The equine laminar junction plays a vital role in transferring the forces of weight-bearing between the epidermal hoof wall and the bone of the third phalanx, but the way in which it performs this function is poorly understood. Using samples from sites varying proximodistally and circumferentially around the hoof, the stress/strain behaviour of this tissue was characterised in three directions: radial tension and proximodistal and mediolateral shear. The influences of toe angle and length were also examined.

For all three test directions, the modulus of elasticity increased with increasing strain magnitude. The mean modulus of elasticity in tension was 18.25±5.38 MPa (mean ±1 S.D., N=116; mean strain 0.25). In proximodistal shear, the mean shear modulus was 5.38±1.49 MPa (N=76; mean shear strain 0.48) and in mediolateral shear 2.57±0.91 MPa (N=66; mean shear strain 0.81). In many cases, the individual hoof or horse from which the samples were taken significantly affected the value of the modulus, suggesting that mechanical history may affect the material properties of this tissue. Few biologically significant variations with location, toe length or toe angle were unambiguously demonstrated, suggesting that the material properties of the laminar junction are independent of position, despite apparent regional variations in function, and that foot shape is not a major determinant of material properties.

The laminar junction of the equine digit forms the interface between the external surface of the third phalanx and the internal surface of the hoof wall, and includes epidermal, dermal and subcutaneous tissues. Both morphologically and mechanically, it is a unique structure. Morphologically, its two-tiered laminar organisation appears to be an adaptation to increase the surface area of the inner aspect of the wall (Linford, 1987), so decreasing stress across the basement membrane. Mechanically, it is positioned to transfer a large proportion of the ground reaction force between the hoof wall and the skeleton. This force is not inconsiderable: Kingsbury et al. (1978) recorded a vertical force of 1700 N (1.7 times body weight) in a 450 kg horse moving at 14 m s−1, and horses landing over fences at speed must routinely exceed this value. The stresses and strains which such forces impose on the junction in vivo are currently unknown. Mechanical failure of the junction can occur as a clinical problem (in the condition known as ‘laminitis’) and may result in palmar rotation or ventral sinking of the third phalanx relative to the hoof wall. The cause of this condition is most frequently toxic or metabolic, although mechanical aetiologies are also documented (Stashak, 1987; Linford et al. 1993; Pollitt, 1995).

The morphology of the laminar region has been well studied (Stump, 1967; Leach and Oliphant, 1983; Linford, 1987; Pollitt, 1994), but we are only aware of one study to date which has documented its mechanical properties (Hallab et al. 1991), and its function has not been quantitatively investigated (we will estimate laminar stresses in the Discussion). Knowledge of the mechanical function of this region is important for several reasons: to increase our understanding of the mechanics of normal hoof function; to understand the ways in which controllable external factors (e.g. hoof shape) may influence the magnitude and direction of forces traversing the junction; and to improve knowledge of the mechanical forces operating during a clinical case of laminitis.

The laminar junction comprises a complex mixture of tissue components. The dermis and subcutis in this region contain significant quantities of collagen and elastin (Stump, 1967) in combination with loose areolar tissue, blood vessels and nerves. In the outer part of the junction, the dermis interdigitates with keratinised epidermis, creating a region with strata of quite different mechanical properties. It is this composite structure that forms the focus of our investigation. This study aims to document the basic material properties of the laminar junction at a number of sites varying proximodistally and circumferentially around the hooves of adult horses. The tensile modulus of elasticity and the shear modulus in two directions (orthogonal to each other and to that of the tensile modulus) were measured at rates of loading which approximate those experienced in vivo. The toe angles and lengths of the feet were measured, and their correlation with these properties was determined. The design of the study was based on inferences about the relative motion of the third phalanx with respect to the hoof wall, as shown in Fig. 1. During normal locomotion and weight-bearing, the proximal dorsal wall rotates caudoventrally about the distal dorsal border. This movement is accompanied by abaxial flaring in the posterior parts of the hoof (Fig. 1A). Tension (Fig. 1B,C) and proximodistal shear (Fig. 1D) in the laminar junction are assumed to result from ventral and caudal movement of the third phalanx with respect to the wall. Abaxial and dorsal movement of the wall with respect to the third phalanx leads to a component of mediolateral or ‘transverse’ shear (Fig. 1B,C). Our study aimed to simulate these three strain conditions and to measure the modulus of elasticity in each direction.

Fig. 1.

Deformation and movement of the hoof wall and third phalanx under load. (A) Dorsolateral view of the hoof wall. The solid line represents the unloaded wall, and the dashed line shows the change in shape that occurs during weight-bearing. Under load, the dorsal wall flattens and moves palmarly, particularly proximally, and there is abaxial movement of the caudal parts of the wall.(B) Lateral view of the hoof wall and third phalanx. The flattening and palmar movement of the dorsal hoof wall occur as a result of tensile forces in the dorsal part of the laminar junction generated by loading of the third phalanx. Abaxial movement of the wall at the quarters generates transverse (mediolateral) shearing between the wall and the bone at this site.(C) Transverse section through the hoof. The arrows show the presumed movement of the sides of the wall relative to the third phalanx. This abaxial movement of the wall will create tension (as well as mediolateral shear) in the laminar junction at the quarters. (D) Frontal plane through the hoof. The suspension of the third phalanx from the hoof wall (*) by the laminar junction creates proximodistal shear at all sites. GFR, ground reaction force.

Fig. 1.

Deformation and movement of the hoof wall and third phalanx under load. (A) Dorsolateral view of the hoof wall. The solid line represents the unloaded wall, and the dashed line shows the change in shape that occurs during weight-bearing. Under load, the dorsal wall flattens and moves palmarly, particularly proximally, and there is abaxial movement of the caudal parts of the wall.(B) Lateral view of the hoof wall and third phalanx. The flattening and palmar movement of the dorsal hoof wall occur as a result of tensile forces in the dorsal part of the laminar junction generated by loading of the third phalanx. Abaxial movement of the wall at the quarters generates transverse (mediolateral) shearing between the wall and the bone at this site.(C) Transverse section through the hoof. The arrows show the presumed movement of the sides of the wall relative to the third phalanx. This abaxial movement of the wall will create tension (as well as mediolateral shear) in the laminar junction at the quarters. (D) Frontal plane through the hoof. The suspension of the third phalanx from the hoof wall (*) by the laminar junction creates proximodistal shear at all sites. GFR, ground reaction force.

Source of material, preparation of test samples and imaging

Forty-six forefeet from 24 horses (Equus caballus L.) were obtained at the time of death from a local equine abattoir, and all tests were performed within 12 h of this time. The horses were all at least 2 years old and in reasonable physical condition. Breed, exact age and clinical history were not available. The toe angle (the angle between the solar surface and the dorsal wall) and toe length (the distance from the hairline to the ground at the dorsal wall) were measured before preparation of test samples. (These variables are routinely monitored and changed by a farrier when trimming or shoeing the horse.) Samples of the laminar junction were prepared to allow testing of stress/strain properties in three orthogonal directions: tension applied at 90 ° to the outer surface of the wall (12 feet; Fig. 2A; Table 1); shear applied so that the wall moved proximally with respect to the bone (proximodistal shear; 12 feet; Fig. 2B; Table 1); and shear applied so that the wall moved in a transverse plane with respect to the bone (mediolateral shear; 10 feet; Fig. 2C; Table 1). For all three directions of testing, samples were taken to investigate differences in the material properties between the dorsum and the quarters. In tension and mediolateral shear, samples were prepared to allow investigation of proximodistal differences at each of these sites. In proximodistal shear, there was insufficient tissue to allow this at the quarters, and two adjacent samples (designated ‘dorsal’ and ‘palmar’) were taken. All sampling sites were defined relative to the proximodistal midpoint of the wall. Dorsal samples were taken at or adjacent to the dorsal midline, and quarters samples were centred one-third of the way between the dorsal midline and the heel. The mediolateral shear samples taken from the quarters were all prepared so that the wall would move dorsally with respect to the bone. For the dorsal mediolateral shear samples, the direction of shear was randomised among feet. An additional 12 feet were used to investigate the relationship between the three directions of testing within individual feet. From these feet, a tensile sample, a proximodistal shear sample and a mediolateral shear sample were taken from the dorsum and from both quarters (Fig. 2D). For all tests, the horn and bone adjoining the piece of junction under test were clamped into the materials testing machine, leaving the entire laminar junction (from the inner aspect of the wall to the outer aspect of the bone) under test.

Table 1.

The dimensions of the samples used in the three tests

The dimensions of the samples used in the three tests
The dimensions of the samples used in the three tests
Fig. 2.

(A) Sampling scheme for tensile tests. A total of 10 samples was taken from each hoof (N=12), six from the dorsum and two from each quarter. (B) Sampling scheme for proximodistal shear tests. A total of eight samples was taken from each hoof (N=12), four from the dorsum and two from each quarter. (C) Sampling scheme for mediolateral shear tests. A total of seven samples was taken from each hoof (N=10), three from the dorsum and two from each quarter. (D) Sampling scheme for feet in which tensile (T), proximodistal shear (PDS) and mediolateral shear (MLS) tests were compared. A total of nine samples was taken from each hoof (N=12), one in each test direction from the dorsum and both quarters (not shown).

Fig. 2.

(A) Sampling scheme for tensile tests. A total of 10 samples was taken from each hoof (N=12), six from the dorsum and two from each quarter. (B) Sampling scheme for proximodistal shear tests. A total of eight samples was taken from each hoof (N=12), four from the dorsum and two from each quarter. (C) Sampling scheme for mediolateral shear tests. A total of seven samples was taken from each hoof (N=10), three from the dorsum and two from each quarter. (D) Sampling scheme for feet in which tensile (T), proximodistal shear (PDS) and mediolateral shear (MLS) tests were compared. A total of nine samples was taken from each hoof (N=12), one in each test direction from the dorsum and both quarters (not shown).

Samples were prepared using a bandsaw, a scalpel being employed where there was danger of overcut with the saw, immediately wrapped in damp paper towels and stored in moisture-proof containers. The cross-sectional area and length (the distance between the inner face of the horn and the outer face of the third phalanx) of the junction were measured using Vernier calipers (Table 1). For each sample, the mean angle of the primary laminae with respect to the cut edges of the horn (and therefore with respect to the direction of the force applied by the materials testing machine) was documented (Fig. 3) using a video digitising system (CCD camera with macro lens: CCD-72, Dage-MTI, Inc., Michigan City, IN, USA; generic 486 PC equipped with a video-digitising card: OFG-640, Imaging Technology Inc., Woburn, MA, USA; image analysis software; Optimas, BioScan, Inc., Edmonds, WA, USA).

Fig. 3.

Determination of the mean angle of primary lamellae with respect to the cut edges of horn. Before materials testing, an image of the cut surface of each sample was captured using a video digitising system. Lines were drawn onto the image parallel to the two cut edges of the outer wall (A). A line was then drawn through each primary lamina, and the angle α at which this line and the mean of the two ‘cut edge’ lines intersected on the dorsal side of the sample was recorded (B). The mean of all the values of α for each sample was then recorded as the ‘mean angle of the primary laminae’. The diagram in A shows a sample prepared for tensile testing. The same procedure was used to calculate the mean laminar angle in proximodistal and mediolateral shear samples. The mean angle of the two cut edges approximated the line of force application of the materials testing machine in tension and was approximately perpendicular to it in shear tests.

Fig. 3.

Determination of the mean angle of primary lamellae with respect to the cut edges of horn. Before materials testing, an image of the cut surface of each sample was captured using a video digitising system. Lines were drawn onto the image parallel to the two cut edges of the outer wall (A). A line was then drawn through each primary lamina, and the angle α at which this line and the mean of the two ‘cut edge’ lines intersected on the dorsal side of the sample was recorded (B). The mean of all the values of α for each sample was then recorded as the ‘mean angle of the primary laminae’. The diagram in A shows a sample prepared for tensile testing. The same procedure was used to calculate the mean laminar angle in proximodistal and mediolateral shear samples. The mean angle of the two cut edges approximated the line of force application of the materials testing machine in tension and was approximately perpendicular to it in shear tests.

Tensile and shear testing

The physical properties of the samples were measured in a materials testing machine (Instron 4024: Instron Canada Ltd, Burlington, Ontario, Canada) using a 1 kN load cell. The horn and bone portions of each sample were clamped directly into the jaws of the Instron, leaving the entire laminar junction free to deform (see Fig. 2). The tissues were clamped sufficiently tightly that rotation of the shear samples did not occur, and this was verified photographically during pilot studies. The sensitivity of the load cell was adjusted to accommodate the relatively low loads sustained by this tissue (maximum force for tensile tests 200 N; maximum force for shear tests 100 N). From the results of pilot studies, it was estimated that these loads would give maximum stresses in the region of 2 MPa in tension and 1 MPa in shear, encompassing strains of over 0.3 in tension, 0.5 in proximodistal shear and 0.8 in mediolateral shear. Samples were not taken to failure as the ultimate properties of this tissue are not relevant to an understanding of its normal mechanical function in vivo and were not of interest in this study. Load data were obtained directly from the Instron. Crosshead displacement was measured using a linear displacement transducer (LSC transducer, Apek Design and Development Ltd, Wimborne, Dorset, UK). Signals from both transducers were digitised at 200 Hz and stored using a data-acquisition system (MacAdios 8ain and Superscope II; GW Instruments Ltd, Somerville, MA, USA). A crosshead displacement rate of 8.3×10−3 m s−1 was chosen as an approximation of the estimated rate of strain experienced by these tissues on impact by a horse moving at a canter or gallop. The maximum time between the death of the animal and testing of the tissues was 12 h.

Calculation of modulus of elasticity, shear modulus and statistical analyses

The general forms of the stress/strain relationships in tension, proximodistal shear and mediolateral shear are shown in Fig. 4A, B and C respectively. For each test, linear regression was used to obtain two values for the modulus: a ‘low-strain’ modulus was determined over the strain range 0–0.05 (tensile tests) and 0–0.1 (shear tests), and a ‘high-strain’ modulus was taken to represent the slope of the second linear/near-linear portion of the stress/strain plot. The range of strain for the high-strain modulus was dictated by the shape of each individual plot. A third-order polynomial was also fitted to each curve using the software package Matlab (The MathWorks Inc, Natick, MA, USA).

Fig. 4.

(A) A representative tensile stress/strain curve. (B) A representative proximodistal shear stress/strain curve. (C) A representative mediolateral shear stress/strain curve.

Fig. 4.

(A) A representative tensile stress/strain curve. (B) A representative proximodistal shear stress/strain curve. (C) A representative mediolateral shear stress/strain curve.

All tensile elastic moduli were calculated using true strain. As lateral displacements could not be measured, stress was defined as applied force divided by the initial cross-sectional area. Shear modulus was calculated as shear stress divided by shear strain. Shear strain was calculated by dividing shear displacement by initial length. The tests were devised to mimic the relative movements of the hoof wall and third phalanx in vivo, and it is recognised that they do not represent a true shear test. The value obtained is therefore not a true shear modulus, but is representative of the mechanics of this tissue in vivo.

Statistical analyses were carried out using the software packages Statview (Abacus Concepts Inc, Berkeley, CA, USA) and DataDesk (Data Description Inc, Ithaca, NY, USA). A P-value of ⩽0.05 was taken as significant. Data points that lay more than two standard deviations from the mean were deleted from the data set. This was performed as an objective method of removing obvious outliers from the data. The data from each mode of testing (tension, proximodistal shear and mediolateral shear) were then analysed separately as follows, with ‘hoof’ nested within ‘horse’ for analyses involving analysis of variance (ANOVA). The dorsal samples were analysed for an effect of hoof, horse and proximodistal sample position using ANOVA. The relationships between toe length and modulus and between toe angle and modulus were investigated using Pearson’s product-moment correlation. The same analysis was performed on the data from the quarters of the hoof, with the addition of paired t-tests to investigate mediolateral, proximodistal and dorsopalmar differences as appropriate. Any difference between the dorsum and the quarters was then investigated using ANOVA, with horse and hoof included in the model and with account being taken of any site differences documented in the earlier analyses.

The strain range over which the high-strain moduli were taken was not standardised because the linear portion of the curve lay in a different strain range for each test. For these data, Pearson’s product-moment correlation was used to investigate the relationship between the modulus and the mean value of the strain range over which that modulus was taken.

We tested for a relationship between the different moduli using the 12 feet from which samples were prepared in all three directions of testing. Pearson’s product-moment correlation was used to compare the data from each pair of test directions (tension versus proximodistal shear, proximodistal shear versus mediolateral shear and tension versus mediolateral shear).

All values are given as the mean ± one standard deviation (S.D.) unless otherwise stated.

Low-strain data

At low strains (0–0.05 in tension, 0–0.1 in shear), the data showed a high degree of variability. No statistically significant differences were found among the different sampling sites in any direction of testing. In tension, the mean value of the modulus of elasticity was 424±459 kPa calculated from 116 tests. In proximodistal shear, a mean value of 396±312 kPa was obtained for shear modulus (N=64), and in mediolateral shear the mean value was 222±104 kPa (N=59).

High-strain data

Fig. 5 shows the results obtained in all three test directions, with the data separated by sample site. Tables 2 and 3 summarise the results of the statistical tests employed.

Table 2.

Effect of sampling site on modulus of elasticity and shear modulus (high-strain data)

Effect of sampling site on modulus of elasticity and shear modulus (high-strain data)
Effect of sampling site on modulus of elasticity and shear modulus (high-strain data)
Table 3.

Effect of factors investigated on modulus of elasticity and shear modulus (high-strain data)

Effect of factors investigated on modulus of elasticity and shear modulus (high-strain data)
Effect of factors investigated on modulus of elasticity and shear modulus (high-strain data)
Fig. 5.

The high-strain results from all sampling sites in the three modes of testing, tension (A), proximodistal shear (B) and mediolateral shear (C), are shown superimposed on a ‘flattened out’ representation of the hoof wall. The data shown represent the mean ± S.D., with the number of tests in parentheses. Sampling sites that showed significantly different values (within each mode of testing) are shown in Table 2.

Fig. 5.

The high-strain results from all sampling sites in the three modes of testing, tension (A), proximodistal shear (B) and mediolateral shear (C), are shown superimposed on a ‘flattened out’ representation of the hoof wall. The data shown represent the mean ± S.D., with the number of tests in parentheses. Sampling sites that showed significantly different values (within each mode of testing) are shown in Table 2.

The mean value of tensile modulus of elasticity at high strains for all the samples combined (N=116) was 18.25±5.38 MPa. The mean value of the mean strain over which these moduli were calculated was 0.25. The only statistically significant effect of sampling site on the value of modulus obtained was at the quarters, in which there was a tendency for the medial proximal site to yield stiffer samples (Table 2). The value of the modulus was also significantly affected by the individual hoof from which the samples were taken (Table 3). Samples that entered the second linear phase of the stress/strain curve at lower strains tended to have a higher modulus of elasticity, although the correlation between the two parameters was rather weak (r=−0.38; Table 3). It is possible that this relationship was a function of the individual hoof involved, rather than being a direct effect: when the mean values for strain range and modulus were taken for each hoof, the correlation coefficient between the two was −0.72 (95 % confidence interval −0.91 to −0.24).

In proximodistal shear, the value of the shear modulus at high strains was 5.38±1.49 MPa for the 76 samples. The mean value of the mean strain over which these moduli were calculated was 0.48. There were no differences in modulus due to sampling site in this direction of testing, but the individual horse from which the samples were taken did have a significant effect (Table 3). Both the strain range over which the stress/strain curve became close to linear and the toe length showed weak inverse correlations with the value of shear modulus obtained (Table 3).

In mediolateral shear, at high strains, the shear modulus at the dorsum had a mean value of 2.57±0.91 MPa calculated from 66 tests. The mean value of the mean strain over which these moduli were calculated was 0.81. The individual horse from which the samples were taken had a consistently significant effect on the value of shear modulus obtained (Table 3) and there were some site-to-site differences, the dorsum being stiffer than the quarters and the dorsal distal samples being stiffer than those from the middle portion of the dorsum (Table 2). Weak negative correlations between toe length and the strain range in which the stress/strain curve became close to linear were also evident (Table 3).

In general, the results show that differences due to sampling site were infrequent and mostly of marginal statistical significance. The most consistently significant factor affecting the results appeared to be the effect of the individual horse or hoof from which the samples were taken. The discrepancy between the number of proximodistal and mediolateral shear tests at high and low strains is partly due to the removal of data points that lay more than two standard deviations from the mean, because this was carried out for the low- and high-strain data separately. In addition, in a small number of tests, it was not possible to achieve a good fit for a regression line at low strains, and these tests were not included in the low-strain data set.

Third-order polynomial results

Fig. 6A, B and C show the mean third-order polynomial fits for the tensile, proximodistal shear and mediolateral shear tests, respectively, together with the curves which represent one and two standard deviations from the mean. Although a third-order polynomial shows a poor fit to the data in the low-strain portion of the graph (below strains of approximately 0.05), it describes the higher-strain regions extremely well. The r2 value for the fit between the data sets and the third-order polynomial had a mean value of greater than 0.99 for all three directions of testing.

Fig. 6.

Mean third-order polynomial fits for tensile (A), proximodistal shear (B) and mediolateral shear (C) tests. In each graph, line 1 represents the mean for the entire data set, line 2 indicates one standard deviation from the mean and line 3 indicates two standard deviations from the mean. The equation for the mean third-order polynomial fit for line 1 is given on each graph. Immediately below the graph, the values of the mean and one standard deviation are given for each coefficient, with the coefficients defined in the following equation: y=ax3+bx2+cx+d, i.e. the equation for the curve one standard deviation from the mean for tensile tests (A) would be: y=309×103x3+52.630×103x2+2868x+75. At low strains (below approximately 0.05 strain in tension and 0.1 in shear), a third-order polynomial clearly gives a poor description of the stress/strain relationship, but at higher strains it describes the relationship extremely well.

Fig. 6.

Mean third-order polynomial fits for tensile (A), proximodistal shear (B) and mediolateral shear (C) tests. In each graph, line 1 represents the mean for the entire data set, line 2 indicates one standard deviation from the mean and line 3 indicates two standard deviations from the mean. The equation for the mean third-order polynomial fit for line 1 is given on each graph. Immediately below the graph, the values of the mean and one standard deviation are given for each coefficient, with the coefficients defined in the following equation: y=ax3+bx2+cx+d, i.e. the equation for the curve one standard deviation from the mean for tensile tests (A) would be: y=309×103x3+52.630×103x2+2868x+75. At low strains (below approximately 0.05 strain in tension and 0.1 in shear), a third-order polynomial clearly gives a poor description of the stress/strain relationship, but at higher strains it describes the relationship extremely well.

Effect of primary laminar orientation and relationship between test directions

The mean angle between the primary laminae in the sample and the line of action of force application was not correlated with the value of modulus obtained for any set of tests. There was no correlation between the values obtained for tension, proximodistal shear and mediolateral shear when these measurements were made on the same feet.

Under the conditions of this experiment, the laminar junction displays a curvilinear relationship between stress and strain in all three directions of testing. In tensile tests (Fig. 4A) there is an initial linear ‘low-strain/low-modulus’ portion, up to a strain of approximately 0.05, followed by a region of rapidly increasing slope which leads into a second near-linear ‘high-strain/high-modulus’ section. The modulus at high strains shows a 43-fold increase over that at low strains (high strain 18.25±5.38 MPa; low strain 424±459 kPa). In both directions of shear (Fig. 4B,C), the stress/strain relationship follows a fairly smooth curve with a 13-fold difference in modulus between the low-and high-strain regions (proximodistal shear, low strain 396±312 kPa; proximodistal shear, high strain 5.38±1.49 MPa; mediolateral shear, low strain 222±104 kPa; mediolateral shear, high strain 2.57±0.91 MPa). The coefficients of variation for the high-strain data are approximately 30 %, but are in the range 50–100 % for the low-strain data. This difference may be partly related to the lack of hydrostatic pressure at low strains, as discussed below.

Comparison with previous work

Comparison of these results with those of Hallab et al. (1991) shows many similarities. Testing dorsal samples in radial tension only, these authors obtained a mean value for the modulus of elasticity of 7.4 MPa for tissues from normal horses. A crosshead displacement rate of 1.42×10−5 m s−1 was used in their experiments. This is approximately two orders of magnitude slower than the rate employed in the present study and may affect the results if the tissue is viscoelastic, as it is likely to be. The stress/strain curves obtained by Hallab et al. (1991) had a similar shape to those we obtained, and their quoted values for modulus of elasticity were taken from the second linear ‘high-strain’ part of the curve. These authors found a trend for decreasing stiffness from proximal to distal at the dorsum. This trend was also present in the current study, but did not reach statistical significance.

Tissue components involved in determining modulus of elasticity

The curvilinear stress/strain relationship displayed by this tissue is likely to be related to the way in which its multiple tissue components are integrated. The laminar junction comprises a complex mixture of epidermal, dermal and subcutaneous tissues, collagen and elastin comprising the major structural elements of the soft tissue regions. Collagen orientation in the laminar dermis has not been studied in detail, but Linford (1987) reported predominantly vertical orientation of the collagen fibres in the dorsal part of the foot, proximal to the tip of the third phalanx. The initial low-modulus region of the stress/strain curve reported in this paper is probably dominated by a combination of elastin and the protein–polysaccharide matrix which characterises dermal tissues. These tissue components are not designed primarily for load-bearing, as the low modulus in this region attests. As the collagen network strains and its fibres reorientate, this protein comes to dominate the stress/strain behaviour, giving a higher modulus at higher strains. The stress/strain curves obtained are similar to those previously obtained for skin (Veronda and Westman, 1970). Collagen has a modulus of elasticity in tension of approximately 1 GPa (Wainwright et al. 1976), elastin being three orders of magnitude less stiff, with a tensile modulus of approximately 1 MPa (Dimery et al. 1985) when loaded parallel to the fibres. The value obtained in the present study for the tensile modulus at high strains (18.25±5.38 MPa) thus corresponds reasonably well with these values, taking into account the fact that the laminar junction comprises a mixture of these and other tissue components and that the collagen and elastin fibres were not necessarily oriented parallel to the direction of loading.

Collagen is structurally optimised as a tensile material, rather than one that acts in shear. It is, however, commonly found in pliant composite materials in conjunction with elastin, which has a shear modulus of approximately 600 kPa (Wainwright et al. 1976). This is close to the values obtained for the shear modulus of the laminar junction at low strains (proximodistal shear modulus 396±312 kPa; mediolateral shear modulus 222±104 kPa). The higher values for shear modulus at high shear strains (proximodistal shear modulus 5.38±1.49 MPa; mediolateral shear modulus 2.57±0.91 MPa) reflect both reorientation and strain of the collagen fibre network and the fact that the test specimens inevitably experienced some tensile stress at these high strains. Some anisotropy of the laminar junction is evident, correlating with the fact that the orientation of collagen fibres within the tissue is not random (Linford, 1987).

One important difference between the mechanical state of the tissue in these experiments and that in vivo is the absence of hydrostatic pressure. The dermis of the laminar junction is a highly vascular tissue, well-endowed with arteriovenous anastomoses (Pollitt and Molyneux, 1990), which presumably allow rapid redistribution of blood on impact. Engorgement of the laminar junction by blood may ‘pre-strain’ and stiffen the tissue considerably at low strains, providing a hydraulic shock-absorptive effect. The high-strain data presented here represent the mechanical properties of the tissue after blood expulsion, which may be similar to the situation at peak digit loading. If this hypothesis is true, the ‘apparent modulus’ in vivo at low strains would be considerably greater than that found in the present tests and might be expected to be somewhat less variable.

Factors that modify modulus of elasticity

The regional differences in the high-strain moduli (Table 2) were few in number and mainly of marginal statistical significance. The relative lack of significant differences between sample sites in this data set may seem surprising in view of the diverse shapes and functions of the different parts of the hoof, but the data suggest that there is no overriding design requirement for different properties in different parts of the foot.

The horse from which the samples were taken had a fairly consistent significant effect on the results obtained, reaching significance in the high-strain data for proximodistal shear and mediolateral shear (Table 3). A related effect of individual hoof was seen in the high-strain tensile data. A difference between horses may be related to exercise history, hoof shape or limb conformation, the soft tissues of the laminar junction presumably remodelling to cope with the forces imposed. The fact that ‘hoof’ was a significant factor in the high-strain tensile data from the dorsum but ‘horse’ was not suggests that both feet of the same horse do not always show the same pattern of stiffness. This could be explained by different hoof shapes or limb conformations as noted above, by consistent favouring of one lead in an asymmetric gait, or by the presence of a unilateral lameness and differential weight-bearing.

Toe angle did not influence the value of modulus obtained, but toe length showed weak significant inverse correlations with the value of modulus in the high-strain proximodistal shear and mediolateral shear data. In the absence of data on the mass of each horse and the surface area of the laminar junction, the relationship between toe length and modulus cannot be interpreted fully, but it may indicate a lower modulus in feet with a greater laminar junction surface area. In all of these significant correlations, there was also a significant effect of individual horse on the values obtained. It is therefore possible that these two findings are related, as the front feet of each horse tended to be similar. Using ANOVA, toe length was found to be significantly related to individual horse (P=0.007), the effect of horse on toe angle having a P-value of 0.19.

In all three test directions at high strain, there was a relationship between the strain range over which the stress/strain curve was linear and the value of the modulus. Contrary to expectation, the relationship was a negative one: given the shape of the stress/strain curves, one would expect higher strains to give higher values for the modulus. The fact that the opposite is true suggests that stiffer samples tended to enter the second linear phase at lower strains than those that were less stiff. In both tension and proximodistal shear, the strain range over which each sample became linear was also significantly related to both the hoof and horse from which the sample was taken, and the values were close to significance in mediolateral shear (tension: hoof, P=0.0001; horse, P<0.0001; proximodistal shear: hoof, P<0.0001; horse, P<0.0002; mediolateral shear: hoof, P=0.08; horse, P=0.08). Thus, not only the value of the modulus at high strains, but also the strain at which it becomes linear, appears to be closely related to the individual hoof and horse.

The interrelationships between many of the factors examined in this study are complex: horse, hoof and toe length, all of which have a demonstrated effect on modulus, are not independent factors. It is therefore difficult to establish the factor that is most influential in determining the properties of the tissues tested, and it is also impossible to rule out the influence of some unmeasured, related but perhaps more significant factor, such as exercise history.

One variable that did not appear to affect the value obtained was the angle of the primary laminae with respect to the line of action of the applied load of the materials testing machine (Fig. 3). This varied from sample to sample, but measurements determined that it had no bearing on the measured value of modulus. Although the orientation of the primary laminae would appear to be an obvious candidate for a relationship between morphology and mechanics, these experiments do not bear that out. Other factors, such as the orientation of collagen fibres, may be more responsible for the mechanical properties of this tissue.

Strains that occur in vivo

The range of strain experienced by the laminar junction in vivo is not known. Fischerleitner (1974) described tests carried out on disarticulated limbs in which the third phalanx and dorsal hoof wall were reported to move more or less as a unit when the limb was loaded. The movement of the bone relative to the wall at extreme loads (as when a horse lands over a fence or at a fast gallop) has not, to our knowledge, been documented. Some estimates can be made, however. The thickness of the solar corium at the distal dorsal margin of the third phalanx, measured parallel to the outer wall, is approximately the same as the thickness of the laminar junction measured at 90 ° to the outer wall. This distance averaged 5.4 mm in the feet used in this study. Assuming minimal movement of the periphery of the sole under load, descent of the third phalanx by this amount would obliterate the solar corium and circumflex artery of the sole, which clearly does not happen. The ‘proximodistal shear’ component of laminar junction strain must therefore be significantly less than 1.0. The mean shear strain for the high-strain proximodistal shear data was 0.48, and it is conceivable that the junction may be strained to this degree in a proximodistal direction at high loads. Much less likely to occur in vivo is the high-strain value for transverse (mediolateral) shear, where a mean shear strain of 0.81 was recorded. A small amount of wall:bone shear will occur in this direction at the quarters when the hoof wall flares abaxially during weight-bearing and, presumably, in the entire hoof during cornering at speed, but the movement that occurs in vivo must be minimal, otherwise the bone would twist substantially inside the hoof capsule on weight-bearing. These high values of strain are thus unphysiological.

The range of strain that occurs in vivo in radial tension is more difficult to estimate. The flattening of the dorsal hoof wall and caudal movement of the dorsal coronary band that occur during weight-bearing are widely believed to be associated with significant radial tensile forces between the hoof wall and third phalanx, but the strains associated with this stress have not been recorded. We know of no data quantifying the expansion that occurs in the hoof capsule at the site of our ‘quarters’ samples, but early experiments of hoof wall deformation demonstrated an outward movement of the ground surface at the heels (at a site caudal to the third phalanx) of up to 1.5 mm per heel at gaits no faster than a trot (Lungwitz, 1891). This movement of the outer wall is presumably accomplished mainly through strain of the soft tissues which lie deep to the wall. A deformation of this magnitude in the laminar junction would represent a strain of 0.27 in radial tension. The mean strain value of 0.25 reported here for the high-strain data is thus not unreasonable.

Estimates of in vivo stress

The stress borne by the laminar junction has not been measured, and even estimation of its magnitude is complex. The direction of the force vector at any one point between the wall and the third phalanx is not known and will undoubtedly vary during different phases of the stride, as well as varying proximodistally and circumferentially at any given instant. Rough estimates of the stresses involved in radial tension and proximodistal shear can be made, however. Using an estimate of 7×10−3 m2 for the surface area of the laminar junction (ignoring the convolutions of the laminae), an estimate of 3.4×10−3 m2 for that area projected vertically downwards, a vertical ground reaction force of 7500 N (Kingsbury et al. 1978) and assuming even distribution of 90 % of that force across the surface, one can estimate the proximodistal and tensile components of force at the dorsum and the quarters. Using a value of 50 ° for the angle made by the dorsum with the ground, one arrives at stress values of 1.28 for radial tension and 1.52 MPa for proximodistal shear at this site. With a quarters-to-ground surface angle of 80 °, values of 0.34 for radial tension and 1.96 MPa for proximodistal shear are obtained. Thus, dorsally, the two components of the resultant vector are fairly similar, whereas at the quarters, proximodistal shear predominates. This estimate ignores any contribution from mediolateral shear and accounts only for the situation at mid-stance when the horse is travelling in a straight line; at other times in the stride, such as breakover and when the animal is moving on a turn, the force vectors will be quite different.

The laminar junction is much less stiff than both the overlying hoof wall and the underlying bone of the third phalanx. Douglas et al. (1996) demonstrated that the outer half of the dorsal hoof wall has a modulus of elasticity of 955±199 MPa, the inner half being less stiff with a value of 502±98 MPa. These values probably represent a gradient of decreasing stiffness across the thickness of the wall. Therefore, at the very innermost part of the wall, the modulus is probably rather less than 500 MPa. The laminar junction, with a tensile modulus of elasticity of 18.25±5.38 MPa at a mean strain of 0.25, is less stiff than even the inner part of the wall. However, the entire laminar junction was included in this measurement, from the inner aspect of the wall to the outer surface of the bone, and no attempt was made to evaluate differences in stiffness within the laminar junction itself. Values are not available for the modulus of elasticity of the equine third phalanx, but mammalian compact bone has a tensile modulus of elasticity of approximately 17 GPa (Wainwright et al. 1976). A difference in stiffness of at least three orders of magnitude thus separates the tissues involved in force transfer between the ground and the skeleton. But, as discussed by Kasapi and Gosline (1997), the stiffnesses of adjacent tissues must be similar at their interface if large strain differentials and high stresses are to be avoided.

The results of the present study provide a basis for estimating the strains that occur in the laminar junction under a given load and yield values which can be incorporated into mathematical models of the equine digit. The difficulties, both technical and humane, of taking measurements from within the feet of live horses make testing in vitro preferable to in vivo for obtaining information about the mechanical function of the digit. In vitro tests can never fully simulate the in vivo situation, however, and some of the caveats involved are noted above. This tissue performs an impressive task, and its complex morphology is paralleled by its complicated function. Further investigation of its role in force transmission is warranted in order better to understand normal and abnormal hoof mechanics, and the factors that lead to digital failure.

This work was supported by the Equine Research Centre, the Ontario Ministry of Agriculture, Food and Rural Affairs grant 15170, the Ontario Racing Commission and NSERC grant OGP0138214.

Dimery
,
N. J.
,
Alexander
,
R. A.
and
Deyst
,
K. A.
(
1985
).
Mechanics of the ligamentum nuchae of some artiodactyls
.
J. Zool., Lond. A
206
341
351
.
Douglas
,
J. E.
,
Mittal
,
C.
,
Thomason
,
J. J.
and
Jofriet
,
J. C.
(
1996
).
The modulus of elasticity of equine hoof wall: implications for the mechanical function of the hoof
.
J. exp. Biol
.
199
,
1829
1836
.
Fischerleitner
,
F.
(
1974
).
Röntgenographische Untersuchungen über den Einfluss der Lageveranderungen des Huf-Strahl und Kronbeines auf die Mechanik der Hornkapsel des Pferdes im Belastungsgerat
.
Vienna, Austria
:
Tierarztlichen Hochschule in Wien. Inaugural dissertation
.
Hallab
,
N. J.
,
Hogan
,
H. A.
and
Hood
,
D. M.
(
1991
).
Determining the mechanical properties of equine laminar corium tissue
.
Equine Athlete
4
,
13
18
.
Kasapi
,
M. A.
and
Gosline
,
J. M.
(
1997
).
Design complexity and fracture control in the equine hoof wall
.
J. exp. Biol
.
200
,
1639
1659
.
Kingsbury
,
H. B.
,
Quddus
,
M. A.
,
Rooney
,
J. R.
and
Geary
,
J. E.
(
1978
).
A laboratory system for production of flexion rates and forces in the forelimb of the horse
.
Am. J. vet. Res
.
39
,
365
369
.
Leach
,
D. H.
and
Oliphant
,
L. W.
(
1983
).
Ultrastructure of the equine hoof wall secondary epidermal laminae
.
Am. J. vet. Res
.
44
,
1561
1570
.
Linford
,
R. L.
(
1987
).
A radiographic, morphometric, histological and ultrastructural investigation of lamellar function, abnormality and the associated radiographic findings for sound and footsore Thoroughbreds and horses with experimentally induced traumatic and alimentary laminitis
.
PhD thesis, University of California, Davis
.
Linford
,
R. L.
,
O’brien
,
T. R.
and
Trout
,
D. R.
(
1993
).
Qualitative and morphometric radiographic findings in the distal phalanx and digital soft tissues of sound Thoroughbred racehorses
.
Am. J. vet. Res
.
54
,
38
51
.
Lungwitz
,
A.
(
1891
).
The changes in the form of the horse’s hoof under the action of the body-weight
.
J. comp. Path. Ther
.
4
,
191
211
.
Pollitt
,
C. C.
(
1994
).
The basement membrane at the equine hoof dermal epidermal junction
.
Equine vet. J
.
26
,
399
407
.
Pollitt
,
C. C.
(
1995
).
Color Atlas of the Horse’s Foot
.
London
:
Mosby-Wolfe
.
Pollitt
,
C. C.
and
Molyneux
,
G. S.
(
1990
).
A scanning electron microscopic study of the dermal microcirculation of the equine foot
.
Equine vet. J
.
22
,
79
87
.
Stashak
,
T. S.
(
1987
).
Adams’ Lameness in Horses
.
Philadelphia
:
Lea & Febiger
.
Stump
,
J. E.
(
1967
).
Anatomy of the normal equine foot, including microscopic features of the laminar region
.
J. Am. vet. med. Ass
.
151
,
1588
1598
.
Veronda
,
D. R.
and
Westman
,
R. A.
(
1970
).
Mechanical characterization of skin – finite deformations
.
J. Biomech
.
3
,
111
124
.
Wainwright
,
S. A.
,
Biggs
,
W. D.
,
Currey
,
J. D.
and
Gosline
,
J. M.
(
1976
).
Mechanical Design in Organisms
.
Princeton
:
Princeton University Press
.