Feeding, fins and braking maneuvers: locomotion during prey capture in centrarchid fishes

Feeding and locomotion are highly integrated among diverse groups of animals (e.g. Hoff et al.,1985; Gorb and Barth,1994; Irschick and Losos,1998; Iwaniuk et al.,1999; Dunbar and Badam,2000; Budick and O'Malley,2000; Domenici,2001; Borla et al.,2002; Alfaro, 2003; Rice and Westneat, 2005; Higham et al., 2005a; Higham et al., 2007; Vincent et al., 2005; Walker et al., 2005). Because feeding is essential for growth and reproduction, the ability of the locomotor system to facilitate feeding contributes meaningfully to fitness. Fishes, in particular, capture prey by suction feeding, a rapid expansion of the mouth cavity to draw in water, but this can only be effective once the predator uses its locomotor system to position its mouth very close to the prey(Day et al., 2005; Higham et al., 2005a; Higham et al., 2006a). Although many fishes utilize suction to capture prey, the extent to which they rely on it depends on the ecology of the predator(Webb, 1984a; Webb, 1984b; Carroll et al., 2004; Higham et al., 2006a). A commonly cited dichotomy separates `ram feeders' from `suction feeders', where the former relies predominantly on swimming speed to overtake evasive prey and the latter relies more on suction to draw non-evasive prey items towards them. These two scenarios, although very different, require a contribution from the locomotor system to be effective. While swimming speeds at the time of prey capture have been measured for many fishes (e.g. Nyberg, 1971; Rand and Lauder, 1981; Norton and Brainerd, 1993; Wainwright et al., 2001; Porter and Motta, 2004; Higham et al., 2005a; Higham et al., 2005b; Higham et al., 2006a; Higham et al., 2007), detailed interspecific examinations of locomotor movements throughout prey capture are rare (Rice and Westneat,2005).

Fishes that rely predominantly on suction to capture prey tend to have small mouths, require a high level of accuracy during feeding, and feed on relatively non-evasive prey (Higham et al., 2006a). By contrast, fishes that rely more on ram tend to have a larger mouth, require less accuracy and feed on relatively evasive prey. In both cases, braking during prey capture is likely to be important. For fishes relying predominantly on suction, braking will enhance accuracy by giving the predator more time to adjust the position of its mouth relative to the prey (Lauder and Drucker,2004; Higham et al.,2006a) and will alleviate the negative effects of swimming speed on suction performance (Higham et al.,2005a). For fishes relying predominantly on ram, braking will enable the predator to be in a good position to follow a prey item that escapes (Webb and Gerstner,2000; Webb, 2006). It remains unclear whether different species employ similar mechanisms of braking and whether the relative timing of braking is similar between species that employ different feeding strategies. Understanding how and when fishes decelerate is central to understanding predator-prey interactions.

Apart from feeding, deceleration is an integral component of animal movement and is ecologically important for intermittent locomotion(Higham et al., 2001; Kramer and McLaughlin, 2001),avoiding obstacles in the environment(Webb and Gerstner, 2000) and arriving predictably at a certain location(Higham et al., 2005b). Despite the pervasiveness of braking among almost all mobile animals, there is a paucity of studies that have addressed the mechanisms underlying deceleration (Drucker and Lauder,2002; Drucker and Lauder,2003; Higham et al.,2005b; McGowan et al.,2005). Fishes have multiple control surfaces that can contribute to braking, including the body, paired fins (pectoral and pelvic), median fins(dorsal and anal) and caudal fin. It is possible that certain fishes use a particular combination of fins to brake while other fishes use a completely different combination. This many-to-one mapping of form to function(Wainwright et al., 2005)could amplify the interspecific diversity of fin movements during braking. However, with the exception of Higham et al.(Higham et al., 2005b) and Rice and Westneat (Rice and Westneat,2005), little is known about the integration of fin movements during deceleration in fishes.

During braking, fishes commonly protract their pectoral fins in order to increase the frontal area of the body(Breder, 1926; Harris, 1937a; Harris, 1937b; Bainbridge, 1963; Videler, 1981; Webb, 1984a; Geerlink, 1987; Webb and Fairchild, 2001; Borla et al., 2002; Drucker and Lauder, 2002; Drucker and Lauder, 2003; Higham et al., 2005b; Rice and Westneat, 2005). Pectoral fins likely enhance stability during braking since they limit side-to-side (yawing) movements by balancing each other if protracted together. In addition, perciform fishes such as centrarchids have pectoral fins that are located laterally on the body and generate a braking force such that the reaction force goes through the center of mass of the fish(Drucker and Lauder, 2002),thus limiting pitching movements. While the median fins can also contribute to braking, moving them to one side can lead to moments of yaw and roll(Drucker and Lauder, 2001; Standen and Lauder, 2005). By moving to the opposite side, the caudal fin might balance the forces from the median fins, possibly preventing a yawing maneuver. Whether species that rely more heavily on the caudal and median fins become less steady during prey capture is not known. If this is the case, then the accuracy of the strike might be compromised since a yawing maneuver would likely shift the location of the mouth relative to the prey.

For aquatic vertebrates, centrarchid fishes have been a model group for studies dealing with feeding (e.g. Nyberg,1971; Lauder,1980; Lauder et al.,1986; Wainwright and Lauder,1986; Norton and Brainerd,1993; Grubich and Wainwright,1997; Sass and Motta,2002; Svanback et al.,2002; Ferry-Graham et al.,2003; Carroll,2004; Carroll et al.,2004; Day et al.,2005; Higham et al.,2005a; Higham et al.,2006a; Higham et al.,2006b; Carroll and Wainwright,2006) and locomotor (e.g. Jayne and Lauder, 1993; Jayne and Lauder, 1994; Jayne and Lauder, 1995; Jayne and Lauder, 1996; Jayne et al., 1996; Gibb et al., 1994; Johnson et al., 1994; Lauder and Jayne, 1996; Drucker and Lauder, 1999; Drucker and Lauder, 2000; Drucker and Lauder, 2001; Standen and Lauder, 2005; Higham et al., 2005b)function. The considerable diversity in both ecology and morphology makes centrarchids an excellent group for addressing questions regarding locomotor function during feeding. For example, species from the genus Micropterus, such as largemouth bass, have large mouths and are thought to rely heavily on high swimming speeds to capture large, relatively evasive prey (Higham et al.,2006a). By contrast, species from the genus Lepomis, such as bluegill, are deeper bodied, have relatively small mouths and are thought to rely heavily on suction to capture small, relatively non-evasive prey.

To gain insights into the mechanisms and timing of deceleration during prey capture, I studied the body and fin kinematics of largemouth bass(Micropterus salmoides) and bluegill sunfish (Lepomis macrochirus) during feeding. I addressed the following three specific questions in this study. (1) How do fishes decelerate? I predicted that both species will increase their frontal area by employing their pectoral, caudal and median fins in order to decelerate. (2) How do fishes modulate the magnitude of deceleration? I predicted that both species would increase the angular excursion of their fins in order to increase frontal area and the speed of the water relative to the fins. (3) How do fishes modulate swimming speed when capturing stationary prey? Because they typically feed on evasive prey, I predicted that largemouth bass would maintain a constant, relatively high ram speed until after the prey was captured. By contrast, I predicted that bluegill would decelerate considerably prior to capturing the prey item in order to maintain strike accuracy. Although comparing two species precludes conclusive interpretations of locomotor behavior based on feeding biology, I chose two closely related centrarchid species that exhibit extremely different morphological and ecological traits related to feeding. Thus, conclusions are likely related to feeding biology rather than other factors such as phylogenetic history.

Largemouth bass (Micropterus salmoides Lacépède) and bluegill sunfish (Lepomis macrochirus Rafinesque), both members of the freshwater family Centrarchidae, were studied. These two species have been the focus of several locomotor (Lauder and Jayne, 1996) and feeding(Higham et al., 2006a; Higham et al., 2006b) studies and are ideal for a comparative study since their ecology and morphology differ considerably (Lauder and Jayne,1996; Carroll et al.,2004; Collar et al.,2005). Furthermore, these genera (Lepomis and Micropterus) are monophyletic and sister taxa with a most recent common ancestor estimated about 24 million years ago(Near et al., 2005). The fish were collected around Davis, CA, USA and were maintained in separate 38-liter aquaria. Data were analyzed from three individuals from each species of similar size. The mean body mass for largemouth bass and bluegill used in these experiments was 195.4±20.9 g and 143.0±24.7 g,respectively. Experiments complied with all guidelines for the use and care of animals in research at the University of California, Davis, where all experiments were conducted.

Each fish used in this study was euthanized using an overdose of MS-222. The body was laid flat on its right side and the dorsal, anal and caudal fins were extended to their maximum position and pinned to a piece of Styrofoam. The left and right pectoral fins were removed carefully, extended to the natural maximum position and pinned to a small piece of Styrofoam. Each fin was brushed with formalin for preservation. With a ruler in the field of view,pictures were taken of each specimen and pectoral fins using a Kodak EasyShare CX7430 digital camera (1728×2304 pixels). The digital images were then imported into ImageJ version 1.33 (NIH, Washington, DC, USA) and the area of each fin was calculated (see Fig. 1). The aspect ratio (AR) of each pectoral fin was calculated by dividing the square of the length of the leading edge by the area.

Experiments were conducted in 200-liter tanks that were divided into two sections, which facilitated obtaining start-stop episodes with a standardized predator-prey distance. A very thin wire, angled away from the predator, was used to suspend small goldfish (approximately 5 cm long) or ghost shrimp(approximately 3-5 cm long). With the predator secured at one end of the tank with a trap door, the prey item was suspended approximately 40 cm from the door [see fig. 1 in Higham et al. (Higham et al., 2005b)]. Once the trap door was removed, the predator was free to swim towards the prey item and capture it. Although only sequences using tethered prey were filmed,freely moving prey were offered to the fish after every two trials involving tethered prey in order to maintain the fish's motivation and avoid any changes in behavior associated with capturing tethered prey. Locomotion and feeding were recorded from each fish using a high-speed NAC Memrecam ci digital system(Tokyo, Japan) operating at 500 images s^-1. Lateral and ventral(via a mirror underneath the prey oriented at 45°) views of the feeding event (Fig. 2) were obtained.

A picture was taken of a ruler that was placed inside the tank at the location of the prey to scale measurements. Two floodlights (600 W) on either side of the camera illuminated the experimental tank. The fish were not fed for a minimum of two days prior to the day of testing and the time between successive trials of a single individual within a single day was approximately 5-10 min. Only those sequences were analyzed in which the fish started from 40 cm away from the prey, the prey was completely consumed and there were no pauses during the prey capture event.

For the 120 ms prior to and the 60 ms following maximum gape, frames were digitized at 250 Hz, since this provides a reliable estimate of acceleration during locomotion (Walker,1998). For each frame, the forward displacement of the fish was calculated by digitizing the anterior margin of the eye in lateral view. Using Igor Pro 5.01 (WaveMetrics Inc., Lake Oswego, OR, USA), the displacement data were smoothed using a cubic spline interpolation (smoothing factor=1,s.d.=0.001) and then the first and second derivative were calculated to obtain velocities and accelerations, respectively. In addition to this measure of instantaneous acceleration, the mean deceleration over the last 60 ms of prey capture was calculated by dividing the change in velocity over this time interval by 60 ms. From the lateral view, gape was the vertical distance between the tip of the lower jaw and the tip of the upper jaw, and maximum gape (MG) occurred when this distance was maximal.

From ventral view coordinates, instantaneous two-dimensional fin angles were calculated every 20 ms starting from 120 ms before MG and ending 60 ms after MG. The pectoral fin and anal fin angles were calculated between lines from the anterior margin of the base of the fin to the distal tip of the fin and a point on the body posterior to the fin. The caudal fin angle was calculated between lines from a point on the body posterior to the base of the fin to the base of the fin and the tip of the fin.

The methods of Geerlink (Geerlink,1987) were used to calculate a theoretical value of maximal drag created by the pectoral fins (D_pec) of bluegill and largemouth bass. The following equation was used: D_pec=½ρS_pecU²C_D,where ρ is the density of the surrounding water (1026 kg m^-3), S_pec is the summed frontal area of the two pectoral fins if they are fully extended in a vertical position, U is the velocity of the fish at the time of MG, and C_D is the drag coefficient [1.17 as in Geerlink(Geerlink, 1987)].

To determine whether the species differ morphologically, an analysis of variance (ANOVA) was performed using species (fixed) as the independent variable. The dependent variables, which were analyzed separately, were pectoral fin aspect ratio (AR), and the log₁₀-transformed residuals from a least-squares regression of body mass and the area of each fin (pectoral, anal and caudal). For pectoral fin area and AR, the average of the left and right fin was used for each individual in the ANOVA. In order to correct for multiple statistical tests, α (0.05) was adjusted using a sequential Bonferroni test(Rice, 1989).

In order to determine how deceleration was modulated, multiple regressions were performed on each species separately, with the mean deceleration over the last 60 ms of prey capture as the dependent variable. The independent variables were ram speed at the time of MG, maximum pectoral fin angle,maximum anal fin angle and maximum caudal fin angle. Additional least-squares regressions were performed in order to determine the correlation between continuous variables. All variables in these analyses were log₁₀transformed to normalize variances, and in each case this allowed the variables to meet the assumptions of the parametric procedures.

Nine variables were included in a principal components analysis (PCA) to reduce dimensionality and search for axes of correlated kinematic variation:magnitude and timing of maximum caudal, anal and pectoral fin angles, maximum average deceleration, angular excursion of the body along the z-axis,and swimming speed at MG. The resulting principal components (1 and 2) became the axes of a multidimensional locomotor kinematic space, and were visualized in graphical form. To determine if the two species occupied different regions of kinematic space, an ANOVA was performed with species (fixed) and individual(nested within species; random) as the independent variables and the PC scores from a particular axis as the dependent variable. In order to properly account for the replication of observations within individuals, the denominator in the F-test for the main effect of species was individual (nested within species) (Zar, 1996). SYSTAT version 9 (SPSS Inc., Chicago, IL, USA) was used for all statistical analyses.

After accounting for body size, bluegill sunfish have significantly larger pectoral (P<0.001) and anal (P<0.001) fins than largemouth bass (Table 1). The two species did not differ in the size of their caudal fins. Bluegill sunfish exhibit a significantly higher pectoral fin aspect ratio (P<0.05)than largemouth bass (1.82 versus 1.34; Table 1).

Throughout the approach to the prey, largemouth bass swam at a much higher speed than bluegill sunfish (Fig. 3; Table 2). Between 120 ms and 40 ms prior to MG, largemouth bass decelerated. During the 40 ms prior to MG, largemouth bass accelerated. Following MG, largemouth bass decelerated continually and substantially. Bluegill exhibited a similar sequence of events, although the acceleration before MG occurred during the 20 ms, rather than 40 ms, prior to MG (Fig. 3).

Ram speed, at the time of MG, was positively correlated with MG for largemouth bass (Fig. 4; r²=0.58; P<0.001). Although both species decelerated maximally following prey capture, the timing of maximum deceleration was much later for largemouth bass (30.3 ms after MG) than bluegill (6.7 ms after MG). Because the magnitude of deceleration was much higher for largemouth bass than bluegill, both species had similar ram speeds 60 ms after MG (Fig. 3A).

Both largemouth bass and bluegill protracted their pectoral fins as they approached the prey item (Fig. 5A). During the 60 ms following MG, largemouth bass protracted their pectoral fins, on average, over 40°. Pectoral fin protraction over the final 60 ms of the strike was significantly greater in bass than bluegill(ANOVA; P<0.05), who first retracted their pectoral fins and then protracted them between 40 and 60 ms after MG(Fig. 5A). Both species abducted their anal fin over 10° during the 60 ms after MG(Fig. 5B). The caudal and anal fins were almost always abducted in opposite directions during braking in both species (Fig. 2).

For both bluegill (r²=0.43; P<0.0001) and largemouth bass (r²=0.31; P<0.01), increased ram speeds 120 ms prior to MG resulted in increased magnitudes of deceleration following prey capture (Fig. 6A). In addition, increased ram speeds 120 ms prior to MG resulted in greater angular excursions (in the z-axis) of the body over the final 80 ms of the feeding event for largemouth bass(r²=0.52; P<0.001) but not bluegill(r²=0.005; P>0.5)(Fig. 6B). The model consisting of maximum fin angles and ram speed at the time of MG explained 84%(P<0.001) and 77% (P<0.001) of the variation in mean deceleration in largemouth bass and bluegill, respectively. However, maximum fin angles alone explain 37% (P>0.05) and 15% (P>0.05)of the variation in mean deceleration in largemouth bass and bluegill,respectively. Thus, ram speed accounts for a substantial amount of variation in mean deceleration. For largemouth bass, the maximum angle of the pectoral fin was also increased in order to increase the magnitude of deceleration. Using the equation from Geerlink(Geerlink, 1987) (see above),largemouth bass and bluegill had theoretical values of maximum drag(D_pec) equal to 0.62 and 0.04, respectively.

From the PCA, swimming (ram) speed at the time of MG, mean deceleration during the final 60 ms, and angular excursion of the body in ventral view were the most highly correlated variables with principal component 1 (PC 1; Fig. 7; Table 3). I will refer to this axis as the `ram speed' axis. Variables describing the angles of the fins loaded strongly on PC 2, and I will refer to this axis as the `fin' axis(Fig. 7; Table 3). Largemouth bass had significantly higher scores on PC 1 than bluegill (P=0.002), but there was no difference between species on PC 2 (P>0.05).

Both largemouth bass and bluegill employ their pectoral, anal and caudal fins during deceleration, although largemouth bass protract their pectoral fins more than bluegill. In addition, largemouth bass decelerate at a much greater magnitude than bluegill and achieve this by swimming faster prior to the onset of deceleration. Both species modulate the magnitude of deceleration by altering swimming speed rather than significantly altering fin kinematics. The greater magnitude of deceleration arises from the exponential increase in drag generated by the fins as swimming speed increases. Largemouth bass swim much faster than bluegill during prey capture and decelerate much later than bluegill. This behavior is likely to be associated with the decreased need for accuracy (because of the larger mouth) and the diet (largely comprised of evasive fishes) of largemouth bass.

Protraction of the pectoral fins during prey capture is a mechanism of deceleration employed by largemouth bass and bluegill(Fig. 5A). However, largemouth bass protract their fins faster and to a greater extent(Fig. 5A), suggesting that they generate more reverse thrust with their fins(Higham et al., 2005b). Although largemouth bass have significantly smaller pectoral fins than bluegill (Table 1), the drag force exerted by a fin is proportional to its velocity squared. Thus, a smaller fin, if protracted at a faster rate, has the potential for generating comparable or greater forces than a larger fin. Using the methods of Geerlink(Geerlink, 1987), the maximum drag forces exerted by the pectoral fins are much greater in largemouth bass than bluegill. Thus, despite having smaller pectoral fins, the swimming speed of largemouth bass appears to result in much higher drag forces from pectoral fins. When the area of the pectoral fins of bluegill sunfish is reduced, the fish compensate by protracting their fins at a higher speed in order to maintain a similar level of deceleration(Higham et al., 2005b). This behavioral modulation, combined with the results in the current study,suggests that fin morphology might not be the only factor that defines braking ability. For example, for a given pectoral fin (shape and size), the ability of the fish to hold that fin against the flow of water would depend largely upon the ability of the abductor muscles to generate and sustain the necessary forces. A fish might then reach a threshold swimming speed where the fin muscles would no longer generate forces great enough to match the force exerted by the water on the fin. Thus, the ability of the pectoral fin abductor muscles to protract the pectoral fins to a position where drag is maximal could potentially limit the forces generated by fins, and ultimately braking performance.

Few studies have quantified pectoral fin angles during prey capture in fishes (Higham et al., 2005b; Rice and Westneat, 2005). Rice and Westneat examined the pectoral fin angles of two herbivorous species of parrotfish, Sparisoma radians and Scarus quoyi, during feeding (Rice and Westneat,2005). As in largemouth bass and bluegill, both parrotfish species sweep their pectoral fins forward to initiate a braking maneuver as they approach the prey item. However, the braking maneuvers by the parrotfishes resulted in a large drop in swimming speed prior to contact with the prey item. This is quite different from the results of the present study in which largemouth bass and bluegill both showed a large increase in swimming speed prior to prey capture. One explanation is that the parrotfishes studied by Rice and Westneat commonly feed from substrate, and braking after prey capture, as in bluegill and largemouth bass, would result in a collision with the substrate and potential injury. Future studies that examine the locomotor kinematics of largemouth bass and bluegill feeding from substrate would provide further insight into the differences between these groups of fishes.

When a fish generates suction, the flow enters the mouth and exits the posterior opercular slits. This allows the fish to ingest water at the same time that water is expelled, resulting in a volume of ingested water far exceeding the volume of the buccal cavity(Day et al., 2005; Higham et al., 2006a). The water that exits the opercular cavities during the feeding event is in close proximity to the anterior surfaces of the abducted pectoral fins(Fig. 2F). Thus, the speed of water passing the pectoral fins might be increased by this added flow out of the opercular cavities, resulting in a further increase in drag generated by the pectoral fins. This might enable fishes that have laterally placed pectoral fins to take advantage of suction feeding for deceleration. Largemouth bass and bluegill might take advantage of this, but bass have a much higher volumetric flow rate during feeding, so their benefit would likely be greater. Future studies that measure the flow exiting the opercular cavities around pectoral fins during feeding would provide insight regarding this potential method of enhancing deceleration.

The caudal and anal fins of both bluegill and largemouth bass were abducted, increasing the frontal area of the body and enhancing deceleration(Fig. 2). The use of caudal and median fins during braking has been observed in many species of fishes (e.g. Videler, 1981; Higham et al., 2005b). The median fins are capable of generating yawing and rolling movements of the body(Jayne et al., 1996; Drucker and Lauder, 2001; Webb, 2004; Standen and Lauder, 2005). During braking, however, both median fins are abducted to a common side(Higham et al., 2005b), thus reducing the possible rolling movement from one of the median fins. In almost all cases, the caudal fin moved to the opposite side of the median fins, which likely limits the yawing movements that might be caused if they were abducted to the same side. However, largemouth bass did exhibit an overall yawing movement during feeding (Table 2), and it was positively correlated with ram speed(Fig. 6B). The caudal fin of largemouth bass is approximately three times the size of the anal fin and is larger than the summed area of the pectoral fins(Table 1). Thus, the yawing movement of largemouth bass likely stems from the disproportionately large caudal fin. The caudal fin of bluegill is less than twice the size of the anal fin and is smaller than the summed area of the pectoral fins(Table 1). This suggests that the pectoral fins are generating more force than the caudal fin, which would help prevent yawing movements. Future studies that quantify the forces, using digital particle image velocimetry (DPIV), generated by the caudal and median fins during braking would provide further insight into their relative contributions to the overall braking force.

The ability of a predator to modulate deceleration is imperative for arriving at a prey item at a predictable speed. I predicted that the angular excursions of the fins would increase in order to increase the magnitude of deceleration. The increased excursions would increase the speed of the surrounding water relative to the fin and thus result in higher drag forces. In addition, I thought that largemouth bass, because they exhibit greater deceleration, would exhibit greater fin excursions compared with bluegill. Given that the magnitude of deceleration of largemouth bass is approximately four times that of bluegill (Table 2), it is surprising that largemouth bass only exhibit greater pectoral, and not other, fin excursions than bluegill during braking. What are largemouth bass doing to achieve this level of deceleration? Within each species, the only variable positively correlated with magnitude of deceleration was ram speed. Since the speed of largemouth bass prior to, and at the time, of prey capture is much greater than that of bluegill(Fig. 3), the drag force(proportional to velocity squared) from the body is much greater in largemouth bass. Given the nature of the aquatic environment, this mechanism for modulating deceleration might be very common among aquatic animals.

Largemouth bass approach their prey at much higher speeds than bluegill(Fig. 3). This result supports the notion that largemouth bass rely more heavily on ram, and bluegill rely more on suction to capture prey (Carroll et al., 2004; Higham et al.,2006a). While both largemouth bass and bluegill decrease their ram speed as they approach the prey item, they both accelerate immediately prior to MG (Fig. 3). This might allow the predator to adjust its location relative to the prey item and then lunge forward at the last moment. Because largemouth bass commonly feed on relatively evasive prey such as fish, it is not surprising that they accelerate more than bluegill immediately preceding MG. This might enhance their chances of capturing a prey that has the potential for initiating an escape response. Alternatively, this acceleration might be a passive result of ingesting water during suction feeding. This is supported by the fact that largemouth bass have a much larger MG and ingest a much larger volume of water than bluegill (Higham et al.,2006a). Given that suction feeding naturally draws the predator forward (Muller et al., 1982),the higher volume flow rate exhibited by largemouth bass might result in higher accelerations of the body than for bluegill. In addition, MG was positively correlated with ram speed at the time of MG(Fig. 4). Given that MG is positively correlated with volume and volumetric flow rate in largemouth bass(Higham et al., 2006a), the correlation between ram speed and MG might simply be a result of increased forces pulling the fish forward. This passive increase in swimming speed due to suction generation might obscure interspecific comparisons of ram speed if the species differ in the size of the ingested volume of water. Future studies that measure ram speed should thus be cautious when interpreting ram speeds as solely reflecting the motivation of the predator.

Largemouth bass typically feed on evasive prey such as fishes, and these prey items likely require higher ram speeds in order to be caught. Largemouth bass decelerate maximally much later than bluegill, supporting the idea that bass maintain speed until the prey item is in the mouth. However, largemouth bass do decelerate following prey capture, suggesting that overshooting the prey item too much might be disadvantageous. Webb and Gerstner suggest that predators feeding on evasive prey benefit from not overshooting the prey in case the prey manages to escape (Webb and Gerstner, 2000). Not overshooting would put the predator in a good position to continue the chase and perhaps capture the prey item. Bluegill also decelerate during prey capture, but the reasons are likely to be different than for largemouth bass. Bluegill are more likely to feed in cluttered habitats or from substrate, and decelerating would prevent a collision in the environment. In addition, bluegill have much smaller mouths and likely depend more on accurately positioning their mouth relative to the prey item (see below for discussion), which might be compromised by swimming fast (Higham et al.,2006a).

An alternative explanation for why largemouth bass swim faster than bluegill during prey capture is that they are simply better swimmers than bluegill. This would suggest that bluegill are relegated to suction feeding as a result of their relatively poor ability to swim. However, several studies have shown that bluegill are morphologically specialized for suction feeding(Carroll et al., 2004; Collar and Wainwright, in press) and are better at generating a suction-induced flow of water than largemouth bass (Higham et al., 2006a; Higham et al.,2006b). In addition, bluegill can reach swimming velocities that are comparable to those of largemouth bass(Domenici and Blake, 1997). These lines of evidence strongly suggest that the locomotor behavior of bluegill and largemouth bass during prey capture is strongly related to their feeding ecology rather than their swimming capabilities.

The prey item in this study was tethered rather than freely moving. While tethering prey could potentially confound the experimental results, it also reflects an ecologically relevant prey capture situation for largemouth bass and bluegill. Both of these species feed on prey that is commonly located on or close to substrate, such as crustaceans(Collar et al., 2005). In addition, capturing prey prior to the initiation of an escape response (i.e. prior to prey movement) has been advocated as an advantage of ram feeding (see Nemeth, 1997) and is common in studies of fish feeding. Future studies that examine locomotor behavior of fish attacking a moving prey item would provide insight into the flexibility of locomotor behavior and whether differences between species persist regardless of prey presentation.

Strike accuracy is imperative for successful prey capture and is an often overlooked aspect of suction feeding performance(Higham et al., 2006a). The constraints imposed by accuracy likely diminish with an increased MG because the ingested volume of water increases considerably with an increase in MG(Higham et al., 2006a; Higham et al., 2006c). Because increases in swimming speed decrease strike accuracy(Webb and Skadsen, 1980; Higham et al., 2006a; Higham et al., 2006c), it is possible that only those species with larger mouths are able to maintain high swimming speeds and still successfully capture prey. Indeed, largemouth bass swim faster and have a much larger MG than bluegill. This idea is corroborated by a study of 18 species of cichlid fishes in which MG was significantly and positively correlated with ram speed (Higham et al.,2006c). Future studies that directly measure accuracy through ontogeny, and over a large range of ram speeds, could provide insight into the relationships between ram speed, MG and strike accuracy. Largemouth bass became increasingly unsteady (increased yawing movements during prey capture) with an increase in ram speed (Fig. 6B). This is probably an additional mechanism for decreased strike accuracy with an increase in ram speed. It is possible that, if bluegill swam faster, they would suffer a reduction in strike accuracy, which might significantly reduce the rate of successful prey capture.

Turning maneuvers are commonly executed using pectoral fins(Breder, 1926; Drucker and Lauder, 2001; Lauder and Drucker, 2004; Walker, 2004; Drucker et al., 2006). One mechanism of turning is to extend the pectoral fin on one side of the body while keeping the other fin against the body. The drag generated by the extended fin on the inside of the turn acts as a pivot and contributes to the rotational torque on the center of mass, resulting in yawing rotation(Breder, 1926; Drucker and Lauder, 2001; Walker, 2004; Drucker et al., 2006). Even with a smaller pectoral fin, an increased rate of fin protraction will result in an exponential increase in the drag force, and ultimately the yawing rotation. Given the relationship between fluid speed and drag, the swimming speed of a fish will likely also have profound effects on turning performance. In addition, the morphology of pectoral fins is likely linked to the swimming behavior of a fish. For example, fishes that typically swim fast, such as largemouth bass, might benefit from having smaller pectoral fins because less muscle power would be required to extend one of them during a turning maneuver.

This manuscript was significantly improved by comments from Peter Wainwright and two anonymous reviewers. This research was supported by NSF grant IOB-0444554 to Peter Wainwright.