Calibration of the mercury-in-silastic strain gauge in tendon load experiments

Calibration of the mercury-in-silastic strain gauge in tendon load experiments

OOZI-9290x48 $3.00 + .w Pergamon Press pk CALIBRATION OF THE MERCURY-IN-SILASTIC STRAIN GAUGE IN TENDON LOAD EXPERIMENTS DIRK J. RIEMERSMA and Jos L...

812KB Sizes 9 Downloads 33 Views

OOZI-9290x48 $3.00 + .w Pergamon Press pk

CALIBRATION OF THE MERCURY-IN-SILASTIC STRAIN GAUGE IN TENDON LOAD EXPERIMENTS DIRK J. RIEMERSMA and Jos L. M. A. LAMMERTINK Department of Functional Morphology, Faculty of Veterinary Medicine, State University. Utrecht. The Netherlands Abstract-A calibration method is presented by which the signals of mercury-in-silastic strain gauges (MISS), implanted in the tendons of in vitro loaded equine hindlegs, were converted to tendon loads. The relationships between MISS-signals and tendon loads were obtained from tensile-force tests applied to the tendons. Special attention was paid to the correction of the MISS-signals for amplitude-shifts resulting from internal repositioning of the MISS after tendon isolation and temperature differences. Shift corrections equivalent to tendon strains up to 2.8 “,; were necessary in the in vitro experiment. The tendon loads deduced from the corrected MlSS-signals were checked by torque analyses of the lower part of the limb. Differences between ComDuted and exnerimentallv obtained values of the torque of the tendon loads with respect to the fetlock joint ianged from’ - 4 to + 13 “d. INTRODUCTION

Deformation of tendon resulting from in vitro or in rir‘o loading can be recorded by liquid metal strain gauges (Lochner et al. 1980; Stone et al. 1983; Riemersma and Schamhardt, 1985; Brown et al. 1986). Conversion of the gauge signals to tendon loads is achieved by submitting the tendons, which contain the gauges, to tensile-force tests. This method requires that the relationship between gauge signals and tendon loads is the same in both the experimental and the tensile test situations. However, it has been observed that the relationship between the gauge signals and tendon loads could change when the tendon was released and reclamped in the draw-bench (Riemersma and Schamhardt, 1985). The fibrous structure of the tendon and the contractile properties of the prestrained elastic transducer were held responsible for considerable variation in initial length of the transducer (initial amplitude) within the unloaded tendon. Changes in shape of the load-gauge signal curves were not observed. Differences of up to 100 “11in deduced tendon loads were likely to occur as a result of this shift. Quantification of the shift in the gauge signal requires one or more reference tendon loads which can be estimated in the experimental situation as well as in the tensile-force test recordings. In a previous in viun experiment (Riemersma and Schamhardt, 1985) the onset of loading was used as a reference load. However, zero load conditions cannot be achieved in the normal gravitational field, and therefore the determination of the onset of loading is a potential source of error. Another change in signal amplitude results from temperature differences between the experimental (in t>ico or in virro) and tensile test (in vitro) environment (Youdin and Reich, 1975; Riemersmaand Schamhardt, 1985; Brown er (11.1986). -

_ Received 20 May 1986: in revised form 8 July 1987.

In this paper a calibration method is described, in which a reference value for tendon load is recorded, using a buckle-like transducer (Barnes and Pinder, 1974). simultaneously with the signals of mercury-insilastic strain gauges (MISS) both in an in vitro limb loading experiment and in the tensile-force tests following tendon isolation. A simple experiment to quantify the influence of temperature on the amplitude of an individual transducer is also described. The procedure was tested by an in vitro experiment in which the digital flexor tendons of isolated equine hindlegs, provided with MISSs, were loaded. The signals of the MlSSs were recorded simultaneously with the ground reaction force. The tendon loads deduced from these signals after calibration were checked by torque analysis of the lower part of the limb. MATERIALS

AND EQUIPMENT

Five equine hindlegs of different size were obtained from routine post mortems. Equine hind limbs can passively resist loading when the noose composed of patella and medial and middle patellar ligaments is hooked over the medial rim of the femoral trochlea (Nickel ef al. 1986). This situation was created in vitro by pressing the patella against the cranial aspect of the femur by means of a large tube-shrink. The digital flexor tendons in the distal part of the limb are loaded when the limb is loaded at the femoral head by a pneumatic limb-loading device (Fig. 1). capable of applying loads up to 4500 N. The superficial digital flexor tendon (SDFT), deep digital flexor tendon (DDFT) and the suspensory ligament (SL) (Nickel et al., 1986; Fig. 1) were provided with Ml%%. The MISSs (Fig. 2) were manufactured in our laboratory from 10 mm Silastic (Dow Corning) tubes (0.3 mm ID, 0.6 mm OD). The tubes were filled with mercury and submerged in xylene, which caused swelling and enlargement of the tubes. The open ends were then closed by 2 mm platinum wire (0.6 mm 469

410

D. J. RIEMERSMA and J. L. M. A. LAMMERTINK

Fig. 1. Equine hindleg in pneumatic loading device and schematic representation of the distal part of the limb. FJ = fetlock joint; SL = suspensory ligament; DDFT = deep digital flexor tendon; SDFT = superficial digital flexor tendon; FP = force plate.

Fig. 3. Typical examples of length/signal amplitude relationships. (A: current = 20 mA, r2 = 0.99995: B: current = 10 mA, rz = 0.99988).

The current through the gauge was fed by one set of lead wires, while the voltage over the gauge was measured using the other two wires. This had the advantage that voltage measurements over the gauge were not affected by temperature and resistance changes of the lead wires (Michaux et al., 1979). The resistance of the MISS was less than 1 Q. A tensile force of approximately 0.05 N was required for 10 ‘A strain of the MISS. The relationship between signal amplitude and gauge length was provided by elongating the MISS by a Mauser mechanical extensometer, while recording length and gauge signal (Fig. 3). A small force plate was constructed in our laboratory and consisted of two steel plates (300 x 350 x 10 mm) separated by three strain gauge equipped steel rings placed triangularly. The strain reading of each ring was individually calibrated to load. The three separate loads were used to calculate the position of the ground reaction force. The vertical ground reaction Fig. 2. Mercury-in-silastic strain gauge. A = stainless steel force was measured with an accuracy of 1 ‘Y,,the point arch; M = mercury; S = silastic tube wall; EB = epoxy bond; of application with an accuracy of 2 mm. Since horiP = platinum wire; L = copper ligatures; LW = lead wires. zontal forces could not be measured by the force plate, care was taken that the point of application of the diameter) to which two lead wires were soldered. After ground reaction force was in a vertical line with the evaporation of the xylene, copper ligatures were tied head of the femur. Reference values for tendon loads were recorded around the tube ends lying over the platinum wires. with a Transverse Tension Transducer (TTT), which is The length of the mercury column was thus decreased to 6-8 mm and the mercury enclosed under a slight based upon the principle of the tendon buckle transducer (Barnes and Pinder, 1974). The TTT (Fig. 4) was overpressure, which prevented subpressure during constructed in our laboratory and consisted of two straining, which might predispose to water penetration through the slightly permeable silicone tube wall. rigid pins (A) and a compression bar (B) which could Small stainless steel arches (height 1 mm, length 3 mm, be moved with respect to the pins by micrometer M. width 4 mm) with projecting wings were attached to The distance between A and B was adjusted to the the soldering points and ligatures using a two com- diameter of the tendon so that an acceptable degree of ponent epoxy glue, which also insulated the wires. bending of the tendon between the pins occurred. The Grooves in the wings were provided, for hooking concavity of A and B resulted in a self-centering of the sutures during implantation. If not overstrained in tendon between them. The transverse forces developed by the loaded tendon were recorded using the strain watery environment these gauges proved to be stable for many weeks. Failure, other than wire break, was gauge equipped (full bridge) steel ring between B and M. The TTT was applied to a tendon while slightly and not observed within 24 h after implantation. steadily pulling the instrument by hand. The MISSs were part of a circuit with constant The TTT signal depended on tendon load (Fig. 5), current (10 or 20 mA). The voltage over the gauge, which is a function of its electrical resistance, varies diameter of the tendon, and distance between A and B. almost linearly with the length of the gauge (Fig. 3). The reproducibility of the TTT signals was tested by

471

Calibration of strain gauge in tendon experiments

Table 1. Coefficients (Cl, C2) of linear approximations [TL = Cl + CZTA; TL = tendon load (N), TA = TTT amplitude (mV)] to 10 sets of TTT-tendon load data Attempt No.

155 mm I

I

Fig. 4. Transverse tension transducer (TTT). For explanation of symbols, see text.

1 2 3 4 5 6 7 8 9 10 Mean S.E.M.

Number of data 349 458 275 393 321 185 465 389 293 357 348

Cl(N)

C2 (N - ‘mV)

- 96.8 -93.5 -91.9 - 96.5 - 85.3 - 71.9 - 88.3 - 84.2 - 88.0 - 80.4 - 88.9 2.2

0.3625 0.3682 0.3707 0.3859 0.3723 0.3557 0.3923 0.3972 0.3893 0.4060 0.3800 0.0052

Data of attempt 3 are shown in Fig. 5.

TTT AtlPLITUDE

(HV,

Fig. 5. Typical recording of transverse tensions by TTI during tendon loading.

the TTT ten times to a marked site of a tendon, which was then loaded and unloaded five to ten times. The TTT was removed and re-applied between each pair of consecutive series of tendon loading. Least squares linear approximations of the data recorded between 100 and 1200 N tendon load (Table 1) enabled comparison of the different sets of data. Tensile-force tests were performed in a hydraulic drawbench. The tendons were transected at the tarsal level. The toe was connected to the draw-bench by a self tightening loop of chain around the hoof. The loose proximal ends of the tendons were gripped in a cryo-jaw (Riemersma and Schamhardt, 1982). Loads were measured with a strain gauge equipped load cell, capable of recording loads up to 10,000 N. applying

Limb loading experiment

MISSs were implanted in longitudinal incisions of the SDFT, DDFT and SL, at the mid-metatarsal site, which has most constant diameter and most ideally

parallel arranged tendon fibres (Riemersma and DeBruyn, 1986). The wings of the arches were caught by the noose of a double suture, which perforated the tendon from the incision to the outside, where it was tied to its fellow suture, holding the opposite wing of the same arch. Embedding the MISS within the tendon in this way protected it from lateral compressive forces, which would produce false signals. The soft tissues surrounding the tendons were removed so that the tendons would give sharp images on radiographs. The tendons were kept wet during implantation, after which the skin was sutured in order to prevent dessication. The internally stabilized hindlegs were loaded with constant loads in the pneumatic loading device (Fig. 1). The signals of the three transducers of the force plate under the hoof and the signals of the three MISSs were recorded simultaneously and, after AD-conversion, stored on diskette. A radiograph was taken of the lower part of the loaded limb. The position of the force plate was imaged on the radiograph by placing steel pins at known positions on the force plate surface. A measuring tape with X-ray absorbing digits was placed in the plane of the tendons for calculation of the magnifications of the radiographs. Torque analyses

The moment arms of the tendon loads to the metatarsophalangeal (fetlock) joint were calculated as the distances between thecentral part ofthe tendons to the centre of rotation of the joint (Reuleau method), which were measured from the radiographs. The point of application and magnitude of the ground reaction force were calculated from the force plate data recorded during limb loading. This vector was constructed in a drawing of the radiograph, using the steel pins indicating the force plate’s position. Its distance to the centre of rotation of the fetlock joint was measured. Torques were calculated as the products of loads and corresponding moment arms. The sum of

472

D. J. RIEMERSMA and J. L. M. A.

torques of the three tendons on the fetlock joint was compared with the torque of the ground reaction force. Determination of the MISS-tendon

----dTa-----!

I--

I

LdTz’

load relationship

After limb loading, the TTT was applied to each tendon, and MISS signals and TTT signals were recorded simultaneously while the limb was moderately loaded. The DDFT was exposed after severing the SDFT, and the SL after severing the DDFT at the tarsal level. Each time the distance between pins A and B of the TTT was recorded. Since the diameter of a tendon varies from proximal to distal (Riemersma and DeBruyn, 1986), it was necessary to mark the site of application of the TTT; a suture was used for this purpose. The toe was then separated from the limb and connected to the draw-bench, while the tendons were successively clamped in the crvo-jaw. The TTT was again applied to the tendon and another simultaneous recording of MISS signals, TTT signals, and (low) tendon loads was made. The rather vulnerable (because of the tiny steel ring) TTT was then removed for a final recording of MISS signals and tendon loads (up to 5000 N). Determination of MISS-temperature

LAMMERTINK

o 15

35

Fig. 6. Typical MISS amplitude/temperature relationships of five prefixations of the MISS. Temperature intervals of about 2,lO and 20°C (dT1, dT2 and dT3 respectively) refer to the data presented in Fig. 7.

relationship

The MISS signal amplitude is temperature dependent (Youdin and Reich, 1975; Brown et al., 1986). All recordings of MISS signals in the current experiment were done at room temperature, but this was not constant. Therefore, during each recording of MISS signals the temperature of the MISS was measured by applying a temperature sensitive resistor (Thermistor) into the incision near the MISS. The MISS amplitude depended on the measure of prestrain and the measure of tendon strain encountered during the limb loading experiment and on the gauge factor and zero setting of the MISS amplifiers. Therefore, the influence of the temperature on the MISS signal was determined after the tensile-force tests, and care was taken not to change the amplifier’s gain and zero settings throughout the experiment. The MISS was attached to a piece of Invar (a metal with temperature coefficient of 0.6 x 10m6) with steel clips, pressing upon the stainless steel arches perpendicularly to the long axis of the MISS. The gauge, thus fixed at given length, was then placed in a waterbath of which the temperature was increased from 10 to 40°C. The expansion of Invar can be neglected in this temperature range, so that the changes in signal amplitude can be attributed purely to the influence of temperature on the MISS. MISS signals were recorded at the temperatures previously measured during MISS recordings and also at increments of about 5°C (Fig. 6). In order to account for variation in length of the MISS during the experiment, this procedure was repeated at several different lengths (initial amplitudes) of the MISS. Amplitude/amplitude correction relationships (Fig. 7) for each relevant temperature interval were deduced from these data.

‘IO

dT3

0

u 0

1000

m

I

2000

p

I

3wO

SIGNAL RMPLITUOE

WOO

dT1 m

tflV1

Fig. 7. Examples of amplitude/amplitude correction relationships for the three different temperature intervals indicated in Fig. 6.

Calibration of MISS signals to tendon loads

The aim of the calibration procedure is to convert the MISS signals recorded during linib loading into tendon loads using the MISS-tendon load relationship resulting from the tensile-force tests. A correction of the latter relationship is necessary because of the shift, resulting from tendon isolation, and the (possible) differences in temperature between the recordings. The following five relationships with MISS signals were used in the procedure: (1) MISS signals and ground reaction forces from the limb loading experiment. (2) MISS signals and TTT signals from the tendon in situ. (3) MISS signals vs TTT signals from the tendon in the draw-bench.

473

Calibration of strain gauge in tendon experiments (4) MISS signals vs tendon loads from the tensileforce tests. (5) MISS signals vs temperature at several lengths of the MISS.

The correction procedure was carried out in two steps: (I) Correction for differences in temperature. (II) Correction for repositioning of the MISS (causing an amplitude shift) as a result of tendon isolation and clamping. The temperature during limb loading was used as a reference temperature. The MISS signals recorded at a temperature different from the reference temperature were corrected using relationship (5). For each given temperature interval the amplitude correction vs signal amplitude (Fig. 7) was determined and applied to the respective recordings. After correction for temperature differences the relationships (1) through (4) were comparable. Relationship (2) was then compared with (3). The difference between the MISS amplitudes of relationships (2) and (3) at a certain TTT signal amphtude (Fig. 8) represented the amplitude shift of the MISS signal resulting from isolation and clamping. The MISS signals of (4) increased (left shift) or decreased (right shift) with the voltage value of the shift, resulting in a MISS signal/tendon load relationship which was directly applicable to relationship (1). In this way, MISS signals recorded during limb loading were converted to tendon loads during limb loading. RESULTS

The influence of temperature on the signal amplitude of the MISS is shown in Fig. 6. The resistance of

the gauge increased with temperature. This increase was slightly greater at higher initial resistance of the gauge, resulting in slightly increasing slopes of the signal amplitude/amplitude correction relationships (Fig. 7) for a given temperature interval. A temperature rise of 1°C resulted in an amplitude increase between 14 and 52 mV (depending on the characteristics and prestrain of the individual gauge and the current, gain and zero settings of the amplifiers), resulting in a mean strain artifact of 0.05 1% (S.D. = 0.015 “/A;n = 15) per “C. The reproducibility of the TTT signals vs tendon loads is shown in Table 1, which contains the coefficients of the lines, fitted through ten sets of TTT-load data by least squares approximations. Each set of data corresponds to one series of tendon loading cycles (Fig. 5) while the TTT was applied. Typical examples of MISSTTT relationships, recorded in situ and during tendon loading in the drawbench respectively, are shown in Fig. 8. The recordings show a high resemblance except for a left shift of the MISS signal. In twelve of fifteen cases the MISS amplitude decreased (left shift) when the tendon was isolated and clamped. The values of the shift ranged from - 1530 mV to +700 mV with a mean value of - 302 mV. These values are equivalent to MISS strains between -2.81% and + 1.54%. A representative example of the relationship between MISS signals and tendon loads is shown in Fig. 9. Length and strain of the MISS are also indicated in Fig. 9, assuming zero strain at the onset of loading. The tendon loads occurring during limb loading, calculated by applying the complete calibration procedure, are listed in Table 2. This table contains also the moment arms of the tendons with

I -1000

I

LssTk_ITZcnv,

1

I Fig. 8. Typical example of MISS-TTT curves, recorded in the intact limb (L) and on the tendon clamped in the drawbench (C) respectively, and corrected for differences in temperature. The shift (Sh) results from tendon isolation and

1

I

woo

so0

I

6.5

-3

I

I

6.6

1 -2

MISS

I

-1

n1ss

I

L&X”

1

1

6.9

7.0

I

6.6

(RHI I I 1 1 OSTFti& P&C. I3

1

1

Y

5

c

Fig. 9. Typical example of MISS-tendon load relationship. Additional scales provide the conversion of the MISS signals into lengths and strains. Zero strain was assumed at the onset of loading (see text).

474

D. J. R~EMERSMA and J. L. M. A. LAMMERTINK Table 2. Tendon loads and ground reaction force and moment arms of their vectors with respect to the fetlock joint of five hind limbs SL Limb No. 1 2 3 4 5

F(N)

Hmm)

2500 4800 4500 1850 3650

37 37 38 29 26

DDFT F(N) r (mm)

F(N)

1900 1500 2650 1130 1150

1650 2800 2100 2120 1510

48 51 48 38 33

SDFT r(mm) 41 50 46 31 32

GRF F(N)

r(mm)

2100 2730 2570 1940 1460

130 128 148 81 121

SL: Suspensory ligament. DDFT: Deep digital flexor tendon. SDFT: Superficial digital flexor tendon. GRF: Ground reaction force. respect to the fetlock joint and the magnitude of the

ground reaction force and its distance to the centre of the fetlock joint. The sum of torques of the three tendons on the fetlock joint and the torque of the ground reaction force (absolute values) are presented in Table 3. The mean difference between these torques, which should be zero in a state of equilibrium, was 5.2 %. DISCUSSION

The influence of temperature on the MISS signal was of minor importance in our in oitro experiment, since all recordings were made at room temperature. Differences in temperature of 15-20°C between in uivo and in vitro environments can be expected though, and may result in an artifact of about 1% strain, which is one third of the maximal tendon strain in the walking horse (Riemersma and Schamhardt, 1985). Youdin and Reich (1975), who used Wheatstone bridge equipment, reported a strain artifact per “C of 0.045 %, which is in the same order of magnitude as our result (0.051%). However, Brown et al. (1986), also using Wheatstone bridge equipment, reported a strain artifact of 0.185 % per “C. This seems significantly higher, but differences in electronic equipment hamper comparison. The shift resulting from isolation of the tendon and clamping it in the draw-bench is an intriguing phenomenon. The influence of clamping is obvious since similar shifts were observed when a clamped tendon was released and consecutively reclamped. This resulted in a shift of unpredictable magnitude (left as well as right shifts), but did not noticeably affect the shape of the curves. Figure 10 shows the result of insufficient clamping in a case where the clamped part of the tendon was not completely frozen solid, as is recommended (Riemersma and Schamhardf 1982). The tendon first slipped and then tore out of the clamp. During slipping, a shift to the right occurred. It has been shown in this investigation and previously (Riemersma and Schamhardt, 1985) that the MISS signal itself is very reproducible. Therefore, the origin of the shift may be found in the non-solid, fibrous character of a tendon. The main course of the tendon fibres is longitudinal, although cross-links between the fibres are present. The different fibres originate from different sarcolemmas of different

Table 3. Torques of the tendon loads and the ground reaction force with respect to the fetlock joint Sum of torques tendons Limb No. Wm) 1 2 3 4 5 Mean:

262 395 395 175 181 282

Torque GRF (Nm) 273 349 380 157 177 267

V0difference -4% 13% 4% 11% 2% 5.2 %

loading of all fibres in uioo is therefore not likely, but an even distribution of loaded fibres in the tendon may be assumed. However, the differences in origin and insertion of different fibres necessitates the possibility of mutual sliding. Associated with the former is the behaviour of a tendon under load. It has been stated (Riemersma and Schamhardt, 1985) that a distinction must be made between deformation of the tendon as a structure and of the collagen fibres as load-bearing components within the tendon. As long as the collagen fibres are not strained, the tendon can be deformed relatively easily. Once the collagen fibres straighten out and begin to strain, the loads necessary for elongation of the tendon will significantly increase. The transition, which is the onset of tendon loading, is associated with the initial length of a tendon (Fig. 9), although the exact point is arbitrary (Fung, 1973; Woo, 1982; Viidik et al. 1982). A technical problem is that elongations of a locally applied gauge, rather than tendon deformations, are recorded. It must therefore be assured that the gauge is attached to a considerable portion of the tendon and that the forces generated by the prestrained elastic transducer can be neglected with respect to the forces generated by the straining tendon. This can safely be assumed to be so when the collagen fibres are strained, but at low tendon loads (below zero strain of the collagen) the tension of the MISS may exceed the tension of the fibres to which the MISS is attached, and contraction of the gauge, rather than relevant tendon deformation, is recorded. It appears from Fig. 9 that a significant portion of the MISS amplitude change is associated with negligible tendon loads. When a tendon is isolated from the limb, the mutual position of the fibres may change because of the lack of motor units. Simultaneous

Calibration

I

of strain gauge

woo TIllE HS

I 0

I lsoo

I 3loo

nrss nv

I

soo

J

aooo

curves corresponding with (b). (b) MISStendon load curves of a deep digital flexor tendon.

Fig. 10. (a) MISS-time

Right shift of the MISS signals resulted clamp (see text).

from slipping off the

deviation around joints and a change in the positions of origin and insertion. The prestrained MISS may exert contractile actions as a result of the absence of relevant loads and may undergo a repositioning which could affect its initial length in the unloaded tendon. This is not, at least entirely, prevented by the sliding fibres to which it is attached, and a shift of the gauge amplitude will be the inevitable result. Regarding the potential repositioning of several tendon fibres, it may be disputed whether tensile-force tests represent the loaddeformation characteristics of a tendon at all. However, many experiments with equine tendons in our laboratory have revealed that the shape of the load-deformation curves and of the MISS-tendon load curves is almost completely reproducible if an even load distribution over the fibres can be assured throughout clamping. It is therefore likely that the mechanical properties of the tendon are not significantly changed by isolating the structure. The shift can be quantified using a reference load related to MISS signals while the tendon is in its natural position and when it is clamped following isolation. It has been shown in this experiment

in tendon

415

experiments

(Table 1) that the signals of the TTT are suitable for this purpose. Although this seems to favour the use of buckle transducers over compliant strain transducers, we feel that the accompanying bending of the tendons is an unacceptable interference for in uioo application. The tendon bending will cause tendon shortening, affecting the natural load distribution over parallel tendons, unless the width of the buckle is adapted to the diameter of each individual tendon (preventing excessive bending). The size of the buckle transducers is another problem which is not solved by using smaller types applied to only parts of the tendon. since the problems of redistribution of tendon fibre loading as described above will then cause additional problems in calibration, It is to be expected that alternative clamping methods, such as the use of bone-(muscle-) tendon-bone specimens, will also be associated with fibre slide. The absence of joint deviations and variable states of rigour in the muscle fibres continuous with the tendon fibres are responsible for this. Considering in vim experiments, additional fibre repositioning may be induced by elements possessing contractile properties, in series with the fibres and in between, e.g. smooth or striated muscle fibres or fibroblasts. The limbs in this experiment were under a constant load which implies static equilibrium and thus the sum of all torques should be zero. This condition is obtained if the sum of torques of the three flexor tendons (SDFT, DDFT and SL) equals that of the ground reaction force, except for the sign. The mean difference of about 50,; indicates a good fit of the figures. Without correction ofthe MISS-load relationship this result could not have been achieved. The load distribution over the three tendons was rather inconstant and completely different from the in uivo load distribution (Riemersma and Schamhardt, 1985). This results from the absence of muscle activity. especially of the deep digital flexor muscle. Variation in degree of rigour of the muscle will result in inconsistent tensions of its tendon. Consequently, loads on the parallel tendons will be inversely proportional to the load of the DDFT. Since the load on the legs relative to body weight was also different for different legs, it is emphasized that the tendon loads here presented are not representative for those occurring in rit,o.

REFERENCES

Barnes, G. R. D. and Pinder, D. N. (1974) In sioo tendon tension and bone strain measurement and correlation. .I. Biomechonics 7, 3542. Brown, T. D., Sigal, L., Njus, G. O., Njus, N. M., Singerman. R. J. and Brand, R. A. (1986) Dynamic performance characteristics J, Biomechanics ;f J6~,4.Quid metal straingage. Fung, Y. C. (1973) Bibrheology of soft tissues. Biorheol. IO, 139-155. Lochner, F. K., Mime, D. W., Mills. E. J. and Groom, J. J. (1980) In riro and in vitro measurement of tendon strain in

416

D. J. RIEMERSMA and J. L. M. A. LAMMERTINK

the horse. Am. J. Vet. Res. 42, 1929-1937. Michaux, B., Vanhoutte, J. J., Jafrin, M. Y. and Fontenier, G. (1979) Calibration-free mercury strain gauge plethysmograph. Med. Biol. Engng Comput. 17, 539-542. Nickel, R., Schummer, A. and Seiferle, E. (1986) Anatomy of the Domestic Animals. (Edited by Nickel, R., Schummer, A., Seiferle, E., Frewein, J., Wilkins, H. and Wille, K. H.) Volume I. The Locomotor System ojthe Domestic Animals. Paul Parey, Berlin/Hamburg. Riemersma, D. J. and DeBruyn, P. (1986) Variations in crosssectional area and composition of equine tendons with regard to their mechanical function. Res. Vet. Sci. 41,7-13. Riemersma, D. J. and Schamhardt, H. C. (1982) The cryo-jaw, a clamp designed for in vitro rheology studies of horse digital flexor tendons. J. Biomechanics 15, 619-620. Riemersma, D. J. and Schamhardt, H. C. (1985) In oitro mechanical properties of equine tendons in relation to cross-sectional area and collagen content. Res. Vet. Sci. 39, 263-270.

Riemersma, D. J., Schamhardt. H. C. and Lammertink J. L. M. A. (1985) In viuo tendon load and tendon strain in the horse. Biomechwzics: Current interdisciplinary Research. (Edited by Perren, S. M. and Schneider, E.) pp. 731-736. Martinus Nijhoff Publ., Dordrecht/Boston/Lancaster. Stone, J. E., Madsen, N. H., Milton, J. L., Swinson, W. F. and Turner, J. L. (1983) Developments in the design and use of liquid-metal strain gages. Exp. Mech. 23, 129-139. Viidik, A., Danielson, C. C. and Oxlund, H. (1982) On fundamental and fenomenological models: structure and mechanical properties of collagen, elastin and glycosaminoglycan complexes. Biorheol. 19,437+#51. Woo, S. L. (1982) Mechanical properties of tendons and ligaments I. Quasistatic and non-linear viscoelastic properties. BiorheoL 19, 385-396. Youdin, M. and Reich, T. (1975) Mercury-in-rubber (Whitney) strain gauge; temperature compensation and analysis of error caused by temperature drift. Ann. biomed. Engng 4, 22&231.