Forelimb Hoof Landing Velocities in Treadmill Trotting and Galloping Horses Raoul F.
Reiser II,1 M.L.
Peterson,2 Christopher E.
Kawcak3 and C.
Wayne McIlwraith3 Department of Health & Exercise Science, Colorado State University, Fort Collins, CO 80523 2 Department of Mechanical Engineering, University of Maine, Orono, ME 04469 3 Department of Clinical Sciences, Colorado State University, Fort Collins, CO 80523 1 ABSTRACT Forelimb injury has been attributed to many factors.
A key component to understanding the injury mechanisms is an accurate knowledge of hoof landing velocity and the ensuing kinematics of hoof strike.
Five seconds of sagittal plane high-speed video were collected at 500 Hz from the right side of 22 mixed breed/grade horses while trotting and galloping on a treadmill.
A reflective marker was placed on the center of the medial side of the hoof of the left forelimb with three hoof strikes digitized and averaged.
Nine led with this limb during the gallop while 13 led with the right limb (treadmill speed trot = 4.4 m/s, gallop = 10.1 m/s).
Differences were assessed at p < 0.05.
Vertical hoof velocities at impact were significantly greater at gallop (3.7+/-0.9 m/s) relative to trot (-1.9+/-0.5 m/s).
No significant differences in vertical hoof velocity were found between lead and trail limb at either speed.
Horizontal hoof velocity at impact was significantly greater in the lead limb (-7.4+/-1.4 m/s) relative to the trail limb (-5.6+/-1.5 m/s) at gallop as well as both being significantly greater than during the trot (-2.5+/-0.7 m/s).
These results provide a starting point for both computer and mechanical models simulating hoof strike with the goal of reducing equine injuries.
INTRODUCTION Musculoskeletal injuries of the lead forelimb are a major problem faced by the equine racing injury.
These injuries are often severe and catastrophic .
A number of factors including unsoundness, conformation, possible inherited traits, and racetrack conditions may be associated with these injuries .
As a first attempt to assess the possible role of the racetrack, Reiser et al.  simulated the surface properties with a relatively simplistic spring-mass-dashpot model.
Results indicated that surface stiffness could significantly alter track deflection (compaction and consolidation), stance phase duration, energy dissipation, peak vertical ground reaction force, and peak loading rate while soil damping ratios had much more subtle effects on the same parameters.
Important to the findings of this model was that the inputs to the model of initial velocity of the hoof as well as the area of contact (which is a combination of hoof size as well as orientation to the soil) were estimated.
Hoof velocity data is limited to the trotting horse  and hoof orientation data limited to two Thoroughbreds galloping on a track at 15.5 m/s while wearing instrumented shoes .
Therefore, much could be improved in this simulation if more precise input parameters were available.
In combination with musculoskeletal modeling, efforts are now underway by the authors of this paper to physically assess the role of the racetrack .
An integral piece of this effort is the development of an impact device that will load the track in the same manner as the racehorse.
Due to soils being non-linear viscoelastic materials, it is critical that the simulated hoof velocity and orientation match that of a racehorse.
This leads to the working hypothesis that accurate hoof velocity and orientation of the lead forelimb at gallop will improve the ability to assess racetrack surfaces as well as improve the understanding of loads on the horse and how they are transmitted through the limb.
Unfortunately, due to the soil compaction that occurs and the ensuing visible obstruction of the hoof, measuring velocity and orientation of the hoof as it strikes the track is not trivial.
A means to follow the hoof as it penetrates the track surface is necessary.
While this device is being designed and tested, it is important to make parallel progress on the track testing device.
This may be accomplished if more confidence could be placed in the relative magnitudes of the horizontal and vertical velocities of the hoof as it strikes the track.
Therefore, the goal of the investigation was to measure hoof velocities while galloping on a treadmill to assist with the design of a racetrack surface testing device.
Even though it is anticipated that horses gallop slightly differently on a treadmill compared to a soil track, the treadmill is a controlled condition where less visual obstruction of the hoof occurs during impact and the ensuing stance phase.
Since data on the trotting horse is currently available, this condition was also measured in order to benchmark the data collection.
METHODS Twenty-two mixed breed/grade horses were analyzed (average age = 2.7 +/- 0.6 yrs and mass = 400 +/- 42 kg).
These horses had been acclimated to trotting and galloping on the treadmill over the previous eight weeks prior to data collection.
They were also part of a larger study that required a bone chip be created surgically in the right or left carpus.
At the time of data collection there was no visible lameness in any of the horses.
Hooves had basic, square trimming with no shoes.
Data was collected during one of their five weekly treadmill sessions where they would start by trotting for two minutes at a speed of 12.9 to 19.3 kph, then galloped for two minutes at a speed of 40.2 to 53.1 kph, and then finished by trotting again for two minutes at the original pace.
Horses trotted and galloped at constant speeds that were comfortable for them, so not all horses trotted and galloped at the same speed. 2 Before starting their treadmill session on the day of data collection, a small piece of retro-reflective tape (~ 1 cm in area) was attached to the medial side of the front left hoof, approximating its center of mass as viewed in the sagittal plane.
A 500 fps high-speed camera with a 20x shutter (Motion Meter, Redlake MASD, Inc., San Diego, CA) was mounted on a tripod approximately three meters to the right of the horse at the level of the forelimb with the focal line perpendicular to the sagittal plane of the horse.
A high-intensity spotlight was then utilized to illuminate the marker in the field of view.
While filming the right hoof from this location would have been more desirable, the bed of the treadmill was slightly below its elevated frame.
This obstructed the line of sight to the right hoof at impact, so it was necessary to take data of the left hoof instead.
Even though the center of mass of the left hoof remained visible at all times, it was not possible to see the orientation of the bottom of the hoof after impact in all horses, so hoof orientation was not examined. After coming to steady state during the initial trot phase of the training session and during the gallop phase, five seconds of digital video were collected and downloaded to S-VHS tape for later processing.
From each video clip three consecutive hoof strikes were isolated in both the trot and gallop and automatically digitized (Motus, Peak Performance Technologies, Inc., Englewood, CO).
Raw coordinate data were low-pass filtered with a cutoff frequency of 180 Hz (4th – order, Butterworth).
Nd Velocities and accelerations were then computed via the 2 – order central difference method.
The orientation of the axes was such that positive was anteriorly directed in the horizontal and upward vertically relative to a fixed reference frame (ie, the motion of the treadmill was not factored into the horizontal data).
Video clips were cropped for each hoof strike so that digitization could begin before the hoof reached its most anterior position in the flight phase of its gait cycle.
At this time the hoof was mainly moving horizontally, so there was no danger of missing the peak vertical velocity at initial hoof strike (Figure 1).
The video clip was again cropped just before the hoof left the field of view of the camera while it was firmly planted on the treadmill.
For the trot this yielded a video clip that was roughly 0.22 seconds long with ~0.09 seconds before initial contact with the treadmill and ~0.13 seconds after.
Since the hoof was moving much faster in the gallop, video clips were only about 0.14 seconds long with ~0.09 seconds before initial contact with the treadmill and ~0.05 seconds after. 0.13 Vertical Position (m) 0.09 0.05 0.01 -0.03 -0.07 -0.2 Trot Vertical Position (m) 0.13 0.09 0.05 0.01 -0.03 -0.07 -0.2 Gallop 0 0.2 Horizontal Position (m) 0.4 0.6 — 0.4 0.6 A B Figure 1.
Exemplar center of hoof marker path during the trot (A) and gallop (B) with an arrow indicating direction of motion along the path during the course of the hoof strike.
Horizontal and vertical positions were referenced relative to a fixed global origin defined by the calibration frame that was digitized to scale the video images appropriately.
The surface of the treadmill moved in the negative horizontal direction.
Note: the treadmill surface deflected slightly during hoof strike and the ensuing load bearing so that the hoof does not travel only in the horizontal direction after coming in full contact with the treadmill.
Since the sole of the hoof was obscured in the video image, a method utilizing either the position, velocity, or acceleration of the center of mass marker was needed to identify initial hoof strike (Figure 2).
While for most horses it appeared that the peak downward velocity occurred right before initial contact, for some horses peak downward velocity occurred slightly sooner.
Horizontal velocity provided no consistent features from which initial hoof contact could be detected.
Fortunately, hoof strike provided a consistently high positive spike in the vertical acceleration during both the trot and gallop.
As a result, initial hoof strike was marked at the time the spike in the vertical acceleration crossed the zero axis before peaking.
The peak value of the vertical acceleration was not chosen because it was probably associated with the time that the hoof made full contact with the treadmill, rather than initial contact.
The value of the vertical and horizontal hoof velocities were then extracted at this time of initial hoof contact.
The treadmill speed for each horse was also extracted from the digitized video by averaging a section of the horizontal velocity after it stabilized from the hoof strike. In order to reduce the variability of the measures, the three velocity parameters from three hoof strikes were averaged to create a representative value from each horse under each condition (trot and gallop).
Even though the main goal was to produce the value that could be used to further the design of the track testing device, which could be accomplished just by compiling the averages of each parameter, statistical comparisons were also conducted to evaluate the differences between lead and trail leg as well as trot to gallop.
Since lead leg was examined on some horses and trail on others, independent ttests were utilized for statistical comparisons of these two relationships.
To compare trot to gallop comparisons were made to the same horse under different conditions, so paired t-tests were utilized there.
All statistical differences were evaluated at p < 0.05. 0.6 0.5 0.4 Position (m) Position (m) 0.3 0.2 0.1 0 -0.1 -0.2 0 0.05 0.1 Time (s) 0.15 0.2 Trot 0.6 0.5 0.4 0.3 0.2 0.1 0 -0.1 -0.2 0 0.02 0.04 0.06 Time (s) 0.08 0.1 0.12 Gallop A D 4 2 Velocity (m/s) Velocity (m/s) 0 0.05 0.1 Time (s) 1000 800 Acceleration (m/s/s) Acceleration (m/s/s) 600 400 200 0 -200 -400 0.15 0.2 0 -2 -4 -6 -8 -10 4 2 0 -2 -4 -6 -8 -10 0 0.02 0.04 0.06 Time (s) 1000 800 600 400 200 0 -200 -400 0 0.05 0.1 Time (s) 0.15 Hor 0.2 Vert 0 0.02 0.04 0.06 Time (s) 0.08 0.1 Hor 0.12 Vert 0.08 0.1 0.12 B E C F Figure 2.
Exemplar vertical and horizontal hoof positions, velocities, and accelerations during the trot (A – C) and gallop (D – F).
The vertical line in each plot represents the start of the spike in vertical hoof acceleration that was used to mark the point of extraction for the vertical and horizontal hoof velocities at initial impact.
RESULTS Nine of the horses lead with their left forelimb during gallop (average age = 2.5 +/- 0.6 yrs and mass = 387 +/- 42 kg) while the left forelimb was the trail leg for the remaining 13 horses (average age = 2.8 +/- 0.7 yrs and mass = 409 +/- 42 kg).
This difference at gallop naturally selected the horses into two groups (Lead Leg and Trail Leg).
Even though there isn’t a distinction between a lead and trail leg during the trot, due to its symmetry, the groupings were maintained to ensure that no differences existed between groups when trotting.
There were no significant differences between age and mass between these two groups (p = 0.347 and 0.240, respectively).
Velocity results are summarized in Table 1.
Within the trot there were no differences between the Lead and Trail Leg groups in treadmill, vertical hoof, or horizontal hoof velocity (p = 0.093, 0.874, and 0.523, respectively).
Within the gallop there were no differences between the Lead and Trail Leg groups in the treadmill and vertical hoof velocity (p = 0.914 and 0.667, respectively).
However, the horizontal hoof velocity was significantly greater in magnitude at impact in the Lead Leg group compared to the Trail Leg (p = 0.011).
Between the trot and gallop, the treadmill and vertical hoof velocity was significantly greater in magnitude during the gallop (combined group, p < 0.001 for both).
Even though there was a difference in horizontal hoof velocity between the Lead and Trail Leg groups, both were significantly greater in magnitude compared to the trot (also p < 0.001 for both). Table 1.
Summary of Hoof Velocity Results (m/s) Trot Gallop Treadmill Vertical Horizontal Treadmill Vertical Horizontal Velocity Velocity Velocity Velocity Velocity Velocity Lead Leg -4.53 -1.87 -2.65 -10.18* -3.63* -7.40* [n=9] (0.25) (0.46) (0.75) (1.21) (0.75) (1.35) Trail Leg -4.31 -1.90 -2.46 -10.12* -3.80* -5.61* [n=13] (0.32) (0.47) (0.61) (1.26) (1.00) (1.54) Combined -4.40 -1.89 -2.54 -10.15* -3.73* [n=22] (0.31) (0.46) (0.66) (1.21) (0.89) () indicate standard deviations * p<0.001 between trot and gallop Note: Since horizontal velocity was different between the lead and trail leg in gallop, they were not combined DISCUSSION This investigation produced a novel set of vertical and horizontal hoof velocities for both the lead and trail limb of galloping horses on a treadmill at the time that the hoof makes initial contact with the treadmill starting the support phase of the limb.
In addition to data for the lead limb, which is more often injured , results were also generated for the trail limb.
While the vertical hoof velocity at impact was not different between the lead and trail limbs, the horizontal velocity was significantly less negative in the trail limb.
When combined with the speed of the treadmill, this indicates that the trail limb is moving faster relative to the ground than the lead limb.
This is a reasonable finding, since the trail limb hits the ground before the lead limb and the horse may be traveling slightly faster at this point.
If this difference is maintained when galloping on a dirt track, it may play a role in the injury process.
The greater horizontal velocity of the trail limb may be protective in nature, since it will help the hoof slide into the dirt rather than come to a more abrupt stop.
The data collected during the trot was for comparison purposes with the existing literature.
While the vertical velocities found here agree highly to those reported by Johnston et al. , the horizontal velocities are quite different.
At a trotting velocity of 4.4 m/s they predicted a horizontal hoof velocity of +0.9 m/s while we found it to be -2.5 m/s.
The difference in sign could result if they were reporting their findings relative to the treadmill rather than a fixed reference frame.
However, there is no indication in their report that this was the case.
Even if it is, there is still a relatively large difference in magnitude, since we found the hoof to be moving forward relative to the treadmill belt at a velocity of +1.86 m/s (-2.54 – -4.40).
Part of the difference in the horizontal velocities between this study and that by Johnston et al.  may be explained by differences in study subjects (18 month old Standardbred colts versus 2.7 yr old mixed breed/grade horses), sampling rate (250 fps versus 500 fps), the use of raw instead of filtered data, the tracking of markers not placed on the center of mass of the hoof, and the use of a linear regression equation working from the time of full support as a reference rather than a time more closely associated with the initial impact of the hoof.
How each of these differences affects the overall results is unclear, as is why the vertical velocities were so highly comparable (-2.0 versus -1.9 m/s) and the horizontal velocities so different.
The fact that our population had a surgical procedure creating a bone chip in the carpus was also considered as a potential factor that would create differences to that of Johnston et al. .
However, it would be anticipated that the bone chip would change the gait in such a way as to make the hoof velocity more comparable to the treadmill speed at impact.
If the hoof met the treadmill with less disparity in horizontal velocity, the impact force should be reduced – protecting the injured joint.
However, the differences with Johnston et al.  were actually in the opposite direction, suggesting that the bone chip did not play a role.
An additional confounding factor lies in the fact that the bone chip was created on the left side for half of the horses and on the right side for the other half.
Therefore, not all hoof velocities measured were on the previously injured side of the horse.
Finally, as previously mentioned, no visible lameness was apparent in any of the horses at the time of data collection.
Therefore, it appears that the bone chip procedure did not play a role in the data collected. It is also interesting to note that Johnston et al.  also trotted their colts at speeds approaching 10 m/s.
The vertical hoof velocities at impact for their trotting horses remains similar to that of our galloping horses (-4 versus -3.7 m/s).
The horizontal hoof velocities remained quite different, as they were in the trotting results.
The direction of motion appears to still be different as well as large differences in absolute magnitudes (+2 versus -7.4 and -5.6 m/s).
This suggests that vertical hoof velocity at impact may be related not to the type of gait, but rather the forward speed (at least when comparing the trot and gallop).
CONCLUSIONS While the results from the gallop provide design guidelines for the track testing device and potential modifications that could be implemented into models simulating hoof strike, hoof strike parameters from actual race speeds and track conditions are still necessary.
Race horses gallop at higher speeds than those examined here, and may use a slightly different gait as a result of breed characteristics and ground conditions (hard treadmill versus soft dirt track).
Therefore, continued efforts should be made to measure hoof velocities of race horses under race conditions.
Only then will both computer and mechanical models of forelimb hoof strike at a gallop be able to most accurately simulate the desired conditions.
ACKNOWLEDGMENTS Special thanks to Brandon Santoni for his assistance with data collection and processing and to the College of Health Sciences at the University of Wyoming (Laramie, WY) for allowing use of the high-speed camera.
REFERENCES  Johnston, C., Hjerten, G.
And Drevemo, S.
Hoof landing velocities in trotting horses.
Equine Exercise Physiology, 3, 167172, 1991.  Mohammed, H.O., Hill, T.
And Lowe, J.
Veterinarian, 82, 331-341, 1992.
Risk of severity of limb injuries in racing Thoroughbred horses.
Cornell  Peloso, J.G., Mundy, G.D.
And Cohen, N.D.
Prevalence and factors associated with musculoskeletal racing injuries of Thoroughbreds.
Assoc. 204, 620-626, 1994.  Peterson, M.L., Reiser, R.F.
II, & McIlwraith, C.W.
System development for in-situ characterization of horse racing track surfaces.
In: The Engineering of Sport 5 (Volume 2).
Edited by Hubbard, M., Mehta, R.D., & Pallis, J.M.
International Sports Engineering Association: Sheffield, England.
Pages 152-158, 2004.  Ratzlaff, M.H., Wilson, P.D., Hyde, M.L., Balch, O.K.
And Grant, B.D.
Relationships between locomotor forces, hoof position, and joint motion during the support phase of the stride of galloping horses.
Acta Anatomica, 146, 200-204, 1993.  Reiser, R.F II, Peterson, M.L., McIlwraith, C.W.
And Woodward, B.
Simulated effects of racetrack material properties on the vertical loading of the equine forelimb.
Sports Engineering, 3, 1-11, 2000.
Read more about Lead : No significant differences in vertical hoof velocity were found between….: