Comparative Biochemistry and Physiology Part A 131 2001 121 Ž .

#133
Review
Comparative aspects of gait, scaling and mechanics in
mammals#
Jean-Pierre Gasc#
UMR 8570 MNHN-CNRS, Laboratoire d’Anatomie comparee, 55 rue Buffon, F-75005, Paris,
France ´
Received 9 March 2001; received in revised form 1 August 2001; accepted 8 August
2001
Abstract
In phylogenetically based systematics, Mammalia is the nomenclatural term which
designates the clade stemming
from the most recent common ancestry of monotremes and theria De Queiroz, Sys.
Biol. 43 1994 497 . Considering # Ž . #
that locomotor performance is a prevalent function to provide the necessary
conditions to survive and transmit genes, it
may be questioned if the diverse types of locomotion exhibited by extant mammals
could have played a role in their
evolution, or have only followed it. We may look after the structural and
behavioural features which are involved in
mammal locomotion compared to other tetrapods and test if they fit with the
proposed phylogeny. Several factors may
be checked: scaling effect in relation to gravitational constraints; geometrical
distribution of masses in the body, and
relative mechanical role of the limbs in the production of the external forces
necessary to forward motion. Classically, it
was thought that the fastest gaits used by terrestrial mammals were based upon a
unique kind of limb motion
co-ordination, called asymmetrical gaits, which in turn may be thought to be
related to a peculiar neuronal wiring.
Kinematic analysis brings an insight to this topic. Is the search for an ancestral
mammalian locomotor pattern judicious?
Notice the small size of many of the first mammals and their probable locomotor
plasticity. relation between grain size Ž
of the elements within the substrate and the organism scale . At a small size, the
gravitational constraint is less .
important, and the distinction between terrestrial and arboreal has probably no
sense when the limbs are the principal
motor elements. There remains the importance of the geometrical distribution of
body elements, the proportions of the
limbs and of the head#neck complex, the tail merely as an appendix, a set of
factors which may have generated the
frame of constraints within which diverse locomotor modes have evolved. # 2001
Elsevier Science Inc. All rights
reserved.
Keywords: Locomotion; Mammals; Evolution; Biomechanics; Neuromotricity
# This paper was originally presented as part of the ESCPB Congress symposium
‘Learning about the Comparative Biomechanics of
Locomotion and Feeding’, Liege July 26 ` #27, 2000. # Tel.: #33-1-40-79-3305; fax:
#33-1-40-79-3299.
E-mail address: gasc@mnhn.fr J.-P. Gasc . Ž .
1095-6433#01#$ - see front matter # 2001 Elsevier Science Inc. All rights reserved.
PII: S 1 0 9 5 - 6 4 3 3 0 1 Ž . 00457-3
122 J.-P. Gasc #Comparati#e Biochemistry and Physiology Part A 131 2001 121 ( )
#133
1. Introduction
The fact that evolution and biomechanics could
be associated would have seemed heretical to
many evolutionists a few decades ago. Darwin’s
assumption in chapter 13 of the ‘Origin of species’
Ž . 1859 that characters related to the habits of life
of animals were useless to disclose the phylogen#etic relationships has been for a
long time clum#sily taken as a dogma. Although this was a misin#terpretation of
Darwin’s text and a long-term
source of confusion in systematics, it was used as
a defence wall against any neo-Lamarckian temp#tation. In the meantime, the
orthogenetic ideas of
Cope 1904 were illustrated in most museums Ž .
organised by his collaborator Osborn whose ex#hibitions showed evolution as a
succession of
events in which an increase of efficiency in func#tion was designated as a main
factor. It may also
be noticed that in reverse most biomechanical
works have been short of any evolutionary con#siderations. More recently, the
change which al#lows an association between evolution and biome#chanics is mainly
due to a more clear definition of
the place functional factors may hold in evolutio#nary processes. In the 1960s it
was still a tradition
to consider de facto every structure as functio#nally adapted. This position was
severely criticised
by Williams 1966 and Gould and Lewontin Ž .
Ž . 1979 who engendered a tempest against what
was called the ‘panadaptationist program’ and the
general tendency to build ad hoc stories to legiti#mate the adaptive role of any
structure. Further#more, Kimura 1968, 1991 claimed that the great Ž .
majority of evolutionary changes at the molecular
level are caused by random fixation of selectively
neutral mutants in the population. This crisis led
to a return toward a position taking into account
the fact that the fitness of one species depends
largely upon the ability of the phenotype to per#form some fundamental activities.
Including ‘per#formance’ in the adaptive chain, as a link between
morphology and fitness which constitutes in itself
a subject for study, Arnold 1983 made an influ- Ž .
ential step. More recently Baum and Larson
Ž . 1991 forged the concept of ‘selective regime’, to
specify the critical aspects of the interaction
between the organism and its environment which
may influence the natural selection of some char#acters. This approach can clearly
be applied to
most biomechanical characters. Once natural se#lection is fully recognised as one
of the major
factors in evolution, it may be reasonably as#sumed that any advantage provided to
phenotypes
for their active displacement in the environment
would be considered as having an evolutionary
meaning. Locomotion indeed may have a decisive
influence toward the survival of individuals and
even if it is not easy to prove that issue see Ž
Arnold, 1994 , it has probably a long-term effect .
on the population balance.
Until the mid-20th century the diverse modes
of locomotion were organised in a succession of
stages illustrating the increase in efficiency, fol#lowing a kind of ‘scala
naturae’ scheme. For in#stance the scenario proposed by Magne de La
Croix Magne de La Croix, 1929 who based his Ž .
classification upon the number and duration of
resting periods for each limb which alternatively
behaves as a pendulum and an inverted pendu#lum. According to this author, a
general evolutio#nary trend would have reduced the number of
feet in contact at a given time of the period to
three, two, a single or even no contact, and also
have introduced a different phase relation within
each pair of limbs. In mammals, there would be a
continuous motion of the limbs and he described
a ‘phylogeny’ which goes from the slow walk when
each limb achieves its complete cycle before an#other comes in action, to the
diverse types of
gallops, from the three times to the rotary type,
which leads to the synchrony of the posterior
limbs ricochet and thus to the bipedal jump. In Ž .
parallel, morphologists recognised the diverse
types of arrangement shown by the limbs in verte#brates and most of them
transformed the mor#phological series obtained by comparison of ex#tinct and living
forms into an evolutionary series.
The most accurate description of the diverse
bones and joints shapes related to locomotion in
vertebrates was the work of an anti-evolutionist,
Vialleton 1924 , who criticised precisely the im- Ž .
proper passage from a series of specialised struc#tures to hypothetical
evolutionary line. He classi#fied the vertebrates limb arrangement in to three
types, based upon the plane stylozeugopodial Ž .
specified in the resting position by the arm thigh Ž .
and forearm leg segments: transversal as in Ž. Ž
Urodeles , horizontal as in lizards and parasagit- . Ž.
tal as in mammals . Boker’s 1935 classification Ž . Ž. ¨ of the terrestrial modes
of locomotion was the
functional image of this morphological series:
crawling, reptation and running. Together with
the renewal of methodology in phylogeny and the
J.-P. Gasc #Comparati#e Biochemistry and Physiology Part A 131 2001 121 ( ) #133
123
increase diffusion of cladistics, similarities in lo#comotor patterns lose any
immediate systematic
value. Rewcastle 1981 made a documented criti- Ž .
cism of what he called the ‘chauvinistic’ i.e. an- Ž
thropocentric aspect of this confusion between .
grade and clade. Even some basic definitions have
been discussed. Contrary to the general assump#tion that all mammals have limbs
which move in a
parasagittal plane, Jenkins 1971 showed by the Ž .
use of X-rays that some species keep their proxi#mal limb segment arm and thigh in
a oblique Ž .
position and may be ascribed as ‘non-cursorial’.
Recent debate about this topic Steudel and Ž
Beattie, 1993; Stein and Casinos, 1997 pointed .
out that we can now have valuable discussion
about the evolutionary meaning of biomechani#cally based adaptive features.
2. What is new in mammals locomotor system?
Many skeletomuscular elements show modifi#cations in Metatherians and Eutherians
when
compared to Monotremes or any Diapsid. For a
long time some of these structural changes were
related to peculiar locomotor capabilities, al#though Bramble and Jenkins 1989
emphasised Ž .
the gradual appearance of these features within
the synapsid clade.
Recent mammals exhibit a striking consistency
in their vertebral axis regionalisation. The head is
articulated on a cervical column consisting of
seven vertebrae, before the first vertebra which is
bound to the pectoral girdle through the sternal
ribs; exceptions are very few. The rib cage forms a
structural unit of 13#22 vertebrae functionally
associated to ventilation which is not in conflict
with locomotion as in most Diapsids Carrier, Ž
1987; Bramble, 1989 . Then a limited number of .
vertebrae, bearing or non-free ribs, form the lum#bar region, often elongated and
flexible owing to
the orientation of the zygapophyseal facets
Ž . Slijper, 1946 . At least two and generally three
ankylosed vertebrae, the sacrals, are fused by
their transverse processes to both ilium, the dor#sal elements of the pelvic girdle
which are di#rected forward and not backward as in Diapsids.
Lastly the caudal region is highly variable in
length but generally limited in diameter at its
basis and its reduction occurs in several lineages.
The platycelic type of articulation between verte#bral centra and the orientation
of the zygapophy#seal facets between the neural arches causes pre#dominantly
sagittal motion, conversely to lateral
predominant motion of the vertebral axis in most
Diapsids, and generally plesiomorphic tetrapods.
Since Thompson 1917 the vertebral column of Ž .
mammals has been compared to a bridge, the
centra being the compressive elements, the neu#ral arches being supports for the
tensile elements
and the girdles and limbs forming the piles. Axial
as well as appendicular musculature orientation
and disposition may be related to the gravitatio#nal constraints induced by such a
posture. How#ever, besides the fact that a bridge must not
move, this image does not take in to account the
dynamic role played by the vertebral axis during
quadrupedal locomotion. Forces are not only
transmitted by this skeletomuscular system but
they are also generated by it to participate to the
stride length when the limbs oscillate Hildebrand Ž
and Hurley, 1985; Gasc, 1993; Fischer, 1986 . The .
limbs have a conservative plan, especially in prox#imal parts stylopodium and
zeugopodium . How- Ž .
ever, the pectoral girdle of marsupials and pla#centals is highly apomorphic. The
shoulder blade
consists only of the scapula, the coracoid bone
being reduced to a simple apophysis near the
joint with the humerus. The clavicle is articulated
to the tip of the scapula spine acromion and Ž .
may or may not make a link with the thoracic
cage. Consequently, the shoulder blade is dorsally
free and bound to the body only through a system
of muscles. Some joints are peculiar in shape and
more sophisticated than their homologues in most
Diapsids. These are the elbow and the ankle
joints, although they are not homologues, which
play the same role in the kinematic chain inside
each pair of limbs. Their articular geometry limits
the number of degrees of freedom to one major:
flexion#extension around a transverse axis, that is
a hinge-like joint between the humerus and ulna
and between the tibia and astragalus. To this
predominant motion is associated a lever system,
force intensifier for the extensor muscles, the
olecranon of the ulna and the calcaneum process
Ž .Ž . tuber calcanei Fig. 1 . This hinge-like joint is
clearly related to the reduction of transverse com#ponents in limb motion. A
supplement of free#dom degree is often provided to the distal au#topodium manus and
pes through a round Ž .
shaped joint between humerus and radius, and
between the astragalus head and the second tarsal
row, permitting accommodation of the contact
124 J.-P. Gasc #Comparati#e Biochemistry and Physiology Part A 131 2001 121 ( )
#133
with complicate substratum. However, since
Jenkins 1971 who demonstrated by the use of Ž .
X-ray motion pictures the abducted position of
the stylopodia during the locomotion of the opos#sum, it is currently admitted that
not every mam#mal moves their limbs in a parasagittal plane. The
stylopodium may present a certain degree of ab#duction during the stepping cycle in
the so-called
non-cursorial species. In my opinion, this assump#tion has not been tested enough
and I suspect
that the definition of cursorial and non-cursorial
types was first coined in a gradistic perspective. I
mean that the ‘non-cursorial’ position of the limbs
yields higher transverse force components and
may be looked as being less efficient, hence ‘lower’
in adaptive value and on the evolutionary scale.
This point has been accurately discussed by Stein
and Casinos 1997 . We may notice that data are Ž .
rather scarce on the forelimbs and comparisons
were often made without regard to the phyloge#netic position of the animals. The
hinge-like shape
of the elbow and ankle joints suggests that the
axis for flexion and extension is kept in a trans#verse position during the support
phase of the
limb. For instance, whenever the humerus is
moved fore and aft in an abducted position, the
hinge axis of the distal epiphysis, defining mainly
the rotation of the ulna for the flexion#extension
of the forearm, becomes oblique. Thus the
humero-radial joint capitulum is brought dor- Ž .
sally, and the palmar surface of the hand cannot
be placed on the ground unless the humerus shaft
is twisted medially in its distal portion. Therefore,
such a torsion is classically related to a transverse
fore limb corresponding to a ‘sprawling’ stance.
However, Gambaryan and Kielan-Jaworowska
Ž . 1997 have shown that this characteristic is not
by itself indicative for a kind of stance. This may
be explained by the fact that the correct place#ment of the palmar surface relative
to the ground
is dependant upon the degrees of freedom at
three articular levels: the shoulder, the elbow and
the wrist. The limitation at one level may be
compensated at another. Such a compensation is
clearly achieved at the humero-radial joint. By its
axial rotation on a distinct condyle capitulum or Ž .
on a convex portion of the trochlea, the radius is
able to give diverse degrees of pronation to the
hand, regardless of the humerus position. Only in
the species most specialised to fast gaits, which
are mainly plant eaters and potential prey within
their habitats, the parasagittal position of the
Fig. 1. Similitude of the elbow a and ankle b joints. Lateral Ž. Ž.
view of right limbs in Dasyprocta leporina ŽCaviomorpha, Ro#dentia ..
humerus is associated with a pure hinge-like el#bow joint to which participates
both the ulna and
radius fused together articulated on a ‘roller’.
Whenever such a ‘runner’ animal needs to rotate
its hands to move or for other purpose, e.g. take
food or to climb, either the radius is kept free, as
in large South American rodents caviomorpha, Ž
Rocha Barbosa, 1997 or a special mobility exists .
in the wrist, as in hyracoids Fischer, 1986 . Ž .
The division of the vertebral column into speci#fied regions and the position of
the paired limbs
along the axis are dependent upon the action of
few regulator genes. It is within the historical
constraints of the resulting architecture peculiar
to mammals that modifications in dimensions or
number of distal parts occurred during the diver#sification of the group since the
Cretaceous
period. As Darwin had already stated 1859, chap. Ž
J.-P. Gasc #Comparati#e Biochemistry and Physiology Part A 131 2001 121 ( ) #133
125
XIII there is no contradiction between the con- .
servative Bauplan still operated during the early
ontogenesis and the huge diversity illustrated by
such diverse limb shapes as those of man, horse,
bat, mole and dolphin. The unity reveals that all
these animals have a common ancestor and that
each of them belongs to a separate phyletic line
which underwent evolutionary adaptive events.
However, mechanical constraints may also inter#vene on the osteomuscular system
during individ#ual lives to shape some bones Carter et al., 1991 Ž .
or to change the proportion of muscle fibre types
Ž . Sartorius et al., 1998 and this may not be forgot#ten in biomechanical
experiments.
2.1. Kinematic characteristics
Motion recording techniques, invented one
century ago, have greatly been improved and digi#tal high-speed pictures are now
available to study
the locomotion of animals. Coupling cinematog#raphy with X-rays has been applied
since the
1970s Jenkins, 1971; Gasc and Jouffroy, 1972 to Ž .
small sized mammals and is now directly digitised.
Marey 1898 had already recognised a striking Ž .
peculiarity of the mammal limb kinematics: the
shoulder blade tilts and its motion may be exactly
compared to that of the femur Fig. 2 . This Ž .
kinematic peculiarity is linked to the structural
design of the pectoral girdle which is formed
almost exclusively by the scapula shaped into a
long blade loosely connected to the thoracic cage.
This shape and position transforms the shoulder
blade into a proximal segment Kuznetsov, 1985 . Ž .
We may resume the kinematic characteristics of
the mammalian limb motion as follows: as in all
tetrapods each limb achieves a cycle consisting of
a stance and a swing Fig. 3 . The stance during Ž .
which the paw is in contact with the ground
begins with a yield which is due to the passive
flexion of the distal joints shoulder and elbow, Ž
knee and ankle ; this is the E2 phase of the .
physiologist Philippson, 1905 during which ex- Ž .
tensor muscles are active although they are
stretched. These muscles may then store elastic
energy Cavagna et al., 1968 . Then the joints are Ž .
extended, this is the E3 phase. The duration of
the contact of each limb with the ground de#creases when the animals speed
increases, which
is marked by the lower value of the duty factor
for higher speed. It is still unclear if this relative
and absolute reduction in time may be ascribed to
the shortening of the yield or to the shortening of
the propulsive phase. The decrease in time does
not have the same mechanical significance for
each phase. Time has no direct incidence on the
efficiency in elastic storage, only compliance of
structures and amplitude of the yield may inter#vene. In contrast, the duration of
the propulsive
phase is related to the time during which struc#tures can be accelerated, that is
to the velocity
reached when the limb takes off. Some results
Ž . Rocha Barbosa et al., 1996 suggest that each
pair of limbs behaves differently to accelerate the
body. The take-off begins the swing and is fol#lowed by an active flexion of the
limb distal joints
Ž . F while the proximal one rotates to protract and
adduct the whole. Next a general extension of the
Fig. 2. Rotation of the rat scapula during the stance phase, comparable to the
rotation of the femur after Kuznetsov, 1985, modified . Ž .
126 J.-P. Gasc #Comparati#e Biochemistry and Physiology Part A 131 2001 121 ( )
#133
Fig. 3. General pattern of the variation in the sum of the main angles of the
mammalian limb during one locomotor cycle. Here the
forelimb of the guinea pig, the summed angles being shoulder and elbow.
joint E1 occurs, the distal parts being slowed Ž .
down before contacting the ground. X-ray motion
pictures Gasc, 1993; Fischer, 1994 have shown Ž .
that in small and medium size mammals running
at fast gait the ventral flexion of the spine in the
lumbar region during the swing phase takes great
part in the hind limb protraction responsible for
the space span from one foot contact to the
following Fig. 4 . To avoid confusion between Ž .
space and time parameters, specific terms must
be used. The space covered by the body centre of
mass during stance is called the step, and the
space covered during one complete cycle one Ž
period is the stride length and can be measured .
on the ground between two successive footfalls of
the same foot. The number of cycles per second is
the frequency. As the speed reached by one ani#mal is the result of the space
covered stride Ž
length in a specified time 1 .Ž . #period#frequency ,
these two parameters are important to consider.
They may be modulated differently by the fore
and the hind limbs Fig. 5 . One may notice that Ž .
the space parameter depends upon structural fac#tors dimensions and geometry of the
animal and Ž .
the time parameter depends upon cell physiology
Žnerve impulse velocity, neuronal communications
and intrinsic rate of muscle fibre contraction ..
The possible motion is the result of the interac#tion of both types of factors.
The periodic motion of the limbs illustrates one
basic principle of locomotion with appendages:
alternate movements of the actuators are trans#formed into continuous displacement
of the whole
body centre of mass. The motion of each of the
four limbs is set in harmony with the motion of
the three others. This co-ordination brings out a
second level of periodic phenomena called gait.
The neural basis of this nested system periodic Ž
motion of the limb into a periodic arrangement of
the four is still disputed. Spinal central pattern .
generator CPG seems to lead the oscillation of Ž .
the individual limb Grillner, 1981 . However, we Ž .
do not know how the complex co-ordination re#vealed by the diverse gaits is
achieved and con#trolled Latash et al., 1999; Bassin et al., 1999 . Ž .
Gaits were probably symmetrical in early te#trapods, in each pair the limbs move in
alterna#tion, and the same interval of time separates their
action in both pairs: the sequence on one side is
the mirror image of the other side. This is not
true in asymmetrical gaits in which the time lag
between the action of fore limbs is different of
J.-P. Gasc #Comparati#e Biochemistry and Physiology Part A 131 2001 121 ( ) #133
127
Fig. 4. Schematic representation of the main kinematic fea#tures of a mammal, here
the guinea pig galloping from Rocha Ž
Barbosa, 1997 . a The left fore limb touches the ground. b . Ž. Ž.
The hind limb touches the ground. Notice the tilt of the free
scapula and the sagittal flexion of the vertebral lumbar region,
motions which increases the stride length of each pair of
limbs.
that for the hind limbs. These gaits were thought
to be characteristic of mammals, because they are
the fastest and necessitate a high precision of
balance. They seemed to be dependant upon an
erect posture enabling large stride lengths, suc#cession of a single foot supports
and dorsoventral
flexion of the spine. However, asymmetrical bouts
were recorded in some Diapsids young crocodiles Ž
and lizards during escape behaviour. .
2.2. Dynamic consequences
The geometry of the terrestrial mammals and
the resulting kinematics of the diverse body parts
have dynamical effects which in turn participate
in the efficiency of locomotion. However, it may
be noticed that some of these effects are experi#enced also by birds and most
probably in the past
by their dinosaur relatives.
Even in cases where the stylozeugopodial plane
is not actually parallel to the sagittal plane, the
feet contact the ground with their axis directed
longitudinally and under the belly or close to the
body wall see for instance, Jenkins, 1971; Goslow Ž
and Van De Graaff, 1982 . The rotation of the .
stylopodium around its own axis is much less
accentuated than described in lizards Renous Ž
and Gasc, 1977 . The predominant motion of the .
limbs is thus parasagittal around their basis, that
is to say in a vertical plane. The dynamic conse#quences have been disclosed for a
long time, and
first of all the Z shaped limbs inherited from
early tetrapods when placed in a parasagittal plane
could exert forces which are directed almost com#pletely in a single plane. They
have essentially
vertical and horizontal components which are used
directly to create corresponding ground reaction
forces to balance gravity and act upon the body
centre of mass. Ground reaction forces recorded
during the passage on a force plate of diverse
domestic species and a few wild ones gave a
quantified amount of stress applied to the limb
bones, and the vertical force peak, which occurs
generally at the take-off, could be largely above
the body weight. Another dynamic consequences
is that the centre of mass is highly elevated above
the ground, increasing the total potential energy,
and the limbs may be assimilated to a pendulum
during swing phase and to an inverted pendulum
during the stance. Length change of the oscillat#ing limb Hildebrand and Hurley,
1985 could Ž .
result in optimisation toward a natural frequency.
Exchanges between potential and kinetic energy
can be more easily tuned in a controlled inverted
pendulum. Attempts have been made to make
models of such systems which may be more en#ergy saving in the transfer of forces
between
actuators and the main body mass. Distribution of
the peripheral masses and the ability to change
their distance one to the others is one of the
major keys of the dynamics of limbed locomotion.
Unfortunately this aspect has not been exten#sively studied probably because
gravity being the
most evident constraint, inertial problems and
dynamic balance were set in the background.
Within a closed mechanical system, the sum of
the inertia momentum of every part is null. Any
displacement of one part relative to the body
centre of inertia would modify the inertial balance
and be regulated by the displacement of another
part. This means that the oscillation of one limb
may not only be looked as a pendulum effect in
the field of gravity, but also, and probably mainly
in the fastest gaits, as a transfer of inertial mo#mentum within the system. We all
empirically
experienced the effect of rapid arm swinging in a
128 J.-P. Gasc #Comparati#e Biochemistry and Physiology Part A 131 2001 121 ( )
#133
Fig. 5. Differences in behaviour of both pairs of limbs when the speed increases,
in the guinea pig running. White squares: fore limbs;
black squares: hind limbs. a The frequency increases more in the forelimbs. b The
stride length increases more in the hind limbs. Ž. Ž.
passive reverse motion of the other arm and the
trunk. This is indeed how the skaters rotate on
ice around a vertical axis. The limb mass distribu#tion was mainly viewed in terms
of energetics
ŽTaylor et al., 1974; Hildebrand and Hurley, 1985;
Steudel, 1990 , with no understanding of the dy- .
J.-P. Gasc #Comparati#e Biochemistry and Physiology Part A 131 2001 121 ( ) #133
129
Fig. 6. Contralateral upper and lower limbs are in phase in a
walking man as in trot . Consequently the ipsilateral limbs Ž .
are in opposition and the inertial variation of the upper
effects the lower during both phases of its cycle, stance and
swing after Preuschoft and Witte, 1991, modified . Ž .
namics of the whole system. The great attention
which was paid to gravitational constraints and
the need for static balance in studying terrestrial
locomotion have cast a shadow on the fact that a
great deal of the efficiency required in natural
conditions is due to the ability to keep a dynamic
balance once the initial acceleration is obtained.
This must modify our conception of the cycle
phases: the swing phase is not only a ‘recovery’
allowing each limb to establish new contact and
to earn a certain amount of space. It has an
actual dynamic effect on the dynamic equilibrium
of the body, the phase during which internal
forces are exchanged, a compromise being
probably set between the need to reduce the cost
of accelerating an inertial mass and the gain in
possible free cost motion in another part of the
body. The stance phase is the phase devoted to
the exchange for external forces. As inertia of
segments is related to their length and their shape,
the allometry studies of long bones have to be
combined with kinematics and anatomy to reach
their functional meaning. Preuschoft and Witte
Ž . 1991 applied such an approach to bipedal condi#tions of the hominids and showed
that the dif#ference in mass distribution in the upper and
lower extremities yields a similar virtual pendu#lum for both pairs of limbs
although they are
different in length. Another consequence of this
pendular effect in the human biped is the inertial
balance between the ipsilateral limbs which move
in the opposite direction Fig. 6 . Ž .
3. Gaits and size
Considering the huge range of sizes among
terrestrial vertebrates and especially mammals
which experienced increasing size during the evo#lution of diverse phylogenetic
lines, scaling effects
on locomotion have been extensively studied. Sev#eral aspects of this general
problem are still in
discussion. On structural grounds, morphometry
has shown that allometric relations between di#mensions of limb bones length and
diameter Ž .
and body mass have been found to be close to
isometry Alexander, 1979 . Only species adapted Ž .
to peculiar kinds of action burrowing or bipedal Ž
jumping may have significant deviation Raich . Ž
and Casinos, 1991 . I am not convinced that the .
slight deviation from the strict geometrical simili#tude or exceptions to the
dynamic similarity hy#pothesis is a real problem. Alexander and Jayes
Ž . 1983 have shown that mammals of any size are
to a large extent actually dynamically similar. In
contrast, it would be surprising that, facing the
physical constraints, one given bone behaves as if
it could resume the whole body as crank-arms
and wheels in one engine. Other factors may
intervene to explain for instance that peak stress
is never exceeded over a large range of sizes
Ž . Biewener, 1989 . Among such factors, some are
structural, such as the internal organisation of
bone tissue, the shock absorption and elastic
storage roles played by the muscles and tendon
proportions; some are behavioural, as the use of
the potential body geometry: the increase of the
mechanical advantage of the musculoskeletal sys#tem in large mammals Biewener, 1989
is Ž .
probably due to the erect posture of most of them
vs. the crouched posture of small ones, especially
for the hind limb which is kept more flexed in
small species, even at fast gaits Gasc, 1993 . On Ž .
the contrary, Fischer 1994 suggests that the Ž .
crouched posture is adaptive for small animals
faced with the irregularities of the ground.
It was demonstrated Heglund et al., 1974 that Ž .
the change from the fastest symmetrical gait, the
trot, to the asymmetrical regime, gallop, is
achieved at frequencies which are correlated to
the mass of the animals. In horses Farley and Ž
Taylor, 1991 this transition is accompanied by a .
noticeable decrease in the peak of ground force
reaction up to 28% for the hind limbs which Ž .
130 J.-P. Gasc #Comparati#e Biochemistry and Physiology Part A 131 2001 121 ( )
#133
seems to result from the relative increase of the
duty factor while the frequency increases. How#ever, the dynamic effects of the
important varia#tion of the moments of inertia were not taken in
account and, for instance, probably would explain
that the in vivo stress recorded on bones varies in
an opposite direction in the trailing and the lead#ing limb when the horse passes
to asymmetrical
gait Biewener et al., 1983 . The leading anterior Ž .
limb executes its stance when the whole
hindquarters are swinging forward, and just be#fore the four limbs will leave any
contact with the
ground.
The way animals proceed to accelerate is also
size dependent. Of course they all increase their
Fig. 7. Some scale effects on locomotor performance after Ž
Strang and Steudel, 1990 . Notice the steeper slope for the .
decreasing in frequency when the animals increase in size.
Fig. 8. Schematic view of the relations within the diverse
morphological parts and levels which are integrated into the
sensorimotor system involved in locomotion.
speed by increasing both parameters involved:
frequency and stride length. However, the small
animals increase their frequency faster than large
animals increase their stride length Strang and Ž
Steudel, 1990 Fig. 7 . As the frequency is bound .Ž .
to the ability to recruit more muscular fibres
possessing a high intrinsic contraction rate and a
higher turn-over of cross-bridges between actin
and myosin molecules, this strategy for increasing
speed is energetically expensive more ATP Ž
molecules used . This is illustrated by the pres- .
ence of only fast fibres in the limb muscles of
shrews Suzuki, 1990; Neveu, personal communi- Ž
cation . On the other hand, the geometry of large .
mammals provides them with long and erect limbs
to make longer strides without any cost supple#ment. I would like to add again that
dynamic
conditions have also to be looked at, because
rotational inertia rises as the fourth power of
linear dimension and could be a benefit in the
costly accelerative phase at the beginning of the
oscillating period of long pendulum-like limbs.
This could be the key to understand that locomo#tion is less costly by unit of mass
for large mam#mals than for small ones.
However, where is the boundary between small
and large terrestrial mammals? Some authors
proposed that mammals over 100 kg may sacrifice
efficiency to safety. The threshold depends much
on the level of integration looked at. Loeb 1989 Ž .
emphasised that delay in the feedback circuits, a
limiting factor in the efficiency of any motor
J.-P. Gasc #Comparati#e Biochemistry and Physiology Part A 131 2001 121 ( ) #133
131
system, is dimension dependent Fig. 8 . In fact, Ž .
except for the synaptic propagation delays which
are similar in large and small animals, every other
functional factor in the neuromotor system is
bound to scaling problems: transmission of the
nerve impulse along axons or potential actions
along muscles fibres.
4. Could we tell anything about the ancestral
locomotion modes of mammals?
Paleontological data and biomechanical con#siderations on living species have been
associated
to reconstruct hypothetically the mode of locomo#tion used by early mammals.
Unfortunately,
postcranial remains are more rarely found than
pieces of skull or teeth. It seems that during more
than 100 million years, from the early Triassic
Cynodonts to the Cretaceous Multituberculates,
the Synapsid line had already inaugurated some
of the mammalian characteristics Bramble and Ž
Jenkins, 1989 , mainly vertebral regionalisation .
and reversal in the direction of the ilium. Even if
lateral flexion of the trunk in the lumbar region is
still observed in the slow walk of most mammals
Ž . Pridmore, 1992 , the change to sagittal axial
flexion may be associated to the capacity to
achieve asymmetrical gaits as a bound or sudden
bibedal jump to escape predators. Parasagittal
hind limbs provide force and directionality in this
action. Gambaryan 1974 has proposed a semi- Ž .
fossorial mode of life for these small sized ani#mals which were contemporary of
the large di#nosaurs. They would have had to dig to forage
their aliments with the forelimbs while firmly
anchored by the hind limbs. In contrast, Boker ¨ Ž . 1935 speculated about the
arboreal creeping
mode for primitive mammals which could have
looked as living didelphids. However, it may be
emphasised that one animal does not locomote in
vacuum. It is necessary to associate musculoskele#tal systems involved and the
‘Umwelt’ of each
species. We must consider the size of the ele#ments which in the environment
constitute possi#ble supports or obstacles for progression. For a
4-cm-long pygmy shrew any herb stem is the
equivalent of a tree, any small stone a cave. Early
true mammals may have been more or less able
to climb, to hide in crevices, to jump and to
proceed to bursts of running.
Further evolution shows the explosive radiation
during the Tertiary period, with the appearance
of large sized herbivorous species and their large
predators. Almost all these forms are extinct and
today only few large terrestrial mammal species
still survive. Is there a biomechanical reason for
this massive extinction and loss in biodiversity?
Probably not, yet we may notice that marine
descendants of the large terrestrial tertiary mam#mals are still present.
5. Conclusions
In 1989, Gerald Loeb concluded a paper on the
neural control of locomotion by these phrases:
‘... the traditionally separate experimental disci#plines of biomechanics and
neurophysiology need
to develop much closer ties so that motor control
problems can be attacked from both perspectives.
Theoreticians in both areas must address not only
the problem of motor control in a given organism,
but also how their proposed solutions would fare
in a world governed by ontogeny and phylogeny.
Locomotion is a particularly appropriate motor
behaviour for such a multidisciplinary approach,
because it is basic to the lives of almost all
animals, assuring a highly constrained and orderly
evolution’.
I shall only add that most of the questions still
addressed are relevant to this multidisciplinary
approach to which must be joined field observa#tions and quantitative data on wild
animals life
history Gasc, 1989 . The adaptive diversity of Ž .
mammals is so high that it seems perilous to
believe that simple general rules could accurately
describe the relation between specified structural
features and a single kind of function. Exceptions
are strongly marked specialisations as in some
diggers or jumpers. The compromising nature of
organ evolution fixes the limits to a purely biome#chanical analysis and forbids
the reduction of the
whole musculoskeletal system to a locomotory
machinery. Nobody would think that the scratch#ing reflex of the dog might have
influenced the
organisation of the hind limb. In contrast,
grooming, searching and catching food items,
touching and grasping, all these mammalian be#haviours in which the forelimbs are
involved, could
have a relation with the structural organisation of
these limbs. In this survey, I tried to point out
some still unclear or neglected aspects of the
mammalian locomotion, especially on dynamic
132 J.-P. Gasc #Comparati#e Biochemistry and Physiology Part A 131 2001 121 ( )
#133
effects. This complex behaviour is supported by
highly integrated structures and functions which
involve two main levels. One is relevant to biome#chanics and may be described and
analysed in
terms of physics, organismal morphology and cell
biology. The other is relevant to neurophysiology.
However, comparative studies including both lev#els are the only way to get a clear
understanding
of the evolutionary patterns of mammalian loco#motion.
References
Alexander, R.McN., 1979. Allometry of limb bones of
mammals from shrews Sorex to elephant Loxo- Ž. Ž
donta . J. Zool. Lond. 189, 305 . #314.
Alexander, R.McN., Jayes, A.S., 1983. A dynamic simi#larity hypothesis for the
gaits of quadrupedal mam#mals. J. Zool. Lond. 201, 135#152.
Arnold, S.J., 1983. Morphology, performance and fit#ness. Am. Zool. 23, 347#361.
Arnold, S.J., 1994. Investigating the origins of perfor#mance advantage:
adaptation, exaptation and lin#eage effects. In: Eggleton, P., Vane-Wright, R. Eds.
, Ž .
Phylogenetics and Ecology. Academic Press, Lon#don, pp. 123#168.
Bassin, P.V., Bernstein, N.A., Latash, L.P., 1999. On
the problem of relation between structure and func#tion in the brain from a
contemporary point of view
1966. Motor Control 3 4 , 332 Ž . #342.
Baum, D.A., Larson, A., 1991. Adaptation reviewed: A
phylogenetic methodology for studying character
macroevolution. Syst. Zool. 40, 1#18.
Biewener, A.A., 1989. Scaling body support in mam#mals: Limb posture and muscle
mechanics. Science
245, 45#48.
Biewener, A.A., Thomason, J., Lanyon, L.E., 1983. Me#chanics of locomotion and
jumping in the forelimb
of the horse Equus : in vivo stress developed in the Ž .
radius and metacarpus. J. Zool. Lond. 201, 67#82.
Boker, H., 1935. Einfuhrung in die vergleichende bi- ¨ ¨
ologische Anatomie der Wirbeltiere. G. Fischer,
Jena, T1.
Bramble, D.M., 1989. Axial-appendicular dynamics and
the integration of breathing and gait in mammals.
Am. Zool. 29, 171#186.
Bramble, D.M., Jenkins, F.A.Jr., 1989. Structural and
functional integration across the reptile#mammal
boundary: the locomotor system. In: Wake, D.B.,
Roth, G. Eds. , Complex Organismal Functions: In- Ž .
tegration and Evolution in Vertebrates. John Wiley
and Sons, Chichester, pp. 13#144.
Carrier, D.R., 1987. The evolution of locomotor stamina
in tetrapods: circumventing a mechanical constraint.
Paleobiology 13, 326#341.
Carter, D.R., Wong, M., Orr, T., 1991. Muskuloskeletal
ontogeny, phylogeny, and functional adaptation. J.
Biomech. 24 1 , 3 Ž . #16.
Cavagna, G.A., Dusman, B., Margaria, R., 1968. Posi#tive work done by a previously
stretched muscle. J.
Appl. Physiol. 24, 21#32.
Cope, E.D., 1904. The primary Factors of Organic
Evolution. The Open Court Publishing Company,
Chicago.
De Queiroz, K., 1994. Replacement of an essentialistic
perspective on taxonomic definitions as exemplified
by the definition of ‘Mammalia’. Syst. Biol. 43 4 , Ž .
497#510.
Farley, C.T., Taylor, C.R., 1991. A mechanical Trigger
for the Trot-Gallop Transition in Horses. Science
253, 306#308.
Fischer, M., 1986. Die Stellung der Schliefer Hy- Ž
racoidea im phylogenetischen System der Eutheria. .
Zugleich ein Beitrag zur Anpassungsgeschichte der
Procaviidae. Cour. Forsch.-Inst. Senckenberg 84,
1#132.
Fischer, M., 1994. Crouched posture and high fulcrum,
a principle in the locomotion of small mammals: The
example of the rock hyrax Ž .Ž Proca#ia capensis Mam#malia: Hyracoidea . J. Human
Evol. 26, 501 . #524.
Gambaryan, P.P., 1974. How Mammals run. Wiley &
Son, New York.
Gambaryan, P.P., Kielan-Jaworowska, K., 1997. Sprawl#ing versus parasagital stance
in multituberculate
mammals. Acta Paleontol Polonica 43, 13#44.
Gasc, J.P., Jouffroy, F.K., 1972. Application de la
cineradiographie a l’etude de la biomecanique des ´ `´ ´ Vertebres. C.R. Acad. Sci.
Paris 275, 763 ´ ´ #766.
Gasc, J.P., 1989. Locomotion studies: evolutionary
viewpoint and multidisciplinary approach. In:
Splechna, H., Hilgers, H. Eds. , Trends in Verte- Ž .
brate Morphology. Fortschritte der Zoologie, 35, pp.
587#593.
Gasc, J.P., 1993. Asymmetrical gait of the Saharian
rodent Ž Ž. Meriones shawi shawi Duvernoy, 1842
Ž . Rodentia, Mammalia : a high-speed cinemato#graphic analysis. Can. J. Zool. 71,
790#798.
Goslow, G.E. Jr., Van De Graaff, K., 1982. Hindlimb
joint angle changes and action of the primary ankle
extensor muscles during posture and locomotion in
the striped skunk Ž . Mephitis mephitis . J. Zool. Lond.
1982, 405#419.
Gould, S.J., Lewontin, R.C., 1979. The spandrels of San
Marco and the Panglossian paradigm: a critique of
the adaptationist programme. Proc. R. Soc. Lond.
205, 581#598.
Grillner, S., 1981. Control of locomotion in bipeds,
tetrapods and fish. In: Brookhart, J.M., Mountcastle,
V.B. Eds. , Handbook of Physiology, 2. Waverly Ž .
Press, Baltimore, pp. 1179#1236.
J.-P. Gasc #Comparati#e Biochemistry and Physiology Part A 131 2001 121 ( ) #133
133
Heglund, N.C., Taylor, C.R., McMahon, T.A., 1974.
Scaling stride frequency and gait to animal size:
Mice to horses. Science 186, 1112#1113.
Hildebrand, M., Hurley, J.P., 1985. Energy of the oscil#lating legs of a fast-
moving cheetah, pronghorn,
jackrabbit, and elephant. J. Morphol. 184, 23#31.
Jenkins Jr., F.A., 1971. Limb posture and locomotion in
the Virginia opossum Ž . Didelphis marsupialis and
other non-cursorial mammals. J. Zool. Lond. 165,
303#315.
Kimura, M., 1968. Evolutionary rate at the molecular
level. Nature 217, 624#626.
Kimura, M., 1991. The neutral theory of molecular
evolution: a review of recent evidence. Jpn. J. Genet.
66, 367#386.
Kuznetsov, A.N., 1985. Comparative functional analysis
of the fore and hind limbs in Mammals in Russian . Ž .
Zool. Zhurn. 64 12 , 1862 Ž . #1867.
Latash, L.P., Latash, M.L., Meijer, O.G., 1999. 30 years
later: the relation between structure and function in
the brain from a contemporary point of view 1966 , Ž .
part 1. Motor Control 3 4 , 329 Ž . #332.
Loeb, G.E., 1989. Neural control of locomotion. Bio#Science 39 11 , 800 Ž . #804.
Magne de La Croix, P., 1929. Filogenia de las locomo#ciones cuadrupedal y bipedal
en los vertebrados, y
evolucion de la forma consecutiva de la evolucion de
la locomocion. Ann. Soc. Cient. Arg. Buenos Aires ´ 108, 383#409.
Marey, E.J., 1898. Analyse des mouvements du cheval
par la chronophotographie. Nature 2, 22#26.
Philippson, M., 1905. L’autonomie et la centralisation
dans le systeme nerveux central. Trav. Lab. Physiol. ` Inst. Solvay 7, 1#208.
Preuschoft, H., Witte, H., 1991. Biomechanical reasons
for the evolution of hominid body shape. Origine sŽ .
de la bipedie chez les Hominides. Cahiers de ´ ´ Ž
Paleoanthropologie , Editions du CNRS, Paris, pp. ´ .
59#77.
Pridmore, P.A., 1992. Trunk movements during loco#motion in the marsupial
Monodelphis domestica
Ž . Didelphidae . J. Morph. 211, 137#146.
Raich, J., Casinos, A., 1991. Limb proportions in terres#trial mammals. Belg. J.
Zool. 121, 295#314.
Renous, S., Gasc, J.P., 1977. Etude de la Locomotion
chez un Vertebre tetrapode. Ann. Sci. nat., Zool., ´´´ Paris 12e ser. 19 2 , 137 ´
Ž . #186.
Rewcastle, S.C., 1981. Stance and gait in Tetrapods: an
evolutionary scenario. In: Day, M.H Ed. , Verte- Ž .
brate Locomotion. Academic Press, London, pp.
239#267.
Rocha Barbosa, O., 1997. Analyse morpho-fonction#nelle des convergences de
l’appareil locomoteur de
grands Rongeurs Caviomorphes d’Amerique du Sud ´ et de petits Artiodactyles
d’Afrique et d’Asie. Doc#torate thesis, Museum national d’histoire naturelle, ´
Paris.
Rocha Barbosa, O., Renous, S., Gasc, J.P., 1996. Com#parison of the fore and hind
limb kinematics in the
symmetrical and asymmetrical gaits of a caviomorph
rodent, the domestic guinea pig, Ca#ia porcellus
Ž .Ž . Linne, 1758 Rodentia, Caviidae . Ann. Sci. nat., ´ Zool. Paris, 13e ser. 17
4 , 149 ´ Ž . #165.
Sartorius, C.A., Lu, B.D., Acakpo-Satchili, L., Ja#cobsen, R.P., Byrnes, W.C.,
Leinwand, L.A., 1998.
Myosin heavy chains Iia and Iid are functionally
distinct in the mouse. J. Cell Biol. 141 4 , 943 Ž . #953.
Slijper, E.J., 1946. Comparative biologic#anatomical
investigations on the vertebral column and spinal
musculature of mammals. Kon. Ned. Akad. Wet. 42
Ž . 5,1#128.
Stein, B.R., Casinos, A., 1997. What is a cursorial
mammal? J. Zool. Lond. 242, 185#192.
Steudel, K., 1990. The work and energetic cost of
locomotion I. The effects of limb mass distribution
in quadrupeds. J. Exp. Biol. 154, 273#285.
Steudel, K., Beattie, 1993. Scaling or cursoriality in
mammals. J. Morph. 217, 55#63.
Strang, K.T., Steudel, K., 1990. Explaining the scaling
of transport costs: the role of stride frequency and
stride length. J. Zool. Lond. 221, 343#358.
Suzuki, A., 1990. Composition of myofiber type in limb
muscles of house shrew Ž . Suncus murinus : lack of
type I myofibers. Anat. Rec. 228, 23#30.
Taylor, C.R., Shkolnick, A., D’Heil, R., Baharav, D.,
Bonut, A., 1974. Running in cheetahs, gazelles, and
goats: energy cost and limb configuration. Am. J.
Physiol. 227, 848#850.
Thompson, d’A.W., 1917. On Growth and Form. Cam#bridge University Press, London.
Vialleton, L., 1924. Membres et ceintures des Vertebres ´ ´ Tetrapodes. Critique
morphologique du Trans- ´
formisme. Doin, Paris.
Williams, G.C., 1966. Adaptation and Natural Selec#tion. Princeton University
Press, Princeton.