Summary

Reasons for performing study: Dressage involves training of thehorse with the head and neck placed in a position defined by therider. The best position for dressage training is currently underdebate among riders and trainers, but there are few scientificdata available to confirm or disprove the different views.

Objective: To evaluate the kinematic effects of different head andneck positions (HNPs) in elite dressage horses ridden at trot.

Methods: Seven high-level dressage horses were subjected tokinetic and kinematic measurements when ridden on atreadmill with the head and neck in 5 different positions.

Results: Compared to free trot on loose reins the HNP desiredfor collected trot at dressage competitions increased T6 verticalexcursion, increased sacral flexion and decreased limbretraction after lift-off. Further increasing head or head andneck flexion caused few additional changes while an extremelyelevated neck position increased hindlimb flexion and lumbarback extension during stance, increased hindlimb flexionduring swing and further increased trunk vertical excursion.

Conclusions: The movements of the horse are significantlydifferent when ridden on loose reins compared to the positionused in collected trot. The exact degree of neck flexion is,however, not consistently correlated to the movements of thehorse’s limbs and trunk at collected trot. An extremelyelevated neck position can produce some effects commonlyassociated with increased degree of collection, but theincreased back extension observed with this position mayplace the horse at risk of injury if ridden in this position fora prolonged period.

Potential relevance: Head and neck positions influencesignificantly the kinematics of the ridden horse. It isimportant for riders and trainers to be aware of these effectsin dressage training.

Introduction

Throughout history the ridden horse has often been depicted witha raised and arched neck, but it is unclear what purpose this

274 EQUINE VETERINARY JOURNALEquine vet. J. (2009) 41 (3) 274-279

doi: 10.2746/042516409X394436

The effect of different head and neck positions on thecaudal back and hindlimb kinematics in the elite dressagehorse at trotM. RHODIN*, C. B. GÓMEZ ÁLVAREZ§#, A. BYSTRÖM†, C. JOHNSTON‡, P. R. VAN WEEREN#, L. ROEPSTORFF† and M. A. WEISHAUPT¶

Department of Clinical Sciences; †Department of Anatomy, Physiology and Biochemistry; and ‡University Hospital of the Swedish Universityof Agricultural Sciences, SE-750 07 Uppsala, Sweden; §Equine Division, Veterinary Medicine, Católica de Temuco University, Chile;#Department of Equine Sciences, Faculty of Veterinary Medicine, Utrecht University, Yalelaan 114, 3584 CM Utrecht, The Netherlands; and¶Equine Hospital, University of Zurich, Switzerland.

Keywords: horse; equestrian dressage; kinematics; rider; head and neck position

position served. Was it used to control the temper of the horse, toinfluence the movements or simply to make the horse look moreimpressive to the observer? While equestrian literature suggestsmany possible answers to these questions, the aims and effects ofdifferent head and neck positions used for dressage are still underdebate among riders and trainers.

In dressage the horse is asked to place the head and neck in aposition defined by the rider, rather than voluntarily chosen by thehorse. The head and neck positions (HNPs) may significantlyinfluence the movements and limb loading and this could be afactor influencing the orthopaedic health of the horse. An extremelylow and hyperflexed HNP is believed to improve the elasticity ofthe horse and reinforce back function (Janssen 2003). This trainingtechnique has been vehemently criticised as being cruel to the horsedue to supposed excessive tension engendered to the structures ofthe neck and back (Balkenhol et al. 2003; Heuschmann 2006).These critics advocate the classical principles of dressage wherecorrect, long-term training eventually should result in the horseacquiring the HNP defined by Fédération Equestre Internationalefor high-level dressage: “the neck raised, the poll high and thebridge of the nose slightly in front of the vertical” (Anon 2006).Heuschmann (2006) states that in this position the horse is then ableto carry its back freely without tension, and that a high neckposition imposed on the horse through the reins, rather thanachieved gradually through training, causes the horse to drop itsback and becoming tense and unwilling.

Riding equipment and training methods can significantly affectthe movements, limb loading and temporal gait characteristics ofthe sport horse (Biau et al. 2002; Roepstorff et al. 2002; Rhodin et al. 2005; Gomez Alvarez et al. 2006; Weishaupt et al. 2006).The relationships between HNP and the movements and wellbeingof the horse claimed by riders and trainers are founded onassumptions or personal experience. There are very few scientificdata available on how different HNPs influence the movement ofthe ridden horse to confirm or disprove the different views.

The purpose of this study was to quantify the effects ofdifferent head and neck positions on the kinematics of the caudalback and hindlimbs in high level dressage horses trotting on a treadmill.

*Author to whom correspondence should be addressed.[Paper received for publication 11.06.08; Accepted 29.10.08]

EVJ 08-236 Rhodin :Layout 1 11/02/2009 14:30 Page 2

M. Rhodin et al. 275

Material and methods

Horses

Seven high-level dressage horses were included in the study. Thehorses were 6 geldings and one stallion, age mean ± s.d. 14 ± 4.3 years, height at withers 1.7 ± 0.07 m and body mass 609 ± 62.3 kg. The horses were examined by an experienced clinicianbefore the study and found to be sound and have normal reactionsto palpation. The horses were fully accustomed to treadmilllocomotion with and without a rider (Buchner et al. 1994) andwere ridden by their own rider wearing their own fitted saddle andbridle with a normal snaffle bit.

Experimental set-up

The horses were measured on a treadmill1 with an integrated forcemeasuring system (Weishaupt et al. 2002) and vertical groundreaction forces were recorded in all limbs simultaneously.Spherical infrared-light reflective markers with a diameter of 19 mm (ProReflex)2 were placed at numerous locations on thehorse. Markers used in the current analysis were placed at thecranial end of the wing of atlas, the spinal processes of T6, L3, L5,S3 and S5, over the joints of the limbs and on the lateral wall ofleft fore and hind hooves.

Twelve ProReflex infrared cameras2 were positioned around andabove the treadmill to create a field of view in which all markerscould be detected by at least 2 cameras. Data were collected for

12 or 15 s at trot at a sampling rate of 240 Hz for 3 horses and theother 4 horses were measured at 140 Hz (due to technical problems).

Head-neck positions

Horses were measured ridden at trot with the head and neck in 5 HNPs (Fig 1): HNP1: Free or natural; voluntarily acquiredposition, unrestrained with loose reins. HNP2: Neck raised, pollhigh and bridge of the nose slightly in front of the vertical;reference position. HNP3: Neck raised, poll high and bridge of thenose slightly behind the vertical. HNP4: Neck lowered and flexed,bridge of the nose considerably behind the vertical. HNP5: Neckextremely elevated and bridge of the nose considerably in front ofthe vertical. HNP4, was achieved with a combination of ordinaryreins and draw-reins and all other positions with ordinary reins.The correctness of each HNP was judged by an internationaldressage judge. The speed was adapted individually to achieveoptimum performance in each HNP. Reference recordings weremade in HNP2 at intervals of 0.2 m/s to have a speed-matchedcontrol for each of the other HNPs.

Kinematic analysis

Dedicated camera software (Q-Track)2 was used to record data andto reconstruct the 3D movement of each marker. The raw x-, y- andz-coordinates were exported to Matlab3 where the followingvariables were calculated: the angle of the neck with respect to thehorizontal plane; the vertical positions of T6 and L5, the angles ofthe lumbar back (L3–L5), sacrum (S3–S5) and femur with respectto the horizontal plane; the stifle, tarsal and hind fetlock jointangles; and the pro- and retraction angles of the left fore- and hindhooves in relation to the elbow joint and L5, respectively. Theelbow joint was selected as reference for forelimb pro- andretraction rather than T6 to avoid influence from longitudinal skindisplacement of the T6 marker between HNPs.

All variables were normalised to 101 points, 0–100% of thestride defined by first contact of either the left fore- or the lefthindlimb using temporal information from the force measuringsystem. Pro- and retraction angles were normalised to 0–100% offore- or hindlimb stance. From the normalised curves stride rangeof motion (ROM) and stride mean was determined for the neckangle, stride ROM and minimum and maximum for T6 and L5vertical positions and stride ROM and stride and stance minimumand maximum for fore- and hindlimb pro- and retraction angles.The resulting values as well as the stride curves were averagedover a minimum of 12 strides for each horse and measurement.

Statistical method

The stride normalised mean curves, ROMs, stride mean, minimumand maximum values listed above were compared pair-wisebetween a speed-matched measurement in the reference position(HNP2) and each of the other HNPs using a paired nonparametrictest (Wilcoxon signed rank). Differences were considered to besignificant at P<0.05 for discrete values and at P<0.05 for aminimum of 5 consecutive data points for the curves.

Ethical review

The experimental protocol was approved by the Animal Health andWelfare Commission of the canton of Zürich.Fig 1: Head-neck positions. Illustration: Matthias Haab.

HNP1 HNP2

HNP3 HNP4

HNP5

EVJ 08-236 Rhodin :Layout 1 11/02/2009 14:30 Page 3

276 The effect of different head and neck positions on the caudal back

Results

Stride normalised group mean curves for the lumbar back, pelvis,femur, stifle, tarsal and hindlimb fetlock joint are displayed in Figure2 and discrete values for neck angle, T6 and L5 vertical height andfore- and hindlimb pro- and retraction angles are in Table 1.

The neck angle was significantly different between the HNPsstudied and ROM decreased in HNP3 and increased in HNP5.

HNP1

The ROM and maximum value for T6 vertical excursion wassignificantly decreased. Forelimb protraction at first contact wasincreased (smaller angle) and retraction was increased after lift-off.The lumbar back was significantly less extended at midstance andthe sacrum significantly less angled during suspension and earlystance. In the hindlimb the fetlock joint was significantly lessextended at midstance, the angle of femur larger during break-overand first part of swing and the stifle and tarsal joints flexed moreslowly in the early swing phase. Hindlimb retraction was increasedafter toe-off, but angle at toe-off was not significantly different.

HNP3

Forelimb retraction after lift-off was significantly reduced and thestifle joint was significantly more flexed in early stance.

HNP4

L5 was slightly but significantly lower at minimum position andthe femur angle was slightly increased in the first half of stance.

HNP5

ROM for T6 and L5 vertical movement was significantly increasedand T6 and L5 minimum position at midstance significantlydecreased but the increase in maximum position during suspensionwas significant only at T6. Forelimb protraction was significantlyincreased during swing but was unchanged at first contact andforelimb retraction was reduced both at lift-off and during earlyswing. The lumbar back was significantly more extended atmidstance and the sacrum significantly more angled during break-over and suspension. The stifle and tarsal joints showed increasedflexion during stance. The angle of the femur was smaller earlyand late in the swing phase and the stifle, tarsal and fetlock jointswere all more flexed during the first two-thirds of swing.

Discussion

In the present study kinematic measurements were conducted ofhigh-level dressage horses ridden by highly experience riders. Themeasurements were made on a treadmill, and treadmill locomotionhas been found to differ significantly from locomotion over groundin that forelimb stance duration increase, diagonal advanceplacement of the hindlimb decrease and fore- and hindlimbretraction increase at trot (Buchner et al. 1994). However, since allmeasurements were conducted under the same conditions witheach horse acting as its own control, this may not have majorinfluence on the results. Further, any effects of speed wereeliminated as the control measurement was speed-matched foreach HNP. Comparing different head and neck positions it isnatural to have the free, unrestrained position (HNP1) as reference.However, as the rider should not have active influence on the horsein HNP1, it was not possible to use it for the speed series andHNP2 was therefore selected for this purpose. Use of skin fixatedmarkers always includes the risk of skin displacement errors.Distal to the elbow in the forelimb and the stifle in the hindlimb,

Tarsal joint

160

150

140

130

120

110

Lumbar back angle Sacral angle

13

12

11

10

9

8

7

0 50 100% Stride

0 50 100% Stride

Ang

le (

degr

ees)

-8

-9

-10

-11

-12

-13

-14

Fetlock joint

0 50 100% Stride

0 50 100% Stride

0 50 100% Stride

0 50 100% Stride

95

90

85

80

75

70

65

160

150

140

130

120

110

100

250

225

200

175

150

Femur angle Stifle joint

Ang

le (

degr

ees)

Ang

le (

degr

ees)

Ang

le (

degr

ees)

Ang

le (

degr

ees)

Ang

le (

degr

ees)

Fig 2: Group mean curves for selected variables with the stride normalisedto 101 points, 0–100%, in horses ridden at trot with the head and neck in 4 different positions HNP1 (solid black line), HNP3 (solid grey line), HNP4(dashed black line), HNP5 (dashed grey line) and a mean of the controlmeasurements in the reference position HNP2 (dash-dotted heavy black line).At top of the graphs horizontal bars are shown indicating significantdifferences (P<0.05) between a speed-matched measurement in the referenceposition HNP2 and HNP1 (upper black bar), HNP3 (upper grey bar), HNP4(lower black bar) and HNP5 (lower grey bar), respectively. At the bottomthere are horizontal bars representing stance from top to bottom of the leftand right forelimbs (black bars) and left and right hindlimbs (grey bars).

EVJ 08-236 Rhodin :Layout 1 11/02/2009 14:30 Page 4

M. Rhodin et al. 277

the skin movement artefacts are small enough to be neglected (vanWeeren et al. 1992). The various HNPs did probably not lead todifferent magnitudes of skin displacement for the proximalmarkers (except for the T6 marker) and as the different positionswere compared pair-wise, with each horse acting as its owncontrol, the skin displacement should not affect the differencesobserved between different head and neck positions. The neckangles may include some skin displacement artefacts but these arerather small due to the long distances between the markers.

It is well documented that the movements and limb loadingpattern of the horse changes when a rider mounts the horse(Schamhardt et al. 1991; Sloet van Oldruitenborgh-Oosterbaan et al. 1995; Clayton et al. 1999). From the present study and thekinetic evaluation of the same horses (Weishaupt et al. 2006) it isknown that additional changes occur when the rider takes the reinsand starts to influence the horse actively. In the current studyhigh-level dressage horses were measured at 3.15 ± 0.15 m/s, aspeed corresponding to the previously published speed forcollected trot of 3.20 ± 0.28 m/s (Clayton 1994), both when riddenon loose reins (HNP1) and in the collected position (HNP2).When changing from free to a collected trot the horse was placedin a shorter frame where the neck was more elevated, the sacrummore angled at hindlimb ground contact and the limbs were more

quickly flexed and protracted after lift-off. The movementsbecame more vertical with increased elevation of the forehandduring suspension and increased suspension duration (Weishauptet al. 2006). Furthermore, forelimb stance duration wassignificantly reduced and vertical impulse tended to beredistributed to the hindquarter (Weishaupt et al. 2006) andhindlimb fetlock extension increased during stance. This suggeststhat the hindlimbs were more engaged in load carrying, probablyin expense of propulsive force as indicated by decreased femurangle during stance (Holmström et al. 1995).

In the international dressage rules the collected trot isdescribed as “the horse moves with the hocks being well engagedand flexed, maintains an energetic impulsion, enables theshoulders to move with greater mobility, thus demonstratingcomplete self carriage. The horse’s steps are shorter than in theother types of trot without losing elasticity and cadence” (Anon2006). In the equestrian literature it is also stated that withincreased collection the hindlimbs should step further underneaththe horse (Anon 1997). The kinematic changes observed betweenfree and collected trot correspond rather well to some of thesedescriptions but the horse’s hindlimbs did not step furtherunderneath the horse. Increased sacral flexion at hindlimb groundcontact and decreased hindlimb retraction after lift-off may

TABLE 1: Mean ± s.d. range of motion (range), maximal (max) and minimum (min) values for selected kinematic variables

HNP1 HNP3 HNP4 HNP5

Neck angle (degrees) Mean -1.03 ± 6.04* -5.15 ± 3.31* 4.42 ± 5.05* -18.39 ± 3.64*(Control) (-8.84 ± 2.94) (-8.11 ± 1.98) (-9.23 ± 2.63) (-9.04 ± 3.15)Range 4.71 ± 0.71 5.18 ± 1.33* 6.30 ± 2.03 6.93 ± 1.08*

(Control) (5.32 ± 1.18) (5.83 ± 1.13) (5.71 ± 1.10) (5.80 ± 0.93)Vertical position T6 (mm) Range 66.18 ± 13.19* 78.06 ± 12.57 77.22 ± 13.40 94.18 ± 19.90*

(Control) (73.75 ± 9.39) (73.47 ± 10.17) (73.45 ± 9.24) (73.05 ± 9.81)Max 1786 ± 117.6* 1799 ± 117.0 1800 ± 119.6 1807 ± 116.25*

(Control) (1792 ± 116.4) (1793 ± 117.7) (1794 ± 116.9) (1794 ± 117.4)Min 1720 ± 114.4 1721 ± 114.9 1723 ± 115.1 1713 ± 114.0*

(Control) (1718 ± 114.9) (1719 ± 115.0) (1721 ± 115.7) (1721 ± 114.7)Vertical position L5 (mm) Range 99.88 ± 10.25 107.45 ± 7.90 112.89 ± 11.28 119.52 ± 13.80*

(Control) (103.51 ± 6.98) (106.08 ± 6.99) (106.54 ± 8.62) (106.08 ± 8.13)Max 1738 ± 103.3 1747 ± 104.1 1750 ± 105.3 1751 ± 103.7

(Control) (1740 ± 102.2) (1745 ± 104.5) (1747 ± 104.2) (1746 ± 103.2)Min 1638 ± 105.8 1639 ± 107.5 1637 ± 107.4* 1631 ± 107.9*

(Control) (1637 ± 106.2) (1639 ± 106.0) (1640 ± 106.6) (1640 ± 105.9)Stride protraction/retraction Range 62.06 ± 2.15 59.96 ± 2.78 60.29 ± 2.95 59.59 ± 2.54fore limb (degrees) (Control) (61.44 ± 2.21) (60.88 ± 2.69) (60.17 ± 2.44) (60.37 ± 2.74)

Max 117.07 ± 2.25* 114.08 ± 2.70* 115.06 ± 2.92 112.29 ± 2.74*(Control) (115.62 ± 1.92) (115.49 ± 2.51) (115.11 ± 2.35) (115.14 ± 2.21)

Min 54.97 ± 1.72 54.13 ± 3.01 54.79 ± 2.04 52.72 ± 3.06*(Control) (54.19 ± 2.05) (54.61 ± 2.09) (54.94 ± 1.67) (54.76 ± 2.23)

Stance protraction/retraction Range 50.81 ± 3.39 48.23 ± 3.62 49.61 ± 2.97 45.25 ± 4.16*fore limb (degrees) (Control) (50.88 ± 2.33) (50.39 ± 2.53) (49.83 ± 2.46) (49.79 ± 2.53)

Max 110.23 ± 3.38 110.01 ± 2.93 110.51 ± 2.88 108.12 ± 4.12*(Control) (111.41 ± 2.52) (111.04 ± 2.46) (110.83 ± 2.42) (110.90 ± 2.26)

Min 59.39 ± 1.44* 61.79 ± 2.01 60.86 ± 1.57 62.89 ± 1.21(Control) (60.52 ± 1.16) (60.67 ± 1.36) (61.02 ± 1.40) (61.14 ± 1.50)

Stride protraction/retraction Range 35.35 ± 1.46 34.73 ± 1.26 34.14 ± 1.30 34.45 ± 2.10hindlimb (degrees) (Control) (35.26 ± 1.43) (34.40 ± 1.37) (34.05 ± 1.43) (33.97 ± 1.79)

Max 119.84 ± 1.76* 118.70 ± 1.46 118.55 ± 1.37 118.46 ± 1.42(Control) (119.34 ± 1.63) (118.83 ± 1.60) (118.52 ± 1.38) (118.56 ± 1.16)

Min 84.47 ± 1.32 83.96 ± 1.86* 84.38 ± 1.90 84.03 ± 2.67(Control) (84.07 ± 1.69) (84.41 ± 1.88) (84.44 ± 1.80) (84.59 ± 2.08)

Stance protraction/retraction Range 31.58 ± 1.90 31.23 ± 1.65 30.73 ± 1.82 31.14 ± 2.21hindlimb (degrees) (Control) (31.72 ± 1.80) (30.87 ± 1.70) (30.65 ± 1.66) (30.65 ± 1.79)

Max 116.40 ± 2.39 115.62 ± 2.33 115.49 ± 2.21 115.48 ± 2.39(Control) (116.15 ± 2.35) (115.56 ± 2.42) (115.45 ± 2.35) (115.54 ± 2.17)

Min 84.80 ± 1.39 84.39 ± 1.94 84.73 ± 1.94 84.34 ± 2.75(Control) (84.45 ± 1.74) (84.68 ± 1.95) (84.79 ± 1.91) (84.89 ± 2.17)

*Significant difference (P<0.05) compared to reference HNP2. A negative value for the neck angle means the neck above the horizontal plane.

EVJ 08-236 Rhodin :Layout 1 11/02/2009 14:30 Page 5

278 The effect of different head and neck positions on the caudal back

perhaps give the impression that the horse’s hindlimbs are furtherunderneath its body. Increased stifle and tarsal joint flexion wasnot observed at collected trot, neither in the present nor in aprevious study (Holmström et al. 1995), but has been documentedin piaffe and passage (Holmström et al. 1995).

Increasing degree of collection is usually associated withreduction of speed (Clayton 1994; Holmström et al. 1995) and someof the previously observed changes, such as increased stride durationand decreased stride length, were not observed in the present studywhen the effect of speed was eliminated and are therefore morelikely an effect of speed rather than degree of collection. However,the decreased femur angle at hindlimb lift-off and decreased fore-and hindlimb retraction in early swing phase found by Holmström et al. (1995) remained in the current study and may therefore be aspecific kinematic characteristic of collection at trot.

When the horse performed the collected trot with a smallerhead (HNP3) or head and neck angle (HPN4) these high-levelhorses showed few additional changes compared to the referenceposition (HNP2). At collected trot, the movements and limbloading of the horse were not consistently correlated to the degreeof flexion of its head and neck. It is not surprising that these highlytrained dressage horses could maintain a similar movement patternbetween different degrees of neck flexion, as one important goal indressage training is that the rider should be able to vary theplacement of the horse’s head and neck. It was not possible toconfirm the gymnastic effects claimed to be achieved with aHNP4-like position (Janssen 2003). The increased ROM in thelumbar back found with a similar position in unridden horses(Gómez Álvarez et al. 2006) could not reproduced in the samehorses when ridden. From a biomechanical point of view, neitherthe fears nor the benefits that has been associated with increasedflexion of the head and neck in dressage horses can therefore beconfirmed based on the current study. However, in the presentstudy it was not possible to make the participating horses accept adegree of neck flexion as extreme as the position referred to as‘rollkur’ (Janssen 2003).

If a more extreme and unnatural head and neck position (HNP5)is imposed on the horse, this may cause profound changes. Ashorter stance duration (Weishaupt et al. 2006) resulted in anincreased flexion of the stifle and tarsal joint during stance andincreased suspension duration. Many of the changes observed inthis extreme position are therefore the same as previously referredto as signs of increased degree of collection. This may be foundawkward at first, since few riders and trainers would accept HNP5as true collection, but, since the extreme elevation of the neck couldtransfer more vertical load to the hindlimbs, it is not illogical thatkinematic changes were observed that related to the increasedweightbearing of the hindlimbs. Even though the vertical impulseswere reduced in the forelimbs, vertical peak force was increased(Weishaupt et al. 2006), increasing the maximal stress on tendonsand ligaments. Further, the increased neck extension seems toincrease back extension at midstance. Because of the saddle it waspossible only to measure the lumbar back, but the same head andneck position caused increased extension of both the thoracic andlumbar back in the same horses when unmounted (Gómez Álvarezet al. 2006). A weighted saddle has been shown to increase maximallumbar back extension (de Cocq et al. 2004) and if back extensionis then even further increased by extreme elevation of the head andneck, this may predispose to back injuries in the horse. Our findingstherefore give some substance to the statement by Heuschmann(2006) that a too high HNP results in the horse dropping its back.

In conclusion, when changing from free to collected trot, thehigh level dressage horse was placed in a shorter frame with theneck more elevated, the movements more vertical and withincreased elevation of the trunk during suspension. The changesobserved included many, but not all, of those described in theequestrian literature for increased collection.

Increased head or head and neck flexion had few additionaleffects compared to the position desired at dressage competitions.At collected trot the degree of neck flexion is thus not consistentlycorrelated to the movements of the horse’s limbs and trunk.

Extreme elevation of the head and neck can produce kinematiceffects indicative of increased degree of collection, but is followedby increased (lumbar) back extension which could place the horseat risk of injury if used for a prolonged period of time. Changes inhead and neck position may influence the young or untrainedhorse more than these highly trained horses but this requiresfurther study.

Acknowledgements

This study was supported by grants from Stiftelsen SvenskHästforskning, the insurance company Sveland, Sweden and UllaHåkansson. The authors wish to thank Sören Johansson, Katja vonPeinen, Nina Waldern and Thomas Wiestner for excellent technicalassistance and all riders and horse owners who generously lent ustheir horses for this study.

Manufacturers’ addresses

1Karga AG, Fahrwangen, Switzerland.2Qualysis, Gothenburg, Sweden.3The Math Works Inc., Natick, Massachusetts, USA.

References

Anon (2006) Dressage. In: Rules for Dressage Events, Fédération EquestreInternationale, Lausanne. pp 7-9.

Balkenhol, K., Müller, H., Plewa, M. and Heuschmann, G. (2003) Zur Entfaltungkommen - statt zur Brust genommen. In: ReiterRevue. pp 46-51.

Biau, S., Couve, O., Lemaire, S. and Barrey, E. (2002) The effect of reins on kineticvariables of locomotion. Equine vet. J., Suppl. 34, 359-362.

Buchner, H.H., Savelberg, H.H., Schamhardt, H.C., Merkens, H.W. and Barneveld, A.(1994) Kinematics of treadmill versus overground locomotion in horses. Vet. Q.16, Suppl. 2, S87-90.

Clayton, H.M. (1994) Comparison of the stride kinematics of the collected, working,medium and extended trot in horses. Equine vet. J. 26, 230-234.

Clayton, H.M., Lanovaz, J.L., Schamhardt, H.C. and van Wessum, R. (1999) Theeffects of a rider’s mass on ground reaction forces and fetlock kinematics at thetrot. Equine vet. J., Suppl. 30, 218-221.

de Cocq, P., van Weeren, P.R. and Back, W. (2004) Effects of girth, saddle and weighton movements of the horse. Equine vet. J. 36, 758-763.

Anon (1997) The Principles of Riding, Completely rev. edn., Deutsche ReiterlicheVereinigung, Kenilworth, Buckingham. pp 140.

Gómez Álvarez, C.B., Rhodin, M., Bobbert, M.F., Meyer, H., Weishaupt, M.A.,Johnston, C. and van Weeren, P.R. (2006) The effect of head and neck position onthe thoracolumbar kinematics in the unridden horse. Equine vet. J., Suppl. 36,445-451.

Heuschmann, G. (2006) Tug of War: Classical versus “Modern” Dressage, Wu WeiVerlag, Schondorf. pp 85-91.

Holmström, M., Fredricson, I. and Drevemo, S. (1995) Biokinematic effects ofcollection on the trotting gaits in the elite dressage horse. Equine vet. J. 27, 281-287.

Janssen, S. (2003) Zur Brust genommen. In: ReiterRevue. pp 41-45.

Rhodin, M., Johnston, C., Holm, K.R., Wennerstrand, J. and Drevemo, S. (2005) Theinfluence of head and neck position on kinematics of the back in riding horses atthe walk and trot. Equine vet. J. 37, 7-11.

EVJ 08-236 Rhodin :Layout 1 11/02/2009 14:30 Page 6

M. Rhodin et al. 279

Roepstorff, L., Johnston, C., Drevemo, S. and Gustas, P. (2002) Influence of draw

reins on ground reaction forces at the trot. Equine vet. J., Suppl. 34, 349-352.

Schamhardt, H.C., Merkens, H.W. and van Osch, G.J.V.M. (1991) Ground reaction

force analysis of horses ridden at the walk and trot. In: Equine Exercise

Physiology 3, Eds: S.G.B. Persson, A. Lindholm and L.B. Jeffcott, ICEEP

Publications, Davis. pp 120-127.

Sloet van Oldruitenborgh-Oosterbaan, M.M., Barneveld, A. and Schamhardt, H.

(1995) Effects of weight and riding on workload and locomotion during treadmill

exercise. Equine vet. J., Suppl. 18, 413-417.

van Weeren, P.R., van den Bogert, A.J. and Barneveld, A. (1992) Correction models for

skin displacement in equine kinematic gait analysis. Equine vet. Sci. 13, 178-192.

Weishaupt, M.A., Hogg, H.P., Wiestner, T., Denoth, J., Stussi, E. and Auer, J.A. (2002)Instrumented treadmill for measuring vertical ground reaction forces in horses.Am. J. vet. Res. 63, 520-527.

Weishaupt, M.A., Wiestner, T., von Peinen, K., Waldern, N., Roepstorff, L., vanWeeren, R., Meyer, H. and Johnston, C. (2006) Effect of head and neck positionon vertical ground reaction forces and interlimb coordination in the dressagehorse ridden at walk and trot on a treadmill. Equine vet. J., Suppl. 36, 387-392.

Author contributions The initiation, conception, planning andexecution for this study were by M.R., C.B.G., C.J., P.R.v.W., L.R.and M.A.W. The statistics were by M.R., A.B. and L.R., and thepaper was written by M.R., C.B.G., A.B. and L.R.

EVJ 08-236 Rhodin :Layout 1 11/02/2009 14:30 Page 7