J Biomerhonicr. Vol. 3. pp. 5 I-61. Pergam~n PRSS. 1970.

Printed in Great B&in

THE MECHANICS OF THE KNEE JOINT IN

RELATION TO NORMAL WALKING*
.I. B. MORRISON?
Department of Mechanical Engineering, Massachusetts Institute of Technology,
Cambridge, Mass. 02 139, U.S.A.

Abstract- Experimental measurements of normal walking were taken using male and female
subjects. The mechanics of the knee joint were simplified and defined in mathematical terms.
By considering the normal knee joint to function according to the mechanical principals thus
defined, the forces transmitted by the joint were calculated from the experimental data.
The general mechanical concepts of knee action are outlined and the assumptions made in
defining the joint ‘model’ described. The results obtained are presented and discussed in relalion
to the assumptions made.

INTRODUCTION joint positions the individual ligaments

ALTHOUGH studies have been carried
many become taut and to which movements of the
out in the field of joint action, investigations joint they exert a restraining force, quanti-
have in the past been mostly confined to the tative assessment of the restraining forces
analysis of movement. Experiments aimed applied to the joint by the individual ligaments
at the analysis of forces acting on the joints cannot be made nor can the relative impor-
and in the surrounding connective tissues tance of their role in walking be deduced.
are few and of a limited nature. The informa- Literature describing the function of the liga-
tion available on the function of the knee ments at the knee, although abundant in
joint is generally derived from two sources, nature, is often in disagreement. Brantigan
namely the study of normal gait, and experi- and Voshell (194 I), review contradictory
ments conducted on cadavers. statements made by different authors and
Information obtained from the study of present the results of experiments on cadavers
normal gait is limited by the difficulty of meas- which assess the validity of these statements.
uring accurately internal movement of the Despite this, contradictory statements re-
joint due to the presence of the surrounding garding the function of the ligaments continue
tissue. For example, while rotations about to be made.
the long axes of the femur and tibia have been The function of the musculature has been
recorded in gait studies, (Inman et al., 1948), studied in greater detail. Much useful infor-
the position of the long axis of rotation of the mation regarding the function of the muscles
joint (taken to be parallel to the long axis and the phasic relationship of their activity
of the tibia) relative to the knee joint is still has been obtained by the development and
obscure. use of electromyography (University of
Observations made by study of knee joint California, 1953; Lippold and Bigland, 1954;
action in cadavers are limited in their relation Basmajian, 1962; Close, 1964). Experiments
to normal joint function in that the joint is involving the electrical stimulation of muscles
not subject to the force systems arising in have greatly clarified the relationship of
normal joint action. While these experiments muscle force to muscle length and velocity
allow the investigator to record at which of shortening (Roberts, 1967). There is yet,

*Received 11 April 1969.

*Present address: R. N. Physiological Laboratory. Alverstoke. Hants.. England.
_(I
52 J. B. MORRISON

however, little information as to the magni- General mechanical concepts of knee action
tude and interrelation of these quantities The knee is extended by the quadriceps
during normal physical activity such as walk- femoris, assisted by the tensor fasciae lame.
ing. Flexion is caused by the hamstrings, assisted
Forces acting across the articulating sur- by gracilis and sartorius. Gastrocnemius
faces of joints have recently received much and plantaris are also flexors of the knee.
attention due to their importance in the Movements of the joint in other directions
design of internal prostheses and in the under- are prevented mainly by the binding effect
standing of joint lubrication problems. The of the ligaments and the geometry of the
external force system acting at the knee joint articular surfaces.
has been measured by several researchers, Backward gliding of the tibia relative to
Elftman (1940), Marks and Hirschberg (1958), the femur is controlled by the posterior cruci-
Bresler and Frankel(l950). Total force acting ate ligment, and forward gliding of the tibia
across the joint, however, has not been is controlled by the anterior cruciate ligament.
investigated in detail prior to the prCsent Most authors state that adduction of the joint
study. In a previous publication by the author is prevented by the lateral collateral ligament
(1968), total force transmitted by the knee and the cruciate ligaments, abduction by the
joint during level walking was stated to have medial collateral ligament and the cruciate
a maximum value of 2-4 times body weight. ligaments (Brantigan and Voshell, 1941;
Forces acting at the hip joint in level walking Gray’s Anatomy, 1962). Steindler (1955),
have been measured in normal subjects by however, maintains that this movement is
Paul (1965) and in subjects having prosthetic checked entirely by the collateral ligaments.
replacement of the femoral head by Rydell The medial collateral ligament and the cruci-
(1966). Rydell measured maximum force at ates, acting together, limit rotations of the
the hip of a male and female subject to be femur on the tibia (i.e. about the long axis
of the order of 1.8 and 3.3 times body weight of the joint)in all positions of the joint. The
respectively. Paul calculated maximum joint lateral collateral ligament resists lateral ro-
force to be in the range 2.3-5-8 times body tation when the joint is in extension. Medial
weight. or lateral moyement of the femur on the tibia
Experiments designed to estimate the is prevented by interaction between the
coeficient of friction in human joints have tibia1 intercondylar eminance and the femoral
been reported by Charnley (1959), McCutchen codyles, and the restraint of the ligaments.
(1962), Bamett and Cobbold (1962), and Joint movement is mainly a relative sliding
Rydell (1966). These experiments indicate motion of the opposing condyles. In the last
the friction coefficient to be in the range 10-20” of extension, however, the femoral
0@02-O-04. A coefficient in this range is condyles roll forwards slightly on the tibia.
superior to that achieved in engineering bear- As the radius of curvature of the femoral
ings of similar structure. condyles decreases from front to back, the
At this point it is appropriate to mention medio-lateral axis of the joint varies in
the controls imposed on joint moverqent by position, depending on the angle of flexion.
the muscles and ligaments and the natural In normal walking, rotations about the long
range of these movements during normal axis of the joint are small, having a mean range
walking. In the following section, there- of about 9” (Inman et al., 1948). Rotation
fore, a description of the fundamental con- occurs in the last few degrees of extension but
cepts of the mechanics of knee action, upon the exact mechanism is uncertain (Gray’s
which knowledge the present work is based, Anatomy, 1962). Steindler (1955) and Morris
is given. (1953) state that the axis of rotation lies
 THE MECHANICS OF THE KNEE JOINT 53

closer to the medial condyles while Gray forces acting across the knee joint were de-
( 1962) states that it passes through the lateral fined in terms of this system of tibial axes
condyles. The knee resists adduction and (see Fig. 2).
abduction in extension, a limited movement The near cylindrical configuration of the
being possible in flexion (Gray’s Anatomy, femoral condyles and the relative flatness of
1962). Greatest flexion of the joint during the opposing tibia1 articulations implies ap-
walking occurs in the swing phase and is proximately a line contact of surfaces in the
of the order of 75” (Berry, 1952). For most of medio-lateral direction. In the analysis a
the stance phase flexion is less than 20” but line contact was assumed and its position on
increases to about 55” at toe off (Berry, the articular surfaces taken to be coincident
1952). with the Z, axis of the tibia; (Fig. 3). Anterior-
posterior displacements of the line of contact
Functional concepts for analysis from the Z, axis due to the rolling of the
In order to calculate the forces transmitted femoral condyles on the tibia1 condyles in
by the joint articulations and the connective extension were neglected. It was further
tissues under dynamic conditions it was assumed that the femoral condyles rotated
necessary to define the joint structure and relative to the tibia about a fixed centre line
the mechanics of its action in mathematical parallel to the Z, axis and intersecting the
terms. Further, the mathematical model as Y, axis of the tibia; (Fig. 1). This centre line
constructed had to be such that a unique soiu- was taken to be coincident with the axis of
tion of force actions could be calculated for
any position and loading of the joint. In order
to satisfy this condition and in view of the
several aspects of joint mechanics not clearly
defined in the literature, the joint structure and
function as defined in mathematical terms in-
volved a degree of mechanical simplification.
The functional concepts adopted are described
as follows.
A set of reference axes X,, Y, and Z, was
adopted in relation to the tibia. These axes
are shown in Fig. 1. The directions of all

Fig. 2. External force system acting at knee-expressed

in terms of tibia1reference axes.

Contact ore0
between opposing
yles

Lotcral aspect Anterior aspect

Fig. 3. Tibial condyles - superior aspect-showing assum-

Fig. 1. Reference axes of femur and tibia. ed contact area of tibial and femoral condyles.
54 J. B. MORRISON

rotation of the knee joint in the position of of these muscle groups and of the method of
180” extension. determining their force vectors relative to the
Muscles and ligaments of the joint were tibia1 axes are given in a previous publication
defined by the position of their attachments (Morrison, 1968).
and for this purpose a further two sets of The force transmitted by the joint articula-
axes were adopted. Axes relative to the pelvis tions was considered as two components, a
X,, YP and Z, were assumed to have origin at direct compressive force R, acting in the
the anterior superior spine of the left hip bone, direction of the Y, axis of the joint, and a side
and to be coincident with the intersections of or shear force R, acting in the medio-lateral
the planes of the body drawn through that direction. Force R, was assumed to be trans-
point. The femur was defined by axes X,, mitted partly as a friction force acting between
Y, and Z,, having origin at the intersection of the faces of the opposing condyles, and partly
the assumed centre line of the femoral as a compressive force acting between the
condyles and the Y8 axis of the tibia (see concave inner boundary of the tibia1 condyles
Fig. 1). The Y, axis represents the mechanical and the inner boundary of the femoral condy-
axis of the femur and the Z, axis is coincident les. The effects of friction in the joint in the
with the centre line of the condyles. The anterior-posterior direction was neglected.
error in the assumption of a fixed point origin It was therefore assumed that an anteriorly
on the femur relative to the axes of the tibia directed force on the tibia was resisted by the
is small, but increases with degree of flexion anterior cruciate ligament whilst a posteriorly
of the joint. The true position of the origin directed force was resisted by the posterior
on the femur subject to 90” of flexion is shown cruciate ligament. The direction of the force
in Fig. 4. A detailed discussion of this move- imposed on the joint by a ligament was
ment is given by Steindler (1955). defined in terms of the positions of the liga-
Extension or flexion of the knee was con- ment’s attachments relative to the tibia1 axes.
trolled by forces acting in the quadriceps Moments of adduction or abduction acting
femoris, hamstrings or gastrocnemius muscle on the joint were equilibrated by a redistri-
groups. As the hamstrings and gastrocne- bution of pressure on the condyles, i.e. a
mius muscles both tend to flex the joint, displacement of the centre of pressure along
electromyographic data describing muscle the line of contact of the condyles from the
activity during the walking cycle was used to joint centre. As pressure on one condyle
decide which of these two muscle groups were tended to zero, further loading in this direction
active at a given instant. Details of the choice was resisted by a reaction in the collateral
ligament of that condyle. Torsional action at
the joint (i.e. about the Y, axis) was neglected.
f’ The effect of torsion on the calculations is
discussed in the presentation of results.

Experimental procedure and analysis

In the following paragraph a brief account
of experimental procedure and analysis is
given. Details of this section of the investiga-
tions are presented in a previous publication
(Morrison, 1968).
Fig. 4. Centre of rotation of femur. CC’-centre of Subjects were filmed from the front and
rotation at 180” extension. d,d*-centre of rotation at
90” flexion. Error in assumed fixed point origin at 90” side whilst walking along an instrumented
flexion-distance CC’. walkway (see Fig. 5). Reaction between
 THE MECHANICS OF THE KNEE JOINT

boom
2 Camera E.M.G coblas

Fig. 5. Diagram of walkway viewed along the Z, axis (above) and y. axis (below).

ground and’foot during one step was measured more than once, the average values obtained
by a force plate. Accelerations of limb from tests on these subjects are considered
segments were calculated from measure- in conjunction with the values obtained in
ments taken from the tine film records. The single tests on the other subjects. In the fol-
externat force system acting at the knee joint lowing discussion the phrase ‘joint force’
was then calculated by summing ground force denotes the compressive force R, acting
and acceleration forces acting on the limb. normal to the articular surfaces of the tibia.
By considering the knee joint to operate Considering component R, to be totally
according to the mechanical principles de- transmitted by the joint surfaces, the resultant
scribed in the previous section and applying values of the two force components R, and
the experimental results to this ‘joint model’, R, are of the order of O-2 per cent greater
a complete force analysis of the joint under than the values of ‘joint force’ (i.e. component
dynamic conditions could be achieved for any R,) quoted in the results. In all cases the
position of the walking cycle. By computing joint force is measured as a fraction or mul-
forces in this manner for each consecutive tiple of the body weight of the subject.
frame of tine film recorded, the force cycles The complete solution of forces is presented
acting at the articular surfaces and in the for three subjects in Figs. 6-10. The results
muscles and ligaments were obtained. shown in these figures are considered to be
representative of the range of results obtained.
In each figure the results obtained for the
RESULTS three subjects are superimposed in order to
The results presented describe fourteen indicate the degree of variation in force sys-
experiments involving 3 female and 9 male tems developed by different subjects perform-
subjects. All subjects were normal adults, ing the same activity.
11 being in the age range 18-24 yr and 1
male of age 38 yr. It should be noted that ( I) Muscle forces
where average figures are presented, to pre- Maximum force values in the region of
vent bias of results towards subjects tested 400 lb were calculated in all three muscle
56 J. B. MORRISON

I I I I i

0
60 60 100 20 40 60 60

FWcentag6 of cycle

Fig. 6. Joint force at knee-level walking. Test No. 11_;2_----; 13-.-.-_.

Ouadriceps
femaris

S 0
*

; 400
S Hamrtrinps
S 200
::
%
0

400
Gastrocnemius
200 . / a’- \
p-Y ‘\
<’
0 \, 1
60 60 100 20 40 60 60
Percentage of cycle

Fig. 7. Muscle force-level walking. Test No. 11_; 2---_--; 13- ._.__

groups. Mean maximum forces developed by groups (Fig. 7) may be explained as follows.
the twelve subjects in the quadriceps femotis, Immediately prior to heel strike, force action
hamstrings and gastrocnemius were 167, in the hamstrings decelerates the forwards
270 and 234 lb respectively. motion of the leg. At heel strike the foot is
Force actions calculated in the three muscle positioned well in front of the knee and hip
 THE MECHANICS OF THE KNEE JOINT 57

80
60 Anterior
40 crwote
20
0
60
= 60 Posterior
- 40 cruciote
z 20
2 0

Medlol
colloterol

60
60 Loterol
40 colloterol
20
0
60 60 100 20 40 60 60
Percentage of cycle

Fig. 8. Ligament forces during level walking. Test No. I 1_;~___-;*~ _._.

50

.c 0

e
f -50

s
P -100
c
-150

-200
80 I00 20 40 60 60
Percentage of cycle

Fig. 9. Torque M, acting at knee joint during level walking. Test No. 1 l-;
2----_: 13_-.-.-_.

i! Loterol
I ,I711
4% condyte

Pwcentoge of cycle Reference axes

Fig. 10. Position of centre of pressure Z, on condyles during level walking. Test No. 11 -:
I----; 13_._.__.
58 J. B. MORRISON

joints and hence vertical force acting on the which neglects the assistance of tensor
foot causes a moment, --M,, to act at both fasciae latae.
joints. In most experiments this moment was
increased by the action of an anteriorly (2) Forces in the ligaments
directed force on the foot at heel strike. This The magnitude and phasing of the forces
moment action is resisted by force action in calculated in the ligaments is shown in Fig. 8.
the hamstrings which, having a biarticular Maximum force recorded in the cruciate liga-
function, stabilizes both the knee joint and ments of the 12 subjects varied from 10 to
the hip joint (assisted by the gluteal muscles). 112 lb. In all tests the posterior cruciate
The advantage of the biarticular muscle in carried the greater force, mean maximum
these circumstances is illustrated by Elftman force being 74 lb compared with 35 lb in the
(1941). Following heel strike the knee is anterior cruciate. In calculating these forces
subject to a moment, +M,, i.e. a tendency the effect of friction was neglected. A friction
to flex the joint, and force action in the quad- force between the articular surfaces acting in
riceps femoris resists this moment and the anterior or posterior direction may either
controls the position of the knee. In the increase or decrease the force required in the
second half of the stance phase, force action in cruciates depending on the direction of rota-
the calf muscles causes plantar flexion of the tion of the femoral condyles on the tibia. As-
ankle and hence produces forwards accelera- suming a value of 0.02 for the coefficient of
tion of the body and a corresponding anterior- friction in the joint, friction forces would
ly directed force acting at the foot. Moment, have a maximum value of the order of 9 lb,
-M,,, acting at the knee tends to extend the this force corresponding to the position of
joint and is resisted by the action of gastro- maximum joint force in the walking cycle.
cnemius. Being a biarticular muscle the gas- Forces calculated to act in the medial
trocnemius both stabilizes the knee and collateral ligament were small, having a
produces plantar flexion of the ankle (assisted maximum value of 29 lb. Forces transmitted
by soleus). At toe-off force action in the quad- by the lateral collateral ligament were much
riceps femoris imparts a forwards acceleration greater, having a mean maximum value of
to the leg. 59 lb, and a maximum value’of 148 lb. These
It should be noted that in calculation of forces were developed to prevent adduction
the force values acting in the hamstrings it at the knee during the stance phase of walking.
was assumed that there was no assistance of According to the literature surveyed
this muscle group from the gracilis or sar- (Brantigan and Voshell, 1941; Gray’s
torius muscles. Gracilis is mainly an ad- Anatomy, 1962: Eds. D. V. Davies and F.
ductor of the hip and its line of action affords Davies), part of the reaction required to
it little leverage at the knee in comparison resist abduction or adduction may be trans-
to the hamstring muscles. Sartorius, also mitted by the cruciates. The experiments of
having less leverage than the hamstrings, is Brantigan and Voshell (1941) show that
a relatively weak muscle. it is reasonable when the knee is in extension the collateral
to assume therefore that the components of ligaments are capable of preventing movement
moment, -M,, transmitted by these two in this direction, no instability occurring when
muscles are of a minor nature and the error the cruciates were severed. In flexion, how-
introduced by assuming the hamstrings ever, due to the slackening of the lateral
group to be the sole flexor of the joint is collateral, absence of the cruciates resulted
small. The same argument may be applied to in increased instability of the joint. This would
justify the assumption of the quadriceps indicate that the assistance of the cruciates in
femoris as the sole extensor of the knee, checking abduction or adduction is more
 THE MECHANICS OF THE KNEE JOINT 59

significant in positions of flexion. From a posterior fibres of the medial collateral liga-
mechanical point of view the collateral liga- ment would be most favourably placed to
ments are best situated to resist abduction resist the inward torque shown in Fig. 9. and
and adduction and, provided there is no initial it is suggested that the greater part of the
slackness in the ligaments, there must conse- torque acting at the knee will be balanced by
quently be more strain in the collateral tension in this ligament. Tension in this
ligament resisting movement than in the cruci- ligament required to produce equilibrium
ates. The large forces calculated to be acting would not significantly increase calculated
in the lateral ligament occurred in the stance values of joint force as force in the posterior
phase of walking and hence at positions of cruciate, which is also in tension at this part
the joint in which this ligament would nor- of the cycle, would tend to be reduced in
mally be taut. It is deduced therefore that in order to maintain equilibrium of forces in the
walking the collaterals rather than the cruci- anterior-posterior direction.
ates provide the major reaction to moments of
abduction or adduction acting on the joint.
With regard to forces calculated in the (3) Forces transmitted by the articular sur-
lateral collateral, it should be noted that the faces
forces ascribed to this ligament for the pur- Joint force results presented in Fig. 6
pose of analysis. will most likely be carried represent the tests in which the greatest,
partly as tension in the ilio-tibia1 tract. The average and smallest values of maximum
distribution of force between these two ele- joint force, R,, were obtained. The three
ments could not be determined from the infor- main peaks in the joint force curve, referred
mation available. It may also be possible that to as peaks a, b and c, in Fig. 6 correspond to
the large adduction force acting on the joint the peaks of muscle force shown in Fig. 7.
during the stance phase is partly equilibrated For the 12 subjects tested the average
by differential action of the lateral and medial values of peaks a, b and c were 2.43, 2-23
hamstrings or of the lateral and medial heads and 2.72 times body weight respectively.
of gastrocnemius. At present however there Maximum joint force measured varied
is no experimental evidence to support this between 2.06 and 4-O with an average value
theory. It has been suggested by several of 3.03 times body weight. Variations in
authors that tension receptors in the articular peak values of joint force and their phasing
ligaments, when activated, instigate a reflex in the walking cycle shown by different
contraction in the muscles capable of protect- subjects are considered by the author to be
ing the particular ligament. This hypothesis is due partly to anthropometric differences
suggested in relation to the ligaments of the between subjects and partly to differing char-
knee joint in particular by Smillie (1946). acteristics in the gait of the subjects. Where
Experiments by Stener (1959) investigating a subject was tested twice, the joint force
the medial collateral ligament of the knee curves obtained from the two tests bore
joint contradict this hypothesis. close comparison, the magnitude of the
Figure 9 shows an increasing inward three peaks and the percentage of the walking
torque, -M,, acting at the knee during cycle at which they occurred being in good
the stance phase of walking. Moment action, agreement.
M,, about the long axis of the tibia must be Tests on male and female subjects revealed
balanced mainly by force actions in the no obvious differences in the magnitude or
ligaments and will alter the distribution of cyclic variation of joint force between the
ligament forces calculated in the analysis. sexes. In Figs. 6-l 0 tests 11 and 13 represent
From mechanical considerations the oblique the results obtained using female subjects
60 J. B. MORRISON

and test 2 represents the results obtained lateral condyle, the stresses acting in the shaft
using a male subject. From the results of of the tibia would be less.
the 14 experiments conducted it would
appear that variations in joint force within SUMhlARY OF RESULTS
the sex groups, due possibly to anthropo- (1) Maximum joint force calculated at the
metric factors and individual gait character- knee during walking was in the range 2-4
istics, are greater than any variations between times body weight, the average value of 12
the sexes due to anatomical differences. subjects being 3.03 times body weight.
The side or shear force, R,,(measured as a (2) When the joint was highly loaded, the
fraction of body weight) acting on the condyles greater portion of the load was transmitted
is shown in Fig. 6. Due to its relatively small by the medial condyles.
magnitude in this graph, the curves of three (3) Forces acting on the joint in the medio-
subjects could not be superimposed clearly lateral direction were generally small. Their
and only the curve of one subject is therefore mean maximum value was calculated to be
shown. Mean maximum value of R,calculated 0.26 times body weight.
for the 12 subjects was O-26 times body (4) The mean maximum forces acting in the
weight. anterior and posterior cruciate ligaments
Experimental results of all subjects indicat- and the medial and lateral collateral ligaments
ed that during the stance phase of walking the were 35,74,14 and 59 lb respectively.
centre of pressure was positioned over the (5) The greatest muscle force calculated was
medial condyles, as shown in Fig. 10. It 405 lb. For the 12 subjects tested the average
therefore follows that the greater part of the values of maximum force developed in the
joint force was transmitted by the medial quardiceps femoris, hamstrings and gastro-
condyles at this part of the cycle. The above cnemius muscle groups was 167, 270 and
statement is at variance with the commonly 234 lb respectively.
held belief that the greater load is trans- (6) No significant difference was apparent
mitted by the lateral condyles. Steindler between the joint forces calculated for male
(1955) states that the greater amount of pres- and female subjects.
sure is borne by the lateral condyles ‘because (7) It is essential that the values quoted
of the obliquity of the anatomical axis of the above are interpreted in relation to the as-
femur’. It is maintained by the author that the sumptions made in defining the mechanics of
obliquity of the axis of the femur cannot the knee joint. The limitations of the analysis
possibly affect the force system acting at the in determining the true value of these quanti-
knee joint, and consequently the distribution ties are discussed in detail with the presenta-
of pressure between the condyles. The force tion of results.
system acting at the knee is dependent, Knowledge of the force values transmitted
rather, on the position of the centre of gravity by the joint tissues is of importance in the
of the body and the external forces acting further development of reconstructive joint
upon it. From a mechanical point of view, a surgery, in the design of mechanical compon-
greater portion of the force transmitted by the ents for partial or total joint replacement, and
medial condyles as opposed to the lateral in the understanding of joint lubrication. It is
condyles would be structurally more favour- hoped that the work described in this paper
able for the following reasons: as the medial will contribute to the solution of problems
condyle probably has a larger bearing surface, in these areas and also to the general under-
the compressive stress at the articular surface standing of the mechanics of the knee.
would be lower; as the medial condyle Acknowledgements-The work described here was
overhangs the shaft of the tibia less than the carried out at the Bio-Engineering Unit, University of
 THE MECHANICS OF THE KNEE JOINT 61

Strathclyde, Glasgow, and financed by the Medical Inman, V. T. er al. (1948) Transverse rotations of the
Research Council. The author wishes to thank Professor segments of the lower extremity in locomotion.
R. M. Kenedi and Dr. J. P. Paul for guidance in the J. Bone Jt Surg. %A, 859.
method of analvsis and Mr. D. N. Condie for assistance Liooold. 0. C. and Bi&nd. B. (1954) The relation
in the execution of experiments and computation of between force, velo&y and integrated electrical
results. activity in human muscles. J. Physiof. 123,2 13.
Marks, M. and Hirschberg. G. (1958) Analvsis of hemi-
plegic gait. Ann. N.Y. A&d. Sci. 74.59. _
REFERENCES McCutchen, C. W. (1962) The frictional properties of
Bamett. C. H. and Cobbold, A. F. (1962) Lubrication animal joints. Wear 5, I- 17.
within living joints. J. BoneJt Surg. 44B, 662. Morris, H. (1953) Human Anatomy. (Edited by J. P.
Basmajian, J. V. (1962) Muscles Alive, Their Functions Schaeffer), 1 Ith Edn. Blakiston, New York.
Revealed by Ekctromyography. Williams and Wilkins, Morrison, J. B. (1968) Bioengineering analysis of force
Baltimore. actions transmitted by the knee joint. Bio-med. Engng
Berry, F. R. (1952) Angle variation patterns of normal 3, 164.
hip, knee and ankle in different operations. Unit. Paul, J. P. (1965). Bioengineering studies of the forces
Calif. Prosth. Dev. Res. Rep. Ser. 2, Issue 2 1. transmitted by joints-II. Biomechanics and Related
Brantigan, 0. C. and Voshell. A. F. (1941) The mechanics Bioengineerina Topics. (Edited by R. M. Kenedi).
of the ligaments and menisci of the knee joint. J. p. 369: Pergamon Press. Oxford. _
Bone J? Surg. 23,44. Roberts, T. D. M. (1967) Neurophysiology of Postural
Bresler, B. and Frankel, J. P. (1950) The forces and Mechanisms. Plenum, New York.
moments in the leg during level walking. Trans. Am. Rydell. N. (1966) Forces Acting on the Femoral Head
Sot. mech. Engrs 72.27. Prothesis. (Edited by A. B. Tryckeri), Litotyp. Goten-
Charnley, J. (1959) The lubrication of animal joints. burg, Sweden.
Proc. Symp. Biomechanics lnstn mech. Engrs, London, Smillie. 1. S. (1946) Iqjuries of the Knee Joint. Living-
p. 12. stone, Edinburgh.
Close, J. R. ( 1964) Motor Function in the Lower Extrem- Steindler. A. ( 1955) Kinesiofogy. Thomas, Springfield,
ity. Thomas, Springfield, Ill. Ill.
Elftman, I-& (1941) The action of muscles in the body. Stener, B. (1959) Experimental evaluation of the hypoth-
Biol. Symp. 3, 19 1. esis of ligamento-muscular protective reflexes. Actu.
Elftman, H. (1940) Forces and energy changes in the phys. stand. 48, Suppl. 166.
leg during walking. Am. J. Physiol. 129,672. University of California (1953) The pattern of muscular
Gray’s Anatomy. (1962) (Edited by D. V. Davies and activity in the lower extremity during walking. Univ.
F. Davies), 33rd Edn. Longmans, London. Cal$ Prosrh. Dev. Res. Rep. Ser. 2, Issue 25.