© 2002 by The Society for Integrative and Comparative Biology
Kinematics, Dynamics, and Energetics of Rowing and Flapping Propulsion in Fishes1
1 Department of Biological Sciences, University of Southern Maine, 96 Falmouth St., Portland, Maine 04103
2 Department of Zoology, Field Museum of Natural History, 1400 S. Lakeshore Dr., Chicago, Illinois 60605
 |
SYNOPSIS |
|---|
TOP
 SYNOPSIS INTRODUCTIONKINEMATICSDYNAMICSPERFORMANCE TRADE-OFFSENERGETICSEVOLUTIONARY AND ECOLOGICAL...References |
|---|
The shape and motion of the pectoral fins vary considerably among fishes that swim in the labriform mode. Pectoral fin motion in fishes is highly variable, but one conspicuous axis of this variation is the rowing-flapping axis. At one extreme of this axis, paddle-shaped fins row back and forth in a plane that is parallel to fish motion, while at the other extreme, wing-shaped fins flap up and down in a plane that is perpendicular to fish motion. We have used two fish, the threespine stickleback (Gasterosteus aculeatus) and the bird wrasse (Gomphosus varius), that fall near the extremes of the rowing-flapping axis to study the dynamic, energetic, and ecological and evolutionary consequences of this kinematic variation. Our work confirms some traditionally held assumptions about rowing and flapping dynamics and energetics but reject others. A computer simulation experiment of virtual rowing and flapping appendages makes several predictions about differences in maneuvering performance and swimming energetics between rowing and flapping, which, in turn, make predictions about the behavior and ecological distribution of fishes that vary along the rowing-flapping axis. Both laboratory and field studies of labrid swimming ability and distribution support these predictions.
|
INTRODUCTION |
|---|
|
TOP
SYNOPSISINTRODUCTIONKINEMATICSDYNAMICSPERFORMANCE TRADE-OFFSENERGETICSEVOLUTIONARY AND ECOLOGICAL...References |
|---|
A conspicuous feature of fish locomotion is the wide diversity of both propulsors and the way these propulsors move and interact with the water to generate propulsive forces (Lighthill, 1969
; Webb, 1982, 1984, 1993). What are the dynamic and energetic differences between these different swimming modes and how do these differences constrain fish ecology (e.g., distribution of species in a river or on a reef) and evolution (e.g., trade-offs between optimizing locomotor and reproductive morphology)?
Much of our collaborative work on the kinematics, dynamics, and energetics of fish swimming has concentrated on variation in pectoral fin, or labriform, propulsion. Most fishes swim with the pectoral fins to maneuver at slow speeds but fishes from several families rely on labriform propulsion for sustained cruising. The kinematics of the pectoral fins in these fishes with sustained labriform propulsion varies tremendously within and among individuals, populations, and species. Much of this variation can be summarized by two variables: stroke plane angle and number of propulsive waves. The number of propulsive waves defines an "oscillation-undulation" continuum and the kinematics, dynamics, and energetics of variation along this axis has been addressed elsewhere (Daniel, 1988; Lighthill and Blake, 1990; Combes and Daniel, 2001; Rosenberger, 2001).
Variation in stroke plane angle defines a "rowing-flapping" axis, which is the subject of this review. Horizontal strokes that are largely parallel with the direction of travel are typically called "rowing" while vertical strokes that are largely normal to the direction of travel are "flapping." It is important to emphasize two points. First, rowing and flapping are idealized extremes of a continuum; real pectoral fin strokes lie somewhere along this continuum (Lauder and Jayne, 1996). Second, the fin strokes of many fishes that use the pectoral fins at slow speeds only do not fall neatly along the rowing-flapping axis (Lauder and Jayne, 1996). But among fishes capable of sustained, labriform propulsion at relatively high speeds, the rowing-flapping axis of variation is quite conspicuous. If we extend beyond the fishes, the rowing-flapping axis is even more striking. Indeed, rowing and flapping occur in diverse taxonomic groups, including at least three different phyla, and have independently evolved numerous times (Fish, 1992, 1993, 1996; Walker and Westneat, 2000; Walker, 2002). Morphologies associated with rowing and flapping illustrate many, beautiful evolutionary convergences (Walker, 2002). Distally tapering, wing-shaped geometries are characteristic of flapping appendages in gastropods, cephalopods, insects, holocephalians, teleosts, and tetrapods. Distally expanded or paddle shaped geometries characteristic of rowing appendages are found in crustaceans, insects, teleosts, and tetrapods. In large animals, paddle-shaped appendages are formed from a continuous membrane while in small animals, the paddle-shape is often formed by hairs radiating from a central shaft (Koehl, 1993; Vogel, 1994; Walker, 2002). Despite this tremendous diversity and evolutionary convergence, the rowing-flapping axis is a largely, if not entirely, aquatic phenomenon and is rarely discussed in the flight literature. There has been some suggestion that the smallest flying insects swim, or row, in air (Thompson, 1917; Horridge, 1956), but the geometry of this rowing should be quite distinct from that occurring in aquatic animals. Stroke plane angle does vary both within individuals and among species in flying insects (Dudley, 2000) but this variation is only superficially related to the rowing-flapping axis found among aquatic animals.
In the following pages, we describe the kinematics of rowing and flapping and how a simple, rowing-flapping model predicts dynamic, energetic and behavioral consequences of the kinematic variation. We focus our description on the rowing stroke of the threespine stickleback, Gasterosteus aculeatus, and the flapping stroke of the bird wrasse, Gomphosus varius. Threespine sticklebacks have several different life history variants; the fish that we have studied are anadromous individuals that are pelagic marine fish that make annual migrations into freshwater streams to breed. The bird wrasse is an active and conspicuous fish found on coral reefs throughout much of the western Pacific and Indian Oceans. Both sticklebacks and the bird wrasse excel at labriform propulsion. Indeed, slow, oxidative fibers are absent from the myotome of both species (te Kronnie et al., 1983; Davison, 1988, 1994) and, consequently, sustained swimming in both species is powered solely by the pectoral fin muscles. Examining the kinematics of fin motion in fishes that maintain labriform propulsion at relatively high swimming speeds is important because no fish presents a stereotypical rowing or flapping motion at low speeds (Lauder and Jayne, 1996). Indeed, some fishes that use a typical rowing stroke at cruising speeds, such as boxfishes (Hove et al., 2001) and sticklebacks (Walker, unpublished data), exploit a more undulatory stroke at very slow or hovering speeds.
|
KINEMATICS |
|---|
|
TOP
SYNOPSISINTRODUCTIONKINEMATICSDYNAMICSPERFORMANCE TRADE-OFFSENERGETICSEVOLUTIONARY AND ECOLOGICAL...References |
|---|
Continued improvement in the resolution of high speed digital video has greatly facilitated our ability to describe the 3-D kinematics of oscillating pectoral fins (Gibb et al., 1994
; Lauder and Jayne, 1996; Walker and Westneat, 1997) and novel methods (Song et al., 2001) have yet to be explored. Despite these technical advances that will ultimately allow estimates of the complete deformation of the fin surface throughout the fin stroke cycle, we want to focus on a more simple kinematic model, Vogel's virtual model of rowing and flapping "appendages" (Vogel, 1994), which provides a useful, heuristic introduction to the major differences between rowing and flapping. Both the rowing and flapping appendage are long, rectangular plates that heave (move back and forth along some axis) and pitch (rotate around a spanwise axis) (Fig. 1). The rowing plate heaves back and forth along an anteroposterior axis and has a feathered orientation (parallel to the flow) during the forward stroke and a broadside orientation (normal to the flow) during the backward stroke (Fig. 1). The force on the plate during the foreward stroke, which is due to viscous stresses, is oriented posteriorly and small in magnitude. By contrast, the force during the backward stroke, which is due to the large separation of the flow from the plate's surface, is oriented anteriorly and large in magnitude. This virtual rowing plate, then, has a distinct recovery stroke in which the resultant force is minimized and a power stroke in which the resultant force is maximized.
|
A rowing fin that oscillates about its root requires slightly different kinematics from the simple heaving and pitching plate. First, a rotation into a feathered orientation during the recovery stroke requires either the trailing edge to elevate, the leading edge to depress, or some combination of both. Given that fish rotate the fin clockwise to begin the power stroke (in left lateral view), simple geometry constrains the leading edge to depress during the recovery stroke. Consequently, the stroke plane angle, which is the angle of the path swept out by the tip of the leading edge relative to the horizontal (Fig. 1), cannot be zero in the root-oscillating fin of a fish. During the clockwise rotation at the start of the power stroke, one would expect the fin to rotate about its trailing edge. At the beginning of the power stroke in G. aculueatus, the fin does not rotate about the trailing edge as a unit, but instead, each ray rotates into a broadside orientation in sequence from leading to trailing edge (this resembles peeling a sticker off of a table top).
Second, in an ideal fin with a fixed, dorsoventrally oriented root, the fin must twist down its span during the recovery stroke in order to feather. Only a small section of the fin can be optimally feathered (perfectly parallel with the local flow), with more proximal sections positively pitched (leading edge up relative to flow) and distal sections negatively pitched. The distal segment of a rowing fin model (Walker and Westneat, 2000) had an optimal maximum pitch during the recovery stroke that ranged between –102° and –114° (i.e., pitched 12° to 24° below the horizontal). In real rowing appendages, feathering is achieved by spanwise twisting in fishes, by rotation of the distal segment about a joint with a proximal segment in mammals, turtles, and beetles, and by the passive flexion of the paddle in mammals, birds, beetles, and many small aquatic invertebrates. G. aculeatus further reduces recovery stroke loading by abducting the fin off the body one ray at a time. As the leading edge rays peel off, the trailing edge rays rotate dorsally along the body so that the ray next in line to peel off is at the dorsoventral level of the more leading edge rays.
Vogel's flapping plate heaves up and down along a dorsoventral axis and pitches nose down during the down stroke and nose up during the up stroke. The resultant force on both strokes is due to bound circulation and flow separation (if the angle relative to the oncoming flow is large enough); it is pointed forward and up on the down stroke and forward and down on the up stroke. The flapping plate generates thrust on both strokes and, unlike the rowing fin, generates and alternating cycle of large, up and down force components.
Unlike the heaving and pitching flapping plate, a root-oscillating flapping fin necessarily experiences variation in flow velocities along its span, with the oscillation component of the resultant velocity becoming more dominant distally. The consequence of this spanwise velocity variation is that the optimal pitch will increase down the span and the fin must twist spanwise, which pitches distal segments more than proximal segments, in order to achieve this optimum. This spanwise twisting is a common feature of the flapping stroke among many aquatic animals other than fishes, including pterapods, finned cephalopods, and sea turtles (Davenport et al., 1984; Satterlie et al., 1985; Davenport, 1987; Renous and Bels, 1993; Vecchione and Young, 1997; Seibel et al., 1998). Although spanwise twisting occurs in both rowing and flapping, the geometry is quite different: an idealized rowing fin during the recovery stroke is characterized by 
m = 90° at the fin root and m < 0° at the distal edge (the value will depend on the fin shape) while the idealized flapping fin during mid down stroke is characterized by m = 0° at the fin root and m << 0° at the distal edge.
How do the kinematics of G. aculeatus and G. varius compare with values expected from the models? The mean stroke plane angle of G. aculeatus is 31.4°. Again, given the clockwise rotation (in left lateral view) of the pectoral fin at the beginning of the power stroke in G. aculeatus, the leading edge must depress during the recovery stroke to allow it to elevate during the power stroke rotation. Were the leading edge depression equal to the trailing edge elevation, so that the fin extended straight laterally at maximum abduction (i.e., no dihedral), the necessary stroke plane angle would be 25.7°. The observed stroke plane angle of G. aculeatus exceeds this value because the distal edge is pitched nose down at maximum abduction, which has the effect of further depressing the leading edge and increasing the stroke plane angle.
The mean stroke plane angle from the horizontal for G. varius is 70.6° (Walker and Westneat, 1997), while the flapping model predicts a stroke plane of 90°. As discussed frequently in the flight literature, a stroke plane angle of less than 90° is expected of flapping flight in negatively buoyant animals because this is one mechanism to generate a net upward force over the stroke cycle (Lighthill, 1975). Inclined stroke planes of less than 90° are the norm in flying insects, birds, and bats. Similarly, the inclined stroke plane of G. varius is expected given that its mass-equivalent weight in water is 0.8 g to 1.3 g (Walker and Westneat, 1997), which is about the same weight of a large insect, such as Manduca sexta, in air.
The morphological angles of attack, m, of the distal fin segment, which is the angle of the distal fin chord relative to the frontal plane (or floor of the flow tank), are compared between G. aculeatus and G. varius in Figure 2. In G. aculeatus, m rapidly falls from 90° to 10° at the beginning of the recovery stroke, gradually drops to –20° at the end of the recovery stroke, and rapidly climbs to 85° at the beginning of the power stroke. The peak negative pitch during the recovery stroke is within the range of the expected value of that predicted by the virtual fin simulation experiment (Walker and Westneat, 2000). In G. varius, m rapidly drops to about –10° during the down stroke and maintains this level until the beginning of the up stroke. About halfway through the up stroke, m reaches its peak of 60°. As predicted by the simple rowing-flapping model, the up stroke m is substantially less than the power stroke m of G. aculeatus. Nevertheless, an asymmetry between down stroke and up stroke m differs from the symmetric angles predicted by the flapping model. Again, this difference from the rowing-flapping model reflects negative buoyancy of G. varius and the necessity of generating a net upward force during the stroke cycle. Because resultant forces are largely normal to the plane of the fin, the resultant force during the down stroke will have a large upward component while that during the up stroke will have only a small downward component.
|
The variation in stroke plane angle and
m between the rowing stroke of G. aculeatus and G. varius has important consequences for the hydrodynamic angle of attack, h, between the local flow and the fin surface (Fig. 2). For the rowing fin of G. aculeatus, h of the distal segment begins the recovery stroke with a very high positive value but rapidly declines to a broad plateau around 20°. Near the end of the recovery stroke, h rapidly drops to about –80°, which is maintained until a rapid rise to high positive angles at the end of the power stroke. The moderately positive h during the recovery stroke in G. aculeatus differs from the rowing model, where values of 0° are expected, and reflects the need to generate a small upward force to balance its negative buoyancy. The high h throughout the power stroke is consistent with the expectations of the rowing model.
In the flapping stroke of G. varius, h is moderately positive (+40° to +50°) at the beginning of the down stroke, steadily declines and becomes negative at about 70% through the down stroke, reaches a negative peak of about –35° about 20% through the up stroke, and steadily increases to its starting, positive value at the end of the up stroke (Fig. 2). The moderately positive values during the down stroke and moderately negative values during the up stroke are consistent with the expectations of the flapping model.
|
DYNAMICS |
|---|
|
TOP
SYNOPSISINTRODUCTIONKINEMATICSDYNAMICSPERFORMANCE TRADE-OFFSENERGETICSEVOLUTIONARY AND ECOLOGICAL...References |
|---|
Recent advances in numerical (Liu, 2002
) and physical (Dickinson et al., 1999) modeling and wake visualization (Gharib et al., 2002; Drucker and Lauder, 2002) allow a highly detailed investigation of the unsteady fluid dynamics of rowing and flapping pectoral fins. One problem in want of a solution, however, is a tool for measuring the real-time force balance on a fish swimming with its pectoral fins. Our solution to this problem is to use the high speed digital camera as a force transducer (see below). We have previously used this method to infer the dynamics of flapping flight in G. varius (Walker and Westneat, 1997; Ramamurti et al., 2002) and we use it here to compare the measured force balance with the expected force balance of idealized rowing and flapping fins. Rowing and flapping are often referred to as drag-based and lift-based propulsion, respectively. This language is a good approximation for the heuristic model of heaving and pitching plates but numerous mechanisms, including viscous stresses, bound circulation, flow separation, fin-vortex interactions, fin inertia, added mass (acceleration reaction) forces, and jet-like (squeeze) forces, potentially contribute to the force balance in either rowing or flapping fins. How accurate, then, is the concept of drag-based rowing and lift-based flapping?
The flow over a flat plate (or any airfoil) at attack angles above the stall angle (>
15°) separates as it passes around the edge, creating a region of low pressure on the downstream surface (Dickinson and Götz, 1993). At very high angles, the separation alternates from both leading and trailing edge (as in the classical example of a sphere in a steady flow) but within some intermediate range of attack angles, separation is limited to the leading edge. The low pressure that results from the separation causes the formation of an attached vortex on the downstream surface as fluid passing over the airfoil reverses direction and is sucked back toward its surface (Dickinson and Götz, 1993).
While the separation vortex remains attached to the plate, two different mechanisms contribute to the hydrodynamic force. First, pressure differences on either side of the plate at some finite angle to the freestream distorts the flow in a way that creates the appearance of a bound circulation superimposed over the freestream (Ellington, 1984). The difference in pressure between sides of the plate creates a circulatory force on the plate; at small attack angles this force is nearly parallel with lift (the force component normal to the local stream), hence this force is usually called circulatory lift. At higher attack angles, this circulatory force simply has a smaller lift component and a larger drag component.
The low pressure in the region of flow separation that creates the attached vortex is the second source of a pressure asymmetry and results in an attached-vortex force, the lift component ("attached-vortex lift") of which has been discussed in the aerodynamic literature. This attached vortex force is the source of pressure drag on a bluff body, such as a sphere or a flat plate normal to a steady flow. At attack angles less than 90°, the attached vortex force has a lift in addition to drag component. The attached vortex force effectively augments the circulatory force (Dickinson and Götz, 1993; Dickinson, 1996) and we refer to the combined force as the vortex-and-bound-circulation force.
Rowing has been classified as "drag-based" because of the resemblance of the power stroke fin and a flat plate oriented normal to a steady flow while flapping has been classified as"lift-based" flapping because of the fin's exploitation of circulatory lift that results from the distorted boundary layer. Clearly, however, the moderate to large h in both G. aculeatus and G. varius suggest that forces resulting from bound circulation ("lift") and flow separation ("pressure drag") occur in both rowing and flapping fins. Indeed, recent computational fluid dynamic models of a robotic rowing fin (Liu and Kato, 2002) and of the flapping fin of G. varius (Ramamurti et al., 2002) show clear regions of flow separation and vortex formation while recent DPIV analysis of the wakes of oscillating pectoral fins demonstrate the repeated cycling of vortex formation and shedding (Drucker and Lauder, 1999, 2000). The drag vs. lift dichotomy, then, while useful as an elementary model, both confounds and obscures the different sources of forces on real oscillating fins, as emphasized by Dickinson (1996). Nevertheless, while a detailed, quantitative comparison between the proportion of thrust derived from drag (the force component parallel to the local flow) and lift (the force component normal to the local flow) might, therefore, seem an exercise in triviality, we suspect there is a relationship between these proportions and fin (or wing) shape.
How well do the dynamics of G. aculeatus and G. varius match the expected dynamics of fish propelled by rowing and flapping fins and what is the influence of the combined vortex-and-bound-circulation force relative to the added mass force and squeeze force? To address these questions, we compared the fin kinematics and instantaneous force balance on each fish throughout the stroke cycle. The net force on a swimming fish can be estimated by measuring the acceleration, dU/dT, of the center of mass, and using the equations
|
|
where T is thrust, L is lift, M is the mass of the system (fish and added mass), D is the drag on the body and fins, and W is the weight of the animal in water. We measured the acceleration of G. aculeatus and G. varius (Walker and Westneat, 1997) by digitizing a point near the center of mass and twice differentiating the x and y coordinates with respect to time using a quintic spline algorithm (Walker, 1998). We measured W using a spring balance and used Hoerner (1965) to estimate D for a body of revolution with the same length, depth, and breadth as the fish. We assume that fin oscillations are the only sources of T and L in the force balance. In order to compare T and L among sequences, we standardized the forces by 
SU2/2, where is the density of the water, S is the surface area of both fins and U is the mean forward speed of the fish.
The major features of the standardized thrust (CT) and lift (CL) curves for both G. aculeatus and G. varius are very similar to those expected from the simple rowing and flapping model in which vortex-circulation forces dominate the force balance (Fig. 3). As predicted by the rowing model, the G. aculeatus CT curve is characterized by a shallow, broad, largely negative plateau during the recovery phase and a very large thrust peak during the power stroke. In contrast to the rowing model, G. aculeatus also presented small positive and negative lift peaks during the recovery and power strokes, respectively. Note, however, that the magnitude of the lift peaks are substantially less than the magnitude of the thrust peak.
|
The occurrence of two positive thrust peaks in G. varius (Fig. 3) is consistent with the flapping model although the asymmetry in the magnitude of the peaks differs from the model. The larger thrust peak during the up stroke reflects the high
m, which produces resultant forces that have a high thrust component. The positive lift during the down stroke and negative lift during the up stroke and the similar magnitudes between the lift and thrust peaks are consistent with the flapping model.
The vortex-and-bound-circulation force should peak when resultant velocities are highest, which should occur halfway through the stroke if the fin is oscillating with simple harmonic motion and if the rotational component to the incident flow is small. The observation that the lift and thrust peaks in both G. aculeatus and G. varius occur about halfway through the strokes supports the hypothesis that the vortex-and-bound-circulation force is an important component in the force balance. In an influential model of rowing propulsion, Blake (Blake, 1979) modeled the acceleration reaction but argued that it contributed nothing to the thrust balance in the power stroke of the angelfish, Pterophylum eimekei, because the negative contribution of the acceleration reaction force in the last half of the stroke canceled the positive contribution in the first half. While it is true that the impulse of the acceleration reaction force should be zero over the power stroke (assuming that m does not vary), the thrust component should be positive (Nachtigall, 1960; Daniel, 1984). Nevertheless, while the acceleration reaction can potentially dominate the force balance of a rowing fin (Daniel, 1984), its influence in the rowing fin of fish is compromised because its feathered orientation at the point of peak negative acceleration (at the end of the recovery stroke) suggests that only a small volume of water will be accelerated. Indeed, the thrust curve of G. aculeatus shows that the acceleration reaction component at the beginning of the power stroke is near zero or is being cancelled by some other component.
The rotational component of an airfoil's kinematics will contribute a rotational component to the circulation resulting in dynamic behavior that resembles (but is not due to) the Magnus force on a rotating sphere or cylinder (Ellington, 1984; Fung, 1993; Dickinson et al., 1999). This rotational force component is an important part of the force balance in Drosophila melanogaster because this animal has very rapid wing rotations at each stroke reversal (Dickinson et al., 1999). Fin rotation in G. varius is largely continuous throughout the stroke cycle but peaks about 25% of the way through the down stroke and up stroke. Peak lift occurs from about 40% to 50% of the way through the down stroke while peak thrust occurs about 50% of the way through the up stroke; both observations are not consistent with a dominant rotational lift mechanism. The rapid change in m at the beginning of the recovery and power strokes in G. aculeatus is not due to the whole fin rotating but to the leading edge peeling off of the body during abduction and peeling up from its feathered orientation during adduction. Not surprisingly, the thrust curve for G. aculeatus shows little evidence of a large rotational force component near the beginning of the power stroke.
Finally, several authors have suggested that a fin closing against the body may result in thrust due to the acceleration of a jet of fluid behind the fish (Blake, 1979; Geerlink, 1983; Daniel and Meyhöfer, 1989). We would expect the squeeze force to peak near the end of fin adduction. The negative thrust on the body at this point of the stroke cycles of both G. aculeatus and G. varius suggests that this squeeze force is not an important component of the force balance.
|
PERFORMANCE TRADE-OFFS |
|---|
|
TOP
SYNOPSISINTRODUCTIONKINEMATICSDYNAMICSPERFORMANCE TRADE-OFFSENERGETICSEVOLUTIONARY AND ECOLOGICAL...References |
|---|
The variation in swimming styles both within individuals and among fishes begs a series of questions. Are there performance differences between swimming modes? What are the trade-offs between energy minimization and other measures of swimming performance, such as mean thrust per stroke cycle? And, can energetic differences and performance trade-offs explain variation in swimming style within and among individuals and species? We explored performance trade-offs between thrust maximization and energy minimization using a computer simulation experiment (Walker and Westneat, 2000
). In this experiment, we modeled the dynamics of rowing and flapping fins using blade-element theory, which essentially divides the fin up into a series of elements along the span and computes a force for each element as a function of its velocity, acceleration, and angle of attack relative to the flow. Features of real fins that confound comparisons of rowing and flapping performance, such as amplitude and fin size and shape, were held constant. Two types of rowing fins were modeled. In the untwisted model, the fin rotated at its base in order to feather during the recovery stroke while in the twisted fin, the fin twisted along its span in order to feather. The untwisted model resembles the distal segments of animals that row with multi-segment appendages, such as freshwater turtles while the twisted model resembles the unsegmented fin of fishes.
Two important results emerged from the simulation. First, a flapping fin is more mechanically efficient than a rowing fin across the entire range of biologically relevant swimming speeds. This result is discussed in the next section. Second, a rowing fin generates more thrust per stroke than a flapping fin at low swimming speeds but this trend is reversed at higher swimming speeds (Vogel, 1994). We found that the difference in thrust generated between rowing and flapping is especially large if only a half-stroke is considered. This result suggests that a rowing stroke is the preferred motion for maneuvering behaviors that require large fore-aft forces, such as starting, stopping, and lateral (yaw) turning.
Do fishes that row maneuver better than fishes that flap? Unfortunately we have very little quantitative data to compare maneuvering performance among rowers and flappers. Two important components of yaw maneuvering are turning radius (minimizing this radius allows greater precision in positional and directional control) and turning rate (or agility). The boxfishes Ostracion meleagris and O. cubicus actuate yaw turns by rowing one pectoral fin while reverse rowing (or braking) the other (Walker, 2000). While boxfishes have the ability to complete 360° (and more) turns with turning radii of effectively zero, these turns are relatively slow (< 30°/sec/total length) (Walker, 2000). In comparison, two similarly-sized fishes that present more of a flapping stroke, the bluehead wrasse (Thalassoma bifasciatum) and the ocean surgeonfish (Acanthurus bahianus) do not perform comparable, "hovering" turns in which all of the power for the turn is generated by oscillating pectoral fins (Gerstner, 1999). Instead, these fishes use the kinetic energy of forward swimming in combination with pectoral fin oscillation to power the turn. Relative to the hovering turns of the boxfish, these "cruising" turns are much faster (> 100°/sec/total length) but with larger radii (> 0.05 total lengths) (Gerstner, 1999).
Rowing fins may also allow fish to hover in still water better than flapping fins. Boxfishes and threespine sticklebacks hover very well by oscillating their pectoral fins with large attack angles on both recovery and power strokes. We have found that the bird wrasse, which flaps its fins, hovers by pitching its body nose-up while oscillating its pectoral fins back-and-forth. Even in this position, a bird wrasse can only hover for a few strokes. Korsmeyer et al. (2002) found that a parrotfish, Scarus schlegeli, which uses a flapping stroke, cannot hover. Finally, Gerstner (1999) found that the flappers, T. bifasciatum and A. bahianus, do not perform hovering (non-translocating) turning maneuvers (see above).
|
ENERGETICS |
|---|
|
TOP
SYNOPSISINTRODUCTIONKINEMATICSDYNAMICSPERFORMANCE TRADE-OFFSENERGETICSEVOLUTIONARY AND ECOLOGICAL...References |
|---|
Rowing vs. flapping
One method of comparing the energetic performance differences associated with rowing and flapping is to compare the mechanical efficiency of rowing and flapping fins, where mechanical efficiency is the ratio of the useful (that which contributes to thrust) to total work done by the oscillating fins on the water. As described above, Vogel (1994)
found that rowing generates more thrust at low speeds. This result has subsequently been interpreted as implying that rowing is more mechanically efficient than flapping at low speeds. Our simulation results (Walker and Westneat, 2000) do not support this interpretation. Instead, the simulated flapping fin had higher mechanical efficiency than the simulated rowing fin at all speeds. The twisted rowing fin was especially inefficient (mechanical efficiency < 0.05) because the fin necessarily generated large drag on the recovery stroke. A stickleback avoids this large drag cost by "peeling" its fin rays off the body as described above. Nevertheless, even the untwisted rowing fin had lower efficiency than the flapping fin. At very slow swimming speeds (<1 L/sec), the efficiencies of the untwisted rowing fin and the flapping fin are sufficiently close to suggest that an optimized rowing fin (say with a more optimal fin shape) could be more efficient than an optimized flapping fin. These virtual model results have subsequently been supported by results on a physical model of a stiff pectoral fin driven by three motors (Kato and Liu, 2003).
The simulation predicts that fishes that flap pectoral fins should be able to maintain higher pectoral-fin-powered swim speeds than fishes that row. One way to test this hypothesis is to compare Ucrit between rowers and flappers. In theory, Ucrit is the minimum speed at which the fish fatigues and should equal the upper bound for the maximum sustainable swimming speed. In practice, measures of Ucrit are dependent on the time interval and velocity increment used to measure the performance but above time intervals of about 15–20 minutes and below velocity increments of about 14Ucrit, there is little variation in measured performance (Hammer, 1995). Although typically measured on fishes swimming by axial undulation, Ucrit is an appropriate measure of pectoral-fin powered swimming ability in Gasterosteus aculeatus and Gomphosus varius because these fishes power their swimming with only the pectoral fins until fatigue velocities (Taylor and McPhail, 1985; Whoriskey and Wootton, 1987; Walker, 1999; Walker and Westneat, 2002).
The Ucrit of G. aculeatus (about 6–8 cm total length) is about 5 lengths/sec (Taylor and McPhail, 1985; Whoriskey and Wootton, 1987) while that of G. varius (12–17 cm total length) is about 5.2 lengths/sec (Walker and Westneat, 2002). While comparisons of swimming speeds are often given in lengths/s, such comparisons are invalid because length-standardized Ucrit is itself a function of body length (e.g., smaller fish have higher Ucrit/length) (Hammer, 1995). A recent review (Hammer, 1995) shows that maximum swimming speeds, such as Ucrit, scale with an exponent of 0.5 (assuming swimming speed scales with muscle power/drag, this exponent suggests that either drag scales with an exponent greater than 2 and/or power scales with an exponent less than 3). Standardizing Ucrit by the square root of length, then, should remove the effect of length. Using this standardization, the mean Ucrit of G. aculeatus is 11.8 root lengths/sec while that of G. varius is 17.9 root length/sec, a difference of about 50%.
While the performance differences between G. aculeatus and G. varius are consistent with the rowing-flapping model, comparisons of Ucrit from different experiments should be interpreted cautiously because many different features of the experimental design of critical swimming speed tests can potentially influence the measured value of Ucrit (Hammer, 1995). To test the hypothesis that fishes that flap pectoral fins can achieve and maintain higher pectoral-fin-powered swim speeds than fishes that row pectoral fins more formally, we compared Ucrit between pairs of closely related fishes from the family Labridae (wrasses), including G. varius (Walker and Westneat, 2002). Within each pair of species compared, one species (G. varius and Cirrhilabrus rubripinnis) flapped each pectoral fin along a steep stroke plane regardless of speed. The other species (Halichoeres bivittatus and Pseudocheilinus octotaenia) oscillated each fin along a shallow stroke plane at slow speeds but steepened the stroke plane with increasing speed. Nevertheless, at fatigue velocities, G. varius and C. rubripinnis had significantly steeper stroke planes than the H. bivittatus and P. octotaenia, respectively. As predicted by the rowing-flapping model, within each pair, the species with the steeper stroke planes had a significantly higher Ucrit than the species with the shallower stroke planes (Fig. 4).
|
Gait transitions
Increasing speed in vertebrates typically requires discrete locomotor modifications, or gait changes (Alexander, 1989
; Webb, 1994). Gait transitions may occur for several reasons: in terrestrial mammals, transitions occur because the gait used at the lower speed is more costly than the gait used at the higher speed (Alexander, 1989) whereas the transition from sustained BCF propulsion to sprinting BCF propulsion in fishes occurs because the muscle powering sustained swimming cannot power sprinting (Rome et al., 1988). Is the gait transition from pectoral fin to BCF swimming in labriform cruisers due to decreased locomotor costs or to lack of power (Korsmeyer et al., 2002)? All four wrasses used for the Ucrit comparisons (see above), presented a gait transition at or near the fatigue velocity in which sustained pectoral fin oscillation was replaced with a "burst and flap gait" in which one to four axial kicks (large amplitude oscillations) alternated with rapid, pectoral oscillation (Walker and Westneat, 2002). Axial kicking always resulted in rapid forward translation of the fish relative to the fixed tank while pectoral fin oscillation resulted in the slow backward translation of the fish relative to the tank. Similar burst-and-flap gaits have been described in other labriform cruisers (Webb, 1973; Drucker and Jensen, 1996a; Korsmeyer et al., 2002). The result that pectoral-fin-powered sprint speeds did not significantly differ from pectoral-fin-powered Ucrits (paired t-test, P = 0.8029, data from Walker and Westneat 2002) suggest that the transition to the burst-and-flap gait reflects a lack of power in the pectoral fin muscles and not a mechanism to reduce locomotor costs.
The energetics of gait transitions was explored in greater detail in the parrotfish Scarus schlegeli (Korsmeyer et al., 2002). Parrotfishes have traditionally been given their own family (Scaridae) but current phylogenetic data suggest that parrotfishes are one radiation within the Labridae (Streelman et al., 2002). S. schlegeli swims with a largely flapping motion, and, like other labrids, swims with its pectoral fins until fatigue speeds, where it switches to a burst-and-flap gait (Korsmeyer et al., 2002). The net cost of swimming (the oxygen consumption measured at speed u minus that measured at rest) increased with speed in S. schlegeli, but the gait transition to a burst-and-flap mode was characterized by an additional cost above that expected if the fish could maintain that speed with the pectoral fins alone. While this does not imply that flapping propulsion is less costly than BCF propulsion within the range of sustained speeds, it does suggest that labriform cruisers switch gaits not to reduce energy costs, as in terrestrial mammals, but to employ a muscle system that has the power to achieve the high swimming speeds. Given the reduced thrust at intermediate and high speeds in rowing relative to flapping fins (Walker and Westneat, 2000), it is not surprising to find lower gait transition speeds in the fishes with fin strokes closer to the rowing extreme (Ucrit results summarized above).
|
EVOLUTIONARY AND ECOLOGICAL MORPHOLOGY |
|---|
|
TOP
SYNOPSISINTRODUCTIONKINEMATICSDYNAMICSPERFORMANCE TRADE-OFFSENERGETICSEVOLUTIONARY AND ECOLOGICAL...References |
|---|
While the rowing-flapping model correctly predicts performance differences between Gasterosteus aculeatus and Gomphosus varius, and between closely related wrasses, can it predict differences in behavior and ecology? (Walker and Westneat, 2000
) review qualitative associations between appendage motion and behavioral and ecological variation across a variety of vertebrate taxa but the rowing-flapping model has been tested more formally among closely related species of Labridae. The rowing-flapping model makes at least two predictions. First, if routine swimming speeds differ between rowers and flappers, then the routines speeds of fishes with flapping, wing-shaped fins should be higher than those of fishes with rowing, paddle-shaped fins. Second, if flow velocities typical of high energy environments exclude species by swimming ability, then fishes that row paddle-shaped fins should be excluded. Recent field studies have shown that labrids with lower aspect ratio, paddle shaped fins tend to swim more slowly (Wainwright et al., 2002) and occupy less energetic zones on the reef (Bellwood and Wainwright, 2001; Fulton et al., 2001) relative to labrids with higher aspect ratio, wing-shaped fins. Our laboratory work has confirmed casual observations that labrids with wing-shaped fins use a flapping stroke while labrids with paddle-shaped fins use a rowing or intermediate stroke (Walker and Westneat, 2002). Additionally, as summarized above, we have shown that the labrids with flapping, wing-shaped fins have higher Ucrit than the labrids with rowing, paddle-shaped fins. The combination of these field and laboratory studies gives broad support for the rowing-flapping model.
Discussions of pectoral fin swimming in fishes have largely focused on the benefits of the fins during hovering, slow swimming and maneuvering. Most fishes that use a labriform gait for slow, steady swimming switch to BCF propulsion for intermediate and high speed, steady swimming, although this transition has been documented in only a few species (Blake, 1979, 1980; Archer and Johnston, 1989; Gibb et al., 1994; Lauder and Jayne, 1996). Most fish recruit a longitudinal band of axial, slow twitch, oxidative (red) fibers to power steady, BCF propulsion (Jayne and Lauder, 1994). Axial red fibers are absent in G. aculeatus and wrasses (te Kronnie et al., 1983; Davison, 1988, 1994). In contrast, the pectoral muscles are invested with high concentrations of red muscle fibers (Davison, 1988, 1994). Does the evolutionary loss of red muscle fibers from the body wall limit the ability of labriform cruisers to achieve and maintain high steady swimming speeds?
This question can be addressed by comparing the Ucrit of fishes that reach their fatigue velocities using either labriform or BCF propulsion. Because of differences in experimental design among tests, comparisons of Ucrit should be limited to general trends. The comparative data show that the labriform cruisers have Ucrits expected of BCF swimmers. More specifically, pectoral fin flappers can achieve and maintain swimming speeds higher than that expected of BCF swimmers at the same size (Uexp–BCF) while the rowers have lower than expected performance (Fig. 5). While certainly not as high as that measured for pelagic and anadromous fishes in the comparative sample, the relatively high, pectoral-fin powered Ucrit of the flappers suggest that the evolutionary loss of axial red muscle fibers has not limited fishes that have lost red axial muscle to a life in slow motion. By contrast, the concentrations of slow oxidative fibers in the pectoral muscles enable these fishes to achieve and maintain relatively high swimming speeds with a labriform gait.
|
|
ACKNOWLEDGMENTS |
|---|
We would like to thank the symposium organizers, Malcolm Gordon, Jay Hove and Ian Bartol, for the invitation to participate in this strongly integrative symposium that brought together scientists with diverse backgrounds. The constructive comments by two anonymous reviewers are greatly appreciated. Ideas in this paper have been developed through conversations with Brad Wright, Rick Blob, Michael Dickinson, Eliot Drucker, and George Lauder although they do not necessarily share all of the views expressed in this paper. This work was funded by a National Science Foundation Postdoctoral Fellowship in the Biosciences Related to the Environment, Office of Naval Research grants N00014-99-0184 and N00014-01-1-0506 and National Science Foundation grants IBN-9407253 and DEB-9815614.
|
FOOTNOTES |
|---|
1 From the Symposium Dynamics and Energetics of Animal Swimming and Flying presented at the Annual Meeting of the Society for Integrative and Comparative Biology, 2–6 January 2002, at Anaheim, California.
2 E-mail: walker{at}usm.maine.edu
|
References |
|---|
|
TOP
SYNOPSISINTRODUCTIONKINEMATICSDYNAMICSPERFORMANCE TRADE-OFFSENERGETICSEVOLUTIONARY AND ECOLOGICAL...References |
|---|
Alexander, R. M. 1989. Optimization and gaits in the locomotion of vertebrates. Physiol. Rev, 69:1199-1227.
Archer, S. D., and I. A. Johnston. 1989. Kinematics of labriform and subcarangiform swimming in the antarctic fish Notothenia neglecta. J. Exp. Biol, 143:195-210.
Bellwood, D. R., and P. C. Wainwright. 2001. Locomotion in labrid fishes: Implications for habitat use and cross-shelf biogeography on the Great Barrier Reef. Coral Reefs, 20:139-150.[CrossRef]
Blake, R. W. 1979. The mechanics of labriform locomotion I. Labriform locomotion in the angelfish (Pterophyllum eimekei): An analysis of the power stroke. J. Exp. Biol, 82:255-271.
Blake, R. W. 1980. The mechanics of labriform locomotion II. An analysis of the recovery stroke and the overall fin-beat cycle propulsive efficiency in the angelfish. J. Exp. Biol, 85:337-342.
Combes, S. A., and T. L. Daniel. 2001. Shape, flapping and flexion: Wing and fin design for forward flight. J. Exp. Biol, 204:2073-2085.
Daniel, T. L. 1984. Unsteady aspects of aquatic locomotion. Amer. Zool, 24:121-134.
Daniel, T. L. 1988. Forward flapping flight from flexible fins. Can. J. Zool, 66:630-638.
Daniel, T. L., and E. Meyhöfer. 1989. Size limits in escape locomotion of carridean shrimp. J. Exp. Biol, 143:245-265.
Davenport, J. 1987. Locomotion in hatchling leatherback turtles Dermochelys coriacea. J. Zool, 212:85-101.
Davenport, J., S. A. Munks, and P. J. Oxford. 1984. A comparison of the swimming of marine and freshwater turtles. Proc. R. Soc. Lond, B, 220:447-475.
Davison, W. 1988. The myotomal muscle of labriform swimming fish is not designed for high speed sustained swimming. New Zealand Nat. Sci, 15:37-42.
Davison, W. 1994. Exercise training in the banded wrasse Notolabrus fucicola affects muscle fibre diameter, but not muscle mitochondrial morphology. New Zealand Nat. Sci, 21:11-16.
Dickinson, M. H. 1996. Unsteady mechanisms of force generation in aquatic and aerial locomotion. Amer. Zool, 36:537-554.
Dickinson, M. H., and K. G. Götz. 1993. Unsteady aerodynamic performance of model wings at low Reynolds numbers. J. Exp. Biol, 174:45-64.[Abstract]
Dickinson, M. H., F.-O. Lehmann, and S. P. Sane. 1999. Wing rotation and the aerodynamic basis of insect flight. Science, 284:1954-1960.
Dorn, P., L. Johnson, and C. Darby. 1979. The swimming performance of nine species of common California inshore fishes. Trans. Am. Fish. Soc, 108:366-372.[CrossRef]
Drucker, E. G., and J. S. Jensen. 1996a. Pectoral fin locomotion in the striped surfperch. I. Kinematic effects of swimming speed and body size. J. Exp. Biol, 199:2235-2242.[Abstract]
Drucker, E. G., and J. S. Jensen. 1996b. Pectoral fin locomotion in the striped surfperch. II. Scaling swimming kinematics and performance at a gait transition. J. Exp. Biol, 199:2243-2252.[Abstract]
Drucker, E. G., and G. V. Lauder. 1999. Locomotor forces on a swimming fish: three-dimensional vortex wake dynamics quantified using digital particle image velocimetry. J. Exp. Biol, 202:2392-2412.
Drucker, E. G., and G. V. Lauder. 2000. A hydrodynamic analysis of fish swimming speed: Wake structure and locomotor force in slow and fast labriform swimmers. J. Exp. Biol, 203:2379-2393.[Abstract]
Drucker, E. G., and G. V. Lauder. 2002. Wake dynamics and locomotor function in fishes: Interpreting evolutionary patterns in pectoral fin design Integr. Comp. Biol, 42:-000.
Dudley, R. 2000. The biomechanics of insect flight. Princeton University Press, Princeton.
Ellington, C. P. 1984. The aerodynamics of hovering insect flight. IV. Aerodynamic mechanisms. Philosophical Trans. R. Soc. Lon., B, 305:79-113.
Fish, F. E. 1992. Aquatic locomotion. In T. E. Tomasi and T. H. Horton (eds.), Mammalian energetics: Interdisciplinary views of metabolism and reproduction, pp. 34–63. Cornell University Press, Ithaca.
Fish, F. E. 1993. Influence of hydrodynamic design and propulsive mode on mammalian swimming energetics. Austral. J. Zool, 42:79-101.
Fish, F. E. 1996. Transitions from drag-based to lift-based propulsion in mammalian swimming. Amer. Zool, 36:628-641.
Fulton, C. J., D. R. Bellwood, and P. C. Wainwright. 2001. The relationship between swimming ability and habitat use in wrasses (Labridae). Mar. Biol, 139:25-33.[CrossRef]
Fung, Y. C. 1993. An introduction to the theory of aeroelasticity. Dover, New York.
Geerlink, P. J. 1983. Pectoral fin kinematics of Coris formosa (Teleostei, Labridae). Neth. J. Zool, 33:515-531.
Gerstner, C. L. 1999. Maneuverability of four species of coral-reef fish that differ in body and pectoral-fin morphology. Can. J. Zool, 77:1102-1110.[CrossRef]
Gharib, M., F. Pereira, D. Daberi, J. R. Hove, and D. Modarress. 2002. Quantitative flow visualization: Toward a comprehensive flow diagnostic tool. Interg. Comp. Biol, 42:-000.
Gibb, A., B. C. Jayne, and G. V. Lauder. 1994. Kinematics of pectoral fin locomotion in the bluegill sunfish Lepomis macrochirus. J. Exp. Biol, 189:133-161.[Abstract]
Hammer, C. 1995. Fatigue and exercise tests with fish. Comp. Biochem. Physiol, 112A:1-20.[CrossRef][Medline]
Hartwell, S. I., and R. G. Otto. 1991. Critical swimming capacity of the Atlantic silverside, Menidia menidia L. Estuaries, 14:218-221.[CrossRef]
Hocutt, C. H. 1973. Swimming performance of three warmwater fishes exposed to a rapid temperature change. Chesapeake Sci, 14:11-16.[CrossRef]
Hoerner, S. F. 1965. Fluid-dynamic drag. Published by the Author, Midland Park, New Jersey.
Horridge, G. A. 1956. The flight of very small insects. Nature, 178:1334-1335.[CrossRef]
Hove, J. R., L. M. O'Bryan, M. S. Gordon, P. W. Webb, and D. Weihs. 2001. Boxfishes (Teleostei: Ostraciidae) as a model system for fishes swimming with many fins: Kinematics. J. exp. Biol, 204:1459-1471.[Abstract]
Jayne, B. C., and G. V. Lauder. 1994. How swimming fish use slow and fast muscle fibers: implications for models of vertebrate muscle rec