Fish Exploiting Vortices Decrease Muscle Activity

See allHide authors and affiliations

Science  28 Nov 2003:
Vol. 302, Issue 5650, pp. 1566-1569
DOI: 10.1126/science.1088295


Fishes moving through turbulent flows or in formation are regularly exposed to vortices. Although animals living in fluid environments commonly capture energy from vortices, experimental data on the hydrodynamics and neural control of interactions between fish and vortices are lacking. We used quantitative flow visualization and electromyography to show that trout will adopt a novel mode of locomotion to slalom in between experimentally generated vortices by activating only their anterior axial muscles. Reduced muscle activity during vortex exploitation compared with the activity of fishes engaged in undulatory swimming suggests a decrease in the cost of locomotion and provides a mechanism to understand the patterns of fish distributions in schools and riverine environments.

Many fishes live in habitats in which they commonly encounter vortices that arise from fluid flow past stationary objects or from the propulsive movements of other animals. Energy extraction from environmental vortices has been consistently implicated as a hydrodynamic mechanism to increase the performance of swimming fishes (18). The preference of fishes to use these unsteady flows has been documented in the field (4, 9, 10) and laboratory (1113). However, the dynamic and transparent nature of flowing water has precluded quantitative visualization of interactions between fishes and vortices and, thus, an understanding of the underlying physical mechanisms involved. Furthermore, the effect of vortical flows on the degree and pattern of axial muscle activity in fishes remains entirely unknown.

We generated periodic vortices of similar strength and size to each other by using a vertically mounted D-section cylinder (i.e., a cylinder bisected along its long axis) placed in water flowing at a known velocity (14) (Fig. 1). These vortices were shed from the D-section cylinder in a staggered array collectively known as a Kármán street (15). A Kármán street is an example of a drag wake (rotation of alternately shed vortices is toward each other upstream), which can form between the thrust wakes (rotation of alternately shed vortices is toward each other downstream) of two fish swimming side by side (5). Vortices generated by the D-section cylinder were similar in strength to those produced by other freely swimming fish (16, 17).

Fig. 1.

Experimental setup to simultaneously visualize columnar vortices shed from a D-section cylinder (using DPIV; camera 1) and the body kinematics of station-holding trout (camera 2).

Compared with swimming in free stream (uniform) flow, there are two hydrodynamic benefits of station holding behind a cylinder. Fish can simply swim against the current in the region of reduced flow, drafting, for example, as a bicyclist would behind another bicyclist, or they can generate lift to move against the flow by altering their body kinematics to synchronize with the shed vortices. Because energy can be captured from cylinder vortices (18), trout that synchronize their body kinematics to vortices appropriately may need to use very little energy and, thus, gain a hydrodynamic advantage beyond that of drafting in the reduced velocity alone.

Simultaneous visualization of a two-dimensional, horizontal slice through the columnar vortices generated by the D-section cylinder (15) using digital particle image velocimetry (DPIV) (19) and the movements of trout with high-speed video (Fig. 1) revealed that trout slalom between vortices rather than through them (Fig. 2, A and B; Movie S1). If the flow is decomposed into a downstream (x axis, Fig. 1) and a lateral (z axis) component, slaloming between vortices occurs when trout move against the downstream flow but with the local lateral flow. Slaloming through oncoming vortices requires opposing the local lateral flow. Actuated foils generate more thrust if they slalom through rather than between Kármán vortices (20), but they require more power input to oppose the instantaneous local flow. Trout slalom between rather than through vortices and, thus, minimize power input rather than maximize thrust output. To quantify the movement of trout near vortices, we described the phase relation between the lateral motions of points along the body relative to the arrival of drifting vortices (Fig. 2C). When the center of a vortex drifted down to the center of mass (COM) of the fish, the COM was at its maximum lateral excursion away from the vortex, indicated by a phase relation of 180°. For all points anterior to the COM, the body moved away from an oncoming vortex (<180°), whereas body points posterior to the COM move toward the vortex (>180°). A phase relation of 0° or 360° would indicate that the fish had intercepted the center of a vortex.

Fig. 2.

Interaction of trout with cylinder vortices. Scale bars represent 2 cm. (A) Time series (ms) of body outlines (black) and midlines (purple) superimposed onto vorticity and velocity vector plots of the cylinder wake. Arrows indicate flow direction and magnitude (scale arrow = 56 cm s–1). Red and blue represent clockwise and counterclockwise vorticity (radians per second), respectively. (B) Midlines for seven consecutive tail-beats, beginning from right to left, of which (A) is a subset. Standard errors in vortex position given as bar lengths within each vortex. (C) Phase between body and vortices, where 180° represents slaloming in between vortices and 0° or 360° represents vortex interception. Fish center of mass shown as checkered circles. Standard error bars are in light gray.

In their preferred position holding station, one to two body lengths (L) downstream from the cylinder and away from the low-pressure suction region (15), trout synchronize their tail-beat frequency to the vortex shedding frequency of the cylinder and display large lateral body movements (Fig. 3, A and C). These kinematic features characterize a mode of locomotion termed the Kármán gait (11). This pattern of movement qualifies as a gait based on previous definitions applied to moving animals, in which body kinematics are observed to change discontinuously (21, 22), usually across a range of speeds or environments. Trout that adopt the Kármán gait display a body wavelength (a description of body shape based on a sine wave) that is longer than the cylinder wake wavelength (downstream spacing of vortices, Fig. 3B). This resulted from the head interacting differently with a vortex than did the tail (Fig. 2C), and ensures that the incident flow is maintained at an angle to the body to generate lift (23, 24). Due to buffeting caused by cylinder vortices, the lateral body motion was much greater than for fish swimming in the free stream (Fig. 3C). Body kinematics have a consistent relation with the frequency and phase of the vortex wake across several cylinder treatments (11), which supports our hypothesis that trout can extract energy from the wake in addition to benefiting from drafting in the reduced flow behind the cylinder. Other fishes with a diversity of body shapes and occupying different ecological niches (brook trout, Salvelinus fontinalis; alewives, Alosa pseudoharengus; bluegill sunfish, Lepomis macrochirus; and yellow perch, Perca flavescens) also exhibited the Kármán gait.

Fig. 3.

Comparison of body kinematics (filled circles) with associated wake variables (filled squares) in the presence of a D-section cylinder and in the absence of a D-section cylinder, or free stream flow (FS). Error bars are shown but may be obscured by data symbols. (A) Tail-beat frequencies for trout behind a cylinder and in the free stream, along with the vortex-shedding frequency of the cylinder. (B) Body wavelength for trout behind the cylinder and in the free stream, along with the cylinder wake wavelength (where L is total body length). (C) Lateral body amplitudes, where open circles, squares, and triangles represent the snout, center of mass, and tail tip, respectively.

Little is known about muscle activity patterns for fish swimming in perturbed environments. Using fine wire electrodes to record electrical muscle activity, we found that trout only activated their anterior, axial red muscles when they adopted the Kármán gait (Fig. 4B). This pattern of isolated muscle activity represents a fundamental departure from the pattern of propagating muscle activity characteristic of propulsive, undulatory locomotion (Fig. 4A) (2527), and further lends support for identifying this mode of locomotion as a distinct gait (22). Isolated, anterior muscle activity during the Kármán gait demonstrates that drafting (i.e., undulatory propulsion in reduced flow) is not the sole mechanism for station holding behind a cylinder. Image analysis of stereotypic body positions during separate DPIV and electromyography experiments illustrates that anterior red muscles become active on the opposite side of the body before a vortex reaches the COM (Fig. 4B). Anterior muscle activity may serve to modulate the angle of attack of the head and body, functioning to sustain body orientation or initiate changes in yaw. Thus, trout can hold station behind a cylinder in fast, down-stream flow by recruiting relatively few muscles to control their body orientation to exploit oncoming vortices instead of recruiting most of their axial muscles to generate thrust via undulation. The small amount of muscle activity seen during the Kármán gait is perhaps not surprising, given that even passive hydrofoils can generate thrust when synchronized to an oscillating flow (23, 24).

Fig. 4.

Time series illustrating that axial red muscle activity differs between trout swimming in free stream flow versus trout holding station behind a cylinder. Circles denote electrode positions with no (open), intermediate (orange), or high (red) muscle activity. (A) A propagating wave of muscle activity for a trout swimming in the free stream. (B) Muscle activity for a trout behind a D-section cylinder with estimated locations of a clockwise vortex.

Our results show a mechanism by which swimming fish can use environmental vortices, which is likely to reduce the cost of locomotion. In a turbulent environment, use of the Kármán gait allows the body of a fish to act as a self-correcting hydrofoil. Design of foil propulsors for biomimetic autonomous underwater vehicles could benefit from the reduced power costs associated with the self-adjusting motions of the Kármán gait. Demonstrating a link between vortex interaction and reduced muscle activity may also provide insight into the dynamics and evolution of aggregation in moving animals. Although trout behind a cylinder slalom in between vortices, vortex interception, although previously discounted (28), can, in theory, facilitate upstream acceleration and explain temporary deviations from the optimal pattern of school formation (5, 29). In addition, a precise understanding of the interactions between muscle activity and vortices across species promises to aid in the design and implementation of fish passageways.

Supporting Online Material

Materials and Methods

Movie S1

References and Notes

View Abstract

Stay Connected to Science

Navigate This Article