In the patient care arena, an understanding of normal locomotion is a prerequisite to knowledge of pathological function in gait. Familiarity with joint motion, ground-reaction forces and muscular activity in normal individuals provides a bedrock of supporting knowledge that serves as a foundation for prosthetists and orthotists who seek to improve the performance of patients with pathological gait deficits. Although awareness of each of these components as it relates to a specific patient offers a revealing perspective by itself, in tandem they enable a three-dimensional differentiation between pathological and compensatory gait patterns. The sum is of far greater value than the individual parts.
In the past, successful orthotic and prosthetic intervention was limited primarily by design and material characteristics. A relatively gross understanding of gait mechanics was sufficient. With advancements in materials science and componentry development, such as the miniaturization of external power sources, the limitations to a patient's performance using a prosthesis or orthosis is more likely than ever to hinge on the practitioner's knowledge of gait mechanics.
This article, the second in a two-part series on normal human locomotion, attempts a narrative description of the dynamic phasic qualities of joint motion, ground-reaction forces and muscular activity.
More than a century ago, the discipline of ambulatory performance assessment emerged as the components of human walking began to be measured and numerically quantified. The French scientist E.J. Marey performed pioneer research in gait assessment technology during the 1870s using multiple camera photography in sequenced series to assess movement, including pathological gait (see Figure 1) . His colossal contributions to gait science are revealed in his development of the first myograph for measuring muscle activity as well as a novel foot-switch measurement system for recording the magnitude and timing of plantar contact (1).
Eadweard Muybridge, an American contemporary and friend of Marey, was supported by Stanford University during the 1880s in using synchronized multiple camera photography with a scaled backdrop to capture movement on film and assess the motion of subjects walking (2). Other major advances into instrumented gait analysis were made by Scherb, who sought to understand the phasic action of muscle activity and performed hand-muscle palpation using a treadmill in 1920, and Adrian, who in 1925 advanced the use of EMG to study the dynamic action of muscles (3).
The applications of engineering and technology to the understanding of human walking received enormous impetus in 1945 when Inman et al. initiated the systematic collection of normal and amputee data on an instrumented walkway in their outdoor gait lab at the University of California--Berkeley (see Figure 2) (4-7). Since that time, a number of researchers and clinicians increasingly have used the growing array of gait technologies to measure and analyze the parameters of human performance in normal and pathological gait (8-12).
A contemporary instrumented walkway is a pathway that contains sensors or other measurement devices in the floor or around the subject's line of progression. The instrumentation is intended to monitor and measure one or more parameters of gait, such as motion (kinematics) or the forces creating motion (kinetics). Instrumentation placed directly on the patient also may measure phasic muscle activity (electromyographics or EMG), pressure against the skin or time-related (temporal) parameters such as velocity and cadence (13).
This article, presented as the second installment in the AAOP Professional Development Certificate Program in Gait and Pathomechanics (14,15), reviews the process of gait in phases from the perspective of kinematics, kinetics and EMG. To facilitate that goal, the basic principles of each are examined, and the current terminology (16) and means of data collection are reviewed.
Kinematics concerns itself with movement without consideration for the cause. The focus in gait analysis is on linear and angular displacements, velocities, accelerations and decelerations. The kinematics of walking can be quantitatively measured by means of instrumentation or qualitatively analyzed by means of observational gait assessment, such as a visual description of an individual's lower extremities, pelvis and trunk motion during ambulation. A qualitative contribution has been made by video technology using slow motion capabilities.
Recent development of inexpensive video gait assessment software packages that require manual measurements has clinical quantitative applications as well, enabling the measurement of joint angles in two dimensions (17). Since walking is a three-dimensional function, however, this type of video software may have limited value for comprehensive assessments or research applications that require a broad span of precise data. Methods that depend on hand measurements against a video image have questionable accuracy and are too labor-intensive for complex multijoint assessments. Also, the data collected from such methods cannot be integrated with kinetic data in real time or accurately displayed in real time or synchronized time with other data.
Most of what practitioners know today about kinematics of normal and pathological gait has been obtained from either an electrogoniometer, which uses electrical transducers attached to adjacent limb segments (see Figure 3) or, more likely, multicamera three-dimensional motion systems that track reflective markers placed on strategic anatomical landmarks.
Motion analysis measures dynamic range of motion. Dynamic range indicates joint motion or excursion from the maximum angle to the minimum angle during a particular phase or phases in the gait cycle. The computerized data obtained from multicamera motion systems can document the motion of an individual's lower and upper extremities, pelvis, trunk, and head during ambulation. Motion analysis markers are small spheres or balls placed at specific bony landmarks that, when tracked by a camera-based video system, can be used to determine body segment and joint position. Active markers are joint and limb segment markers used during motion analysis that emit a light signal (see Figure 4) . Passive markers are markers that reflect visible or infrared light (see Figure 5) .
Kinetics is the general term given to the study of forces that cause movement. Force may be defined as a push or a pull and is produced when one object acts on another. The units used to measure force are Newtons (N). Forces in walking can be internal (such as muscle activity, ligamentous constraint, or friction in muscles and joints) or external (such as ground-reaction forces created from external loads).
The rotational potential of the forces acting on a joint is called torque, moment or moment of force. The internal joint moment is the net result of all of the internal forces acting about the joint, including moments due to muscles, ligaments, joint friction and structural constraints. The joint moment usually is calculated around a joint center. When we think in terms of internal moments, for example, a net knee extensor moment means the knee extensors (quadriceps) are dominant at the knee joint, and the knee extensors are creating a greater moment than the knee flexors (hamstrings and gastrocnemius). The units used to express moments or torques are Newton-meters (N-m) and for research purposes usually are normalized to the subject's body mass. Normalization is the process by which a relationship is established between initially collected data (raw data) and some other basic reference data. Normalized to the subject's body mass, Newton-meters are expressed as N-m/kg.
The term joint power is used to describe the product of a joint moment and the joint angular velocity. Joint power is said to be generated when the moment and the angular velocity are in the same direction and said to be absorbed when they are in opposite directions. The units used to measure joint power are Watts (W).
Engineers and researchers studying gait attempt to measure the moment of force produced by muscles crossing a joint, the mechanical power flowing to and from those same muscles, and the energy changes of the body that result from this power. This requires the integration of both kinematic and kinetic data using very specialized data collection and processing systems (see Figure 6) .
The external ground-reaction force line is a familiar concept to most clinicians trained in orthotics and prosthetics. Understanding its spatial relationship relative to the location of primary joints in normal gait is intuitively helpful in understanding the pathomechanics of a given patient.
A reaction force is the force that an initial body (A) exerts on a second body (B) in response to a force exerted by B on A. The reaction force has equal magnitude but opposite direction relative to the force exerted on A by B. Ground-reaction force is comprised of three components: 1) vertical force, 2) fore-aft shear and 3) medial-lateral shear. Information on these forces is obtained from a force platform or force plate, which is a transducer set into the floor to measure the forces and torques applied by the foot to the ground (see Figure 7 and Figure 8 ).
These devices provide quantified measures of the three components of the resultant ground-reaction force vector and the resultant torque vector about a given joint. The ground-reaction force line essentially is the vector summation of the three reaction forces resulting from the interaction between the foot and ground. The moment of force or torque is the cross product of the radius vector and the force. The radius vector, traditionally assigned the variable r, is a position vector from the point around which the calculations are made to the line of action for the force being considered, traditionally assigned the variable F. The length of r is the moment arm of the force F. In two dimensions, the moment of force about a point is the product of a force and the perpendicular distance from the line of action of the force to the point. Typically, the moments of force are calculated about the center of rotation of a joint and are expressed in Newton-meters (N-m).
We have seen the resultant ground-reaction force (GRF) vector is the mean load bearing line, which takes into account both gravity and momentum (see Figure 9) . It has magnitude as well as directional qualities. The spatial relationship between this line and a given joint center influences the direction in which the joint will tend to rotate. This has enormous implications in understanding what orthotic or prosthetic component or alignment variant might be used to stabilize a joint during ambulation. The ground-reaction force line and the external moments or torques created at the major joints are presented in the pages that follow.
Electromyographic (EMG) data provide important information in terms of understanding the direct physiological effect of prosthetic or orthotic design variants. Measuring muscle activity is like going straight to the mouth of the horse. Knowledge of the timing and intensity of the muscles throughout gait may suggest alterations in gait training and orthotic or prosthetic alignment or componentry to reduce excessive, ill-timed or prolonged muscle activity.
Electromyographic information is obtained by inserting fine wire electrodes directly into the muscle belly or by placing noninvasive surface electrodes over the muscle apex. Wire electrodes have the advantage of precise placement and are less likely to register "cross-talk" from adjacent muscles (see Figure 10) . Wire electrodes are essential for measuring deep muscles. Surface electrodes provide a noninvasive alternative for measuring muscle activity of superficial groups (see Figure 11) .
Inman et al. (7) and Perry (8) have presented comprehensive models of human locomotion based on kinematic, kinetic and EMG collections, which have been drawn upon for the preparation of this article. The timing of the first five phases of gait identified by Perry can be precisely identified by the magnitude and timing of the vertical force graph of the stance limb. The typical free walk vertical force graph reveals itself as a curve with two peaks and a valley (see Figure 12) . By contrast, slow walk velocity and running do not display the midstance valley (F2) typical of normal locomotion in free walk velocity. The final three phases of gait that occur during swing can best be identified by kinematic positioning.
The beginning of the gait cycle is referred to as initial contact. At the moment the foot strikes the ground, the ankle is at the neutral position, and the knee is close to full extension. In the sagittal plane, the alignment of the ground-reaction force vector at initial contact is posterior to the ankle joint, creating a plantarflexion moment (see Figure 13) . The three pretibial muscles (tibialis anterior, extensor digitorum longus and extensor hallicus longus), all of whose line of pull is anterior to the ankle joint, maintain the ankle and subtalar joint in neutral through eccentric contraction. The function of the peroneus tertius is considered identical to the extensor digitorum longus; they share the identical lateral tendon, and their muscle bellies blend into each other.
At the knee the vector is anterior to the joint axis, creating a passive extensor torque. Activity of the quadriceps and hamstring muscle groups continues from the previous terminal swing to preserve and stabilize the neutral position of the knee joint.
The hip and pelvis are emerging from a function of swing limb advancement with significant flexion, about 30 degrees. In normal gait, maximum hip flexion occurs during terminal swing and initial contact. A rapid high-intensity flexion moment thus is created at the hip as the vector falls anterior to the joint, placing great demand on the hip extensors. To restrain this impending flexion torque created by the anterior position of the vector, both the gluteus maximus and the hamstrings are activated. In the coronal plane, the gluteus medius is active preparing to stabilize the pelvis.
To absorb the impact force of loading and to maintain forward momentum, the eccentric action of the pretibial muscles regulates the ankle plantarflexion rate. A heel rocker action occurs as the pretibials pull the tibia forward over the fulcrum of the os calcis even as the foot is moving into a plantargrade position. This movement enables forward momentum of the tibia relative to the foot, but it also flexes the knee (see Figure 14) . During the peak of loading response, the magnitude of the vertical ground-reaction force exceeds body weight. The pretibials (tibialis anterior, extensor halicus longus and extensor digitorum longus) act as a shock absorber during loading response.
As a shock-absorbing mechanism and for energy efficiency, the knee flexes under the eccentric action of the quadriceps to about 15 to 18 degrees. During the stance phase of gait, the maximum knee-flexion angle usually is reached at foot flat. The quadriceps muscle group following this plantargrade posture controls the degree of knee flexion. Just as the pretibials advance the tibia forward over the foot in the rocker mechanism described, the quadriceps advance the femur over the tibia. This integrated action provides controlled forward movement of the entire lower-extremity unit.
The hip maintains its posture of about 30 degrees of flexion, creating a rapid, high-intensity flexion torque, the second-highest joint torque in normal gait after the dorsiflexion torque, which occurs at the talocrural joint during terminal stance. During loading response the hip extensors act as a shock absorber around the hip joint. Although not solely identified as hip extensors, the hamstrings nevertheless act as hip extensors as well as limit forward flexion of the pelvis and trunk. The function of the hamstrings when the hip is in flexion during stance is taken over by the gluteus maximus as stance progresses. Hip extensors prevent further flexion at the hip, and shock absorption is provided by the gluteus maximus, hamstrings and adductor magnus. The medial-lateral control function of the hip adductors occurs as body weight is assumed by the stance leg.
The momentum of forward progression over a stable foot with tibial stability maintained is referred to as the ankle rocker. The ankle rocker movement that progresses the tibia over a stationary foot is controlled early in midstance by the eccentric contraction of the soleus and is assisted by the gastrocnemius as the knee nears extension (see Figure 15) . At the beginning of midstance, the ankle is in a posture of 10 degrees of plantarflexion and moves through a range of more than 15 degrees to arrive at 5 to 7 degrees of dorsiflexion by the end of this phase. As the lower limb rolls forward over the stance foot, the body weight vector becomes anterior to the ankle joint, creating an increasing dorsiflexion moment. Activity of the soleus assisted by the gastrocnemius controls the rate of dorsiflexion. Action of the plantarflexors is crucial in providing limb stability as the contralateral toe-off transfers body weight onto the stance foot.
At the beginning of midstance, the vector is posterior to the knee joint but moves anterior as midstance progresses. The knee extends from 15 degrees of flexion to a neutral position. This is particularly mechanically efficient since plantarflexion of the ankle is most forceful with the knee in extension. The quads are active as knee extensors in early midstance only. Momentum of the contralateral swing leg creates an extension torque on the ipsilateral knee that decreases demand on the quadriceps and extends the knee without muscle action. By the end of midstance, the vector is anterior to the knee, creating passive stability. In the coronal plane, the ground-reaction force line is medial to the anatomical knee joint on the stance side, creating a varus moment. The moment is restrained by the capsular structures of the knee, especially the lateral collateral ligaments.
The hip joint is in a flexed posture of 30 degrees, which is reduced to 10 degrees as midstance progresses. The vector is anterior to the hip in early midstance and moves increasingly posterior to the hip, gradually reducing the flexion torque and diminishing the demand on the hip extensors. The gluteus maximus, active in early midstance, yields to this passive hip extension as the hip nears vertical alignment over the femur. Vertical ground-reaction force is reduced in magnitude at midstance due to the upward momentum of the contralateral swing limb. This upward momentum improves stability at the ipsilateral hip. The gluteus maximus, at this point not needed for sagittal stability, is active as an abductor rather than a hip extensor.
In the coronal plane, activity of hip abductors during midstance is essential to provide hip stability and avoid excessive pelvic tilt. In the frontal view the body mass and the ground-reaction force are quite medial to the structural support point at the head of the femur. At the time of midstance during gait it has been estimated that the vertical loading on the head of the femur on the stance side reaches a magnitude approximately equal to 21/2 times body weight (18). This creates a strong tendency toward excessive pelvic tilt (positive trendelenberg). The gluteus medius responds to limit pelvic tilt and stabilize the pelvis.
In terminal stance, forward fall of the body moves the vector further anterior to the ankle, creating a large dorsiflexion moment (see Figure 16) . Stability of the tibia on the ankle is provided by the eccentric action of the calf muscles. The plantarflexors are more active during this heel-off period than any other period of gait. The soleus and gastrocnemius prevent forward tibial collapse and allow the heel to rise over the metatarsal heads as the center of mass of the HAT (head, arms and trunk) advances over the foot. This is referred to as the forefoot rocker.
The forefoot rocker is comprised of two components, and some believe there are two distinct forefoot rockers. The initial forefoot rocker (third rocker) begins at heel off and ends when the contralateral limb contacts the ground. The mechanics are much different in the terminal forefoot rocker (fourth rocker), which occurs in preswing as body weight rapidly is unloading the ipsilateral limb and shifting to the contralateral side. The initial forefoot rocker (third rocker) serves as an axis around which progression of the body vector advances beyond the area of foot support, creating the highest demand of the entire gait cycle on the calf muscles. Minimal ankle movement of 5 degrees is required to reach 10 degrees of dorsiflexion, which then is maintained. The maximum amount of dorsiflexion of the anatomical ankle joint occurs during heel off.
The knee achieves an angular position of full extension accompanied by a mild extension torque that diminishes in the latter part of terminal stance. Joint stability and forward progression at the knee are achieved without muscle action.
While it once was believed the hip underwent up to 10 degrees of hyperextension during this period, it actually is likely to be less. The accuracy of early goniometric measurements at the hip is suspect. Electrogoniometers are not well-suited for measurement around the hip where they may be prone to reflect lumbar motion as well as soft tissue displacement. At any rate, hip extension combined with 5 degrees of pelvic rotation provides a smooth progression and facilitates an increased step length. A mild hip-extension torque is present. The trailing posture of the limb and the presence of the vector posterior to the hip provide passive stability at the hip joint. The tensor fascia lata serves to restrain the posterior vector at the hip. At the end of terminal stance, the magnitude of the vertical force reaches a second peak greater than body weight similar to that which occurred at the end of loading response.
During preswing, the ankle moves rapidly from its dorsiflexion position at terminal stance to 20 degrees of plantarflexion (see Figure 17) . Although the ankle reaches its angular peak of plantarflexion during this period, actual plantarflexor activity is decreased in intensity as the limb is unloaded. In late preswing, the vertical force is diminished, and the plantarflexors are quiescent. There is no "push off" in normal reciprocal free walk bipedal gait. The dorsiflexion torque present at the beginning of preswing diminishes rapidly as the metatarsophalangeal joints extend to 60 degrees.
Passive knee flexion is created by planted hyperextended toes, advancement of the body past the metatarsal heads and contralateral loading. An early extension torque at the knee quickly gives way to a flexion torque. With the vector posterior to the knee, the knee flexes rapidly to achieve 35 degrees of flexion by the end of preswing, more than half the requirement for toe clearance in swing phase.
The hip flexes to a neutral position initiated by the rectus femoris, sartorius and adductor longus and assisted by momentum. The sagittal vector extends through the hip as the hip returns to a neutral posture. The adductor longus also decelerates the passive abduction created by contralateral body weight transfer. The continuing backward rotation of the pelvis effectively lengthens the trailing limb and counteracts hip flexion.
Action of the pretibial muscles and long toe extensors begins to lift the foot and the ankle, which initially is at approximately 20 degrees of plantarflexion, its maximum achieved at any period in the gait cycle (see Figure 18) . By the end of initial swing, however, plantarflexion position is reduced to about 5 to 10 degrees, providing foot clearance for the midswing phase.
Although the knee began initial swing in only 30 degrees of flexion, the momentum from hip flexion assisted by the short head of the biceps femoris, sartorius and gracilis creates further rapid knee flexion to 60 degrees with the goal of providing limb advancement and foot clearance.
The hip is flexed 20 degrees initiated not only by the iliacus but by activity of both the gracilis and sartorius, which contribute to flexion of both the hip and knee joints.
The knee extends as the ankle dorsiflexes, contributing to foot clearance while advancing the tibia (see Figure 19) . Pretibial muscle activity continues to preserve foot clearance as the ankle moves further toward dorsiflexion to reach a neutral position. Movement from plantarflexion toward dorsiflexion during the swing phase is referred to as dorsiflexion recovery.
Rapid knee extension, a passive event created by momentum, moves the knee from 60 to 30 degrees of flexion. Half of the knee extension needed for subsequent step length is achieved. The tibia assumes a relatively vertical position.
The hip flexors continue to preserve 30 degrees of hip flexion with mild EMG activity. The foot achieves ground clearance by 1 cm. The gracilis, sartorius and iliacus cut off in early midswing, and the hamstrings begin midway to decelerate the thigh. Additional limb advancement is created largely by momentum. Pelvic rotation is now neutral. The gluteus medius is quiescent on the ipsilateral side.
During terminal swing, the function of pretibial activity changes from one of foot clearance in swing to more appropriate limb placement and positioning for initial contact. A neutral position prepares the foot for the heel rocker function, assuring a heel first posture (see Figure 20) .
In the second half of terminal swing, the quadriceps extend the knee concentrically in a shortening contraction to facilitate full knee extension, which, assisted by pelvic rotation, accomplishes a full step length.
Eccentric contraction of both the hamstrings and the gluteus maximus is critical to accomplish deceleration of the thigh segment and restrain further hip flexion, which remains at 30 degrees. The long hamstrings have multiple roles of decelerating the leg, stabilizing the knee and limiting hip flexion in an eccentric or lengthening contraction. The gluteus maximus prepares for the impending forces of loading.
This review of human walking has explained several consistent patterns. With a few exceptions, muscular activity will oppose the external mechanical moment. Efficient body mechanics favors lengthening contractions, and agonists and antagonists active in opposition to each other actually are more the exception than the rule.
Normal human locomotion requires a complex interactive control between multiple limb and body segments that work congruently to provide the most shock-absorbing and energy-efficient forward movement possible. Gait characteristics are influenced by muscle strength, dynamic range of motion, and shape, position and function of numerous neuromuscular and musculoskeletal structures--as well as the ligamentous and capsular constraints of the joints. The primary goal is energy efficiency in progression using a stable kinetic chain of joints and limb segments that work congruently to transport the passenger unit forward.
The author would like to express appreciation to Ken Hudgens, program manager of the prosthetic and orthotic department, California State University--Dominguez Hills, for the preparation of Figure 9 .