The 3 most common gaits of equids (walk, trot, and canter or gallop) have been extensively evaluated.1 Various other gaits of horses selectively bred for their natural, characteristic additional gaits (ie, gaited horse breeds) have been scientifically analyzed only in the last decade.2–7 Although footfall pattern, limb timing, and foot placement are often specific to each breed, these factors might also vary naturally with StrD in and between individual horses of the same breed. Little is known about the extent of limb loading that occurs with various alternative gaits, particularly compared with loading at the 3 main gaits.
The tölt of Icelandic horses belongs to the category of 4-beat ambling gaits and ideally consists of a symmetric, regular footfall pattern with the forelimb landing after the ipsilateral hind limb at approximately 25% of StrD. Scientific data regarding temporal and spatial kinematic variables as well as ground reaction force of this gait are available2,4,5,7–9; however, to the authors’ knowledge, a direct comparison of tölting and trotting at equal speeds has not yet been performed. Such an assessment would allow for comparison between gaits with and without a suspension phase and would be important to understand the effect of differences in gait mechanics and limb loading. Estimates of the degree of Fzpeak achieved at a tölt versus a trot have been reported, with conflicting findings. One group of investigators5 speculated that an extremely brief StDrel at a tölt might lead to higher Fzpeak than at a trot, whereas another group7 speculated that the greater overlap of limb support might result in lower Fzpeak.
Riders claim that to ride an Icelandic horse at a tölt, they need to rebalance the horse on the hind limbs and raise its head and forehand. To induce and facilitate this presumed weight shift, the saddle, rider, or both are traditionally placed farther back on the horse in Icelandic horse riding than in other riding styles.10 However, whether and to which extent this technique induces an alteration of the horse's balance has never been confirmed or quantified scientifically to the authors’ knowledge.
The purpose of the study reported here was to assess differences in ground reaction forces, limb movement, and posture of Icelandic horses at a tölt versus trot at equal velocities (V1 and V2). We believed that knowledge of limb loading at both gaits would be important for evaluation of their impact on orthopedic health. Our hypotheses were that, compared with at a trot, the higher HNP of an Icelandic horse at a tölt would lead to a weight shift to the hind limbs and that the higher protraction arc of the forelimbs in tölting horses would lead to shorter contact durations and therefore to higher Fzpeak.
Materials and Methods
Animals
The study group consisted of 12 sound privately owned Icelandic horses (4 stallions, 3 geldings, and 5 mares), with a mean ± SD age of 10 ± 3 years, mean body weight of 354 ± 24 kg, and mean height at the withers (highest point of the shoulders) of 1.37 ± 0.03 m. Owner consent was obtained prior to the trials. Horses were judged to be free of lameness and back pain on the basis of findings of a thorough clinical examination. Horses were shod in accordance with standard shoeing principles with steel shoes (20 × 8 mm), without the application of pads or padding material. All were introduced to tölting and trotting on a high-speed treadmilla for a period of at least 4 days before the study began. The study was performed with the approval of the Animal Health and Welfare Commission of the Canton of Zurich (permission No. 206/2010).
Experimental design
All horses were ridden by 1 of 2 riders experienced in riding Icelandic horses. Measurements were carried out in random order with horses moving at a tölt and trot at 2 speeds each (3.4 m/s [V1] or 39 m/s [V2]). Tölt at both speeds and trot at V1 were performed with riders fully seated; trot at V2 was performed with riders seated forward (seat slightly raised from the saddle, full weight in the stirrups, upper body tilted slightly forward). Bridles with snaffle bits were used, and a dressage-style saddle for Icelandic horses was fitted to each horse in a standard saddle position.11 For each trial, gait accuracy and horse posture were assessed on-site by an experienced judge for Icelandic horse competitions.
Data acquisition
Ground reaction force and temporal variables for each limb were measured with a treadmill-integrated force-measuring system.12 Synchronized kinematic data were simultaneously acquired by tracking of reflective spherical markers (diameter, 19 mm) placed over anatomic landmarks on both sides of horse and rider. Markers placed on the horse were located at the wings of the atlas; the spinous processes of L4, L5, and S3; the sacrum (tuber sacrale); the lateral aspect of both forelimbs over the center of rotation of the shoulder (tuberculum majus), carpal, and fetlock (metacarpophalangeal) joints as well as both hind limbs over the tuber coxae; stifle, tarsal, and fetlock (metatarsophalangeal) joints; and the lateral hoof wall at the level of the coffin (distal interphalangeal) joint for all hooves. Markers placed on the rider were located at both shoulders (acromion) and hips (tuber coxae). Motion-capture softwareb was used to control 9 high-speed infrared camerasc and to calculate 3-D kinematic data. The left-handed coordinate system was aligned with the treadmill, with the x-axis pointing toward the horse's head, the y-axis pointing toward the horse's right side, and the z-axis pointing upward. For both systems, a sampling frequency of 480 Hz was applied. Recording was performed for 15 seconds, which resulted in > 25 strides for each gait.
Data analysis
Force curves and limb positional data acquired with the treadmill-integrated force-measuring system were used to compute StrD, stride rate, stride length, stance duration, lateral step duration, suspension duration, time of advanced placement (interval representing hind hoof impact prior to diagonal forehoof impact), time of advanced completion (interval representing hind hoof lift off prior to diagonal forehoof lift off), stride impulse (impulse of all 4 limbs during stride), and percentage of total impulse carried by both forelimbs as well as impulse and Fzpeak for each limb. All time variables were standardized as percentage of StrD; both absolute stance duration and StDrel (ie, duty factor) were calculated only for stance duration. Force and impulse values were additionally standardized to the combined mass of horse and rider (N/kg and N·s/kg, respectively). Values of multiple strides in a record were averaged for each variable. Velocity data were derived from treadmill belt speed.
Kinematic variables were determined for each limb, including proretraction angle, proretraction length, protraction length, retraction length, and caudal and cranial overswing lengths as well as height and point of maximum height of the coffin joint flight arc. All kinematic variables and calculations referred to the sagittal plane of the horse. Proretraction angle was determined during stance as the angle of the coffin joint marker rotated around an imagined mathematical center between both shoulders (calculated spatial midpoint of both shoulder markers) in the forelimbs or as the angle of the coffin joint marker rotated around an imaginary center between both tuber coxae markers in the hind limbs. Proretraction length was the distance that the horse's hooves traveled backward on the treadmill while the respective limb rotated during stance, with the zero reference point defined as the position of vertical orientation of the metacarpus (line through the lateral marker at the level of the fetlock joint and the marker at the level of the carpal joint) or metatarsus (line through the lateral marker at the level of the fetlock joint and the marker at the level of the tarsal joint). Cranial overswing length was defined as the amount of maximal protraction at the end of swing minus the amount of retraction at the beginning of stance, and caudal overswing length was defined as the amount of maximal retraction at the beginning of swing minus the amount of retraction at the end of stance.
Also determined were vertical movement of the head (midpoint between the left and right atlas marker), shoulder (midpoint between left and right shoulder marker), and sacrum of each horse as well as angulation and movement of the caudal aspect of the back and croup region during the limb-loading phase. Caudal back angle was calculated as the pitching rotation of L4 around the sacrum with reference to a horizontal line through the sacrum; a negative value indicated that L4 was lower than the sacrum. Croup angle was measured as the pitching rotation of S3 around the sacrum, with reference to a horizontal line through the sacrum; a negative value indicated that S3 was lower than the sacrum. Croup shape was defined as the angle at the sacrum formed by the 3 markers (L4, sacrum, and S3). To quantify the extent of the so-called Hankenbeugung, heights of the sacrum, stifle joint, and fetlock joint as well as the respective vertical segment lengths (height of the sacrum – height of the stifle joint; height of the stifle joint – height of the fetlock joint) were measured at the points of limb first contact and at Fzpeak.
Position and movements of the rider's hip and rider's position in relation to the horse's back as well as forward tilt of the rider's upper body were calculated. For this purpose, rider's hip location was determined as the spatial mean between the left and right tuber coxae markers and, likewise, the rider's shoulder location by the spatial mean of the left and right acromia markers. Rider position was defined as the x-axis position of the midpoint of the rider's hip in relation to horse's L5. Rider tilt was defined as the angle formed by a line through the midpoints of the rider's shoulder and hip and a vertical line passing through the midpoint of the hip; a positive value indicated that the rider was tilting forward.
Time series of kinematic variables and discrete limb contact durations from the treadmill-integrated force-measuring system were imported into a statistical software programd for additional analysis. Stridecycle times of the left forelimb were used to split time series into strides. Data for each stride cycle were time standardized to 501 points (0% to 100% of StrD), and all strides of a recording were averaged to a standardized stride by which the mean value, ROM, maximum value, and minimum value of the entire stride for the x- and z-dimensions were calculated.
Values of corresponding kinetic and kinematic variables for the contralateral limbs were pooled for each test condition (tölt or trot at V1 or V2) and horse and reported as forelimb and hind limb values. Finally, group means and SDs for each test condition were calculated for each variable from the values of all 12 horses.
Statistical analysis
Statistical analysis was performed with commercially available software.e First, normality (Kolmogorov-Smirnov test) and equal variance of data were evaluated. Then, differences between values obtained during tölting and trotting at each speed were tested via 1-factor repeated-measures ANOVA. Values of P < 0.05 were considered significant.
Results
All 12 horses completed all trials at a tölt or trot at treadmill velocities of 3.4 or 3.9 m/s. Mean treadmill velocities did not differ significantly between the tölt and trot at both speeds. Results of gait analysis for horses at a tölt and trot at the 2 speeds were summarized (Tables 1 and 2). At both speeds, footfall pattern at a tölt was characterized by lateral coupling in that the relative lateral step duration was clearly shorter (< 18% of StrD) than expected for a perfectly regular 4-beat rhythm (25% of StrD). At a trot, half of the horses (6/12) had an airborne phase at V2, whereas only 2 horses had an airborne phase at V1.
Mean ± SD values of kinetic variables for orthopedically normal Icelandic horses (n = 12) ridden at a tölt and trot at 2 treadmill velocities (V1 and V2).
| Variable (unit) | Trot at V1 | Tölt at V1 | Trot at V2 | Tölt at V2 |
|---|---|---|---|---|
| Velocity (treadmill speed; m/s) | 3.42 ± 0.02 | 3.44 ± 0.04 (0.6) | 3.90 ± 0.11 | 3.91 ± 0.09 (0.1) |
| StrD (ms) | 602 ± 20 | 550 ± 24 (–8.7)* | 574 ± 20 | 523 ± 23 (–8.9)* |
| Stride rate (strides/min) | 99.8 ± 3.3 | 109.4 ± 4.7 (9.6)* | 104.7 ± 3.8 | 115.0 ± 4.9 (9.8)* |
| Stride length (mm) | 2,058 ± 61 | 1,891 ± 79 (–8.1)* | 2,237 ± 91 | 2,042 ± 90 (–8.7)* |
| Stance duration, absolute (ms) | ||||
| Forelimbs | 294 ± 10 | 260 ± 17 (–11.5)* | 268 ± 10 | 237 ± 11 (–11.7)* |
| Hind limbs | 278 ± 8 | 282 ± 15 (1.7) | 251 ± 9 | 258 ± 15 (2.7) |
| Stance duration, relative (%StrD)† | ||||
| Forelimbs | 48.8 ± 1.8 | 47.3 ± 3.0 (–3.1)* | 46.9 ± 2.3 | 45.4 ± 2.2 (–3.2)* |
| Hind limbs | 46.1 ± 1.8 | 51.3 ± 1.5 (11.3)* | 43.9 ± 2.2 | 49.4 ± 1.9 (12.7)* |
| Ipsilateral step duration (%StrD)‡ | 47.5 ± 2.3 | 18.0 ± 2.2 (–62.2)* | 46.7 ± 3.4 | 17.9 ± 2.6 (–61.7)* |
| Suspension duration (%StrD) | 0.1 ± 2.2 | — | 1.5 ± 4.3 | — |
| Time of advanced placement (%StrD)§ | −2.5 ± 2.3 | −3.3 ± 3.4 | — | |
| Time of advanced completion (%StrD)‖ | 0.2 ± 2.2 | — | −0.3 ± 3.9 | — |
| Stride impulse (N·s/kg)¶ | 5.90 ± 0.19 | 5.39 ± 0.24 (–8.7)* | 5.63 ± 0.20 | 5.13 ± 0.22 (–8.9)* |
| Limb impulse (N·s/kg)¶ | ||||
| Forelimbs | 1.70 ± 0.06 | 1.55 ± 0.07 (–8.9)* | 1.66 ± 0.08 | 1.47 ± 0.07 (–11.4)* |
| Hind limbs | 1.25 ± 0.05 | 1.15 ± 0.07 (–8.3)* | 1.16 ± 0.05 | 1.10 ± 0.06 (–5.4)* |
| Forelimb impulse (%)# | 57.5 ± 1.1 | 57.3 ± 1.2 (–0.3) | 58.8 ± 1.4 | 57.2 ± 1.1 (–2.7)* |
| Fzpeak (N/kg)** | ||||
| Forelimbs | 8.88 ± 0.46 | 9.52 ± 0.63 (7.2)* | 9.58 ± 0.66 | 9.96 ± 0.61 (4.0)* |
| Hind limbs | 7.49 ± 0.41 | 6.61 ± 0.29 (–11.8)* | 7.57 ± 0.35 | 6.93 ± 0.28 (–8.5)* |
Values in parentheses represent the mean percentage difference between tölt and trot for the given variable and velocity.
Difference is significant (P < 0.05) between values for tölt and trot at equal velocity.
Stance duration as percentage of StrD (ie, duty factor).
Interval between hind hoof impact and ipsilateral forehoof impact,
Interval for hind hoof impact prior to the diagonal forehoof impact.
Interval for hind hoof lift off prior to diagonal forehoof lift off.
Mass-standardized vertical impulse.
Percentage of stride impulse carried by both forelimbs.
Mass-standardized Fzpeak.
— = Not applicable. %StrD = Percentage of StrD.
Group mean ± SD values of kinematic variables for 12 orthopedically normal Icelandic horses ridden at a tölt and trot at 2 treadmill velocities (V1 and V2).
| Variable | Value | Trot at V1 | Tölt at V1 | Trot at V2 | Tölt at V2 |
|---|---|---|---|---|---|
| Forelimbs | |||||
| Proretraction angle (°)* | ROM | 52.1 ± 1.7 | 46.0 ± 3.2 (–11.6)† | 53.7 ± 2.3 | 47.6 ± 2.1 (–11.4)† |
| Proretraction length (mm)‡ | x-ROM | 948 ± 28 | 847 ± 46 (–10.6)† | 981 ± 28 | 877 ± 32 (–10.6)† |
| Protraction length (mm) | x-Max | 472 ± 23 | 399 ± 34 (–15.3)† | 494 ± 40 | 424 ± 29 (–14.1)† |
| Retraction length (mm) | x-Min | −476 ± 42 | −448 ± 39 (–5.9)† | −487 ± 41 | −453 ± 40 (–7.1)† |
| Cranial overswing length (mm)§ | Δx | 35 ± 13 | 35 ± 24 (–1.1) | 44 ± 17 | 43 ± 27 (–0.9) |
| Caudal overswing length (mm)‖ | Δx | −5 ± 3 | −3 ± 1 (–34.7) | −6 ± 3 | −5 ± 4 (–11.8) |
| Coffin joint flight arc | |||||
| Maximum height (mm) | z-ROM | 168 ± 23 | 199 ± 50 (18.9)† | 185 ± 36 | 219 ± 56 (18.1)† |
| Time of maximum height (%SwD) | — | 33.8 ± 2.6 | 44.8 ± 5.9 (32.8)† | 35.3 ± 4.1 | 46.8 ± 6.7 (32.6)† |
| Hind limbs | |||||
| Proretraction angle (°)* | ROM | 39.7 ± 1.5 | 40.2 ± 1.4 (1.3) | 40.8 ± 1.5 | 41.4 ± 1.2 (1.5) |
| Proretraction length (mm)‡ | x-ROM | 871 ± 27 | 881 ± 38 (1.2) | 894 ± 34 | 907 ± 29 (1.5) |
| Protraction length (mm) | x-Max | 513 ± 34 | 525 ± 54 (2.3) | 520 ± 40 | 542 ± 45 (4.1)† |
| Retraction length (mm) | x-Min | −358 ± 21 | −356 ± 29 (–0.5) | −373 ± 25 | −365 ± 29 (–2.2) |
| Cranial overswing length (mm)§ | Δx | 4 ± 5 | 6 ± 7 (79.1)† | 5 ± 5 | 7 ± 8 (46.6) |
| Caudal overswing length (mm)‖ | Δx | −65 ± 18 | −8 ± 4 (–87.7)† | −66 ± 18 | −11 ± 7 (–83.3)† |
| Coffin joint flight arc | |||||
| Maximum height (mm) | z-ROM | 73 ± 20 | 72 ± 14 (–1.3) | 82 ± 18 | 76 ± 11 (–8.2)† |
| Time of maximum height (%SwD)¶ | — | 56.7 ± 25.9 | 41.2 ± 27.3 (–27.4) | 42.9 ± 21.9 | 34.6 ± 21.5 (–19.4) |
| Caudal back configuration | |||||
| Caudal back angle (°)# | at Fzpeak | −9.4 ± 1.6 | −6.2 ± 2.4 (–34.3)† | −8.9 ± 2.4 | −5.2 ± 2.5 (–42.0)† |
| Croup angle (°)** | at limb FC | −14.1 ± 2.8 | −18.2 ± 3.4 (29.2)† | −14.6 ± 3.0 | −19.0 ± 3.1 (30.3)† |
| at Fzpeak | −14.0 ± 3.1 | −20.3 ± 3.1 (44.7)† | −14.2 ± 2.7 | −20.8 ± 2.8 (46.1)† | |
| Croup shape (°)†† | at limb FC | 156.3 ± 2.4 | 152.8 ± 2.3 (–2.2)† | 156.7 ± 2.7 | 152.8 ± 2.7 (–2.5)† |
| at Fzpeak | 156.6 ± 2.9 | 153.6 ± 2.6 (–2.0)† | 156.9 ± 2.9 | 154.0 ± 2.8 (–1.8)† | |
| Height at marker (mm) Head‡‡ | z-sMean | 1,408 ± 38 | 1,528 ± 50 (8.5)† | 1,400 ± 48 | 1,525 ± 30 (9.0)† |
| Shoulder§§ | z-sMean | 1,000 ± 27 | 1,026 ± 35 (2.7)† | 1,000 ± 31 | 1,025 ± 29 (2.5)† |
| z-ROM | 58 ± 8 | 54 ± 8 (–7.7)† | 56 ± 8 | 54 ± 7 (–3.6) | |
| z at Fzpeak | 999 ± 29 | 1,010 ± 31 (1.1)† | 998 ± 29 | 1,004 ± 27 (0.7)† | |
| Sacrum | z-sMean | 1,282 ± 37 | 1,295 ± 33 (1.0)† | 1,282 ± 34 | 1,292 ± 34 (0.8)† |
| z-ROM | 47 ± 7 | 31 ± 6 (–34.3)† | 43 ± 6 | 31 ± 5 (–27.1)† | |
| z at Fzpeak | 1,261 ± 37 | 1,284 ± 33 (1.8)† | 1,263 ± 34 | 1,280 ± 34 (1.3)† | |
| Stifle joint | z at Fzpeak | 635 ± 19 | 637 ± 18 (0.3) | 638 ± 17 | 637 ± 17 (–0.2) |
| Fetlock joint, forelimbs | z at Fzpeak | 89 ± 11 | 92 ± 9 (3.2)† | 86 ± 10 | 88 ± 9 (2.7)† |
| Fetlock joint, hind limbs | z at Fzpeak | 74 ± 11 | 80 ± 11 (8.3)† | 73 ± 9 | 76 ± 9 (4.1)† |
| Hind limb segmental heights (mm) | |||||
| Sacrum – stifle joint | z at Fzpeak | 626 ± 26 | 646 ± 24 (3.3)† | 625 ± 24 | 643 ± 25 (2.9)† |
| Stifle joint – fetlock joint | z at Fzpeak | 562 ± 15 | 558 ± 12 (–0.7) | 565 ± 15 | 561 ± 13 (–0.8) |
| Rider | |||||
| Height of rider's hip (mm)‖‖ | z-ROM | 65 ± 9 | 29 ± 9 (–56.0)† | 46 ± 9 | 25 ± 5 (–46.6)† |
| Rider position (mm)¶¶ | x-sMean | 276 ± 37 | 274 ± 44 (–0.8) | 343 ± 34 | 275 ± 39 (–19.8)† |
| Rider tilt (°)## | sMean | 11.0 ± 1.8 | 8.6 ± 2.0 (–22.2)† | 27.3 ± 3.1 | 10.0 ± 3.7 (–63.5)† |
Kinematic values pertain to a left-handed coordinate system, with the x-axis pointing toward the horse's head and the z-axis pointing upward, and spatial data are projections onto the sagittal plane. The ROM was measured during stride.
Angle calculated as y-rotation during stance either of the forelimb coffin joint marker around the mean of both shoulder markers or of the hind limb coffin joint marker around the mean of the tuber coxae markers.
Difference is significant (P < 0.05) between values for tölt and trot at the same velocity.
Distance the horse's hoof traveled backward on the treadmill belt while the respective limb rotated, with the zero reference point defined as the position of vertical orientation of the metacarpus or metatarsus.
Amount of maximal protraction minus amount of retraction at the beginning of stance.
Amount of maximal retraction minus amount of retraction at the end of stance.
Time of occurrence of maximal height of coffin joint trajectory during swing.
Pitching rotation of L4 around the sacrum with reference to a horizontal line through the sacrum; negative value indicates that L4 was lower than the sacrum.
Pitching rotation of S3 around the sacrum, with reference to a horizontal line through the sacrum; negative value indicates that S3 was lower than the sacrum.
Angle at the sacrum formed by the 3 markers (at L4, sacrum, and S3).
Midpoint between the left and right atlas marker.
Midpoint between the left and right shoulder marker.
Midpoint between the left and right tuber coxae marker.
Position of the rider's hip (midpoint between both hip markers) in relation to the horse's L5 regarding x-axis coordinates.
Angle formed by a line through the rider's shoulder (midpoint of both shoulder markers) and hip (midpoint between both hip markers) and a vertical line passing through the hip position; positive value indicates rider tilting forward.
%SwD = Percentage of swing duration. Δ = Difference (regarding given coordinate). FC = First contact. Max = Maximum during stride. Min = Minimum during stride. sMean = Mean value for entire stride. x = x-axis. z = z-axis.
See Table 1 for remainder of key.
Values of most kinetic and kinematic variables differed significantly between tölting and trotting at the same speeds, and differences between the 2 gaits at the 2 speeds were similar. Compared with at a trot, StrD and stride length were shorter and impulse of the forelimbs and hind limbs was lower at a tölt. The mean position of the rider in relation to the horse's back as well as the impulse balance between fore- and hind limbs was similar at both speeds at a tölt and at V1 at a trot. At the faster trot speed, rider position was approximately 68 mm farther cranially and led to a 2.7% increase of impulse carried by both forelimbs of the horse. At a tölt versus trot, horses carried their heads higher and forelimbs had a shorter StDrel and higher Fzpeak; in hind limbs, longer StDrel was associated with lower Fzpeak. The shape of the vertical force curves for the forelimbs had a spiked apex, and the shape of curves for the hind limbs had a blunter apex (Figure 1). Proretraction angle and proretraction length were smaller in the forelimbs at a tölt than at a trot but remained unchanged between gaits in the hind limbs. In the forelimbs, maximal protraction length and retraction length at a tölt were shorter than at a trot. In the hind limbs, maximal protraction length was longer at V2 at a tölt and retraction length remained unchanged at both speeds. Height of the flight arc of the hooves at a tölt was higher in the forelimbs at both speeds than at a trot and lower in the hind limbs at V2 (Figure 2). The point of maximal height of the flight arc was shifted toward the moment of ground impact in the forelimbs at a tölt and in the hind limbs at a trot. Mean height at the horse's sacrum and shoulder was higher and range of vertical excursion of the sacrum was lower at a tölt than at a trot.
Representative plots for body weight–standardized vertical ground reaction forces (A and B) and mean vertical height (C and D) throughout a stride for an orthopedically normal Icelandic horse ridden on a treadmill at a trot (A and C) and tölt (B and D) at the same velocity (3.4 m/s). In panels A and B, values for the left forelimb (blue), right forelimb (red), left hind limb (cyan), and right hind limb (green) are displayed separately. In panels C and D, y-axis values represent mean vertical height and ROM for the rider's hip (magenta; mean of values for both tuber coxae markers) and horse's height at the sacrum (black; tuber sacrale) and shoulder (green; mean values for both markers placed at the left and right tuberculum majus), each centered at the given stride mean value. Stride-standardized curves represent a mean of 25 strides for an interval of 1.5 strides (150% of StrD). Fore- and hind limbs behaved in a similar fashion for both gaits with respect to Fzpeak and vertical movement of the trunk: at Fzpeak of each limb, the associated proximal body markers (shoulder or sacrum) were at their lowest position, whereas when limb support changed to the contralateral limb, respective markers were at their highest position.
Citation: American Journal of Veterinary Research 76, 12; 10.2460/ajvr.76.12.1031
Representative plots for body weight–standardized vertical ground reaction forces (A and B) and mean vertical height (C and D) throughout a stride for an orthopedically normal Icelandic horse ridden on a treadmill at a trot (A and C) and tölt (B and D) at the same velocity (3.4 m/s). In panels A and B, values for the left forelimb (blue), right forelimb (red), left hind limb (cyan), and right hind limb (green) are displayed separately. In panels C and D, y-axis values represent mean vertical height and ROM for the rider's hip (magenta; mean of values for both tuber coxae markers) and horse's height at the sacrum (black; tuber sacrale) and shoulder (green; mean values for both markers placed at the left and right tuberculum majus), each centered at the given stride mean value. Stride-standardized curves represent a mean of 25 strides for an interval of 1.5 strides (150% of StrD). Fore- and hind limbs behaved in a similar fashion for both gaits with respect to Fzpeak and vertical movement of the trunk: at Fzpeak of each limb, the associated proximal body markers (shoulder or sacrum) were at their lowest position, whereas when limb support changed to the contralateral limb, respective markers were at their highest position.
Citation: American Journal of Veterinary Research 76, 12; 10.2460/ajvr.76.12.1031
Representative plots for body weight–standardized vertical ground reaction forces (A and B) and mean vertical height (C and D) throughout a stride for an orthopedically normal Icelandic horse ridden on a treadmill at a trot (A and C) and tölt (B and D) at the same velocity (3.4 m/s). In panels A and B, values for the left forelimb (blue), right forelimb (red), left hind limb (cyan), and right hind limb (green) are displayed separately. In panels C and D, y-axis values represent mean vertical height and ROM for the rider's hip (magenta; mean of values for both tuber coxae markers) and horse's height at the sacrum (black; tuber sacrale) and shoulder (green; mean values for both markers placed at the left and right tuberculum majus), each centered at the given stride mean value. Stride-standardized curves represent a mean of 25 strides for an interval of 1.5 strides (150% of StrD). Fore- and hind limbs behaved in a similar fashion for both gaits with respect to Fzpeak and vertical movement of the trunk: at Fzpeak of each limb, the associated proximal body markers (shoulder or sacrum) were at their lowest position, whereas when limb support changed to the contralateral limb, respective markers were at their highest position.
Citation: American Journal of Veterinary Research 76, 12; 10.2460/ajvr.76.12.1031
Mean segmental lengths of hind limbs for 12 horses ridden on a treadmill at a tölt and trot at the same velocity (3.4 m/s) at time of limb first contact (FC) and at time of Fzpeak. Lengths represented are based on the height of the markers placed at the sacrum (light gray), stifle joint (medium gray), and fetlock joint (black). Absolute heights of the indicated marker locations at time of Fzpeak and the differences between the gaits are provided in Table 2. All absolute heights decreased during load onset (limb FC to Fzpeak). This decrease was primarily attributable to flexion of the fetlock joint and, to a minor extent, flexion of the tarsal joint (height difference of stifle joint – fetlock joint). Conversely, the length of the segment representing flexion of the hip joint (height difference of sacrum – stifle joint) slightly increased. *Indicated value is significantly (P < 0.05) larger than the value at limb FC. †Indicated value is significantly (P < 0.05) smaller than the value at limb FC.
Citation: American Journal of Veterinary Research 76, 12; 10.2460/ajvr.76.12.1031
Mean segmental lengths of hind limbs for 12 horses ridden on a treadmill at a tölt and trot at the same velocity (3.4 m/s) at time of limb first contact (FC) and at time of Fzpeak. Lengths represented are based on the height of the markers placed at the sacrum (light gray), stifle joint (medium gray), and fetlock joint (black). Absolute heights of the indicated marker locations at time of Fzpeak and the differences between the gaits are provided in Table 2. All absolute heights decreased during load onset (limb FC to Fzpeak). This decrease was primarily attributable to flexion of the fetlock joint and, to a minor extent, flexion of the tarsal joint (height difference of stifle joint – fetlock joint). Conversely, the length of the segment representing flexion of the hip joint (height difference of sacrum – stifle joint) slightly increased. *Indicated value is significantly (P < 0.05) larger than the value at limb FC. †Indicated value is significantly (P < 0.05) smaller than the value at limb FC.
Citation: American Journal of Veterinary Research 76, 12; 10.2460/ajvr.76.12.1031
Mean segmental lengths of hind limbs for 12 horses ridden on a treadmill at a tölt and trot at the same velocity (3.4 m/s) at time of limb first contact (FC) and at time of Fzpeak. Lengths represented are based on the height of the markers placed at the sacrum (light gray), stifle joint (medium gray), and fetlock joint (black). Absolute heights of the indicated marker locations at time of Fzpeak and the differences between the gaits are provided in Table 2. All absolute heights decreased during load onset (limb FC to Fzpeak). This decrease was primarily attributable to flexion of the fetlock joint and, to a minor extent, flexion of the tarsal joint (height difference of stifle joint – fetlock joint). Conversely, the length of the segment representing flexion of the hip joint (height difference of sacrum – stifle joint) slightly increased. *Indicated value is significantly (P < 0.05) larger than the value at limb FC. †Indicated value is significantly (P < 0.05) smaller than the value at limb FC.
Citation: American Journal of Veterinary Research 76, 12; 10.2460/ajvr.76.12.1031
Discussion
The present study involved comparisons of ground reaction force, limb movement, and posture of Icelandic horses at a tölt and trot on a treadmill at V1 (3.4 m/s) and V2 (3.9 m/s). The purpose was to characterize various properties of each gait as well as to evaluate loading of the distal portion of the limbs to provide data by which its impact on orthopedic health could be estimated. Compared with results of a study7 involving overground gait analysis (with force plates) of Icelandic horses tölting at V1, values obtained via treadmill in the present study for limb forces, impulses, stance durations, and balance between the fore- and hind limbs were similar.
Studies conducted with horses trotting on the ground13,14 and on treadmills15,16 have revealed that the trot is a 2-beat gait with a suspension phase, whereas the tölt has a 4-beat rhythm and the airborne phase is missing or is only observed at high speeds.2,5,6,9 In the present study, generally large kinetic and kinematic dissimilarities were evident between the 2 gaits. Differences were comparable at V1 and V2, even though most horses had a suspension phase only when trotting at V2. Rider position was the same for both gaits at V1; therefore, for purposes of gait comparison, we chose to use data obtained at V1. With horses trotting at V2, the rider was positioned farther cranially than in the other test conditions.
Stride rate at a tölt was higher than at a trot at the same treadmill speed. This resulted in a shorter StrD and stride length and consequently in lower total impulse during a stride cycle. Limb stance durations and Fzpeak differed not only between gaits but also between fore- and hind limbs within both gaits. At tölt, a mean forelimb StDrel of 47% with a transverse suspension (suspension of contralateral limbs) in the present study indicated a running gait; whereas for the hind limbs, an StDrel of 51% indicated a walking gait.17 The present study, however, was the first to reveal evidence that at a tölt, vertical movement of the horse's trunk segments in relation to its limb positions was consistent with the bouncing mechanism of a running gait. The lowest height at the shoulder and sacrum of horses in the present study was observed concurrently with Fzpeak (at approx midstance; Figure 1). The spring-like mechanisms of the fore- and hind limbs were phase-shifted to each other in time by approximately 25% of StrD, resulting in a slight rocking type of back movement (ie, pitching motion).18 This observation contrasted with that for the trot, during which the spring-like movement of the fore- and hind limb was synchronous. The pitching motion of the horses’ backs at a tölt was different from the motion observed at a walk, which has been characterized by a high shoulder and high pelvis position at midstance of the forelimb and hind limb, respectively (inverted pendulum).18 Because use of StDrel frequently fails to reliably discriminate between walking and running dynamics, differentiation between the 2 gaits should be preferably based on the phase difference of kinetic and potential energies.17
At a tölt versus trot, forelimb Fzpeak was notably higher (by 7.2%), despite lower limb impulses resulting from the shorter StrD in the present study. This could be explained, in part, by the shorter StDrel of the forelimbs, which was owing to a transverse forelimb suspension more than twice as long at a tölt than at a trot (2.7% of StrD vs 1.2% of StrD). For similar force profiles, Fzpeak is known to change inversely proportionally to StDrel.19,20 Therefore, for the observed 3.1% shorter StDrel at a tölt, an increase in force of the same amount would be expected because the balance between forelimb and hind limb impulse remained constant. The much higher Fzpeak might be explained by the different forelimb force profiles of the 2 gaits. At a tölt, profiles were slender and had a spiked apex, compared with the blunter shape of profiles obtained at a trot. Because the relationship between Fzpeak and StDrel is stronger when force profiles appear as spikes,20 Fzpeak at a tölt was overproportionally higher than at a trot. Taking into account the shorter stride cycle in tölting horses, force onset and decline were faster at a tölt (Figure 1). A possible explanation for these differences might be the higher carriage of the horse's head and neck. A higher HNP is known to restrict craniocaudal forelimb mobility,21,22 which, in the present study, resulted in shorter forelimb total proretraction angle and length. This in turn resulted in a higher stride rate and a prolonged phase of relative forelimb transverse suspension. Additionally, forelimb stiffness at a tölt was higher than at a trot because there was less limb compression despite the higher Fzpeak. Higher limb stiffness and shorter ground contact duration transmit the impulse more directly to the ground, generating force profiles with more spikes and thus higher peaks.
Higher forelimb flight arc in combination with vertical displacement of the trunk caused by the longer transverse suspension phase at a tölt led to the imposing forelimb protraction arc during the swing phase in the present study. This higher flight arc at a tölt required faster hoof protraction velocities because swing duration was approximately 6% briefer. However, vertical stride impulse in horses when tölting was not greater than what would be expected from a horse's body mass and StrD, making a considerable active component of downward hoof acceleration before impact unlikely. Hind limb force profiles at a tölt revealed an 11.3% longer StDrel and 11.8% lower Fzpeak than at a trot and thus a proportionally inverse relationship to that observed for the forelimbs.
Comparison of forces and stance durations within both gaits revealed that Fzpeak was remarkably higher in the forelimbs than in the hind limbs at a tölt (Fzpeak at the forelimbs/Fzpeak at the hind limbs = 1.44) than at a trot (1.19). This difference was mainly attributable to the lower hind limb Fzpeak at a tölt and those differences, in turn, to the longer StDrel (≥ 50% of StrD). At a tölt, forelimb stance durations were shorter than in the hind limbs; conversely, at a trot, stance durations were shorter in the hind limbs.
Compared with at a trot, dorsoventral excursion of the horse's sacrum at a tölt was reduced by approximately a third and was even smaller than that at a walk, as described by Waldern et al.8 Concurrently, ROM (z-axis) of the rider's hip at a tölt was smaller than that of the horse's sacrum and only half as much as that previously reported for a trot,8 which explained the high comfort factor of the tölt for the rider, even at fast speeds (Figure 1 and Table 2). This finding corresponded with the smaller vertical excursion of the calculated center of mass at a tölt versus a trot reported by other investigators.6,23
Tölting of Icelandic horses requires increased collection and can be achieved by simultaneously applying encouraging and restraining riding aids in combination with raising of the horse's head and neck, thereby inducing a weight shift to the hind limbs.24 In this context, collection is defined as a slight elevation of the head and neck combined with a lowering of the hind quarters caused by an increased flexion of tarsal, stifle, and hip joints during the stance phase (Hankenbeugung).13,14,25,26 With respect to posture of the hind quarters in the horses of the present study, stifle height at mid stance (as a measurement of the combined effect of fetlock joint hyperextension and tarsal joint flexion) did not differ between both gaits. Concomitantly, the height of the sacrum was even higher at a tölt versus trot, which was attributable to the considerably lower hind limb Fzpeak (ie, less-loaded limb spring). Moreover and functionally even more important, there was no shift of impulse to the hind limbs, even when horses carried their heads approximately 125 mm higher at a tölt than at a trot. This contrasted with reported measurements for warmblood horses at a trot, for which the influence of the HNP position has been investigated.21,22 In warmblood horses, elevation of the head and neck leads to a weight shift toward the hind limbs when ridden (1.8%) and when not ridden (1.0%).21,22 Compared with those findings, no evidence of increased collection at a tölt could be confirmed in the present study. However, tölting Icelandic horses stepped farther toward their center of mass, resulting in longer hind limb protraction length at V2 and a steeper croup angle at both V1 and V2.
The longer StDrel and total proretraction length in the hind limbs and the lack of a hind limb transverse suspension at slow tölting speeds, compared with at a trot, might resemble so-called Groucho running. This manner of running is characterized by an increase in the degree of stifle joint flexion, resulting in a decrease in spring stiffness of the limbs, increase in contact durations, prolongation of stance length, decrease in the duration of or lack of the suspension phase, and increase in energy requirements.27 However, there was no difference in stifle joint height at midstance between the tölt and trot in the present study, and total hind limb compression during load onset at a tölt was only approximately two-thirds as much as at a trot (Figure 2). The hind limb spring constant at a tölt, as estimated from Fzpeak, was consequently approximately 32% (V2) to 50% (V1) higher, caused by a stiffer proximal segment (hip joint) than at a trot. Conversely, the spring constant of the fetlock joint remained equal for all test conditions. Therefore, despite the longer relative ground contact durations for the hind limbs at a tölt (V1, 51.3% of StrD; V2, 49.4% of StrD) than at a trot (V1, 46.1% of StrD; V2, 43.9% of StrD), there was no evidence that the tölt was more similar to a Groucho run than the trot in the Icelandic horses of the present study. However, a difference might be expected in comparison to warmblood horses, which have considerably shorter hind limb contact durations (< 40% of StrD) than do Icelandic horses, at a trot.16,18
At medium trotting speed (V2), riders assumed a more forward seat position than they did during the other 3 test conditions in the study reported here. This position was chosen to maintain a maximally stable rider position with least disturbance of the horse, particularly when horses had high forelimb protraction arcs. The more forward position likely caused the significant shift of impulse to the forelimbs (2.7% increase of impulse carried by the forelimbs), compared with at a tölt at the corresponding speed and with the impulse balance at V1 at both gaits. This observation corresponds to results of a previous study,28 in which even slight shifts (some centimeters in magnitude) of the rider's position influenced impulse balance between the fore- and hind limbs in Icelandic horses.
In the present study, the trot of Icelandic horses was different in some respects from the gait pattern of 3-gaited horses such as warmblood breeds. The brief or lacking airborne phase of most of the Icelandic horses was conspicuous, given that both trotting speeds were rather high in relation to the small size of the horses. Icelandic horses trotting at 3.4 m/s had comparable dynamics (equal Froude numbers)29 to 1.7-m-tall warmblood horses at 3.8 m/s.16 In warmblood breeds, durations of suspension phases increase with increasing velocity and have been observed even at slow relative speeds for horses trotting without16,30 or with a rider.31 Given the smaller body size of Icelandic horses relative to warmblood breeds, their briefer or missing suspension phases are a consequence of the relatively longer StDrel and total proretraction length. As has been reported for warmblood horses,15 forelimb stance durations in Icelandic horses were longer than in the hind limbs, but their durations were closer matched to each other than in warmblood breeds.16,21
The trot of the Icelandic horses on the treadmill was not clearly 2-beated in the present study. Fore- and hind limb impact was dissociated at V1 and V2 because of delayed placement of the diagonal hind limb (delay of 2.5% of StrD at V1 and 3.3% of StrD at V2). In contrast, warmblood horses trotting at V2 on a treadmill have a clear acoustic 2-beat footfall rhythm and a perfect synchronous impact of diagonal limbs (delay of 0.07% of StrD).16 Mean values of temporal variables representing footfall patterns (ie, StDrel, suspension duration, timing of advanced placement, timing of advanced completion, and ipsilateral step duration) of trotting Icelandic horses in the present study had SDs between 1.8% and 4.3% of StrD, which represented almost twice the variation of values for warmblood horses (1.1% to 2.3% of StrD).16
A large degree of variation in gait pattern is common in Icelandic horses, not only for the trot but even more for the tölt.2,9 Particularly in 5-gaited horses, correctness of footfall rhythm is not as stable as in 3-gaited horses. This difference might be explained by findings of genetic research32 that a premature stop codon exists in the DMRT3 gene, which has been linked to additional gaits but also to inferior scores in gait competitions for the trot and canter. Footfall rhythm in Icelandic horses is also highly sensitive to external influences, such as speed, horse imbalance, riding errors, and shoeing changes.9,24 Despite these differences between warmblood and Icelandic horses trotting at equal relative velocity, similarities existed in the percentage of total impulse carried by both forelimbs, but Fzpeak values were lower (decrease of 21% to 23%) for the Icelandic horses in the present study because of a longer StDrel (increase of 21% to 26%) in fore- and hind limbs.15,16,30
Investigation of whether the tölt was energetically more advantageous than the trot at the same speed was not a primary purpose of the present study. However, the study data allowed estimation of energetic differences between the 2 gaits. The rate of energy consumption (Ėrun; power) per Newton of body weight for running is inversely proportional to the mean limb contact duration (tc; calculated from the 4 individual limb contact durations within a stride) such that Ėrun is proportional to 1/tc.33 From the overall mean stance durations of fore- and hind limbs in the present study (trot at V1, 0.286 seconds; trot at V2, 0.260 seconds; tölt at V1, 0.271 seconds; and tölt at V2, 0.248 seconds), the metabolic power needed can be calculated, revealing it was approximately 5.5% higher at a tölt than at a trot at V1. This drawback became slightly smaller (4.8% higher than at a trot) for the faster speed.
Additional energy is required to compensate for horizontal deceleration, caused by the braking forces of the leading limb during limb impact, and subsequent reacceleration of the body by the trailing limb.34 This component of transport energy, which is much smaller than Ėrun because of the considerably lower horizontal ground reaction forces (10% to 20% of Fzpeak),7 is higher for smaller angles of attack of the landing limb.34 In the present study, horizontal energy loss might have been lower at a tölt as a result of the smaller proretraction angle of the forelimbs. In total, these calculations might contradict the expectation that the tölt gait may have developed because it confers an energetic advantage.35 However, they are in agreement with rider observations that over long distances, even 5-gaited horses appear to prefer the trot over the tölt when becoming fatigued.36
The high HNP characteristic of tölting horses is believed to predispose horses to back pain because of an extension of the thoracolumbar portion of the spinal column as observed with warmblood horses.37,38 Our data showed that Icelandic horses had a less-extended caudal back angle when tölting versus trotting at the same speed, despite the higher HNP at the tölt. This finding is in agreement with that of another study28 involving Icelandic horses, which revealed less bridging pressure patterns under the saddle at a tölt (high HNP) than at a walk (low HNP). It follows that elevation of the head and neck might not lead to the same degree of back extension in Icelandic horses as in other breeds, possibly because of the shorter back and higher muscle tension in Icelandic horses. In accordance with our hypothesis, the high forelimb action combined with the higher HNP at a tölt led to shorter StDrel and thus to an increase in Fzpeak of approximately 7.2%. In addition, the faster force onset at a tölt versus trot led to an increase in loading of the forelimbs and might be associated with a higher risk of injury, particularly in the long term and at high speeds.
The fairly high prevalence of bone spavin in Icelandic horses has led to the hypothesis that overload of the hind limbs at a tölt might be a contributing factor. However, findings in the present study suggested that the tölt rather than trot at the same speeds induced a 7.2% increase in loading of the forelimbs, whereas the hind limbs were even less loaded with regard to Fzpeak (11.8% decrease) and limb impulses (8.6% decrease). These data were in agreement with a previous estimation7 and might explain the reported lack of a link between the ability to tölt and the prevalence of bone spavin in Icelandic horses.39
In the study reported here, the higher HNP of Icelandic horses at a tölt versus trot, together with the shorter StDrel, resulted in increases in forelimb flight arc and Fzpeak; conversely, longer StDrel in the hind limbs led to lower Fzpeak. Despite the higher HNP in horses at a tölt, no measurable weight shift to the hind limbs was identified, nor was there any clear evidence that tölting required more collection than trotting. The extent to which tölting of Icelandic horses with a low HNP similar to that used when trotting40 may influence the measured kinetic and kinematic variables remains to be evaluated. Overall limb timing varied considerably among and within Icelandic horses at both gaits, which was characterized by the propensity toward lateral coupling at a tölt and a slight 4-beat rhythm at a trot, compared with patterns in warmblood horses. Additional differences from trotting warmblood horses included a longer StDrel in Icelandic horses, resulting in lower Fzpeak and a briefer or lacking suspension phase. Energetic estimations suggested that tölting at V1 and V2 was energetically less advantageous than trotting at equal speeds.
Acknowledgments
This manuscript represents a portion of a thesis submitted by Dr. Waldern to the Graduate School for Cellular and Biomedical Sciences, University of Bern, Switzerland as partial fulfillment of the requirements for a Doctor of Philosophy degree.
Supported by the Stiftung Forschung für das Pferd, Haldimann Stiftung, Swiss Metall Union, Stiftung Temperatio, Swiss Veterinary Office, and Islandpferde Vereinigung Schweiz. Presented in abstract form at the 7th International Conference of Canine and Equine Locomotion, Strömsholm, Sweden, June 2012.
The authors thank Bea Rusterholz, Bernhard Podlech, and Margrit Rusterholz for technical assistance.
ABBREVIATIONS
| Fzpeak | Peak vertical force |
| HNP | Head-neck position |
| ROM | Range of motion |
| StrD | Stride duration |
| StDrel | Stance duration relative to stride duration |
| V1 | Slow speed |
| V2 | Medium speed |
Footnotes
Mustang 2000, Graber AG, Fahrwangen, Switzerland.
Qualysis Track Manager, version 2.8, Qualisys, Gothenburg, Sweden.
Oqus 300, Qualysis, Gothenburg, Sweden.
MathWorks Inc, Natick, Mass.
SigmaStat, version 3.5, SPSS Inc, Chicago, Ill.
References
1. Barrey E., Gaits and interlimb coordination. In: Back W, Clayton HM, eds. Equine locomotion. 2nd ed. Philadelphia: Saunders Elsevier, 2013;85–98.
2. Zips S, Peham C, Scheidl M, et al., Motion pattern of the toelt of Icelandic horses at different speeds. Equine Vet J Suppl 200;(33): 109–111.
3. Rumpler B, Riha A, Licka T, et al. Influence of shoes with different weights on the motion of the limbs in Icelandic horses during toelt at different speeds. Equine Vet J Suppl 2010;(38): 451–454.
4. Nicodemus MC, Clayton HM. Temporal variables of four-beat, stepping gaits of gaited horses. Appl Anim Behav Sci 2003; 80: 133–142.
5. Robilliard JJ, Pfau T, Wilson AM. Gait characterisation and classification in horses. J Exp Biol 2007; 210: 187–197.
6. Biknevicius AR, Mullineaux DR, Clayton HM. Locomotor mechanics of the tölt in Icelandic horses. Am J Vet Res 2006; 67: 1505–1510.
7. Biknevicius AR, Mullineaux DR, Clayton HM. Ground reaction forces and limb function in tolting Icelandic horses. Equine Vet J 2004;36: 743–747
8. Waldern NM, Wiestner T, Ramseier LC, et al., Effects of shoeing on limb movement and ground reaction forces in Icelandic horses at walk, tolt and trot. Vet J 2013; 198 (suppl 1): e103–e108.
9. Weishaupt MA, Waldern NM, Amport C, et al., Effects of shoeing on intra- and inter-limb coordination and movement consistency in Icelandic horses at walk, tolt and trot. Vet J 2013;198, (suppl 1):e109–e113.
10. Feldmann A, Rostock A., Sattelung und zäumung. In: Feldmann A, Rostock A, eds. Islandpferde reitlehr. Bonn, Germany: Thenee Druck AG, 1986;32–41.
11. Harman J. Tack and saddle fit. Vet Clin North Am Equine Pract 1999; 15: 247–261.
12. Weishaupt MA, Hogg HP, Wiestner T, et al., Instrumented treadmill for measuring vertical ground reaction forces in horses. Am J Vet Res 2002; 63: 520–527.
13. Holmström M, Fredricson I, Drevemo S. Biokinematic effects of collection on the trotting gaits in the elite dressage horse. Equine Vet J 1995; 27: 281–287.
14. Clayton HM. Comparison of the stride kinematics of the collected, working, medium and extended trot in horses. Equine Vet J 1994; 26: 230–234.
15. Weishaupt MA, Wiestner T, Hogg HP, et al., Vertical ground reaction force-time histories of sound warmblood horses trotting on a treadmill. Vet J 2004; 168: 304–311.
16. Weishaupt MA, Hogg HP, Auer JA, et al., Velocity-dependent changes of time, force and spatial parameters in warmblood horses walking and trotting on a treadmill. Equine Vet J Suppl 2010;(38): 530–537.
17. Biknevicius AR, Reilly SM. Correlation of symmetrical gaits and whole body mechanics: debunking myths in locomotor biodynamics. J Exp Zool A Comp Exp Biol 2006; 305: 923–934.
18. von Peinen K, Wiestner T, Bogisch S, et al., Relationship between the forces acting on the horse's back and the movements of rider and horse while walking on a treadmill. Equine Vet J 2009; 41: 285–291.
19. Alexander RM, Maloiy GMO, Hunter B, et al., Mechanical stresses in fast locomotion of buffalo (Syncerus caffer) and elephant (Loxodonta africana). J Zool 1979; 189: 135–144.
20. Witte TH, Knill K, Wilson AM. Determination of peak vertical ground reaction force from duty factor in the horse (Equus caballus). J Exp Biol 2004; 207: 3639–3648.
21. Weishaupt MA, Wiestner T, von Peinen K, et al., Effect of head and neck position on vertical ground reaction forces and interlimb coordination in the dressage horse ridden at walk and trot on a treadmill. Equine Vet J Suppl 2006;(36)387–392.
22. Waldern NM, Wiestner T, von Peinen K, et al., Influence of different head-neck positions on vertical ground reaction forces, linear and time parameters in the unridden horse walking and trotting on a treadmill. Equine Vet J 2009; 41: 268–273.
23. Buchner HH, Obermuller S, Scheidl M. Body centre of mass movement in the sound horse. Vet J 2000; 160: 225–234.
24. Feldmann A, Rostock A., Tölt. In: Feldmann A, Rostock A, eds. Islandpferde reitlehre. Bonn, Germany: Thenee Druck AG, 1986;249–271.
25. Weishaupt MA, Bystrom A, von Peinen K, et al., Kinetics and kinematics of the passage. Equine Vet J 2009; 41: 263–267.
26. Egenvall A, Byström A, Weishaupt MA, et al., Horse-rider interaction. In: Back W, Clayton HM, eds. Equine locomotion. 2nd ed. Philadelphia: Saunders Elsevier, 2013;341–368.
27. McMahon TA, Valiant G, Frederick EC. Groucho running. J Appl Physiol 1987; 62: 2326–2337.
28. Ramseier LC, Waldern NM, Wiestner T, et al., Saddle pressure distributions of three saddles used in Icelandic horses and their effects on ground reaction forces, limb movements and rider positions at walk and tölt. Vet J 2013; 198 (suppl 1): e81–e87.
29. Hof AL. Scaling gait data to body size. Gait Posture 1996; 4: 222–223.
30. Dutto DJ, Hoyt DF, Wickler SJ. Ground reaction forces during level, incline, and decline trotting. Comp Integr Biol 2002; 42: 1223.
31. Bogisch S, Geser-von Peinen K, Wiestner T, et al., Influence of velocity on horse and rider movement and resulting saddle forces at walk and trot. Comp Exerc Physiol 2014; 10: 23–32.
32. Andersson LS, Larhammar M, Memic F, et al., Mutations in DMRT3 affect locomotion in horses and spinal circuit function in mice. Nature 2012; 488: 642–646.
33. Kram R, Taylor CR. Energetics of running—a new perspective. Nature 1990; 346: 265–267.
34. Bertram JE. Gait as solution, but what is the problem? Exploring cost, economy and compromise in locomotion. Vet J 2013; 198 (suppl 1): e3–e8.
35. Wickler SJ, Hoyt DF, Cogger EA, et al., The energetics of the trot-gallop transition. J Exp Biol 2003; 206: 1557–1564.
36. Ingolfsdottir IB. Training im Gelände - Sich die Natur zu Nutze machen. Eidfaxi 2013; 4: 70–75.
37. Rhodin M, Gomez Alvarez CB, Bystrom A, et al., The effect of different head and neck positions on the caudal back and hindlimb kinematics in the elite dressage horse at trot. Equine Vet J 2009; 41: 274–279.
38. Gómez Alvarez CB, Rhodin M, Bobbert MF, et al., The effect of head and neck position on the thoracolumbar kinematics in the unridden horse. Equine Vet J Suppl 2006;(36): 445–451.
39. Axelsson M, Bjornsdottir S, Eksell P, et al., Risk factors associated with hindlimb lameness and degenerative joint disease in the distal tarsus of Icelandic horses. Equine Vet J 2001; 33: 84–90.
40. Schwörer-Haag A, Haag T., Die Erfolgstreppe. In: Schwörer-Haag A, Haag T, eds. Islandpferde besser reiten und richtig ausbilden. Stuttgart, Germany: Franckh-Kosmos, 2003;95–115.
