiing on an incline.

The preference of DP for skiing on the level is consistent with the lower cost of this technique com-pared to the others; however, this parameter cannot be considered a valid determinant of techniqueselection. Indeed no differences in cost between techniques were found for skiing at a speed of 14 km/h on a gentle slope, where DP is definitely preferred. Since no differences in the cost of locomotionwere observed to correspond to the gait change from one technique to another, this parameter doesnot meet the first and second criteria fixed by Hreljac (1995). Furthermore, according to other

Please cite this article in press as: Pellegrini, B., et al. Biomechanical and energetic determinants of techniqueselection in classical cross-country skiing. Human Movement Science (2013), http://dx.doi.org/10.1016/j.humov.2013.07.010

14 B. Pellegrini et al. / Human Movement Science xxx (2013) xxx–xxx

researchers (Farley & Taylor, 1991), the capacity of humans to directly perceive the metabolic cost ofthe activity performed is rather low, and thus metabolic cost does not comply with the third criterion.It is therefore unlikely that the selection of skiing gaits would be determined by metabolic reasons, aswas hypothesized and observed in early studies of human walking and running (Margaria, 1938) andin studies of animal gaits (Dawson & Taylor, 1973; Hoyt & Taylor, 1981; Wickler, Hoyt, Cogger, &Myers, 2003). Hanna and coworkers (Hanna, Abernety, Neal, & Burgess-Limerick, 2000) suggested thatthe energetic optimization observed in some investigations may in fact be a consequence of the opti-mization of other parameters, rather than an indication that oxygen consumption is the direct deter-minant of transition. In addition, a study designed to ascertain the influence of attentional processeson the transition have demonstrated that cognitive processes play an important role in mediating therelative significance of one or more contributing physical factors (Daniels & Newell, 2003). It is thenpossible to hypothesize that gait transitions are not triggered by just one factor but are possibly theresult of an interaction of different psychophysiological stimuli.

It is thus difficult to draw a single conclusion about the determinants of technique selection incross-country skiing and, on the basis of our results, it is not possible to single out a unique parameterthat acts as a determinate for technique selection in classical cross-country skiing. We can, however,suggest that transitions in cross-country skiing are the result of a pool of determinants and are deter-mined by the activation of the most ‘‘sensible’’ trigger/variable. We can suggest that limits in the forceexerted through the poles and in the duration of leg thrust time are the most important determinantsof the preferred technique selection and that skiing economy could explain the residual situations.

Acknowledgements

The authors gratefully acknowledge the athletes who volunteered to participate in thisinvestigation.

References

Alexander, R. M. (1989). Optimization and gaits in the locomotion of vertebrates. Physiological Reviews, 69, 1199–1227.Åstrand, P. O., Rodahl, K., Dahl, H. A., & Strømme, S. R. (2003). Textbook of work physiology: Physiological basis of exercise (4th ed.).

Champaign, Illinois: Human Kinetics.Bartlett, J. L., & Kram, R. (2008). Changing the demand on specific muscle groups affects the walk-run transition speed. The

Journal of Experimental Biology, 211, 1281–1288.Bellizzi, M. J., King, K. A., Cushman, S. K., & Weyand, P. G. (1998). Does the application of ground force set the energetic cost of

cross-country skiing? Journal of Applied Physiology, 85, 1736–1743.Biewener, A. A., & Taylor, C. R. (1986). Bone strain: A determinant of gait and speed? The Journal of Experimental Biology, 123,

383–400.Biewener, A. A., Thomason, J., Goodship, A., & Lanyon, L. E. (1983). Bone stress in the horse forelimb during locomotion at

different gaits: A comparison of two experimental methods. Journal of Biomechanics, 16, 565–576.Bilodeau, B., Rundell, K. W., Roy, B., & Boulay, M. R. (1996). Kinematics of cross-country ski racing. Medicine and Science in Sports

and Exercise, 28, 128–138.Carson, R. G., Riek, S., & Shahbazpour, N. (2002). Central and peripheral mediation of human force sensation following eccentric

or concentric contractions. The Journal of Physiology, 539, 913–925.Cignetti, F., Schena, F., Zanone, P. G., & Rouard, A. (2009). Dynamics of coordination in cross-country skiing. Human Movement

Science, 28, 204–217.Daniels, G. L., & Newell, K. M. (2003). Attentional focus influences the walk-run transition in human locomotion. Biological

Psychology, 63, 163–178.Dawson, T. J., & Taylor, C. R. (1973). Energetic cost of locomotion in kangaroos. Nature, 246, 313–314.di Prampero, P. E., & Ferretti, G. (1999). The energetics of anaerobic muscle metabolism: A reappraisal of older and recent

concepts. Respiration Physiology, 118, 103–115.Diedrich, F. J., & Warren, W. H. (1998). The dynamics of gait transitions: Effects of grade and load. Journal of Motor Behavior, 30,

60–78.Farley, C. T., & Taylor, C. R. (1991). A mechanical trigger for the trot-gallop transition in horses. Science, 253, 306–308.Fewster, J. B., & Smith, G. A. (1996). The influence of midstance vertical ground reaction force on the human walk-run transition.

Medicine and Science in Sports and Exercise, 28, 87.Gruber, M., & Gollhofer, A. (2004). Impact of sensorimotor training on the rate of force development and neural activation.

European Journal of Applied Physiology, 92, 98–105.Hanna, A. J., Abernety, A. B., Neal, R. J., & Burgess-Limerick, R. J. (2000). Triggers for the transition between human walking and

running. In W. A. Sparrow (Ed.), Energetics of human activity (pp. 124–164). Human Kinetics: Champaign Ill.Harridge, S. D., Bottinelli, R., Canepari, M., Pellegrino, M. A., Reggiani, C., Esbjornsson, M., et al (1996). Whole-muscle and single-

fibre contractile properties and myosin heavy chain isoforms in humans. Pflugers Archives, 432, 913–920.

Please cite this article in press as: Pellegrini, B., et al. Biomechanical and energetic determinants of techniqueselection in classical cross-country skiing. Human Movement Science (2013), http://dx.doi.org/10.1016/j.humov.2013.07.010

B. Pellegrini et al. / Human Movement Science xxx (2013) xxx–xxx 15

Hill, A. V. (1938). The heat of shortening and the dynamic constants of muscle. Proceedings of the Royal Society of London Series B,Biological Sciences, 126, 136–195.

Holmberg, H. C., Lindinger, S., Stoggl, T., Bjorklund, G., & Muller, E. (2006). Contribution of the legs to double-poling performancein elite cross-country skiers. Medicine and Science in Sports and Exercise, 38, 1853–1860.

Holmberg, H. C., Lindinger, S., Stoggl, T., Eitzlmair, E., & Muller, E. (2005). Biomechanical analysis of double poling in elite cross-country skiers. Medicine and Science in Sports and Exercise, 37, 807–818.

Hoyt, D. F., & Taylor, C. R. (1981). Gait and energetics of locomotion in horses. Nature, 291, 239–240.Hoyt, D. F., Wickler, S. J., Dutto, D. J., Catterfeld, G. E., & Johnsen, D. (2006). What are the relations between mechanics, gait

parameters, and energetics in terrestrial locomotion? Journal of Experimental Zoology Part A: Comparative ExperimentalBiology, 305, 912–922.

Hreljac, A. (1993). Preferred and energetically optimal gait transition speeds in human locomotion. Medicine and Science inSports and Exercise, 25, 1158–1162.

Hreljac, A. (1995). Determinants of the gait transition speed during human locomotion: kinematic factors. Journal ofBiomechanics, 28, 669–677.

Hreljac, A., Arata, A., Ferber, R., Mercer, J. A., & Row, B. R. (2001). An electromyographical analysis of the role of the dorsiflexor onthe gait transition during human locomotion. Journal of Applied Biomechanics, 17, 287–296.

Komi, P. V., & Norman, R. W. (1987). Preloading of the thrust phase in cross-country skiing. International Journal of SportsMedicine, 8(Suppl 1), 48–54.

Kram, R., & Taylor, C. R. (1990). Energetics of running: a new perspective. Nature, 346, 265–267.Lindinger, S. J., Gopfert, C., Stoggl, T., Muller, E., & Holmberg, H. C. (2009). Biomechanical pole and leg characteristics during

uphill diagonal roller skiing. Sports Biomechanics, 8, 318–333.Malcolm, P., Segers, V., van Caekenberghe, I., & de Clercq, D. (2009). Experimental study of the influence of the m. tibialis

anterior on the walk-to-run transition by means of a powered ankle-foot exoskeleton. Gait & Posture, 29, 6–10.Margaria, R. (1938). Sulla fisiologia, e specialmente sul consumo energetico, della marcia e della corsa a varie velocità ed

inclinazioni del terreno. Atti Accademia Nazionale dei Lincei Classe Scienze Fisiche Matematiche e Natural Serie VI, 7, 299–368.Mercier, J., Le Gallais, D., Durand, M., Goudal, C., Micallef, J. P., & Prefaut, C. (1994). Energy expenditure and cardiorespiratory

responses at the transition between walking and running. European Journal of Applied Physiology and OccupationalPhysiology, 69, 525–529.

Minetti, A. E. (2004). Passive tools for enhancing muscle-driven motion and locomotion. The Journal of Experimental Biology, 207,1265–1272.

Minetti, A. E., Ardigo, L. P., & Saibene, F. (1994). The transition between walking and running in humans: Metabolic andmechanical aspects at different gradients. Acta Physiologica Scandinavica, 150, 315–323.

Nilsson, J., Tveit, P., & Eikrehagen, O. (2004). Effects of speed on temporal patterns in classical style and freestyle cross-countryskiing. Sports Biomechanics, 3, 85–107.

Pellegrini, B., Bortolan, L., & Schena, F. (2011). Poling force analysis in diagonal stride at different grades in cross country skiers.Scandinavian Journal of Medicine & Science in Sports, 21, 589–597.

Peronnet, F., & Massicotte, D. (1991). Table of nonprotein respiratory quotient: An update. Canadian Journal of Sport Sciences, 16,23–29.

Prilutsky, B. I., & Gregor, R. J. (2001). Swing- and support-related muscle actions differentially trigger human walk-run and run-walk transitions. The Journal of Experimental Biology, 204, 2277–2287.

Raynor, A. J., Yi, C. J., Abernethy, B., & Jong, Q. J. (2002). Are transitions in human gait determined by mechanical, kinetic orenergetic factors? Human Movement Science, 21, 785–805.

Rotstein, A., Inbar, O., Berginsky, T., & Meckel, Y. (2005). Preferred transition speed between walking and running: Effects oftraining status. Medicine and Science in Sports and Exercise, 37, 1864–1870.

Rubenson, J., Heliams, D. B., Lloyd, D. G., & Fournier, P. A. (2004). Gait selection in the ostrich: Mechanical and metaboliccharacteristics of walking and running with and without an aerial phase. Proceedings of the Royal Society B, 271, 1091–1099.

Saibene, F., Cortili, G., Roi, G., & Colombini, A. (1989). The energy cost of level cross-country skiing and the effect of the friction ofthe ski. European Journal of Applied Physiology and Occupational Physiology, 58, 791–795.

Saibene, F., & Minetti, A. E. (2003). Biomechanical and physiological aspects of legged locomotion in humans. European Journal ofApplied Physiology, 88, 297–316.

Segers, V., Lenoir, M., Aerts, P., & de Clercq, D. (2007). Influence of M. tibialis anterior fatigue on the walk-to-run and run-to-walktransition in non-steady state locomotion. Gait & Posture, 25, 639–647.

Smith, G. A. (1992). Biomechanical analysis of cross-country skiing techniques. Medicine and Science in Sports and Exercise, 24,1015–1022.

Taylor, C. R. (1985). Force development during sustained locomotion: A determinant of gait, speed and metabolic power. TheJournal of Experimental Biology, 115, 253–262.

Thorstensson, A., & Roberthson, H. (1987). Adaptations to changing speed in human locomotion: Speed of transition betweenwalking and running. Acta Physiologica Scandinavica, 131, 211–214.

Weyand, P. G., Sandell, R. F., Prime, D. N., & Bundle, M. W. (2010). The biological limits to running speed are imposed from theground up. Journal of Applied Physiology, 108, 950–961.

Wickler, S. J., Hoyt, D. F., Cogger, E. A., & Myers, G. (2003). The energetics of the trot-gallop transition. The Journal of ExperimentalBiology, 206, 1557–1564.

Please cite this article in press as: Pellegrini, B., et al. Biomechanical and energetic determinants of techniqueselection in classical cross-country skiing. Human Movement Science (2013), http://dx.doi.org/10.1016/j.humov.2013.07.010