JOURNAL OF EXPERIMENTAL ZOOLOGY 305A:923934 (2006)

Correlation of Symmetrical Gaits and Whole Body Mechanics: Debunking Myths in Locomotor Biodynamics
AUDRONE R. BIKNEVICIUS1 AND STEPHEN M. REILLY2 1 Department of Biomedical Sciences, Ohio University College of Osteopathic Medicine, Athens, Ohio 45701 2 Department of Biological Sciences, Ohio University, Athens, Ohio 45701

ABSTRACT

Independent maturation of gait (Hildebrand) and whole body mechanics (Cavagna et al.) traditions in locomotor analyses has led to conicting terminology. Re-evaluation of these traditions yields three primary insights. First, walking and running should be recognized by their fundamentally different mechanics. Because duty factor fails to consistently distinguish these mechanics, its use in discriminating walks from runs should be abandoned in preference to parameters that more accurately reect the movements of the center of mass (COM; phase difference in external mechanical energy or Froude number). Second, trot should be reserved as a descriptor of a particular footfall pattern. This and all gait terms lack explicit information about limb compliance and thus COM movements. Third, symmetrical gait denitions should be broadened to reect the four primary footfall patterns: the lateral-couplet dominated pattern of the pace, the diagonal-couplet dominated pattern of the trot and the more independent sequencing of footfalls of the two singlefoots. Intermediate gaits (perennially confusing and a mouthful to pronounce) are thereby subsumed by these four discrete gaits. Confusion between gait terminologies would be avoided r 2006 Wiley-Liss, Inc. if limb phase were consistently reported. J. Exp. Zool. 305A:923934, 2006.

How to cite this article: Biknevicius AR, Reilly SM. 2006. Correlation of symmetrical gaits and whole body mechanics: debunking myths in locomotor biodynamics. J. Exp. Zool. 305A:923934.

Scientic discovery is disseminated through published reports or professional presentations. The common feature of both venues is the reliance on wordswords that link the thoughts of the author with the imagination of the audience. While science prides itself in its precision, sometimes the same words are applied to different physical entities. For example, a navicular bone is one of the tarsal bones in humans but it is also the distal sesmoid bone in the feet of ungulates (Riegel and Hakola, 99). Adoption of common words to reect discrete physical entities in science is perhaps even more likely to yield confusion. Hence, reference to a knee conjures up the image of the femurtibia articulation unless, of course, the audience is composed of equine researchers for whom knee implies the carpus (Riegel and Hakola, 99). Independent development of terminology can occur because of disciplinary isolation or methodological differences between avenues of research (Reilly and
r 2006 WILEY-LISS, INC.

Biknevicius, 2003). One consequence of autonomously maturing research elds is that walks and runs have been dened using different parameters that signify overlapping, but not equivalent, physical realities.

Walking and running footfall patterns Hildebrand gait model

The rst method to distinguish walks from runs is based exclusively on duty factor (the ratio of support duration to stride duration), a parameter
Grant sponsor: National Science Foundation; Grant number: IBN 0080158. Correspondence to: A.R. Biknevicius, Department of Biomedical Sciences, 125 Life Sciences Building, Ohio University College of Osteopathic Medicine, Athens, OH 45701. E-mail: biknevic@ohio.edu Received 12 February 2005; Accepted 28 June 2006 Published online 6 October 2006 in Wiley InterScience (www. interscience.wiley.com). DOI: 10.1002/jez.a.332.

924

A.R. BIKNEVICIUS AND S.M. REILLY

that is frequently used as a proxy for locomotor speed (duty factor most commonly scales inversely with speed; Demes et al., 94). As codied by Hildebrand (65), with inspiration from earlier workers (e.g., Howell, 44), running occurs when duty factor falls below 0.5; conversely, duty factors Z0.5 are considered to be walks. Walks and runs are further categorized into gaits (timing of footfalls), a tradition that stems from at least the time of Goiffon and Vincent (1779). About 200 years later, Hildebrand (65, 76) systematized symmetrical gaits according to limb phase, dened as the phase difference between the touchdowns of ipsilateral hind- and forelimbs as a ratio of stride duration (Fig. 1A). A forelimb lands progressively later in the stride duration relative to its ipsilateral hindlimb as limb phase increases, i.e., ipsilateral fore- and hindlimbs land synchronously at a limb phase of 0 whereas forelimb touchdown is half a stride duration after the ipsilateral hindlimb when limb phase is 0.5. The discrete symmetrical gaits are pace (limb phase 5 0), trot (0.5) and the two singlefoots (0.25 and 0.75), allowing approximately 70.06 limb phase about these values. The remainder of the limb phase range can either be considered to represent intermediate gaits or less synchronized versions of the discrete gaits (e.g., a 0.41 limb phase is either deemed a lateralsequence diagonal couplets or a 4-beat trot, respectively). [We refer the readers to Reilly and Biknevicius (2003) or Stevens (this volume) for a more detailed explanation of how symmetrical gaits are identied.] The classic Hildebrand symmetrical gait plot (Fig. 1A) charts limb phase (gait) against duty factor (speed); the footprintshaped area on the plot circumscribes symmetrical gaits in walks and runs (as per Hildebrand, 76). By this method, some gaits are theoretically possible as both walks and runs (trots, singlefoots) whereas other gaits occur primarily as runs (pace) or as walks (all other symmetrical gaits).

The popularity of this footfall-kinematics approach to dening walks and runs (as well as gaits) is its ease of data capture: all that is required is a camera, preferably high-speed. As early as 1976, a rich database already existed on the gaits of at least 143 non-avian tetrapod genera (seven amphibians, four lizards, two crocodilians and 130 mammals) plus 37 dog breeds and 13 horse breeds (Hildebrand, 65, 67, 68, 76). Hildebrands work recorded gaits across a wide array of locomotor behaviors, including linear and turning trials as well as steady and non-steady speed. Although subsequent studies largely focused on steady speed and level terrestrial locomotion (e.g., White and Anderson, 94; Parchman et al., 2003), some have branched out to other substrates (e.g., arboreal locomotion: Lammers and Biknevicius, 2004; Stevens, 2006). The underlying importance of understanding footfall gaits is that they are one way to quantify the output of the neuromotor system to move the body in response to environmental demands.

Walking and running center of mass (COM) mechanicsCavagna et al. model

Another method for distinguishing walks from runs involves the movements of the COM of the entire body (Fig. 1B) and is referred to here as the whole body mechanics approach (Cavagna et al., 77). Because slow moving tetrapods travel with limbs that have a high level of stiffness, their COM is at its highest position near midstance when the forward velocity of the COM is low. Consequently, gravitational potential energy (Ep) cycles out of phase with total kinetic energy (Ek) and the phase difference between minima of Ep and Ek (phase shift) is near 1801. This pendulum-like exchange of energies provides an opportunity to recover external mechanical energy with every step, thereby reducing muscular effort and the metabolic cost of locomotion at slow speeds. At faster speeds, the limbs of tetrapods function with

Fig. 1. Synthesis of symmetrical gaits and whole body mechanics. (A) Kinematic (footfall) approach to terrestrial locomotion. Shaded area represents symmetrical gaits. Gaits are determined by limb phase; duty factor distinguishes walking (Z0.5) from running (o0.5) gaits. Modied from Hildebrand (76). (B) Whole body mechanics approach to terrestrial locomotion. Total kinetic and gravitational potential energies uctuate out of phase with one another during walking (pendulum-like or vaulting mechanics; left). In-phase uctuations of these external mechanical energies occur during running (spring-mass or bouncing mechanics; right). Modied from Farley and Ko (97). (C) If the two denitions of walking and running in (A) and (B) were completely congruent, then vaulting mechanics would be found only on the high duty factor (walking) side of the gait plot and bouncing mechanics only on the low duty factor (running) side of the gait plot. (D). Empirical evidence refutes the idealistic synthesis (i.e., complete congruence between footfall and whole body mechanics denitions) primarily by the common nding of bouncing mechanics used at high duty factors. Gait space occupied by quadrupeds indicated by the ellipses (Sources: lizards and salamander, E.J. McElroy, unpublished data; all others, see Table 1). (E) Symmetrical gaits redened as four primary gaits (pace, lateral-sequence singlefoot, trot, diagonal-sequence singlefoot), each occupying a range of 0.25 limb phase. Intermediate gaits are subsumed into the primary gaits.

J. Exp. Zool. DOI 10.1002/jez.a

UNIFYING LOCOMOTOR PERSPECTIVES

925

J. Exp. Zool. DOI 10.1002/jez.a

926

A.R. BIKNEVICIUS AND S.M. REILLY

greater compliance so that the COM no longer rises during the rst half of stance but rather drops to its lowest position near midstance. The resulting in-phase uctuations in Ep and Ek are inconsistent with pendulum-like mechanics (phase shift near 01) but some recovery of mechanical energy is still, at least theoretically, possible via the storage and release of elastic strain energy (another form of potential energy) in the ligaments, tendons and muscles of the limbs and back. Therefore, from the perspective of whole body mechanics, walks and runs are synonymous with inverted pendulum (vaulting) and springmass (bouncing) mechanics, respectively. There is a limit to how quickly tetrapods can move and still take advantage of inverted pendulum mechanics while retaining constant contact with the ground (Alexander, 76). As the COM moves in an arc over a stiff limb, it resembles a mass attached to the end of a string moving in a circle. The mass will continue to move about the circle as long as the centripetal force required to maintain the movement of the body mass along the arc (mv2/l where m is mass, v is forward velocity and l is length of the string or, in the case of walking, hip height at midstance) does not exceed gravitational force (mg where g is the gravitational acceleration or 9.81 m sec1). The ratio of these forces is known as Froude number (Fr 5 v2/gl). In theory, animals may continue to move with inverted pendulum mechanics as long as Fro1 but empirical data show that animals typically switch to bouncing mechanics at much slower speeds (Fr$0.5 for bipeds, $0.35 for quadrupeds; Alexander, 77; Alexander and Jayes, 83; Gatesy and Biewener, 91; Grifn et al., 2004), so that the trigger for a vaultbounce transition is determined not by the limits of centripetal force but rather by some metabolic, mechanical and/or perceived exertion factors (Hoyt and Taylor, 81; Farley and Taylor, 91; Raynor et al., 2002; Grifn et al., 2004). Because Fr also reects the ratio of Ek to Ep, this approach is actually an extension of the whole body mechanics approach. The dataset for whole body mechanics is less extensive than that for gaits. Nearly 30 years later after Cavagna et al.s (77) study, whole body mechanics have been recorded for only 18 nonavian genera (one amphibian, three lizards, one crocodilian and 13 mammals; Table 1). The count rises to 29 genera when birds (eight genera) and invertebrates (three) are included, still a far cry from the sampling of gaits in tetrapods. FurtherJ. Exp. Zool. DOI 10.1002/jez.a

more, whole body mechanics studies tend to explore a small subset of terrestrial locomotor behavior, namely, steady-speed locomotion on level substrates. When an animal moves in bursts, Ek is lost and the work required for locomotion increases (Alexander, 89).

Integrating walking and running gaits with vaulting and bouncing COM mechanics
How well do the footfall and whole body mechanics approaches to dening walks and runs coincide? The most simplistic (and optimistic) synthesis (Reilly and Biknevicius, 2003) would map vaulting mechanics (out-of-phase uctuations of Ep and Ek) on the walking side (duty factor Z0.5) of the Hildebrand gait space and bouncing mechanics (in-phase Ep and Ek) on the running side (duty factor o0.5; Fig. 1C). Locomotor data on tetrapods that include hindlimb duty factor, whole body mechanics and gait (footfall patterns of all limbs) are surprisingly limited (12 nonhopping quadrupeds: Table 1); nevertheless, a clear pattern has emerged (Fig. 1D). At rst, the idealistic synthesis appears to have some support: vaulting mechanics is limited to high duty factors and only bouncing mechanics occurs at low duty factors. But on closer inspection, one nds that the 0.5 duty factor line does not cleanly separate vaulting and bouncing strides because bouncing mechanics repeatedly breaches the 0.5 walkrun duty factor boundary (e.g., trotting opossums, singlefooting horses; Fig. 1D). In fact, the entire mechanical repertoire of some species, such as rats, occurs at duty factors greater than 0.5! The idealistic synthesis is, therefore, soundly rejected, and one is led to the conclusion that the Hildebrand (footfall kinematics) and Cavagna et al. (whole body mechanics) denitions of walks and runs are not interchangeable. BRIDGING THE VOCABULARY GAP Recognizing that some terminology (walk, run, trot) is used inconsistently within the locomotor biodynamics community is but the rst and easiest step in resolving semantic conicts. Much more difcult is implementing a change in the culture so that these inconsistencies and misunderstandings are minimized. In the subsequent sections, we offer several compromises that are supported by empirical evidence from both the Hildebrand and Cavagna et al. traditions. These are organized as

UNIFYING LOCOMOTOR PERSPECTIVES TABLE 1. Gaits used during whole body mechanics studies Gaits1 Common name Quadrupeds Anuran Lizards Frog Skink Gecko Lizard Alligator Opossum Rat Chipmunk Ground squirrel Macaque Dog (domestic) Dog (domestic) Horse (domestic) Ram (domestic) Bipeds Mammalian Kangaroo Kangaroo rat Kangaroo rat Springhare Human (normal loco) Human (Groucho) Tinamous Rhea Ostrich Quail Quail Chicken Turkey Penguin Cockroaches Cockroaches Crab Genus species Kassina maculata Eumeces skiltonianus Coleonyx variegatus Sceloporus magister Alligator mississippiensis Monodelphis domestica Rattus norvegicus Tamias striatus Spermophilus tridecemlineatus Macaca speciosa Canis familiaris Canis familiaris Equus caballus Ovis musimon Megaleia rufa Dipodomys merriami Dipodomys spectabilis Pedetes cafer Homo sapiens Homo sapiens Eudromia elegans Rhea americana Struthio camelus Excalfactoria chinensis Colinus virginianus Gallus gallus Meleagris gallopavo Aptenodytes forsteri Blaberus discoidalis Periplaneta americana Ocypode quadrata Vaulting LSDC T T LSDC, T LSDC, T T G G W W LSLC LSSF W T T T, G T H H H H R GR GR, R R GR, R R R GR, R R Bouncing LSDC, T T T T Source Ahn et al. (2004) Farley and Ko (97) Farley and Ko (97) Our dataunpubl. Willey et al. (2004)

927

Crocodilian Mammalian

Parchman et al. (2003) Our dataunpubl. Heglund et al. (82) Heglund et al. (82) Cavagna et al. (77) Cavagna et al. (77) Grifn et al. (2004) Minetti et al. (99) Cavagna et al. (77) Cavagna Heglund Heglund Cavagna Cavagna et et et et et al. al. al. al. al. (77) (82) (82) (77) (77)

W GW W W W W W W W W

Avian

McMahon et al. (87); Bertram et al. (2002) Hancock et al. (2004) Cavagna et al. (77) Rubenson et al. (2004) Heglund et al. (82) Heglund et al. (82) Muir et al. (96) Cavagna et al. (77) Grifn and Kram (2000) Full and Tu (90) Full and Tu (91) Blickhan and Full (87)

Polypeds Invertebrate

ATe

ATr ATr ATe

Blank cells in the Vaulting and Bouncing columns indicate that either vaulting or bouncing mechanics have not been recorded (human studies listed are representative only). 1 Abbreviations for gaits: ATe, alternating tetrapod gait (Barnes, 75); ATr, alternating tripod (polypedal version of the trot); G, gallop; GR, bipedal grounded run (signies Groucho run in humans); GW, groucho walk (humans only); H, bipedal hop; LSDC, lateral-sequence diagonal couplets (4-beat trot); LSLC, lateral-sequence lateral couplets (4-beat pace); LSSF, lateral-sequence singlefoot; R, bipedal run; T, trot; W, bipedal walk; W, unspecied walking gait.

mythspopular beliefs that are unsupported by facts and that result in distorted truths. The rst involves the belief, held by some kinematically oriented biomechanists, that runs necessarily include a ight phase. The second myth explores the assumption, held by some biomechanists of the whole body mechanics tradition, that trots are always runs.

Myth 1: Runs contain an aerial phase

We propose that walk and run be used to explicitly describe vaulting and bouncing COM mechanics, respectively (a convention that will be used hereafter in this paper). There is no simple equivalence between the Hildebrand (footfall) and Cavagna et al. (whole body mechanics) denitions
J. Exp. Zool. DOI 10.1002/jez.a

928

A.R. BIKNEVICIUS AND S.M. REILLY

of walks and runs, and in this section we present supportive evidence that duty factor fails to provide consistent and meaningful mechanical distinctions between locomotor behaviors. At the core of the rst myth (runs contain an aerial phase) is the 0.5 duty factor boundary between walks and runs sensu Hildebrand (65). The concept that running requires a period of suspension between steps is virtually dogma in common parlance. It is fostered by Muybridges (1887) sensational serial snapshot photographs that revealed aerial phases in trotting horses or the fact that human racewalkers fault when an aerial phase is detected (Rogers, 2000). Yet, data from both the Cavagna et al. and Hildebrand traditions have repeatedly shown that a ight phase is not a prerequisite for running. Bipeds provide the simplest case in point. Although the natural vaultbounce transition in humans is coincident with an abrupt drop in duty factor and an incorporation of an aerial phase (Gatesy and Biewener, 91; Kram et al., 97), humans can modify hindlimb dynamics to run without an aerial phase. This was rst demonstrated with Groucho running, an unnatural gait in humans characterized by a high degree of limb compliance, elongated step lengths and no period of suspension (McMahon et al., 87). Vertical oscillations of the COM are reduced during Groucho running but still include a small drop near midstance (relative to its vertical position at touchdown and liftoff) that is consistent with bouncing mechanics. Increased substrate compliance may work together with limb compliance to produce similar results (McMahon and Greene, 79). While Groucho running is a contrived gait in humans, terrestrial locomotion with bouncing mechanics in the absence of an aerial phase occurs naturally within another group of striding bipeds, namely birds. Ground-dwelling birds smoothly pass across the 0.40.6 Fr boundary traditionally set for walkrun transitions without incorporating a period of suspension between steps (Gatesy and Biewener, 91). Nor do these birds display discrete changes in limb kinematics as they cross this boundary. Although the birds traveling at these intermediate speeds move with high stride frequencies typical of running, the lack of an aerial phase has led many researchers to label this locomotor behavior as a walk (e.g., Reilly, 2000). However, recent whole body mechanics studies veried that Ep and Ek of the COM uctuate in phase with one another (a hallmark
J. Exp. Zool. DOI 10.1002/jez.a

of bouncing mechanics) when Fr is greater than $0.6 even if the birds have not crossed the 0.5 duty factor boundary to an aerial run (ostriches Struthio camelus, Rubenson et al., 2004; tinamous Eudromia elegans, Hancock et al., 2004). Thus, Fr is more accurate than duty factor in predicting the transition from vaulting mechanics to bouncing mechanics in striding birds that employ grounded running. Grounded running has also been observed across quadrupedal tetrapods: primitive mammals (opossum Monodelphis domestica, Parchman et al., 2003; rat Rattus norvegicus, pers. obs.), amphibians (frog Kassina maculata, Ahn et al., 2004) and lizards (Sceloporus magister, pers. obs.). Moreover, it is a natural component of locomotion in invertebrate polypeds (cockroaches Blaberus discoidalis and Periplaneta americana and ghost crab Ocypode quadrata; Blickhan and Full, 87; Full and Tu, 90, 91). In other words, grounded running is a common feature of terrestrial locomotion at intermediate to fast speeds. What predisposes some tetrapods to use grounded running? Once again, a comparison with the articial running gait of humans is instructive. Two key parameters in Groucho running are a high degree of limb compliance (bent knee running) and prolonged support durations (McMahon et al., 87; Bertram et al., 2002). The sprawled and crouched postures of small tetrapods may provide multiple links along the limb to enhance limb compliance. For example, birds that move with a more crouched posture (kiwi, guinea fowl, quail) rarely demonstrate an aerial phase when running after crossing the vaultbounce Fr boundary, rather they preferentially stay with a grounded run (Gatesy and Biewener, 91; Reilly, 2000). By comparison, the stretched limb cursorial birds (tinamous, rhea, ostrich), that presumably have greater limb stiffness, transition to an aerial run following a period of grounded running (Abourichid and Renous, 2000; Abourichid, 2001; Hancock et al., 2004). Foot length may be another factor determining grounded running: the relatively long feet of birds (Gatesy and Biewener, 91) as well as the plantigrade feet of many non-avian tetrapods may predispose these animals to grounded running by incurring relatively high support durations. This may explain why, for example, the desert spiny lizard Sceloporus magister, which has especially long hindfeet and toes, transitions from an aerial trot to a grounded trot with slower speeds (all the while using bouncing mechanics; A.R.B and S.M.R., pers. obs.).

UNIFYING LOCOMOTOR PERSPECTIVES

929

Fig. 2. Symmetrical gait space used by Monodelphis domestica when moving with bouncing mechanics (open symbol, grounded run; lled symbol, aerial run). Footfall patterns are indicated by gait diagrams (LH, left hindlimb; LF, left forelimb; RF, right forelimb; RH, right hindlimb). Slight asymmetries in contralateral footfall or support durations (indicated by ) may result in aerial phases. Modied from Parchman et al. (2003).

Gait choice is also a potent factor in determining whether or not a tetrapod uses grounded running. For example, while a crouched posture (and presumably higher limb compliance) may contribute to grounded trots in the short-tailed gray opossum M. domestica, a speed-related shift in limb phase may be equally important in explaining its other grounded runs (Fig. 2). Monodelphis tends to shift from a trot to a lateral-sequence diagonal couplets (4-beat trot) as speed decreases (Parchman et al., 2003). Support durations of each limb are more evenly distributed over the stride during a 4-beat gait than they are in a trot. Consequently, there is a range of duty factors at which the lateral-sequence diagonal couplets lack periods of suspension whereas the trots at these duty factors are already aerial. Finally, slight shifts in the timing of touchdown or liftoff can result in the inclusion or exclusion of an aerial phase during the stride (illustrated by in Fig. 2). Any discussion on grounded running would not be complete without highlighting the observation that some symmetrical gaits virtually never display periods of suspension even at their highest speeds and lowest duty factors. These are the singlefoot gaits, in which feet land sequentially approximately 0.25 limb phase after one another. Animals that use the lateral-sequence singlefoot across a large range of speeds (walking through running) include elephants (Hutchinson et al., 2003) and breeds of gaited horses (Hildebrand, 65). For example, the tolt, the lateral-sequence singlefoot gait characteristic of Icelandic horses, is used at speeds and duty factors that overlap with fast walking to cantering (albeit leaning towards

the lateral-sequence lateral couplets gait space, or 4-beat pace, at higher speeds; Zips et al., 2001). Although tolting Icelandic horses rarely obtain a suspension phase, all tolts for which whole body mechanics were recorded displayed in-phase uctuations Ep and Ek (Biknevicius et al., 2006), thus tolts are grounded runs. The lack of a period of suspension in lateral-sequence singlefoot gaits in either elephants or gaited horses is related neither to posture nor foot length because these are extremely erect and/or cursorial animals. Rather, an aerial phase in a pure lateral-sequence singlefoot gait is simply impossible unless duty factors o0.25, i.e., speeds at which singlefoot gaits have never been recorded. In summary, the exploration of the rst myth draws several conclusions about locomotor biodynamics. First, the myth is soundly debunked: both bouncing mechanics (in-phase Ep and Ek) and Hildebrand-dened running gaits (duty factor o0.5; singlefoots) have the capacity to occur without a period of suspension. The widespread representation of grounded running at intermediate speeds signies that grounded running is a common locomotor behavior in terrestrial tetrapods. Therefore, running does not require an aerial phase. Second, the Hildebrand denition of walks and runs based solely on hindlimb footfall patterns (duty factors o0.5) is inadequate for discriminating the mechanics of the COM. Indeed, an arbitrary designation of any other duty factor would similarly fail to distinguish vaulting and bouncing mechanics across all tetrapods (Fig. 1D). Therefore, we propose that Hildebrands walk run denitions be abandoned and that the
J. Exp. Zool. DOI 10.1002/jez.a

930

A.R. BIKNEVICIUS AND S.M. REILLY

terms walk and run explicitly imply whole body mechanics (vaulting and bouncing mechanics, respectively). This proposal creates some practical difculties because few labs are equipped to verify whole body mechanics, it is difcult to apply in eld situations and, thus far, it is only applicable to steady-state locomotion. How then can students of locomotion infer vaulting and bouncing mechanics when whole body mechanics data are impractical to capture? Fortunately, Froude number (Fr) may provide an acceptable and accessible alternative. The same equipment that is used to determine footfall patterns can be used to measure forward velocity and midstance hip height required for calculating Fr. Currently, walkrun boundaries determined empirically for bipeds and quadrupeds using Fr (Alexander and Jayes, 83) are better indicators of vaultbounce boundaries than of transitions between walking and running sensu Hildebrand based on the few studies that report both Fr and COM mechanics (e.g., Farley and Ko, 97; Rubenson et al., 2004). Unfortunately, the ability of Fr to serve as a truly robust measure for distinguishing vaulting (walking) and bouncing (running) mechanics across tetrapods awaits a broader survey of tetrapods.

Myth 2: A trot is always a run

Cavagna et al.s (77) seminal study gured the external mechanical energy uctuations of run 5 hop 5 trot trials separately from walk trials. Although run clearly refers to the bouncing mechanics of the striding bipeds (humans, birds) in their study, its equivalency to the hop of ricochetal mammals or trot of quadrupedal mammals was set in the minds of many comparative biomechanists. The labeling shorthand has led to the assumptions that all trots are necessarily runs (bouncing mechanics) and that trotting gait and vaulting mechanics are mutually exclusive. At rst glance, these assumptions appear reasonable. There is a historical bias toward studying a subset of trotters: large, primarily cursorial and often domesticated, mammals. These models cross the walkrun transition by switching between discrete gaits, from a non-trotting walking gait to a trot for running. Although the precise walking gaits of the quadrupeds were not identied by Cavagna et al. (77), inferences are possible by surveying studies that explicitly recorded footfall patterns in these or similar animals. Among the
J. Exp. Zool. DOI 10.1002/jez.a

many walking gaits reported for the domestic dog Canis familiaris (including lateral-sequence lateral couplets, singlefoot and diagonal couplets represented on the high duty factor half of the Hildebrand, 68 gait plot), they choose the lateralsequence lateral couplets (4-beat pace) gait when walking at steady speed (Grifn et al., 2004). Because steady speed is a requirement for whole body mechanics studies, it is likely that the Cavagna et al.s dogs similarly moved in this gait. Gait data for the ram Ovis musimon are less accessible but can be expected to parallel bovids and equids that walk using one of two lateralsequence gaits (singlefoot or lateral-couplet; Hildebrand, 76). Finally, both lateral-sequence singlefoots and diagonal-sequence diagonal couplets have been reported for walking adult macaques (Hildebrand, 67). In other words, the quadrupedal mammals in Cavagna et al. (77) typically use nontrotting gaits to walk using vaulting mechanics. However, for many terrestrial tetrapods, a trot represents a viable gait for walking (vaulting mechanics) and running (bouncing mechanics). Indeed, many tetrapods use trots and 4-beat trots across large range of speeds and duty factors (from less then 0.2 to greater than 0.8; Fig. 1A; Hildebrand, 76). While it is likely that trots occurring with duty factors o0.5 are all runs (bouncing mechanics), and it is conceivable that at least some the trots with duty factors of 0.50.6 are also operating under bouncing mechanics (grounded runs), under which whole body mechanics model do the remaining trots operate? Slow trots have been reported in frogs, turtles, lizards and crocodilians as well as in an array of mammals as different as hedgehogs and hippopotamus (Hildebrand, 76; White and Anderson, 94; Ahn et al., 2004; McElroy et al., 2004; Willey et al., 2004), but data on whole body mechanics of slow trots are limited to a few amphibians and reptiles (Fig. 1D; Farley and Ko, 97; Ahn et al., 2004; McElroy et al., 2004; Willey et al., 2004). In these trials, phase shifts between minima of Ek and Ep hover closer to 1801 than 01, so that these slow trots are inverted pendulum (walking) gaits. Thus, vaulting mechanics can and does occur with a trotting footfall pattern. That a walking trot exists (and functions as an inverted pendulum gait) should not be particularly surprising. Indeed, the walk of humans is akin to a bipedal trot, with contralateral upper and lower limbs swinging in unison. Both bipedal and quadrupedal trots are dynamically stable gaits in spite of their small base of support. However,

UNIFYING LOCOMOTOR PERSPECTIVES

931

while walking bipeds can maximally recover $7080% of their external mechanical energy through pendulum-like exchange of Ep and Ek (Cavagna et al., 77; Grifn and Kram, 2000), the maximum energy recoveries of walking trots range lower, from 51% in the western skink Eumeces skiltonianus (Farley and Ko, 97) to a mere 32% in the American alligator Alligator mississippiensis (Willey et al., 2004). Possible detrimental factors limiting the pendulum-like exchange of Ek and Ep in primitive tetrapods may include their lumbering locomotor behavior paired with a more sprawling posture, a high degree of asynchrony in the footfalls of the diagonal limbs, the use of non-optimal speeds and interference due to tail dragging via its effect on limb retraction dynamics (Willey et al., 2004; Reilly et al., 2005; Reilly et al., 2006). In summary, a trot is a footfall pattern observed in quadrupedal tetrapods in which diagonalcouplet limbs land more-or-less synchronously. Trot alone does not specify a movement pattern of the COM because trotting may occur with bouncing or vaulting mechanics. Confusion arising from equating trots with runs would dissolve if trot were reserved for describing a diagonalcouplet gait, as per its original sense for hundreds of years (Reilly and Biknevicius, 2003), without presumption of how the COM is moving. Thus, we propose that trot be reserved as a designation for a gait (a footfall pattern) with no a priori expectation of whole body mechanics patterns. REVISITING HILDEBRANDS SYMMETRICAL GAIT DEFINITIONS Hildebrands symmetrical gait denitions, based on limb phase, remain credible descriptors of footfall patterns over the range of observed speeds (reected by duty factor). Even more modern gait syntheses use limb phase as their foundation (e.g., Abourichid, 2003). Hildebrand envisioned symmetrical gaits as a continuum of limb phases along the ordinate of the symmetrical gait plot but he somewhat arbitrarily organized the named gaits around the octiles of limb phases (Fig. 1A). Designation of the four primary symmetrical gaits is largely uncontested, and the pace, lateral-sequence singlefoot, trot and diagonalsequence singlefoot are easily distinguished as the four fast gaits represented by the ngers on the symmetrical gait space. By contrast, the slow symmetrical gaits represent a progressive desynchronization in the touchdowns of the forelimb

relative to its ipsilateral hindlimb that lacks instantly recognizable boundaries. Thus, following Hildebrands (65) terminology, increasing limb phase shifts gaits seamlessly from the lateralsequence lateral couplets through the lateralsequence singlefoot to the lateral-sequence diagonal couplets to the trot and nally to the diagonal-sequence diagonal couplets (Fig. 1A). Clearly, some of the slow gaits are the same as those listed as fast (running) gaits (trot, lateral-sequence singlefoot). The remaining gaits represent the intermediate footfall patterns: lateral-sequence lateral couplets, lateral-sequence diagonal couplets, diagonal-sequence diagonal couplet. Just how real are these intermediate gaits? To answer this question, it helps to begin by examining the discrete gaits, beginning with the trot. How synchronously must contralateral limbs land for a gait to still qualify as a trot? Simultaneous footfalls of diagonal couplets are rare even among the premier trotters (e.g., dogs; Bertram et al., 97). Hildebrand (76) allowed for some variation in footfall timing when he dened the trot as the middle octile on his plot, so that 0.5 limb phase 76% (range: 0.440.56) are the traditional boundaries for the true trot. Flanking the traditional trot boundaries lay what Hildebrand described as non-trot walks (lateral-sequence diagonal couplets and diagonal-sequence diagonal couplets). These are often termed the 4-beat trot, trot-like walk or even dirty trot (Zips et al., 2001; Reilly and Biknevicius, 2003; Lammers and Biknevicius, 2004). These alternative labels include reference to the trot because the contralateral fore- and hindlimbs are landing more in sync than are the ipsilateral limbs, i.e., these gaits are essentially diagonal-couplet dominated as their formal Hildebrand titles indicate (Fig. 1D). Similarly, the lateral-sequence lateral couplets can be regarded as poorly synchronized version of the pace (4-beat pace, pace-like walk). Complete incorporation of the intermediate gaits into the trot and pace demonstrates a bias away from the, admittedly less common but still legitimate, singlefoot gaits. For example, the lateral-sequence lateral couplets and lateral-sequence diagonal couplets intermediate gaits could also be considered to represent poorly coordinated variants of the lateral-sequence singlefoot; a similar linkage may be drawn for the diagonal-sequence diagonal couplets and diagonal-sequence singlefoot. We recommend a more generous interpretation of the four primary gaits. We suggest that the limb
J. Exp. Zool. DOI 10.1002/jez.a

932

A.R. BIKNEVICIUS AND S.M. REILLY

phase axis of the symmetrical gait plot be subdivided into quartiles (rather than octiles; Fig. 1E), thereby subsuming the intermediate gaits into the four primary gaits. The effect of this consolidation of gait terms is two fold. First, it emphasizes that trotting (diagonal-couplet dominated behavior) is the most common gait among tetrapods and that most of what we know about whole body mechanics (both vaulting and bouncing) actually involves trots. Secondly, it draws our attention to the lateral-sequence singlefoot as a viable alternative to the trot for walking slowly. It becomes the gait of choice for domestic dogs moving at steady speed (0.15 limb phase) when optimizing pendulum-like recovery of external mechanical energy (Grifn et al., 2004). It is also the gait of choice when stability needs, and not energy economy, are to be maximized. Of the creeping gaits that retain at least three feet on the ground at all times, the most stable is the lateralsequence singlefoot (McGhee and Frank, 68). Not surprisingly, therefore, studies of postnatal gait development in placental carnivorans (cats Felis catus, Peters, 83; Howland et al., 95; dogs Canis domestica, Hildebrand, 68), rodents (jird Meriones tristrami and dormouse Eliomys meanurus; Eilam, 97) and even primates (baboon Papio cynocephalus and macaque Macaca mulatta; Shapiro and Raichlen, 2005) have reported lateral-sequence walking gaits in the youngest animals. Finally, confusion in describing symmetrical gaits (using either Hildebrands seven gaits or our proposed four gaits) can be avoided if workers specify limb phase rather than simply using previous gait terms. In summary, there is much to be gained by integrating the Hildebrand and Cavagna et al. perspectives on locomotion. First, walking and running should be recognized by their fundamentally different mechanics. Because duty factor fails to consistently distinguish these mechanics, its use in discriminating walks from runs should be abandoned in preference to phase difference in external mechanical energy or Froude number. Duty factor remains useful as a measure of the relative time allotted for reaccelerating and redirecting the COM (when the foot is in contact with the substrate). Furthermore, the yielding of limbs necessary for bouncing mechanics can occur with or without an aerial phase. Second, trots (and all other gait terms) should be reserved as descriptors of footfall patterns. Generalizing ndings on cursorial mammals to all tetrapods has led to the misconception that trots are only used
J. Exp. Zool. DOI 10.1002/jez.a

with bouncing mechanics. Gait terms lack explicit information about limb compliance and thus COM movements. Finally, symmetrical gait denitions should be broadened to reect the four primary footfall patterns: the lateral-couplet dominated pattern of the pace, the diagonal-couplet dominated pattern of the trot and the more independent sequencing of footfalls of the two singlefoots. Tetrapods consciously or subconsciously employ gaits that balance the stability demands of the terrain with acceptable levels of energy expenditure. Neuromuscular control determines both gait (footfall patterns interacting with the substrate) and limb compliance (capacity for pendulum-like or spring-like recovery of energy). Knowledge of only gait or only limb compliance inadequately summarizes locomotor effort. Decisions to employ a particular gait or whole body mechanics may reect other needs than simply energy conservation (e.g., intermittent locomotion associated with foraging for food; Reilly and Biknevicius, in press). ACKNOWLEDGMENTS We thank the International Society of Vertebrate Morphology for hosting the symposium Integrating approaches to the study of terrestrial locomotion at its Seventh Congress and symposium co-organizer Nancy Stevens. Special thanks go to Beth Brainerd (ICVM Program Ofcer and associated editor for this issue) and the Journal of Experimental Zoology for publishing the symposium. Eric McElroy kindly shared his lizard and salamander data for Fig. 1; Ohio University undergraduate thesis researchers Amy Back and Kristin Hickey collected the Rattus and Sceloporus data. Jennifer Hancock, Eric McElroy, Nancy Stevens and two anonymous reviewers provided valuable feedback. LITERATURE CITED
Abourichid A. 2001. Kinematic parameters of terrestrial locomotion in cursorial (ratites), swimming (ducks), and striding birds (quail and guinea fowl). Comp Biochem Physiol Part A 131:113119. Abourichid A. 2003. A new way of analysing symmetrical and asymmetrical gaits in quadrupeds. C R Biol 326:625630. Abourichid A, Renous S. 2000. Bipedal locomotion in ratites (Paleognatiform): examples of cursorial birds. Ibis 142: 538549. Ahn AN, Furrow E, Biewener AA. 2004. Walking and running in the red-legged running frog, Kassina maculata. J Exp Biol 207:399410.

UNIFYING LOCOMOTOR PERSPECTIVES Alexander RMcN. 1976. Mechanics of bipedal locomotion. In: Davies PS, editor. Perspectives in experimental biology, Vol. 1. Zoology. Oxford: Pergamon Press. p 493504. Alexander RMcN. 1977. Mechanics and scaling of terrestrial. In: Pedley TJ, editor. Scale effects in animal locomotion. New York: Academic Press. p 93110. Alexander RMcN. 1989. Optimization and gaits in the locomotion of vertebrates. Physiol Rev 69:11991227. Alexander RMcN, Jayes AS. 1983. A dynamic similarity hypothesis for the gaits of quadrupedal mammals. J Zool Lond 201:135152. Barnes WJP. 1975. Leg co-ordination during walking in the crab, Uca pugnax. J Comp Physiol 96:237256. Bertram JEA, Lee DV, Todhunter RJ, Foels WS, Williams AJ, Lust G. 1997. Multiple force platform analysis of the canine trot: a new approach to assessing basic characteristics of locomotion. Vet Comp Orthop Traumatol 10:160169. Bertram JEA, DAntonio P, Pardo J, Lee DV. 2002. Pace length effects in human walking: Groucho gaits revisited. J Motor Behav 34:309318. Biknevicius AR, Mullineaux DR, Clayton HM. 2006. Locomo tor mechanics of the tolt in Icelandic horses. Am J Vet Res 67:15051510. Blickhan R, Full RJ. 1987. Locomotion energetics of the ghost crab. II. Mechanics of the centre of mass during walking and running. J Exp Biol 130:155174. Cavagna GA, Heglund NC, Taylor CR. 1977. Mechanical work in terrestrial locomotion: two basic mechanisms for minimizing energy expenditure. Am J Physiol 233:243261. Demes B, Larson SG, Stern JT Jr, Jungers WL, Biknevicius AR, Schmitt D. 1994. The kinetics of primate quadrupedalism: hindlimb drive reconsidered. J Hum Evol 26: 353374. Eilam D. 1997. Postnatal development of body architecture and gait in several rodent species. J Exp Biol 200:13391350. Farley CT, Ko TC. 1997. Mechanics of locomotion in lizards. J Exp Biol 200:21772188. Farley CT, Taylor CR. 1991. A mechanical trigger for the trotgallop transition in horses. Science 253:306308. Full RJ, Tu MS. 1990. Mechanics of six-legged runners. J Exp Biol 148:129146. Full RJ, Tu MS. 1991. Mechanics of a rapid running insect: two-, four- and six-legged locomotion. J Exp Biol 156: 215231. Gatesy, SM, Biewener AA. 1991. Bipedal locomotion: effects of speed, size and limb posture in birds and humans. J Zool Lond 224:127147. Goiffon T, Vincent G. 1779. Memoire articeille des preincipes realtifs a la de le representation des animaux tant en peiture, quen sculpture: I. Partie concernant le chaval Alfort. Paris: Ecol Royale Veterinaire. Grifn TM, Kram R. 2000. Penguin waddling is not wasteful. Nature 408:929. Grifn TM, Main RP, Farley CT. 2004. Biomechanics of quadrupedal walking: how do four-legged animals achieve inverted pendulum-like movements? J Exp Biol 207: 35453558. Hancock JA, Biknevicius AR, Earls KD, Stevens NJ, Reilly SM. 2004. Groucho running in tinamous. J Morphol 260:297. Heglund NC, Cavagna GA, Taylor CR. 1982. Energetics and mechanics of terrestrial locomotion III. Energy changes of the centre of mass as a function of speed and body size in birds and mammals. J Exp Biol 79:4156.

933

Hildebrand M. 1965. Symmetrical gaits in horses. Science 150: 701708. Hildebrand M. 1967. Symmetrical gaits in primates. Am J Phys Anthropol 26:119130 Hildebrand M. 1968. Symmetrical gaits of dogs in relation to body build. J Morphol 124:353360. Hildebrand M. 1976. Analysis of tetrapod gaits: general considerations and symmetrical gaits. In: Herman RM, Grillner S, Stein PSG, Stuart DG, editors. Neural control of locomotion. New York: Plenum Press. p 203236. Howell AB. 1944. Speed in animals. Chicago: Univ of Chicago Press. Howland DR, Bregman BS, Goldberger ME. 1995. The development of quadrupedal locomotion in the kitten. Exp Neurobiol 135:93107. Hoyt DF, Taylor CR. 1981. Gait and the energetic of locomotion in horses. Nature 292:239240. Hutchinson JR, Famini D, Lair R, Kram R. 2003. Are fast-moving elephants really running? Nature 422: 493494. Kram R, Domingo A, Ferris DP. 1997. Effect of reduced gravity on the preferred walkrun transition speed. J Exp Biol 200:821826. Lammers AR, Biknevicius AR. 2004. The biodynamics of arboreal locomotion: the effects of substrate diameter on locomotor kinetics in the gray short-tailed opossum (Monodelphis domestica). J Exp Biol 207:43254336. McElroy EM, Biknevicius AR, Reilly SM. 2004. Mechanics of locomotion in lizards. J Morphol 260:311. McGhee RB, Frank AA. 1968. On the stability properties of quadrupedal creeping gaits. Math Biosci 3:331351. McMahon TA, Greene PR. 1979. The inuence of track compliance on running. J Biomech 12:893904. McMahon TA, Valiant G, Frederick EC. 1987. Groucho running. J Appl Physiol 62:23262337. Minetti AE, Ardigoo LP, Reinach E, Saibene F. 1999. The relationship between mechanical work and energy expenditure of locomotion in horses. J Exp Biol 202:23292338. Muir GD, Gosline JM, Steeves JD. 1996. Ontogeny of bipedal locomotion: walking and running in the chick. J Physiol 493:589601. Muybridge E. 1887. Animal locomotion. Phildelphia: University of Pennsylvania Press. Parchman AJ, Reilly SM, Biknevicius AR. 2003. Whole-body mechanics and gaits in the gray short-tailed opossum Monodelphis domestica: integrating patterns of locomotion in a semi-erect mammal. J Exp Biol 206:13791388. Peters SE. 1983. Postnatal development of gait behavior and functional allometry in the domestic cat (Felis catus). J Zool Lond 199:461486. Raynor AJ, Yi CJ, Abernethy B, Jong QJ. 2002. Are transitions in human gait determined by mechanical, kinetic or energetic factors? Hum Movement Sci 21: 785805. Reilly SM. 2000. Locomotion in the quail (Coturnix japonica): the kinematics of walking and increasing speed. J Morphol 243:173185. Reilly SM, Biknevicius AR. 2003. Integrating kinetic and kinematic approaches to the analysis of terrestrial locomotion. In: Bels VL, Gasc JP, Casinos A, editors. Vertebrate biomechanics and evolution. Oxford: Bios Scientic Publishers Ltd. p 243265. Reilly SM, Willey JS, Biknevicius AR, Blob RW. 2005. Hindlimb function in the alligator: integrating movements,

J. Exp. Zool. DOI 10.1002/jez.a

934

A.R. BIKNEVICIUS AND S.M. REILLY Shapiro LJ, Raichlen DA. 2005. Lateral sequence walking in infant Papio cynocephalus: implications for the evolution of diagonal sequence walking in primates. Am J Phys Anthropol 126:205213. Stevens NJ. 2006. Stability, compliance and limb coordination: the ecological relevance of gait. J Exp Zool (ICVM symposium issue). White TD, Anderson RA. 1994. Locomotor patterns and costs as related to body size and form in teiid lizards. J Zool Lond 233:107128. Willey JS, Biknevicius AR, Reilly SM, Earls KD. 2004. The tale of the tail: Limb function and locomotor mechanics in Alligator mississippiensis. J Exp Biol 207:553563. Zips S, Peham C, Scheidl M, Licka T, Girtler D. 2001. Motion pattern of the toelt of Icelandic horses at different speeds. Equine Vet J Suppl 33:109111.

motor patterns, ground reaction forces and bone strain of terrestrial locomotion. J Exp Biol 208:9931009. Reilly SM, McElroy EJ, Odum RA, Hornyak VA. 2006. Tuataras and salamanders show that walking and running mechanics are ancient features of tetrapod locomotion. Proc R Soc B 273:15631568. Riegel RJ, Hakola SE. 1999. Illustrate atlas of clinical equine anatomy and common disorders of the horse, Vol. 1. Marysville, OH: Equistar Publications, Limited. Rogers JL. 2000. USA track & eld coaching manual. Champaign, IL: Human Kinetics. Rubenson J, Heliams DB, Lloyd DG, Fournier PA. 2004. Gait selection in the ostrich: mechanical and metabolic characteristics of walking and running with and without an aerial phase. Proc R Soc Lond B 271: 10911099.

J. Exp. Zool. DOI 10.1002/jez.a