Skeletal muscle mechanics: implications for rehabilitation.

Skeletal muscle biomechanical properties have been widely studied since the 1600s. Because the primary purpose of skeletal muscle is to generate force and movement, it is important to understand the mechanical basis of these properties. When muscle force or movement is lost due to disease, trauma, or disuse, an understanding of muscle properties can improve our ability to restore function.

Ideally, muscle contraction would be studied during normal movement, that is, under conditions of physiological activation, experiencing physiological loads, and contracting at physiological velocities. Clearly, this is technically not possible in humans. Thus, descriptions of muscle biomechanical properties have relied on experimental studies in model systems. In this review, we describe muscle contractile properties during either constant-length (isometric) or constant-force (isotonic) contractions. We will then use this information to speculate as to how muscles operate in vivo.

Length-Tension Relationship: Isometric Muscle Contraction 

The original biological experiments performed by Blix[1] demonstrated that the force developed by a muscle during isometric contraction (ie, when the muscle is not allowed to shorten) varies with the muscle's length. Experimentally, isometric contractions are performed at different lengths (Fig. 1), and peak isometric tension is measured at each length. These tensions are then plotted against length, and a relationship such as that shown in Figure 2 is obtained. It has thus been demonstrated that at very long and very short lengths, muscles generate low tension, whereas at intermediate or "optimal" lengths, muscles generate higher tension. Although a general description of this relationship was presented in the 1800s, the precise structural basis for the length-tension relationship in skeletal muscle was not elucidated until the sophisticated mechanical experiments of the early 1960s were performed.[2,3] It was these experiments that defined the precise relationship between myofilament overlap and tension generation, which we refer to today as the sarcomere length-tension relationship. In its most basic form, the length-tension relationship illustrates that tension generation in skeletal muscle is a direct function of the magnitude of overlap between the actin and myosin filaments.

Sarcomere Length-Tension Relationship

Sophisticated experiments by Gordon et al[2] defined what might be one of the most explicit structure-function relationships in all of biology. In these experiments, a small segment of muscle fiber was held at a constant length (and therefore a region of the fiber was held at a constant sarcomere length).

The results of these experiments[2] are summarized in Figure 2. In this figure, muscle relative tetanic tension (as a percentage of maximum) is plotted as a function of sarcomere length (in micrometers). This was one instance in which anatomy met physiology in dramatic fashion, because knowledge of the precise anatomic lengths of the myosin and actin filaments was crucial for understanding the basis of the sarcomere length-tension relationship.

Descending Limb of the Length-Tension Curve

As a muscle was highly stretched by the investigators[2] to a sarcomere length of 3.65 [mu]m, the muscle developed no active tension. Why? The answer lay in the observation that, as the myosin filament was 1.65-[mu]m long and the actin filament was 2.0 [mu]m in length, at a sarcomere length of 3.65 [mu]m, there was no overlap (interdigitation) between the actin and myosin filaments. Therefore, although fiber excitation might permit actin-myosin interaction by removing the inhibition of the actin filament, because no myosin cross-bridges are in the vicinity of the actin active sites, no tension generation occurred.

As muscle length decreased, overlap between actin and myosin was possible, and the amount of tension generated by the muscle increased as sarcomere length decreased. Increasing tension with decreasing sarcomere length occurred until the muscle reached a sarcomere length of 2.2 [mu]m, because, as sarcomere length decreased, the number of cross-bridge connections between actin and myosin increased, resulting in increased tension. This region of the length-tension curve is known as the descending limb.

Plateau Region of the Length-Tension Curve

As sarcomere length changed from 2.0 [mu]m to 2.2 [mu]m, muscle tension remained constant. Again, this was a direct result of thick filament structure. Because the myosin filament is a polymeric arrangement of myosin molecules arranged in an antiparallel fashion, many myosin "backbones" come together in the center of the myosin filament. Thus, there exists a bare region of the myosin molecule that is devoid of cross-bridges. The length of this bare region was 0.2 [mu]m. Even though sarcomere shortening over the range of 2.2 to 2.0 [mu]m resulted in greater filament overlap, it did not result in increased tension generation because no additional cross-bridge connections were made. The region of the length-tension curve over which length change results in no change in tension is known as the plateau region. The muscle length at which the maximum tetanic tension ([P.sub.o]) is attained is known as optimal length ([L.sub.o]).

Ascending Limb of the Length-Tension Curve

At a sarcomere length of 2.0 [mu]m, the actin filaments from one side of the sarcomere juxtapose the actin filaments from the opposite side of the sarcomere (Fig. 2). It might be predicted that shortening past this point would be impossible. As sarcomere length decreases below the plateau region, however, actin filaments from one side of the sarcomere overlap with the actin filaments on the opposite side of the sarcomere. That is, at these lengths, actin filaments overlap both with the opposing actin filament and with the myosin filament. Under these conditions, the actin filament from one side of the sarcomere interferes with cross-bridge formation on the other side of the sarcomere, resulting in decreased muscle force output. This decrease occurs from 2.0 to 1.87 [mu]m, and this region is known as the shallow ascending limb of the length-tension curve. The word "shallow" distinguishes this region from the next portion of the length-tension curve, which is known as the steep ascending limb, because at these very short lengths, the myosin filament actually begins to interfere with shortening as it abuts the sarcomere Z-disk, reducing force precipitously.

The length-tension relationship illustrates that muscle force varies as a function of sarcomere length (myofilament overlap). This is a physiologic property of the force-generating system and should not simply be viewed as an anatomic artifact.

Structural Basis of the Passive Length-Tension Curve

The thin line in Figure 2 represents the tension recorded if a muscle is stretched to various lengths without stimulation. Note that near the optimal length, passive tension is almost zero. As the muscle is stretched to longer lengths, however, passive tension increases dramatically. These relatively long lengths can be attained physiologically, and therefore, passive tension can play a role in providing resistive force even in the absence of muscle activity. The structure(s) responsible for passive tension are obviously outside of the cross-bridge itself, because muscle activation is not required. Recent studies have shown that the origin of passive muscle tension is actually within the myofibrils themselves.[4] Interestingly, a new structural protein has been identified that is the source of this passive tension. The very large protein, creatively named "titin" (and formerly termed "connectin") connects the thick myosin filaments end to end (Fig. 3). This very large protein is also relatively fragile and thus has probably been missed in earlier studies because laboratory techniques destroyed the protein. Experiments have now been performed in which isolated fibers were bathed in a solution that selectively extracted the titin molecule.[5] Under these conditions, the time course of loss in muscle passive tension is exactly correlated with the time course of appearance of titin protein in the extracting solution. In addition to passively supporting the sarcomere, titin stabilizes the myosin lattice so that high muscle forces do not disrupt the orderly hexagonal array. If titin is selectively destroyed, normal muscle contraction causes significant myofibrillar disruption.[6] Finally, it appears that titin may act as sort of a "molecular ruler" during the formation of sarcomeres (sarcomerogenesis), which occurs during development and following chronic muscle stretch during immobilization.[7]

Force-Velocity Relationship: Isotonic Muscle Contraction

In contrast to the sarcomere length-tension relationship, the force-velocity relationship does not have a precise, anatomically identifiable basis. The force-velocity relationship illustrates that the maximum force generated by a muscle is a function of its velocity. This relationship can also be stated in the reverse; that is, muscle contraction velocity is dependent on the force resisting the muscle. Historically, the force-velocity relationship was used to define the kinetic properties of the cross-bridges as well as the precise form of the force-velocity relationship itself. The form of this relationship has been shown to explain the behavior of whole muscle[8,9] and isolated single muscle fibers.[10]

Experimental elucidation of the force-velocity relationship was first presented by Hill[8] and Katz,[9] but the current description of the force-velocity relationship has been ascribed to Hill.[11]

The force-velocity relationship, like the length-tension relationship, is a curve that actually represents the results of many experiments plotted on the same graph. Experimentally, a muscle is stimulated maximally and allowed to shorten (or lengthen) against a constant load (Fig. 4A). The muscle velocity during shortening (or lengthening) is measured and then plotted against the resistive force. The general form of this relationship is plotted in Figure 4B. On the horizontal axis, we have plotted muscle velocity relative to maximum contraction velocity ([V.sub.max]), and on the vertical axis, we have plotted muscle force relative to maximum force ([P.sub.o]).

Concentric Contractions--Muscle Actively Shortening

When a muscle is maximally electrically activated and required to lift a load that is less than its maximum tetanic tension, the muscle begins to shorten. Contractions that permit the muscle to shorten are known as concentric contractions. In concentric contractions, the force generated by the muscle is always less than the muscle's maximum force ([P.sub.o]). As the load the muscle is required to lift decreases, contraction velocity increases. This occurs until the muscle finally reaches its [V.sub.max] at which force generation is zero. Maximum contraction velocity is a parameter we can use to characterize muscle that is related to both fiber type distribution and architecture. The force-velocity relationship is characterized by a rectangular hyperbola and is expressed mathematically as (1) (P + a)v = b([P.sub.o] - P) where a and b are constants derived experimentally (usually about 0.25), P is muscle force, and v is muscle velocity. This equation can be used to determine the relative muscle force that occurs as a muscle is allowed to shorten. Some of these values are presented in the Table. The force-velocity relationship is a steep rectangular hyperbola. In other words, force drops off rapidly as velocity increases. For example, in a muscle that is shortening at only 1% of its maximum contraction velocity (extremely slow), tension drops by 5% relative to maximum isometric tension (Table). Similarly, as contraction velocity increases to only 10% of maximum (easily attainable physiologically), muscle force drops by 35%. Even when muscle force is only 50% of maximum, muscle velocity is only 17% of [V.sub.max] As shortening speed increases, force drops precipitously.

Eccentric Contractions-Muscle Actively Lengthening

As the load imposed on the muscle increases, it reaches a point at which the external load is greater than the load the muscle can generate. Thus, the muscle is activated, but it is forced to lengthen due to the high external load. This is referred to as an eccentric contraction (contraction in this context does not necessarily imply shortening). There are two main features to note regarding eccentric contractions. First, the absolute muscle tensions are very high relative to the muscle's [P.sub.o]. Second, unlike concentric contractions, the absolute tension is relatively independent of lengthening velocity.[9] This suggests that skeletal muscles are very resistant to lengthening, a property that comes in very handy for many normal movement patterns in which muscles function as "brakes" to decelerate a limb (for example, the hamstring muscles "brake" the tibia during the swing phase of gait) or to absorb the momentum of the body during stance (for example, the quadriceps femoris muscles absorb body momentum during heel-strike).

Eccentric contractions are currently a subject of great interest for two main reasons. First, much of a muscle's normal activity occurs while it is actively lengthening; thus, eccentric contractions are physiologically common. Second, muscle injury and soreness are thought to be associated more with eccentric contractions.

The Cross-Bridge Cycle

We alluded to cyclic interaction between actin and myosin in explaining the force-velocity curve. Much of our understanding of the mechanism of muscle contraction has come from excellent biochemical studies performed from the 1950s to the mid-1970s.[12] It was during this period that methods for isolating specific muscle proteins were developed as well as the methods for measuring their physicochemical and biochemical properties.

In its simplest form, biochemical experiments on muscle contractile proteins have shown that during the cross-bridge cycle, actin (A) combines with myosin (M) and adenosine triphosphate (ATP) to produce force, adenosine diphosphate (ADP), and inorganic phosphate ([P.sub.i]). This can be represented as a chemical reaction in the form: (2) A + M + ATP [arrow right] A + M + ADP + [P.sub.i] + Force

However, we also know that death of a muscle produces a rigor state whereby actin and myosin interact to form a very stiff connection. This connection can be represented as (3) A + M [arrow right] A [multiplied by] M "rigor" complex

If actin and myosin can interact by themselves, where does ATP come into the picture during contraction? Experiments have demonstrated that the myosin molecule can hydrolyze ATP into ADP and [P.sub.i]. In other words, (4) M + ATP [arrow right] M + ADP + [P.sub.i]

Scientists now agree that ATP serves at least two functions in skeletal muscle systems. First, ATP disconnects actin from myosin. Second, ATP is hydrolyzed by the myosin molecule to produce the energy required for muscle contraction. This description of the different biochemical steps involved in muscle contraction is referred to as the Lymn-Taylor actomyosin adenosine triphophatase hydrolysis mechanism.[12,13]

The relationship between the Lymn-Taylor kinetic scheme and the mechanical cross-bridge cycle is not fully known. Lymn and Taylor,[13] however, proposed that their biochemical data could be incorporated into a four-step cross-bridge cycle that could be envisioned as follows (Fig. 5): 1. The actin-myosin bridge very rapidly

dissociates due to ATP binding

to myosin. 2. The free myosin bridge moves into

position to attach to actin, during

which ATP is hydrolyzed. 3. The free myosin bridge along with

its hydrolysis products rebinds to

the actin filament. 4. The cross-bridge generates force,

and actin displaces the reaction

products (ADP and [P.sub.i]) from the

myosin cross-bridge. This is the

rate-limiting step of contraction.

The actin-myosin cross-bridge is

now ready for the ATP binding of

step 1.

This is currently an active area of muscle biophysical research.[12] One might imagine the difficulty in confirming these elementary mechanical and biophysical reactions. A recent advance in biochemistry, however, has allowed direct testing and manipulation of this scheme. The advance involves the development of "caged" compounds--compounds that are inactive in their caged form and become active when the cage is instantaneously removed by a pulse of high-energy laser light.[14] Using caged ATP, single muscle fibers have been subjected to experiments such as those described earlier and found to behave much as predicted based on the biochemical data.[15]

Skeletal Muscle Architecture

All muscles are made of similar contractile units (sarcomeres), but the arrangement of these sarcomeres dramatically affects whole muscle contractile properties. The fundamental relationship between muscle architecture and muscle function is that muscle velocity and excursion are proportional to the number of sarcomeres in series, whereas muscle force is proportional to the total cross-sectional area of sarcomeres.[16,17] In practice, it is difficult to determine these values precisely, so we often say that muscle velocity and excursion are proportional to muscle fiber length, whereas muscle force is proportional to total fiber cross-sectional area (PCSA). Physiological cross-sectional area is a calculated variable[18] that estimates the total muscle fiber cross-sectional area. It is not necessarily the same as any area calculated in one of the traditional anatomical planes as would be obtained, for example, from computed tomography or magnetic resonance imaging. This is because the PCSA is a calculated amount that attempts to sum the entire muscle fiber cross-sectional area. Depending on the architecture of a particular muscle, all fibers will not be present in a given anatomical plane.

Muscle fiber length must be determined experimentally by microdissection of isolated skeletal muscles. Such microdissection has been performed for numerous animal and human muscles.[19,20] As architecture is a major determinant of muscle function, it helps to have a "feel" for the architectural properties of various human muscles. These properties have been measured in the lower extremity[21,22] and upper extremity[23-26] and are presented in graphical form in Figure 6. Because muscle fiber length is proportional to muscle excursion and muscle PCSA is proportional to [P.sub.o], the architectural specialization shown in Figure 6 suggests a wide range of forces and excursions that can be produced by the various skeletal muscles. This range of forces and excursions indicates that muscles are constructed of various designs, which enable them to produce high forces and/or high excursions. For example, the vastus muscles, with their relatively short, pennated fibers, appear to be specialized for high force production, whereas the sartorius muscle, with its longer fibers and smaller cross-sectional area, appears to be better suited for high excursions and low force. Such specialization is not necessarily an either/or proposition, as some muscles generate large forces (due to large size) and produce high excursion.

In the lower limb, inspection of functional muscle groups (eg, hamstrings, quadriceps femoris, dorsiflexors, plantar flexors) permits a number of generalizations to be made (Fig. 7). The quadriceps femoris muscles are characterized by their relatively high pennation angles, large PCSAs, and short fibers. In terms of design, these muscles appear suited for the generation of large forces. The hamstring muscles, however, by virtue of their relatively long fibers and intermediate PCSAs, appear to be designed for large excursions (because excursions are proportional to fiber length). The same appears to be true of the plantar flexors and dorsiflexors.

Force-Generating Properties of Muscles With Different Architectures

If we had two muscles with dramatically different designs but identical muscle mass, as illustrated in Figure 8, both muscles would have the same amount of contractile material (mass), but the arrangements would be quite different. The muscle in Figure 8B has relatively long fibers that extend almost the entire length of the muscle and are parallel to the muscle's force-generating axis (ie, the axis formed by connecting the line from muscle origin to insertion). This is the classic parallel-fibered muscle. The muscle shown in Figure 8A has relatively short fibers that extend a very short length relative to the muscle and are tilted by about 30 degrees to the muscle's force-generating axis. This is the classic pennated muscle.

In Figure 9A, we have plotted the length-tension curves of the two muscles. Note that the muscle with the longer fibers has a greater absolute working range than the muscle with the shorter fibers. This is because, for a given length change, each sarcomere in the long-fibered muscle lengthens less, the length change being distributed over a greater number of sarcomeres. However, note also that the long-fibered muscle generates a lower tension than the short-fibered muscle because the muscle with the shorter fibers contains a much greater PCSA. Generally, muscles with short fibers and a large PCSA are designed for force production, whereas muscles with long fibers are designed for excursion. This concept is well illustrated by the force-velocity curves plotted in Figure 9B. Note that the muscle with long fibers has a [V.sub.max] that is much greater than that of the muscle with short fibers. Each sarcomere within the fiber contracts at the same velocity, regardless of fiber length. However, by placing more sarcomeres in series in the muscle with long fibers, the overall muscle velocity is greater. Again, note that the P, for the muscle with short fibers is much greater than that observed for the muscle with long fibers due to its greater PCSA

Torque Definition

We have discussed isolated muscle properties, but in the musculoskeletal system, muscles generate force and transmit the force, via tendons, to the bones. If muscles generate sufficient force, bones rotate about joint axes. We thus have a mechanical system that can be quantitatively described using the equation for torque ([tau]): (5) [Mathematical Expression Omitted] where r is the moment arm, F is the applied force (due to muscle tension), and x represents the vector cross-product of the two quantities Fig. 10A). Thus, a force (F) is applied a distance (r) away from an axis (open circle, Fig. 10A) that is free to rotate. The arrows above the moment arm and force variables signify that they are vector quantities. That is, both have magnitude as well as direction. Ibis is another way of saying that the orientation between r and F is important. If we expand this equation, we can express torque as (6) [tau] = ~r~ [multiplied by] ~F~ [multiplied by] sin[theta] where the vertical bars around r and F represent the vector magnitudes, and [theta] is the angle between the direction of force application and the axis of rotation. The basis for this expression can be seen in Figure 10B, where the line of force (F) is applied at an angle ([theta]) relative to the axis of the moment arm. The component of force that causes rotation is the short side of the triangle (F [multiplied by] sin[theta]).

The moment arm is measured in distance units such as meters. Because the sin[theta] term is dimensionless, torque has units of newton-meters (force-distance units). Another common unit of torque is foot-pounds (such as are used in the automobile industry and, interestingly, on many isokinetic dynamometers used in rehabilitation).

The word "strength," as we use it, clinically actually represents torque. All real-world performance (eg, sprinting, lifting weights, getting out of bed, writing) represents manifestations of torque by the musculoskeletal system. If we grasp the implications of this statement, it is clearly incorrect to conclude that a person is strong only because his or her muscles generate large forces. It would be just as ridiculous to conclude that a person is strong only because that person has large moment arms. Based on this discussion, it is clear that at least three strategies exist for changing torque: (1) changing force magnitude (tension-producing capability of a muscle), (2) changing moment arm magnitude (changing the insertion of a muscle), and (3) changing the angle between force magnitude and moment arm magnitude. Two of these strategies require surgical changes, whereas the third (changing tension) is amenable to training.

Muscle Force Producing the Moment

Determination of the muscle force that produces a joint moment is simple in principle but extremely difficult in practice. As discussed previously, if muscle is fully activated, muscle force varies with muscle length (length-tension relationship) and muscle velocity (force-velocity relationship). If we know the muscle's length and velocity, we theoretically can predict muscle force.

Moment Arm Producing the Moment

Determination of joint moment arm requires an understanding of the anatomy and movement (kinematics) of the joint of interest. For example, some joints can be considered to rotate about a fixed point. A good example of such a joint is the elbow.[27] At the elbow joint, where the humerus and ulna articulate, the resulting rotation occurs primarily about a fixed point, referred to as the center of rotation. In the case of the elbow joint, this center of rotation is relatively constant throughout the joint's range of motion (ROM).[27] In other joints (eg, the knee), however, the center of rotation moves in space as the knee joint rotates, because the articulating surfaces are not perfect circles.[28] In the case of the knee, it is not appropriate to discuss a single center of rotation--rather, we must speak of a center of rotation corresponding to a particular joint angle, or, using the terminology of joint kinematics, we must speak of the instant center of rotation (ICR), that is, the center of rotation at any "instant" in time or space.

Having defined a joint ICR, the moment arm is defined as the perpendicular distance from the line of force application to the axis of rotation. The moment arm is illustrated in Figure 11 for a simulated elbow joint. In Figure 11A, the elbow joint is almost fully extended. The angle ([theta]) between the brachialis muscle and the ulna is relatively small (ie, [theta]=20[degrees]). The distance between the brachialis muscle insertion site and the elbow ICR is 5 cm. In this example, the perpendicular distance between the line of force application and the elbow ICR is shown by the dashed line and is equivalent to 5 cm x sin(20[degrees]) = 1.7 cm. Thus, because the joint is nearly fully extended, this presents an unfavorable mechanical advantage to the muscle--the moment arm is relatively small. Much of the force generated by the muscle will simply compress the joint surface, not rotate it. Contrast this situation with the conditions shown in Figure 11B, in which the joint has now been flexed so that [theta]=50 degrees. Now, the moment arm equals 5 cm x sin(50[degrees]) = 4.3 cm. We see that for a simple hinge joint (a joint with a fixed ICR), the maximum moment arm is attained at [theta]=90 degrees, If we plotted moment arm versus joint angle for this simple hinge joint, we would obtain a simple sine function that has a maximum of 5 cm occurring at [theta]=90 degrees. Such a curve can be generated for any joint. In general, the experimental curves are not quite as simple as the one previously discussed.

Joint Angle Corresponding to Maximum Muscle Force

Examination of current physiology texts reveals a good deal of uncertainty regarding the definition of the joint angle that corresponds to maximum muscle force. Recent studies of torque generation in animals and humans have generally agreed that the joint angle at which the muscle generates maximal force is not the same angle at which the moment arm is maximum. Thus, during normal joint rotation, both the moment arm and muscle force are constantly changing, which results in the "shape" of the strength or torque curve. This concept has recently been addressed in detail by experimental and theoretical modeling.[29-31]

Range of Motion as a Function of Architecture

We can now "combine" the muscle architectural discussion with the joint moment arm concept. We stated that muscles with longer fibers have a longer functional ROM than muscles with shorter fibers. Does this imply that muscles with longer fibers are associated with joints that have larger ROMs? No! A muscle with longer fibers does have a longer working ROM. The amount of muscle length change that occurs as a joint rotates, however, is very strongly dependent on the muscle's moment arm (the perpendicular distance from the muscle insertion to the axis of joint rotation). Ibis idea is illustrated in Figure 12, in which we have attached a simulated "muscle" using two different moment arms. In Figure 12A, the moment arm is much less than in Figure. 12B. This means that in Figure 12A, the muscle will change length much less for a given change in joint angle compared with the same change in joint angle in Figure 12B. As a result, the active ROM for the muscle-joint system shown in Figure 12A will be much greater than that shown in Figure 12B in spite of the fact that their muscular properties are identical. In this example, increasing the moment arm decreased the ROM from 40 degrees (Fig. 12A) to only 25 degrees (Fig. 12B).

We should now qualify the previous statement about muscle design and architecture. Muscles that are, for example, designed for speed because of their very long fibers may not actually produce large velocities of contraction if they are placed in position with a very large moment arm. The increased moment arm causes a greater joint moment, and the muscle is actually best suited for torque production. Similarly, a muscle that appears to be designed for force production due to the large PCSA, if placed in position with a very small moment arm, may actually produce high joint excursions or angular velocities. Thus, muscle design may or may not be a reflection of its actual use in the physiologic muscle-joint torque-generating system. In general, muscle fiber length and muscle moment arm are positively correlated.32 Thus, muscles with long fibers tend to have long moment arms, but this is not necessarily the case. Muscle architectural features may represent muscle adaptation to kinematic criteria.

Summary

Skeletal muscles represent a clear example of a structure function relationship. Isometric contractile properties result from the relative overlap of actin and myosin contractile filaments. Isotonic force results from the timing of the interaction between actin and myosin during the cross-bridge cycle. Sarcomere arrangement in whole muscles dramatically affects contractile properties and is termed "architecture." Muscle architecture is highly specialized for the various upper- and lower-extremity muscles.

The interaction between these muscles and the joint provides flexibility and specialization, which are required for normal movement. it can now be seen that normal movement represents the culmination of complex interaction between the nervous system, muscles, and joints. Evaluation of a phenomenon as complex as "movement" or "strength" requires an understanding of the role of each component of this system in producing the movement or the force. As diagnostic procedures improve in their specificity, rehabilitation may be directed toward the particular physiological component that does not function as desired. In this way, treatments can be developed that have a rational physiological basis and that result in measurable physiological changes.

Acknowledgment

We thank Becky Chamberlain for her artistic work.

References

[1] Blix M. Die lange und die spannung des muskels. Vierte Abbandlung Skand Arch Physiol. 1895;5:173-206. [2] Gordon AM, Huxley AF, Julian FJ. The variation in isometric tension with sarcomere length in vertebrate muscle fibers. J Physiol (Lond). 1966;184:170-192. [3] Edman KAP. The velocity of unloaded shortening and its relation to sarcomere length and isometric force in vertebrate muscle fibres. J Physiol (Lond). 1979;246:255-275. [4] Magid A, Law DJ. Myofibrils bear most of the resting tension in frog skeletal muscle. Science. 1985;230:1280-1282. [5] Funatsu T, Higuchi H, Ishiwata S. Elastic filaments in skeletal muscle revealed by selective removal of thin filaments with plasma gelsolin. J Cell Biol. 1990;110:53-62. [6] Horowits R, Podolsky RJ. The positional stability of thick filaments in activated skeletal muscle depends on sarcomere length: evidence for the role of titin filaments. J Cell Biol. 1987;105:2217-2223. [7] Williams P, Goldspink G. The effect of immobilization on the longitudinal growth of striated muscle fibers. J Anat. 1973;116:45-55. [8] Hill AV. The heat of shortening and the dynamic constants of muscle. Proc R Soc Lond [Biol]. 1938;126:136-195. [9] Katz B. The relation between force and speed in muscular contraction. J Physiol (Lond). 1939;96:45-64. [10] Edman KAP. The velocity of unloaded shortening and its relation to sarcomere length and isometric force in vertebrate muscle fibres. J Physiol (Lond). 1979;246:255-275. [11] Hill AV. First and Last Experiments in Muscle Mechanics. New York, NY: Cambridge University Press; 1970. [12] Webb MR, Trentham DR. Chemical mechanism of myosin-catalyzed ATP hydrolysis. In: Peachy ID, Adrian RH, Geiger SR, eds. Handbook of Physiology. Bethesda, Md: American Physiological Society; 1983:237-255. [13] Lymn RW, Taylor EW. Mechanism of adenosine triphosphate hydrolysis by actomyosin. Biochemistry. 1971;10:4617-4624. [14] McCray JA, Herbette L, Kihara T, Trentham DR. A new approach to time-resolved studies of ATP-requiring biological systems: laser flash photolysis of caged ATP. Proc Natl Acad Sci USA 1980;77:7237-7241. [15] Goldman YE. Kinetics of the actomyosin ATPase in muscle fibers. Annu Rev Physiol. 1987;49:637-654. [16] Powell PL, Roy RR, Kanim P, et al. Predictability of skeletal tension from architectural determinations in guinea hindlimbs. J Appl Physiol. 1984;57:1715-1721. [17] Bodine SC, Roy RR, Meadows DA, et al. Architectural, histochemical, and contractile characteristics of a unique biarticular muscle: the cat semitendinosus. J Neurophysiol. 1982; 48:192-201. [18] Lieber RL. Skeletal Muscle Structure and Function. Baltimore, Md: Williams & Wilkins; 1992. [19] Sacks RD, Roy RR. Architecture of the hind limb muscles of cats: functional significance. J Morphol. 1982;173:185-195. [20] Roy RR, Medows ID, Baldwin KM, Edgerton VR. Functional significance of compensatory overloaded rat fast muscle. J Appl Physiol. 1982;52:473-478. [21] Wickiewicz TL, Roy RR, Powell PL, Edgerton VR. Muscle architecture of the human lower limb. Clin Orthop. 1983;179:275-283. [22] Friedrich JA, Brand RA. Muscle fiber architecture in the human lower limb. J Biomech. 1990;23:91-95. [23] Brand PW, Beach RB, Thompson DE. Relative tension and potential excursion of muscles in the forearm and hand. J Hand Surg [Am]. 1981;3:209-219. [24] An KN, Hui FC, Morrey BF, et al. Muscles across the elbow joint: a biomechanical analysis J Biomech. 1981;14:659-669. [25] Lieber RL, Fazeli BM, Botte MJ. Architecture of selected wrist flexor and extensor muscles. J Hand Surg [Am]. 1990; 15:244-250. [26] Lieber RL, Jacobson MD, Fazeli BM. Architecture of selected muscles of the arm and forearm: anatomy and implications for tendon transfer. J Hand Surg [Am]. 1992; 17:787-798. [27] London JT. Kinematics of the elbow. J Bone Joint Surg[Am]. 1981;63:529-535. [28] Nisell R. Mechanics of the knee. Acta Orthop Scand. 1985;56:1-42. [29] Lieber RL, Boakes JL. Sarcomere length and joint kinematics during torque production in the frog hindlimb. Am J Physiol. 1988;254: C759-C768. [30] Hoy MG, Zajac FE, Gordon ME. A musculoskeletal model of the human lower extremity: the effect of muscle, tendon, and moment arm on the moment-angle relationship of musculotendon actuators at the hip, knee, and ankle. J Biomech. 1990;23:157-169. [31] Lieber RL, Shoemaker SD. Muscle, joint, and tendon contributions to the torque profile of frog hip joint. Am J Physiol. 1992;263:R586-R590. [32] McClearn D. Anatomy of raccoon (Procyon lotor) and caoti (Nasua narica and N nasua) forearm and leg muscles: relations between fiber length, moment-arm length, and joint excursion. J Morphol. 1985;183:87-115.

RL Lieber, PhD, is Associate Professor, Department of Orthopaedics, Biomedical Sciences Graduate Group, University of California at San Diego School of Medicine and Veterans Administration Medical Center, 3350 La Jolla Village Dr, San Diego, CA 92161 (USA). Address all correspondence to Dr Lieber.

SC Bodine-Fowler, PhD, is Assistant Professor, Department of Orthopaedics, Biomedical Sciences Graduate Group, University of California at San Diego School of Medicine and Veterans Administration Medical Center.

This work was supported by the Veterans Administration and National Institutes of Health Grants AR34192 and AR40050.