Clinical Anatomy 16:215–223 (2003)

ORIGINAL COMMUNICATION

Force-Length Characteristics of the In Vivo Human

Gastrocnemius Muscle
CONSTANTINOS N. MAGANARIS*
Centre for Biophysical and Clinical Research into Human Movement, Manchester Metropolitan University, Alsager, United Kingdom

In this study, the force-length characteristics of the in vivo medial (GM) and lateral (GL)
heads of the human gastrocnemius muscle were estimated from measurements in eight
healthy male subjects. This involved: 1) dynamometry-based measurements of the moment
generated during maximal isometric plantar flexion; 2) ultrasound-based measurements of
fascicular length and pennation angle; and 3) ultrasound-based calculations of moment arm
lengths. All measurements were taken over the ankle angle range from 20° of dorsiflexion to
30° of plantar flexion. Tendon forces were calculated by dividing the moments recorded by
the muscle moment arm lengths, and fascicular forces were calculated by dividing the tendon
forces estimated by the cosine of pennation angle. In the transition from 30° of plantar flexion
to 20° of dorsiflexion, the GM muscle fascicular length and force increased linearly from 24
to 39 mm and from 222 to 931 N, respectively. Over the same ankle angle range, the GL
muscle fascicular length and force increased linearly from 30 to 47 mm and from 139 to 393
N, respectively. Estimates of the sarcomeric lengths corresponding to the fascicular lengths
measured indicated that the two muscles operated in the range 1.4 –2.2 ␮m, below the optimal
length region for force generation according to the cross-bridge mechanism of contraction.
These results indicate that the force–length relation of the in vivo human gastrocnemius
muscle is limited to the ascending limb of the bell-shaped force–length curve obtained from
experiments on isolated material. Clin. Anat. 16:215–223, 2003. © 2003 Wiley-Liss, Inc.

Key words: ultrasound; length-tension; muscle contraction; moment arm; sar-

comere

INTRODUCTION mechanical behavior of a pennate structure to that of

The classic experiments by Gordon et al. (1966)
showed that the force–length relation of isolated mus-
cle fibers approximated a bell-shaped pattern. The
level of contractile force elicited was related to the
extent of myofilament overlap in the sarcomere, a
relationship that gave support to the cross-bridge the-
ory of contraction introduced by Huxley (1957).
As opposed to isolated fiber experiments, the op-
erating length range of muscle fibers under in vivo
conditions is confined by the anatomical constraints of
the skeleton and the architectural features of the mus-
cle. Due to lack of technology for measuring muscular
architecture in vivo, determining the force–length
characteristics of an intact muscle has necessitated
either presuming that cadaver-based measurements of
fascicular length and pennation angle are representa-
tive of those during muscle contraction (Cutts, 1988)
or neglecting pennation effects and approximating the
a parallel-fiber structure (Herzog and ter Keurs,
1988a,b; Herzog et al., 1991). The assumptions in both
approaches, however, are clearly unrealistic. The ar-
chitecture of a muscle changes during contraction due
to a stretch in the serial tendons and aponeuroses
(Narici et al., 1996; Kawakami et al., 1998; Maganaris
et al., 1998; Martin et al., 2001) and the fascicles in a
pennate muscle shorten upon contraction less than
those in a parallel fiber muscle (for review see Gans,
1982). Thus, the force–length relations reported
*Correspondence to: Constantinos N. Maganaris, Centre for Bio-
physical and Clinical Research into Human Movement, Manches-
ter Metropolitan University, Alsager ST7 2HL, UK.
E-mail: C.N.Maganaris@mmu.ac.uk
Received 23 November 2001; Revised 19 April 2002
Published online in Wiley InterScience (www.interscience.wiley.
com). DOI 10.1002/ca.10064

© 2003 Wiley-Liss, Inc.

216 Maganaris
should be viewed with caution when referring to in
vivo muscle function.
Ultrasonography, however, has recently made it
possible to measure in vivo muscular architecture with
high accuracy and reproducibility (Narici et al., 1996;
Kawakami et al., 1998; Maganaris et al., 1998). We
used ultrasonography and dynamometry to estimate
the force–length characteristics of the intact human
gastrocnemius muscle.
MATERIALS AND METHODS
Experimental Design
The force–length relation of the gastrocnemius
muscle of the right limb was examined in eight
healthy male volunteers (age, 22 ⫾ 2 years; height,
173 ⫾ 3 cm; and body mass, 75 ⫾ 3 kg; mean ⫾ SD)
with approval from the institutional ethics committee.
Measurements were taken at ankle angles of ⫺20°
(dorsiflexion direction), ⫺10°, 0° (the sole of the foot
at right angles to the longitudinal axis of the tibia),
⫹10° (plantar flexion direction), ⫹20°, and ⫹30°. All
subjects could passively rotate the foot from ⫹30 to
⫺10°, but a slight force was applied externally to bring
the foot to ⫺20°. At each ankle angle studied, the
following parameters were either measured or calcu-
lated: 1) the moment elicited upon isometric plantar
flexion maximal voluntary contraction (MVC); 2) the
gastrocnemius muscle fascicular length and pennation
angle; and 3) the gastrocnemius muscle moment arm
length.
Measurement of MVC Moment
MVC contractions were elicited in the prone posi-
tion on the bench of an isokinetic dynamometer (Lido
Active, Loredan Biomedical, Davis, CA). The pivot
point of the dynamometer lever was aligned with the
lateral malleoli and the foot was firmly secured at the
dynamometer footplate with Velcro straps. Measure-
ments were taken after compensating for gravitational
and passive moments about the ankle. After a warm-
ing up protocol consisting of three submaximal con-
tractions, the subjects generated four MVCs with the
knee fully extended (0°) and two MVCs with the knee
flexed at 120°. MVCs were elicited randomly for knee
and ankle angles, and were separated by a 2-min rest.
Measurements were repeated until obtaining the
MVC moments established in a familiarization trial
(⫾5%) 2 to 3 days earlier.
Conventional dynamometers record moments in
only one plane in each test. In the present experi-
ment, moment measurements were taken in the sag-
ittal plane. In vitro and in vivo experiments have
shown that the lateral side of the tibiotalar joint axis
points downwards and posteriorly by 10° relative to
the medial side (Isman and Inman, 1969; van den
Bogert et al., 1994). This geometric arrangement re-
sults in a triplanar orientation in the moment pro-
duced about the tibiotalar joint; however, ankle mo-
ment measurements in the sagittal plane include only
the cosine component of this moment. Trigonometry-
based calculations using the above axis angle values
indicate that the measurable component of moment
accounts for 97% of the entire moment vector. Thus,
the present MVC moment measurements were con-
sidered to be valid.
Measurements of Fascicular Length and
Pennation Angle
Fascicular lengths and pennation angles were mea-
sured in vivo from sonographs taken during MVCs
elicited at the knee extension angle (Narici et al.,
1996; Kawakami et al., 1998; Maganaris et al., 1998;
Martin et al., 2001). A linear-array, 7.5 MHz, B-mode
ultrasound probe (Esaote Biomedica, Florence, Italy),
with width and depth resolutions of 1 mm and 0.62
mm, respectively, was coated with transmission gel
and placed over the midlongitudinal axes of the me-
dial (GM) and lateral (GL) heads of the gastrocnemius
muscle, 5 cm above (first and third MVCs) and 5 cm
below (second and fourth MVCs) the midlength of the
muscle belly during MVC (Fig. 1). The architecture of
the two muscle heads in the above regions has been
shown to be representative of that along and across the
muscle belly (Narici et al., 1996; Maganaris et al.,
1998). In the muscle scans recorded, fascicular, inter-
fascicular, and aponeurotic echoes were identified.
The criterion adopted for ensuring that the scanning
plane coincided with the plane in which the fascicles
lay was that the echo of the entire fascicular path
could be seen in the scans (Narici et al., 1996;
Kawakami et al., 1998; Maganaris et al., 1998). This
criterion was met in most of the scans taken with the
scanning head placed in the midsagittal plane of the
muscle. If, however, this probe position generated
echoes in only a part of the fascicular path, the probe
was oriented manually to a plane that included the
entire length of the fascicles. In each scan, the length
of one to three fascicles was digitized using comput-
erized image analysis (NIH Image, National Institute
of Health, Bethesda, MD), taking into account any
visible curvature along the fascicular path. Pennation
angle measurements were taken at the fascicular in-
sertions in the deep aponeurosis (Fig. 2).

Fig. 1. Schematic representation of the ultrasound scanning sites.
The muscle belly shown refers to either of the two muscles examined
in the right limb as seen from the back.
Calculation of the Moment Arm Length
Moment arm lengths were obtained using the ex-
cursion method (An et al., 1984), adapted under in
vivo conditions as described by Ito et al. (2000) and
Maganaris (2000). This method necessitates identify-
ing a reference landmark along the action line of the
muscle and measuring displacement of the landmark
over a given rotation in the joint that the muscle
spans. Because the action lines of the GM and GL
muscles do not coincide and the axis of rotation in the
tibiotalar joint is not perpendicular to the sagittal
plane, separate measurements were taken to quantify
the displacements of the reference landmarks in the
two muscles. The reference landmark in each muscle
was the end point of the echo generated by the ten-
don in the distal myotendinous junction of the mus-
cle. To locate the myotendinous junction, five axial-
plane sonographs were recorded (with the scanning
probe described above) along the central portion of
the GM and GL muscle bellies at 2-cm intervals. The
In Vivo Force–Length Characteristics 217
medial and lateral borders of each muscle in each scan
were marked on the skin to identify the midlongitu-
dinal axis of the muscle, which was also drawn on the
skin. The probe was positioned in the sagittal plane
along the midlongitudinal axis of each muscle to lo-
cate the exact position of the myotendinous junction.
The probe was oriented in a plane reflecting clear
echoes and secured with adhesive tape on the skin.
Scans were first recorded having the ankle plantar
flexed by 5° more than the angle at which the moment
arm length was to be obtained (Figs. 2, 3A). The same
procedure was repeated with the ankle plantar flexed
by 5° less, compared to the angle at which the mo-
ment arm length was to be obtained. The displace-
ment dx of the tendon endpoint in each myotendi-
nous junction over the 10°-rotation (d␪) was then
digitized in the scans taken. The absence of medio-
lateral shifts in the gastrocnemius muscle during plan-
tar flexion-dorsiflexion rotation (Fukunaga et al.,
2001) indicates that all dx measurements in each mus-
cle were taken in a single muscular section. The
moment arm lengths of the GM and GL muscles were
calculated from the ratios dx/d␪ (d␪ in rad).
Data Analysis
The moment-producing potential of the gastrocne-
mius muscle at each ankle angle was estimated by
subtracting the MVC moment at the knee flexion
angle from that at the knee extension angle. For each
knee angle, the MVC that generated the highest mo-
ment was considered. The above calculation assumes
that at any given ankle angle: 1) the gastrocnemius
muscle is fully active during an MVC with the knee
extended; 2) the biarticular gastrocnemius musculo-
tendon unit is slack at a knee flexion angle of 120° and
thus it transmits upon MVC negligible contractile
forces to the calcaneus; and 3) the moment produced
by other muscles crossing the ankle joint is the same
between MVCs generated at knee angles of 0 and
120°. The validity of the first assumption could be
tested by superimposing electrical current upon the
MVCs carried out, referred to as the “twitch superim-
position technique” (Belanger and McComas, 1981;
Allen et al., 1995). Unfortunately, we found the appli-
cation of this method in the experimental trial imprac-
tical: placement of stimulating electrodes over the
muscle prohibited muscle scanning, and placement of
stimulating electrodes over the tibial nerve in the
popliteal fossa proved to be ineffective at the knee
flexion angle due to loose contact between the elec-
trodes and the skin. However, to assess whether our
subjects were capable of activating fully the gastroc-
nemius muscle upon volition, the twitch superimpo-
sition technique was applied in the familiarization
218 Maganaris
Fig. 2. Left: Typical sonographs 5 cm above the midlength of the
GM (top) and GL (bottom) muscle bellies at an ankle angle of ⫺20°.
The horizontal white striations are aponeurotic echoes, the oblique
white striations are interfascicular echoes, and the oblique black stri-
ations are fascicular echoes. The superimposed white line at the top
indicates the path of a fascicle between the superficial and deep
aponeuroses. ␾, the angle of pennation; SF, subcutaneous fat; GM,
trial. For these measurements, twitch doublets of
150 –175 V were applied to the GM and GL muscles
through two self-adhesive rubber electrodes of dimen-
sions 7.5 ⫻ 12.5 cm. Our pilot measurements showed
no evidence of incomplete motor unit activation dur-
ing MVC. The validity of the second assumption is
supported by previous EMG-based studies (Hof and
van den Berg, 1977; Gravel et al., 1987) and by recent
results showing that knee flexion above 115° yields no
measurable decrease in the moment generated during
plantar flexion contraction (Maganaris, 2001). The
third assumption has been discussed in detail by Her-
zog and ter Keurs (1988a) and Herzog et al. (1991):
because all the synergists and antagonists of the gas-
trocnemius muscle span the ankle joint only, their
moment contribution to the net MVC moment at any
given ankle angle would not differ between knee
angles. This view has been challenged by moment
calculations from in vivo measurements of Achilles
medial head of the gastrocnemius muscle; GL, lateral head of the
gastrocnemius muscle; SOL, soleus muscle; FHL, flexor hallucis lon-
gus muscle. Right: Typical sonographs over the medial head of the
gastrocnemius myotendinous junction at ankle angles of ⫺5° (top) and
⫹5° (bottom). M, muscle; DA, deep aponeurosis; MTJ, myotendinous
junction; T, tendon. Note the negligible angulation between the ori-
entations of the deep aponeurosis and tendon.
tendon force in a recent study (Arndt et al., 1998);
however, these measurements were taken in only one
subject. More importantly, the authors assumed that
the length of the Achilles tendon moment arm does
not change as a function of muscle-tendon length and
contraction-state, which later proved to be incorrect
(Maganaris et al., 1998).
The moment contribution of the GM and GL mus-
cles to the entire gastrocnemius muscle moment was
obtained from the relative physiological cross-sec-
tional areas of the GM and GL muscles to that of the
entire gastrocnemius muscle: 70 and 30%, respec-
tively (Huijing, 1985; Fukunaga et al., 1992). Tendon
forces in each muscle were calculated by dividing the
estimated moment by the moment arm length of the
muscle (Fig. 3B). Fascicular forces were estimated by
representing the muscle as a parallelogram with apo-
neuroses and tendons operating over the same line
(Gans, 1982; Cutts, 1988; Narici et al., 1996;
 In Vivo Force–Length Characteristics 219

Kawakami et al., 1998; Maganaris et al., 1998). The

fascicular force elicited upon contraction was there-
fore obtained by dividing the tendon force by the
cosine of pennation angle (Fig. 3B). The selection of
the above model as opposed to that developed by
Huijing and Woittiez (1985) was based on visual in-
spection of the aponeuroses and tendons in the scans
taken over the myotendinous junction region (Fig. 2).
Nevertheless, it should be recognized that although
simple unipennate muscle models (with or without an
angle between the tendons and the aponeuroses) are
computationally efficient for estimating changes in
forces when altering muscle length, they are mechan-
ically unstable. Modelling mechanically stable archi-
tectures requires detailed information about the arcs
over which the fascicles and the tendinous sheets
operate (Van Leeuwen and Spoor, 1992), which can-
not be obtained in vivo at present.
Direct measurements on cadaveric human muscles
have previously shown that ultrasonography provides
valid measurements of fascicular length and penna-
tion angle (Kawakami et al., 1993; Narici et al., 1996).
Inter- and intraobserver variations of the scanning
procedures and image analyses involved have been
confirmed to be less than 8% (Maganaris et al., 1998;
Ito et al., 2000; Maganaris, 2000, 2001). In the present
study, all scan morphometrics were carried out three
times and mean values were used for further analysis.
In each muscle, average values of pennation angle
across all fascicles traced and both muscular regions
scanned were considered in the analysis.

RESULTS
Tables 1 and 2 show the measured and calculated
values of the parameters examined. The following
observations can be made from the data shown:

1. The MVC moment at the knee extension angle

Fig. 3. A: Illustration of the method used to estimate moment
arm lengths. The lower extremity model shown refers to either of the
two muscles examined for estimating the moment arm length at the
ankle angle of 0° (the sole of the foot at right angles to the longitudinal
axis of the tibia). The excursion dx of the myotendinous junction
(MTJ) in the transition from ⫹5° (top) to ⫺5° (bottom) was measured
in the ultrasound scans taken. This excursion was then divided by the
10° (0.175 rad) ankle rotation carried out to obtain the moment arm
length of the muscle. B: Functional line diagram of the musculoskel-
etal model used to calculate forces at the ankle angle of 0°. The model
shown refers to either of the two muscles examined. An MVC gener-
ates the fascicular force vector F1, part of which (F2) is transmitted
along the tendon to produce a plantar flexion moment M about the
ankle. According to the moment equilibrium equation, F2 ⫽ M 䡠 t⫺ 1,
where t is the moment arm length of the muscle estimated as shown in
the above panel. According to the vectorial analysis of forces, F1 ⫽
F2 䡠 cos⫺1␾, where ␾ is the pennation angle of the muscle.
was higher by 57– 65% than that at the knee
flexion angle. This difference is in line with
previous results (Sale et al., 1982; Kawakami et
al., 2000a) and reflects the ineffectiveness of the
gastrocnemius musculotendon unit to transmit
contractile forces and generate calcaneal dis-
placements at knee angles above 90° (Hof and
van den Berg, 1977; Gravel et al., 1987; Maga-
naris, 2001).
2. Both muscles exhibited similar architectural fea-
tures between the two regions scanned, in agree-
ment with previous studies reporting a uniform
geometry along the muscle belly (Narici et al.,
1996; Maganaris et al., 1998).
3. The two muscles exhibited similar patterns of
220 Maganaris

TABLE 1. Parameters Measured as a Function of Ankle Angle*

Ankle angle (degrees)

Dorsiflexion 7 Plantar flexion
⫺20 ⫺10 0 ⫹10 ⫹20 ⫹30
MVC moment at knee extension (Nm) 201 ⫾ 17 191 ⫾ 17 172 ⫾ 15 157 ⫾ 13 117 ⫾ 11 86 ⫾ 8
MVC moment at knee flexion (Nm) 124 ⫾ 11 117 ⫾ 11 109 ⫾ 9 99 ⫾ 8 73 ⫾ 6 52 ⫾ 4
GM muscle fascicular length A (mm) 39 ⫾ 6 34 ⫾ 5 33 ⫾ 5 29 ⫾ 4 28 ⫾ 4 25 ⫾ 3
GM muscle fascicular length B (mm) 39 ⫾ 6 36 ⫾ 5 31 ⫾ 4 30 ⫾ 4 25 ⫾ 4 24 ⫾ 4
GL muscle fascicular length A (mm) 46 ⫾ 7 45 ⫾ 7 41 ⫾ 6 38 ⫾ 5 32 ⫾ 4 28 ⫾ 4
GL muscle fascicular length B (mm) 47 ⫾ 8 44 ⫾ 7 39 ⫾ 6 34 ⫾ 6 33 ⫾ 5 32 ⫾ 4
GM muscle pennation angle A (deg) 30 ⫾ 4 36 ⫾ 5 41 ⫾ 5 45 ⫾ 6 48 ⫾ 7 50 ⫾ 7
GM muscle pennation angle B (deg) 32 ⫾ 5 36 ⫾ 5 43 ⫾ 6 43 ⫾ 6 48 ⫾ 6 53 ⫾ 7
GL muscle pennation angle A (deg) 20 ⫾ 3 24 ⫾ 3 27 ⫾ 4 29 ⫾ 5 31 ⫾ 5 32 ⫾ 5
GL muscle pennation angle B (deg) 22 ⫾ 3 25 ⫾ 4 27 ⫾ 5 30 ⫾ 5 30 ⫾ 5 33 ⫾ 5

*MVC, maximum voluntary contraction; GM and GL, medial and lateral heads of the gastrocnemius muscle, respectively; A and B, 5 cm
above and 5 cm below the midlength of the muscle belly, respectively. Fascicular lengths and pennation angles are average values from
one to three fascicles. Values are mean ⫾ SD (n ⫽ 8).

changes in any given parameter examined as a
function of ankle angle, not surprising given that
both muscles are anatomic subunits of the same
actuator (the gastrocnemius muscle).
4. Neither muscle exerted its maximal force at the
ankle angle of 0°, in contrast with the general
notion that skeletal muscles generate maximal
forces at neutral anatomical positions (see also
Lieber et al., 1994).
The observed changes in all parameters examined
as a function of ankle angle agree with several previ-
ous studies (Sale et al., 1982; Rugg et al., 1990; Herzog
et al., 1991; Narici et al., 1996; Kawakami et al., 1998;
Maganaris et al., 1998, 2000).
As shown in Figure 4A, the contractile force of each
muscle increased as a function of fascicular length. In
either muscle, fitting the force–length data with linear
models produced R2 ⬎ 0.98. No further improvement
in R2 (P ⬎ 0.05; Student’s t-test) was obtained by
fitting the data using second and third order polyno-
mials. The starting point and the slope of the force–
length relation were different between the two mus-
cles, but the normalized force–length curves were
very similar (Fig. 4B), thus indicating that the two
muscles operated over similar relative length regions.
DISCUSSION
Determination of force-length characteristics in in-
tact human muscles has often been attempted; how-
ever, the following unrealistic assumptions have been
made: 1) the architecture of cadaveric muscles repre-
sents muscular architecture during contraction (Cutts,
1988); 2) pennate muscles behave mechanically like
parallel fiber muscles (Herzog and ter Keurs, 1988b;
Herzog et al., 1991); 3) antagonist muscles are inactive
during contraction (Ichinose et al., 1997); and 4) the
operating range of a muscle during submaximal and
maximal contractions is the same (Leedham and
Dowling, 1995). In the present experiment, these lim-
itations have been eliminated. An additional advan-
tage of the present methodology is that it allows esti-
mates to be derived of the theoretical operating range
of the average sarcomere. Dividing the GM and GL
fascicular lengths in the present study by 17,600 and

TABLE 2. Parameters Calculated as a Function of Ankle Angle*

Ankle angle (degrees)

Dorsiflexion 7 Plantar flexion
⫺20 ⫺10 0 ⫹10 ⫹20 ⫹30
GM muscle muscle moment (Nm) 54 ⫾ 4 52 ⫾ 4 44 ⫾ 4 41 ⫾ 3 31 ⫾ 3 23 ⫾ 2
GL muscle moment (Nm) 23 ⫾ 2 22 ⫾ 2 19 ⫾ 2 17 ⫾ 2 13 ⫾ 2 11 ⫾ 1
GM muscle moment arm length (mm) 51 ⫾ 4 54 ⫾ 4 60 ⫾ 4 63 ⫾ 4 68 ⫾ 4 70 ⫾ 4
GL muscle moment arm length (mm) 55 ⫾ 4 58 ⫾ 4 62 ⫾ 4 64 ⫾ 4 67 ⫾ 4 69 ⫾ 4
GM tendon force (N) 1055 ⫾ 82 960 ⫾ 73 731 ⫾ 57 650 ⫾ 45 455 ⫾ 35 327 ⫾ 29
GL tendon force (N) 416 ⫾ 35 379 ⫾ 31 306 ⫾ 24 265 ⫾ 20 194 ⫾ 15 160 ⫾ 13
GM fascicular force (N) 931 ⫾ 95 810 ⫾ 79 577 ⫾ 52 500 ⫾ 48 331 ⫾ 35 222 ⫾ 20
GL fascicular force (N) 393 ⫾ 42 350 ⫾ 38 278 ⫾ 26 235 ⫾ 23 171 ⫾ 18 139 ⫾ 15

*GM and GL, medial and lateral heads of the gastrocnemius muscle, respectively. Values are mean ⫾ SD (n ⫽ 8).

Fig. 4. A: Fascicular force-length relations of the GM and GL
muscles. The linear models used to fit the data in each muscle are also
shown. Data are mean ⫾ SD (n ⫽ 8). B: Normalized fascicular
force–length relations of the GM and GL muscles. Average values
from all eight subjects are presented.
21,300 in-series sarcomeres, respectively (Huijing,
1985), indicated that the average sarcomere operated
at between 1.4 and 2.2 ␮m. The optimal sarcomeric
length region of 2.6 –2.8 ␮m (Walker and Schrodt,
1974; Lieber et al., 1994) would be obtained at fascic-
ular length regions between 46 and 49 mm in the GM
muscle and between 55 and 60 mm in the GL muscle.
Extrapolating the fascicular length–ankle angle rela-
tions in the present experiment (GM muscle: y ⫽
⫺3.4x ⫹ 110.4, R2 ⫽ 0.99; GL muscle: y ⫽ ⫺2.859x ⫹
114.61, R2 ⫽ 0.99) indicated that the above fascicular
lengths would be reached over the ankle angle range
between ⫺43° and ⫺57°. These angles far exceed the
physiological end-range dorsiflexion ankle angle in
In Vivo Force–Length Characteristics 221
man, ⬃25° (Sammarco et al., 1973; Nigg et al., 1992).
It must be emphasized, however, that the above cal-
culations assume that the lengths of a fascicle and the
fibers contained in it coincide. Indeed, there is evi-
dence from staining-based studies on the number and
position of motor end-plates in cadaveric specimens
that the gastrocnemius muscle fibers span the entire
fascicular length (Kawakami et al., 2000b).
The fascicular force–length relation and the theo-
retical sarcomeric operating range obtained indicate
that the gastrocnemius muscle has more in-series sar-
comeres than those required to generate maximal ac-
tivation forces at optimal lengths. This finding, how-
ever, should not be taken to imply that physiological
function is associated with operation of the gastrocne-
mius muscle outside its plateau region. Physiological
functioning requires generating smaller muscular
forces than those produced during maximal isometric
contraction, thus resulting in less myofilament overlap
and longer sarcomeric lengths. During locomotion, for
example, the human gastrocnemius muscle contracts
near-isometrically, close to its theoretical optimal
length (sarcomeric operating range: 2.75–2.92 ␮m;
Fukunaga et al., 2001). This propels the body ahead at
minimal metabolic cost. Similar results have been
reported from sarcomeric length measurements dur-
ing physiological movement in several other species
(Burkholder and Lieber, 2001). Numerous experi-
ments have shown that the number of serial sarco-
meres is highly adaptable to changes in muscle length
(Tabary et al., 1972; Williams and Goldspink, 1973;
Herring et al., 1984; Williams, 1990). For many mus-
cles, the muscle length that dictates sarcomere num-
ber seems to be the length corresponding to maximal
isometric force (Herring et al., 1984; Williams, 1990).
The above experiments on sarcomeric length, how-
ever, indicate that in some muscles adjustments in
sarcomeric number may be made to optimize habitual
function rather than maximize contractile force gen-
eration (see also Herring et al., 1984).
The present results are of clinical relevance. Infor-
mation about force–length properties of in vivo mus-
cles is important in surgical procedures involving ten-
don transfer/lengthening (Delp et al., 1995; Loren et
al., 1996). Laser diffraction has allowed direct length
measurements to be taken in sarcomeres of the af-
fected muscle (Lieber and Friden, 1998); however,
this intraoperative method is highly invasive, thus
impractical for measuring sarcomeric length in a
healthy muscle. Therefore, providing reference values
according to which corrective operation would be
planned may not be always feasible. Moreover, laser
diffraction necessitates making appropriate correc-
tions in the length measurements taken to account for
222 Maganaris
the stretch of the series elastic component during
contraction (Lieber et al., 1994). The present meth-
odology allows noninvasive measurements to be taken
in a healthy muscle during contraction, e.g., in the
contralateral unaffected limb of a patient. From these
measurements, the length-range over which the mus-
cle develops force can be determined and guide the
decision making of tendon transfer in the affected
limb. If, for example, a muscle operates during con-
traction in the ascending limb of the force–length
relation as shown here, lengthening the in-series ten-
don or transferring its insertion proximally will result
in shifting the operating length of the sarcomeres
further away from the optimal range for force gener-
ation. In cases where it is anatomically possible, the
resultant reduction in contractile force could be com-
pensated for by an increase in the force moment arm
length by inserting the tendon at a site allowing op-
eration over a greater distance from the center of
rotation in the joint spanned. This would not only
achieve the desired range of movement in the joint
but also approximate the moment-producing potential
of the contralateral healthy muscle.
REFERENCES
Allen GM, Gandevia SC, McKenzie DK. 1995. Reliability of
measurements of muscle strength and voluntary activation
using twitch interpolation. Muscle Nerve 18:593– 600.
An KN, Takahashi K, Harrigan TP, Chao EY. 1984. Determi-
nation of muscle orientations and moment arms. J Biomech
Eng 106:280 –282.
Arndt AN, Komi PV, Bruggemann GP, Lukkariniemi J. 1998.
Individual muscle contributions to the in vivo Achilles ten-
don force. Clin Biomech 13:532–541.
Belanger AY, McComas AJ. 1981. Extent of motor unit activa-
tion during effort. J Appl Physiol 51:1131–1135.
Burkholder TJ, Lieber RL. 2001. Sarcomere length operating
range of vertebrate muscles during movement. J Exp Biol
204:1529 –1536.
Cutts A. 1988. The range of sarcomere lengths in the muscles
of the human lower limb. J Anat 160:79 – 88.
Delp SL, Statler K, Carroll NC. 1995. Preserving plantar flex-
ion strength after surgical treatment for contracture of the
triceps surae: a computer simulation study. J Orthop Res
13:96 –104.
Fukunaga T, Kubo K, Kawakami Y, Fukashiro S, Kanehisha H,
Maganaris CN. 2001. In vivo behaviour of human muscle
tendon during walking. Proc R Soc Lond B Biol Sci 268:
229 –233.
Fukunaga T, Roy RR, Shellock FG, Hodgson JA, Day MK,
Lee PL, Kwong-Fu H, Edgerton VR. 1992. Physiological
cross-sectional area of human leg muscles based on mag-
netic resonance imaging. J Orthop Res 10:926 –934.
Gans C. 1982. Fiber architecture and muscle function. Exerc
Sport Sci Rev 10:160 –207.
Gordon AM, Huxley AF, Julian FJ. 1966. The variation in
isometric tension with sarcomere length in vertebrate mus-
cle fibers. J Physiol 184:170 –192.
Gravel D, Arsenault AB, Lambert J. 1987. Soleus-gastrocne-
mius synergies in controlled contractions produced around
the ankle and knee joints: an EMG study. Electromyogr
Clin Neurophysiol 27:405– 413.
Herring SW, Grimm AF, Grimm BR. 1984. Regulation of
sarcomere number in skeletal muscle: a comparison of hy-
potheses. Muscle Nerve 7:161–173.
Herzog W, Read LJ, ter Keurs HEDJ. 1991. Experimental
determination of force-length relations of intact human gas-
trocnemius muscles. Clin Biomech 6:230 –238.
Herzog W, ter Keurs HEDJ. 1988a. A method for the determi-
nation of the force-length relation of selected in vivo human
skeletal muscles. Pflugers Arch 411:637– 641.
Herzog W, ter Keurs HEDJ. 1988b. Force-length relation of in
vivo human rectus femoris muscle. Pflugers Arch 411:642–
647.
Hof AL, van den Berg JW. 1977. Linearity between the
weighted sum of the EMGs of the human triceps surae and
the total torque. J Biomech 10:529 –539.
Huijing PA. 1985. Architecture of the human gastrocnemius
muscle and some functional consequences. Acta Anat
(Basel)123:101–107.
Huijing PA, Woittiez RD. 1985. Notes on planimetric and
three-dimensional muscle models. Neth J Zool 35:521–525.
Huxley AF. 1957. Muscle structure and theories of contraction.
Prog Biophys Chem 7:255–318.
Ichinose Y, Kawakami Y, Ito M, Fukunaga T. 1997. Estimation
of active force-length characteristics of human vastus late-
ralis muscle. Acta Anat (Basel) 159:78 – 83.
Isman RE, Inman VT. 1969. Anthropometric studies of the
human foot and ankle. Bull Prosthet Res 10/11:97–129.
Ito M, Akima H, Fukunaga T. 2000. In vivo moment arm
determination using B-mode ultrasonography. J Biomech
33:215–218.
Kawakami Y, Abe T, Fukunaga T. 1993. Muscle-fiber penna-
tion angles are greater in hypertrophied than in normal
muscles. J Appl Physiol 74:2740 –2744.
Kawakami Y, Amemiya K, Kanehisa H, Ikegawa S, Fukunaga
T. 2000a. Fatigue responses of human triceps surae muscles
during repetitive maximal isometric contractions. J Appl
Physiol 88:1969 –1975.
Kawakami Y, Kumagai K, Huijing PA, Hijakata T, Fukunaga
T. 2000b. The length-force characteristics of human gas-
trocnemius and soleus muscles in vivo. In: Herzog W, edi-
tor. Skeletal muscle mechanics: from mechanisms to func-
tion. Chichester: John Wiley & Sons Ltd. p 327–341.
Kawakami Y, Ichinose Y, Fukunaga T. 1998. Architectural and
functional features of human triceps surae muscles during
contraction. J Appl Physiol 85:398 – 404.
Leedham JS, Dowling JJ. 1995. Force-length, torque-angle and
EMG-joint angle relationships of the human in vivo biceps
brachii. Eur J Appl Physiol 70:421– 426.
Lieber RL, Friden J. 1998. Musculoskeletal balance of the
human wrist elucidated using intraoperative laser diffrac-
tion. J Electromyogr Kinesiol 8:93–100.
Lieber RL, Loren GJ, Friden J. 1994. In vivo measurement of
human wrist extensor muscle sarcomere length changes.
J Neurophysiol 71:874 – 881.

Loren GJ, Shoemaker SD, Burkholder TJ, Jacobson MD, Fri-
den J, Lieber RL. 1996. Human wrist motors: biomechani-
cal design and application to tendon transfers. J Biomech
29:331–342.
Maganaris CN. 2000. In vivo measurement-based estimations
of the moment arm in the human tibialis anterior muscle-
tendon unit. J Biomech 33:375–379.
Maganaris CN. 2001. Force-length characteristics of in vivo
human skeletal muscle. Acta Physiol Scand 172:279 –285.
Maganaris CN, Baltzopoulos V, Sargeant AJ. 1998. In vivo
measurements of the triceps surae complex architecture in
man: implications for muscle function. J Physiol 512:603–
614.
Maganaris CN, Baltzopoulos V, Sargeant AJ. 2000. In vivo
measurement-based estimations of the human Achilles ten-
don moment arm. Eur J Appl Physiol 83:363–369.
Martin DC, Medri MK, Chow RS, Oxorn V, Leekam RN, Agur
AM, McKee NH. 2001. Comparing human skeletal muscle
architectural parameters of cadavers with in vivo ultrasono-
graphic measurements. J Anat 199:429 – 434.
Narici MV, Binzoni T, Hiltbrand E, Fasel J, Terrier F, Cerre-
telli P. 1996. In vivo human gastrocnemius architecture with
changing joint angle at rest and during graded isometric
contraction. J Physiol 496:287–297.
Nigg BM, Fisher V, Allinger TL, Ronsky JR, Engsberg JR.
1992. Range of motion of the foot as a function of age. Foot
Ankle 13:336 –343.
In Vivo Force–Length Characteristics 223
Rugg SG, Gregor RJ, Mandelbaum BR, Chiu L. 1990. In vivo
moment arm calculations at the ankle using magnetic reso-
nance imaging (MRI). J Biomech 23:495–501.
Sale D, Quinlan J, Marsh E, McComas AJ, Belanger AY. 1982.
Influence of joint position on ankle plantar flexion in hu-
mans. J Appl Physiol 52:1636 –1642.
Sammarco GJ, Burstein AH, Frankel VH. 1973. Biomechanics
of the ankle: a kinematic study. Orthop Clin North Am
4:75–96.
Tabary JC, Tabary C, Tardieu C, Tardieu G, Goldspink G.
1972. Physiological and structural changes in the cat’s soleus
muscle due to immobilization at different lengths by plaster
casts. J Physiol 224:231–244.
van den Bogert AJ, Smith GD, Nigg BM. 1994. In vivo deter-
mination of the anatomical axes of the ankle joint complex:
an optimization approach. J Biomech 27:1477–1488.
Van Leeuwen JL, Spoor CW. 1992. Modelling mechanically
stable muscle architectures. Philos Trans R Soc Lond B Biol
Sci 336:275–292.
Walker SM, Schrodt GR. 1974. I segment lengths and thin
filament periods in skeletal muscle fibers of the Rhesus
monkey and the human. Anat Rec 178:63– 81.
Williams PE, Goldspink G. 1973. The effect of immobilization
on the longitudinal growth of striated muscle fibers. J Anat
116:45–55.
Williams PE. 1990. Use of intermittent stretch in the preven-
tion of serial sarcomere loss in immobilized muscle. Ann
Rheum Dis 49:316 –317.