7. Morphology and Functional Ecology of the Fins and Axial Skeleton

The body of a fish essentially consists of a compression resistant notochord or vertebral column, surrounded by lateral musculature, and wrapped in a complex arrangement of connective tissue and skin (Danos et al. 2008). In contrast to terrestrial locomotion, where the limbs involved in locomotion must also support the body, fishes can use a variety of mechanisms for locomotion, both independently and in concert, and can employ a variety of control surfaces such as scales, body projections, and fins to affect their posture and position in the water column (Webb 1994, 2006).

Forces to Overcome

To achieve forward motion, the force generated by a swimming fish must equal (constant swimming speed) or exceed (acceleration) the resistance to movement caused by drag (Webb 1975; Blake 1983a). The two components of drag are friction drag and pressure drag, both of which can best be understood by boundary-layer theory. Water moving across the body of a fish, either by the fish moving through water or holding position in flowing water, has a gradient in relative velocity that increases from 0, where water molecules are in contact with the fish, to that of the free-stream velocity, the velocity of the undisturbed water at some distance from the fish (Blake 1983a). The region between the free-stream velocity and the velocity at the fish is referred to as the boundary layer (Figure 7.1). Flow in the boundary layer can be laminar, resulting in low friction drag, or turbulent, where the resultant eddies form a thicker boundary layer compared to laminar flow and overall friction drag is increased (Webb 1975; Blake 1983a). The change from laminar to turbulent flow is predicted by the Reynolds number, a hydrodynamic measure calculated as

where L = fish length; U = speed; and ν = the kinematic viscosity of water, which is approximately 0.01 cm²s⁻¹ (Webb 1975; Purcell 1977).

Friction drag arises from the viscosity of water in the boundary layer. The greater the surface area of the body, the greater the friction drag. Friction drag also increases exponentially with swimming speed. For laminar flow, the exponent is 1.5, rising to 1.8 for turbulent flow in the boundary layer (Alexander 1967a). Pressure drag is caused by eddies generated along and behind the body by the separation of the boundary layer from the body of the fish (Figure 7.1A). The farther back that boundary separation occurs, the lower the underpressure and the size of the wake. Because turbulent boundary layers separate farther back than laminar boundary layers (Figure 7.1B), the pressure drag resulting from separation of a laminar boundary layer is higher than that for a turbulent one (Blake 1983a). Streamlining also reduces boundary layer separation and thus lowers pressure drag. Other things being equal, pressure drag increases at approximately the square of velocity (Alexander 1967c). Because of how friction and pressure drag are formed, a body shape that reduces friction drag has the opposite effect on pressure drag. Friction drag is related to surface area, so a body shape that minimizes the surface-to-volume ratio, such as a sphere, would have the lowest friction drag. Among freshwater fishes, a more globular shape, such as shown by some sunfishes, would have lower friction drag but a higher pressure drag in contrast to a more elongate, streamlined fish such as a trout, which would have higher friction drag but a lower pressure drag (Alexander 1967c).

FIGURE 7.1. Flow separation around a fish holding position in flowing water.

A. Flow lines, friction and pressure drag, and the boundary layer at a point tangential to the body. The relative thickness of the boundary layer is greatly exaggerated. The length of the arrows indicates the relative velocity of water, ranging from zero in contact with the body of the fish to the free-stream velocity indicated by the arrows of identical length on the right.

B. Changes in flow separation from the body in laminar (dashed lines, black arrows) and turbulent (dotted lines, white arrows) flow. Based on Webb (1975) and Blake (1983a).

Generated Forces

Water flowing over the body and fins of a fish can generate lift because the shapes are acting as hydrofoils—such lift is often referred to as dynamic lift. Bernoulli’s equation predicts that pressure will decrease as the velocity of fluid increases across a surface, so lift for a hydrofoil occurs when flow across the upper surface exceeds that of the lower, resulting in a pressure differential (Webb 1975). Such conditions occur when the angle incidence (α) of the hydrofoil increases from zero (Figure 7.2). The lift generated by a hydrofoil acts normal to the drag force and increases with the angle of incidence up to a point where flow lines begin to separate from the hydrofoil (usually about 15°), resulting in a sudden increase in pressure drag and a sudden decrease in lift so that a stall occurs. Because the amount of lift generated by turbulent flow is greater than for laminar flows, as a consequence of later separation of flow lines as described previously, higher values of lift occur at higher Reynolds numbers (Webb 1975; Blake 1983a).

FIGURE 7.2. Flow lines, lift, drag, and the resultant pressure force at three angles of incidence (α) of a hydrofoil. Drag is parallel to the axis of flow (or motion) while lift is normal to the axis of flow or motion. Based on Webb (1975) and Blake (1983a).

Freshwater fishes occupy a wide range of habitats with a correspondingly high range of current speeds and degrees of turbulence. To maintain hydrodynamic stability, change posture, initiate changes in course, or change location, fishes must control translational and rotational forces. Translational forces refer to movement of a body from one point in space to another without rotation and occur in three planes: surge, slip, and heave (Figure 7.3). Surge refers to movement forward or backward, slip refers to sideways movement, and heave refers to movement up or down. Rotational forces refer to movement around the center of mass and occur along three axes: yaw, pitch, and roll (Figure 7.3). Yaw describes the rotation about the center of mass from side to side, pitch is the rotation up or down, and roll is the rotation along the horizontal axis of the body. Some actions do not result in a change of rotational or translational state because they result in keeping the body in the same location (e.g., hovering) (Alexander 1967c; Webb 2006).

Body Shape, Fin Location, and Maneuverability

Control and maneuverability during hovering or active movement are related closely to fin placement relative to the center of mass, the control of fin rays and fin area by muscles, and swimming speed (Alexander 1967c; Webb 2006). Four zones are recognized relating to fin placement and function (Figure 7.4): (1) an anterior body zone of rudders and lift surfaces positioned anterior to the center of mass that are important in translational forces; (2) a zone of keels located at the center of mass that are particularly important in controlling roll; (3) a zone of stabilizers located immediately posterior to the center of mass and important in controlling yaw, pitch, and roll; and (4) a zone of locomotion and rudders located well posterior to the center of mass that is again important in translational forces (Aleev 1969; Gosline 1971). Anterior control surfaces (zone 1) can include pectoral fins, the head, or the anterior part of the spinous dorsal fin, with the head particularly important in turning motion in elongate body shapes (Webb 2006). A fin, such as the spinous dorsal in zone 1, acts to deflect the fish away from its forward course, but during rapid forward progress in a straight line, it is advantageous for it to be folded down, which also helps to reduce drag. Pectoral fins can also be furled during high swimming speeds (Webb 2006). A single dorsal fin located over the center of mass (zone 2) serves as keel but does not stabilize or deflect the forward course of the body. Many lower teleosts, such as herrings, minnows, suckers, catfish, and trout (groups in the Clupeomorpha, Ostariophysi, and Protacanthopterygii; Figure 7.5), have dorsal fins in this general position or in a position slightly posterior to the center of mass where the fin can also function as a stabilizer (rudder) or aid in propulsion (Figure 7.4B) (Aleev 1969; Gosline 1971). In higher teleosts, such as Moronidae, Centrarchidae, and Percidae (groups in the Acanthomorpha; Figure 7.4A), the dorsal fin consists of two parts, the more anterior spinous dorsal fin and the more posterior soft dorsal fin. The spines can be raised or lowered depending on need. It is important to remember, however, that fins can serve multiple purposes, including camouflage, communication, and in the case of spines, defense.

FIGURE 7.3. Terms used in describing translational (black font and arrows) and rotational (gray font and arrows) changes in state about the center of mass in fishes. Photograph of Colorado Pikeminnow (Ptychocheilus lucius) courtesy of Tom Kennedy. Based on Alexander (1967a) and Webb (2006).

FIGURE 7.4. Potential fin functions relative to the center of gravity in (A) higher teleosts illustrated by the Freckled Darter (Percina lenticula), and (B) lower teleosts illustrated by the Blacktail Shiner (Cyprinella venusta). Based on Aleev (1969) and Gosline (1971).

FIGURE 7.5. Major levels of fish evolution. Names at the base of the cladogram define inclusive groups (e.g., Osteoglossomorpha to Tetraodontiformes are included within the Teleostei). Names at the ends of branches refer to particular lineages. Black text identifies groups that have, or had, representation in North American freshwater habitats. The Sarcopterygii includes lobefin fishes as well as tetrapods. Based on Nelson (2006).

Many freshwater fishes achieve static lift (=buoyant lift) by having air bladders or low-density fatty inclusions within the body cavity so that the mass of water displaced approaches the mass of the fish (Gee 1983). However, because the vertebral column bounds the upper extent of the abdominal cavity, low-density inclusions result in the center of buoyancy being beneath the center of mass (Eidietis et al. 2003). The difference between the center of mass and the center of buoyancy is termed the metacentric height, and a negative value, typical of most fishes, results in a rolling torque (a reason why a recently dead or an incapacitated fish turns belly up). Fishes must use behavioral changes, such as resting on the bottom or leaning against structures, or fin movements, to compensate for this inherent instability. To a certain extent, this rolling torque likely was reduced by the location of the swimbladder dorsal to the gut in actinopterygians compared to the ventral position of the lung (the precursor to the swimbladder) in early bony fishes (the lobefin fishes within the Sarcopterygii) such as lungfishes (Lauder and Liem 1983; Webb 2002). Some actinopterygians also have a more anterior location of gas volume such that the pitching torque generated by the mass of the head skeleton is reduced (Webb 2006).

Types of Locomotion

Fish swimming modes can be divided into those involving the body and caudal fin (BCF) and those using various combinations of paired or median fins for locomotion (MPF) (Blake 2004). BCF locomotion is undulatory, involving alternate waves of contractions on either side of the body, because of sequential innervation of lateral body muscles (serial myomeres) that are three-dimensionally folded and divided into blocks by connective tissue (myosepta) (Danos et al. 2008). Furthermore, BCF swimming can be subdivided into steady, continuous swimming versus unsteady, transient (burst and coast) swimming (Blake 2004). Burst-and-coast propulsion occurs in many pelagic and nektonic fishes, and in fishes with streamlined bodies, it can provide considerable energy savings per distance traveled in contrast to steady swimming (Blake 1983b).

BCF swimming typically is categorized into three to five modes: anguilliform, subcarangiform, carangiform, thunniform, and ostraciiform (Breder 1926; Webb 1975; Lindsey 1978). The modes are named after exemplar species and characterized by increasing concentration of the propulsive force in the caudal fin, although they do not imply phylogenetic relationships (Webb 1975; Blake 2004). The ostraciiform mode has a complete, or nearly complete, absence of body undulation with all propulsive power generated by oscillation of the caudal fin. Because the ostraciiform swimming mode is exemplified by tropical marine box fishes, marine electric rays, and tropical African freshwater elephant fishes (Lindsey 1978; Helfman et al. 2009), and is not represented by any North American freshwater fish group, it will not be discussed further.

The remaining four modes were originally defined by perceived differences in swimming based on morphology and not on hydrodynamic analyses and, among other things, overlooked the three-dimensional geometry of the body during swimming. Recent research indicates that two-dimensional views of dorsal midline profiles of anguilliform, subcarangiform, carangiform, and thunniform modes are essentially indistinguishable, at least during certain swimming speeds. Because of this, the traditional modes of BCF swimming in fishes are not always representative of hydrodynamic differences and lack a functional basis (Blake 2004; Lauder and Tytell 2006). Current research suggests that thunniform and carangiform modes are quite similar in most, although not all, features. Because of the high similarity between the carangiform and thunniform modes (Blake 2004), and because I know of no North American freshwater fish using a thunniform swimming mode, it is not treated further. The remaining three BCF modes are not distinct in all attributes and are grouped differently based on different functional and morphological criteria, including propulsive wavelengths, wake patterns, tendon lengths, and red muscle activity (Table 7.1) (Lauder and Tytell 2006; Danos et al. 2008). Thus, although useful as general shorthand descriptors of BCF swimming, the taxon-named swimming modes are not totally distinct but share various features.

ANGUILLIFORM BCF LOCOMOTION In anguilliform swimming, which is ontogenetically and phylogenetically the basal mode of BCF swimming in ray-finned fishes, the Actinopterygii (Figure 7.5), the entire body is employed to generate thrust through a series of waves moving from head to tail (Gosline 1971). In contrast to early studies indicating that large amplitude undulations occurred all along the body over a range of swimming speeds, recent work indicates that body waves have increasing amplitude posteriorly, thus increasing water displacement toward the tail, and that the anterior body region only shows strong undulation during acceleration and not during steady swimming (Müller et al. 2001; Lauder and Tytell 2006). Fishes using anguilliform swimming are elongate and flexible, such as freshwater eels, lampreys, some catfishes, and the larvae of most fishes (Blake 1983a). In contrast to nonanguilliform swimming, anguilliform swimmers are also generally adept at backward locomotion (Webb 2006).

TABLE 7.1 Similarities and Differences among Commonly Recognized Modes of Body and Caudal Fin (BCF) Locomotion

Anguilliform swimming, at least as shown by eels, does differ from other swimming modes in several ways (Table 7.1). Red muscle activation tends to occur in short blocks ipsilaterally, in contrast to long blocks in the carangiform mode and intermediate blocks in the subcarangiform mode (Danos et al. 2008). One of the original descriptors of swimming modes, the propulsive wavelength adjusted for body length, is still useful, being short in anguilliform swimming, intermediate in subcarangiform modes, and high in carangiform modes (Tytell and Lauder 2004; Danos et al. 2008). Even though it tends to increase posteriorly, wave amplitude is also somewhat greater anteriorly in anguilliform swimming, in contrast to the other modes that are highly similar in this regard (Lauder and Tytell 2006). Wake form differs in anguilliform swimmers, with wakes having lateral momentum but not substantial downstream flow momentum (the momentum opposite the line of thrust of the body), in contrast to other swimming modes. The difference most likely is caused by the absence of a distinct caudal fin structure in eels in contrast to fishes having caudal fins that are distinct from the body (Lauder and Tytell 2006). In five other features, anguilliform and subcarangiform modes do not differ (Table 7.1). These include four features of the myosepta (the sheets of connective tissue separating blocks of myomeres and onto which muscle fibers insert) involving the lateral myoseptal tendon length, the presence of epineural (located on the dorsal surface of the vertebral centrum) and epipleural (located above the abdominal ribs) tendons, and the shape of the myosepta; the fifth similarity is in the firing duration of red muscle fibers (Danos et al. 2008). Red muscle fibers are oxidative and used in slow, prolonged swimming; as such, they are highly vascularized and contain abundant myoglobin, a red oxygen-binding pigment characteristic of muscle (Syme 2006).

LARVAL FISHES AND ANGUILLIFORM LOCOMOTION During their larval period, the majority of all North American freshwater fishes use anguilliform locomotion in the sense of generating more than one complete propulsive wavelength within the length of the body (Webb and Weihs 1986). Anguilliform swimming in larvae occurs because the musculature and axial skeleton are not sufficiently developed to use lift-based subcarangiform or carangiform modes, both of which would place greater compressive force on the axial skeleton and require more muscular power. In addition, because of their small size and speed, larval fishes operate in an environment dominated by viscous rather than inertial forces so that any cessation of swimming movement stops forward progress—there is no coasting in the absence of inertial forces. The balance between viscous and inertial forces is determined by the Reynolds number (Re), the same equation described previously for prediction of laminar versus turbulent flow in a boundary layer. Re<1 indicates a totally viscous environment and Re >1,000 indicates a totally inertial environment; at intermediate values both forces are represented (Lauder and Tytell 2006), but for values of Re below 300–450, viscous forces predominate over inertial forces (Webb and Weihs 1986; Fuiman 2002). It is difficult for us to really imagine life at low Reynolds numbers. Purcell (1977), in discussing swimming in microorganisms and the impact of the primacy of viscous over inertial forces, said: “If you are at [sic] very low Reynolds number, what you are doing at the moment is entirely determined by the forces that are exerted on you at that moment, and by nothing in the past.”

Once the yolk sac is absorbed, larval fishes generally swim at 1–3 body lengths per second (Fuiman 2002). Thus R_e of a 5-mm larval fish would be 25–75, at the lower end of the intermediate range, and subject to viscosity effects. In a viscosity-dominated environment, pushing against the water by an elongate body is more effective than using caudal fin propulsion (Webb and Weihs 1986), but because of the unimportance of inertial forces, larvae must swim continuously to move. (Recall that the law of inertia, or Newton’s first law, states that a particle at rest or moving in a straight line with a constant velocity will continue to do so, provided the particle is not subject to an unbalanced force.) As soon as the larvae stop actively swimming, they come to a halt (Purcell 1977; Blake 1983a). As fishes increase in size, the importance of inertial forces increase relative to viscous forces so that once Re reaches 300–450, they can employ an energy-saving burst-and-glide approach to locomotion (Fuiman 2002).

NON-ANGUILLIFORM BCF LOCOMOTION Increased posterior localization of body undulation and power and the development of distinct caudal fins characterize the traditional modes of subcarangiform and carangiform locomotion (Table 7.1). Subcarangiform swimming occurs in the majority of nonlarval North American freshwater fishes, including salmonids, cyprinids, catostomids, centrarchids, and percids; however, specific studies on swimming are limited to relatively few species of salmonids, cyprinids, and centrarchids (Blake 1983a; Lauder and Tytell 2006). Subcarangiform fishes typically have fairly flexible but low-aspect-ratio caudal fins, such as the fins of many minnows, suckers, catfishes, sunfishes, and darters (Figure 7.4). Aspect ratio expresses the amount of lift generated by a hydrofoil, with lift increasing with aspect ratio, and is determined by the square of fin span divided by fin area. Examples of carangiform swimmers within North American freshwater fishes are less common but potentially include herring and shad (family Clupeidae). Carangiform swimming also is likely approached by two large cyprinid fishes endemic to the Colorado River system, Humpback Chub (Gila cypha) and Bonytail Chub (G. elegans); both have high-aspect-ratio caudal fins and narrow caudal peduncles, although there are no supporting biomechanical or hydrodynamic studies on these species. Consequently, in terms of BCF swimming, the vast majority of North American freshwater fishes occupy the anguilliform-subcarangiform range of swimming modes.

FIGURE 7.6. A. Bluegill (Lepomis macrochirus), with the dotted line showing the outline of the pectoral fin used in labriform locomotion.

B. Gait change and relative metabolic power and cost as a function of swimming speed in Bluegill (mean length 19.5 cm). The gait transition occurs at approximately 1.3 body lengths per second. Based on data from Kendall et al. (2007).

MPF LOCOMOTION Similarly to categorization of BCF swimming modes, Breder (1926) recognized six undulatory modes of MPF locomotion and one oscillatory mode, based principally on the median or paired fins involved, the general appearance of the waveforms, and the length of the fin relative to body length, but not on functional aspects of fin kinematics (Blake 1980, 1983a). In a simplified system, Blake (1983a; 2004) recognized the distinction between undulatory and oscillatory modes and divided undulatory modes into two groups based on fin kinematics. Type I includes fishes using fins with high amplitude, low frequency, and long wavelengths, such as dorsal fin locomotion in the Bowfin (Amia calva). Amiiform locomotion is also advantageous in allowing backward as well as forward locomotion (Webb 2006). Type II includes fishes using fins with low amplitude, high frequency, and short wavelengths, such as pipefishes (Syngnathidae), a primarily marine group but with some species, such as the Gulf Pipefish (Syngnathus scovelli), entering into fresh water. Because thrust is achieved most efficiently by accelerating a large mass of water to a low velocity rather than the reverse, type I locomotion is more efficient than type II (Blake 1983a).

Oscillatory fin locomotion, also referred to as labriform swimming, is shown by fishes using pectoral fins for locomotion, such as mudminnows (Umbra spp.), sticklebacks (Gasterosteidae), and certain centrarchids (Lepomis and Pomoxis) (Figure 7.6; Drucker and Lauder 2000; Walker 2004; Jones et al. 2007). Fishes using MPF locomotion are common in complex habitats such as weedy ponds, lake margins, or streams with abundant submerged or emergent vegetation or woody debris and are adept at backward as well as forward locomotion (Webb 2006).

GAITS, MANEUVERABILITY, AND SPECIALIZATION

Swimming fishes can accommodate a wide range of speeds and maneuverability, with different speeds referred to as “gaits” or “gears.” Not all fishes or all life stages have all gaits, and potentially different structures may be employed at different gaits (Webb 1994, 2006; Blake 2004). Slow swimming in juvenile and adult fishes can occur via median and paired fins (MPF), which are gradually supplemented and then replaced by BCF swimming as speed increases or when increased acceleration is needed (Webb 1994). For instance, Bluegill (Lepomis macrochirus) switch from labriform, pectoral swimming, to undulatory body swimming as speed and power requirements increase (Figure 7.6) (Kendall et al. 2007). Initially, the cost of locomotion decreases while power output increases until approaching a speed of 25 cms⁻¹, corresponding to about 1.3 body lengths per second. Bluegill can achieve swimming speeds above 25 cms⁻¹ by involving increased muscle mass, albeit by increased cost, by changing from labriform to BCF locomotion. MPF locomotion also affords a high degree of maneuverability at a low metabolic cost (Kendall et al. 2007).

Gait changes also can involve different muscle groups, obviously in the change from MPF to BCF swimming, but also within BCF swimming. At low speeds, oxidative red muscle, which is aerobic, supplies the power. Red muscle is slow contracting but capable of maintaining cruising speeds without fatigue. As speed increases, fast oxidative glycolytic (pink) and/or fast glycolytic (white) muscle fibers are added (Webb 1994). White fibers have greater power output than red or pink muscle, are good for sprints and fast starts, but have shorter times to fatigue. White and red fibers do not overlap in use in less derived teleosts, including bowfin, gars, hiodontids, and clupeids; in more derived teleosts having multiple innervation of white fibers, graded responses are possible with overlap in firing of red and white muscles at intermediate swimming speeds (Webb 1994).

Loss of Gaits and Specialization in Water-Column Fishes

Fishes that primarily use BCF cruising or sprinting gaits are adapted to exploit widely distributed food resources and are well represented in the marine environment but less common in freshwater habitats (Webb 1994). Fishes having cruising or sprinting gaits are characterized by high-aspect-ratio caudal fins; relatively stiff fins; large anterior body masses; and rigid, streamlined bodies (Webb 1984). Streamlined body shapes are defined as having fineness ratios >2, where fineness ratio is body length divided by maximum diameter (Blake 1983a). Fishes with specialized accelerator gaits, such as ambush predators, have enhanced fast-start gaits, resting on the bottom or hovering in the water column when not actively pursuing prey. Fishes with specialized accelerator gaits entrain a large amount of water along the body to assure maximum thrust during fast starts. As such, they are characterized by having a large caudal area and a low aspect ratio caudal fin, a deep caudal peduncle, a relatively flexible body, and with the dorsal and anal fins located or extended posteriorly. Fishes using accelerator gaits are adapted to take locally abundant, evasive prey (Webb 1984). Two variants of accelerator gaits are characterized by (1) resistance minimizers, such as esocids, which have a large caudal area and elongated, circular anterior body shape; and (2) thrust maximizers, such as cottids, which have a dorso-ventrally flattened body and large fin depth along most of the body length as well as enlarged pectoral fins that also contribute to fast starts (Webb 1984). Esocids, a group of fishes represented in North America since at least the Cretaceous Period of the late Mesozoic (Chapter 2), have the highest acceleration rates recorded to date (Webb 1994). As a trade-off, MPF gaits usually are poorly developed in accelerators (Webb 1994).

Maneuverers are found in complex structured habitats, such as macrophyte beds, but also occur in open water where their maneuvering ability allows them to position their body to capture small prey items. These fishes, such as sunfishes, have large, flexible MPF and short, deep bodies (Gosline 1971). Not surprisingly, MPF gaits, which excel in low-speed swimming, are emphasized in maneuverers, whereas station-holding gaits and BCF sprinting, cruising, and fast-start gaits may be suppressed (Webb 1994). Prey used by maneuverers tend to be relatively nonevasive (Webb 1984).

Loss of Gaits and Specialization in Substratum Fishes

Burrowers are characterized by having elongate bodies with a loss or reduction of projecting appendages, as seen in larval lampreys (ammocoetes) and in American Eels (Anguilla rostrata). Specialized burrowers are slow swimmers and show the suppression of fast-start, cruising, or sprinting gaits. Some burrowers might use burst-and-coast swimming to compensate for high-speed swimming costs (Webb 1994). Fishes having flow-refuging gaits live on the bottom of swiftly flowing streams and take advantage of the reduced current flow in the boundary layer. Many stream fishes rely on frictional contact with the substratum to minimize the swimming energy required to maintain position. Examples include the cottid form with low lift and high drag (Webb 2006). Flowrefuging fishes offset drag by having structures to increase friction with the stream bottom or to create negative lift, especially with enlarged pectoral fins as in catostomids, salmonids, darters, and cottids. Such frictional forces are not high, but because of similarities in densities of fish and water, they can be important. As current speed increases, many lotic fishes, such as Longnose Dace (Rhinichthys cataractae), Logperch (Percina caprodes), and River Darter (Percina shumardi) increase their frictional contact with the substratum by releasing gases from their swimbladders and thus increasing their density (Gee 1983). High drag–low lift cottiform fishes suppress MPF and BCF sustainable gaits (Webb 1994).

In addition to morphological and physiological features associated with flow-refuging gaits, a number of North American freshwater fishes have distinctive behaviors that are initiated as flow increases and fishes risk downstream displacement. Oral grasping, literally grabbing a stationary object with the mouth and hanging on to remain stationary in flowing water, has been documented in 10 genera and 19 species of minnows, but these observations have been in a laboratory stream environment and not in nature (Table 7.2). Oral suction for the purpose of remaining stationary, closely related to oral grasping, is known for two species of suckers, and the Bluehead Sucker (Catostomus discobolus) has been observed using oral suction on rocks to maintain position in a natural stream (Table 7.2). Studies investigating whether there is oral grasping and suction in minnows and suckers for this purpose, rather than for feeding, have also included species in the families Characidae, Ictaluridae, Fundulidae, Poeciliidae, Atherinopsidae, Centrarchidae, Percidae, and Cichlidae, none of which showed this kind of oral grasping or suction (Leavy and Bonner 2009).

Another flow-refuging behavior is referred to as “parr posture” (Arnold et al. 1991). In this behavior, fishes orient into the current, supporting themselves on the tips of their pectoral fins, with the center of mass located in the middle of a triangle formed by the anterior support of the two pectoral fins and the posterior support of the pelvic fins or body (Webb 2006). This behavior allows fishes such as cyprinids (Ward et al. 2003), salmonids (Arnold et al. 1991), and cottids (Webb et al. 1996) to maintain position in a current, where drifting prey are likely available, at a low metabolic cost. At increased current speeds, fishes lower themselves onto the bottom, lower the dorsal fin, and use the downward force of water on the angled pectoral fins to maintain position. A similar behavior, sometimes accompanied by arching of the back as current flow increases, has been documented in darters (Etheostoma and Percina) (Matthews 1985a).

TABLE 7.2 Oral Grasping and Oral Suction as Station-Holding Behaviors in North American Freshwater Fishes

Another pattern of station holding is to minimize drag by having a flattened (depressed) body form such as flounders, soles, and rays. This pattern is common in areas where there is insufficient surface roughness to increase friction. However, space for organ systems results in a “blister-shaped” dorsal profile and relatively large surface area, both of which result in increased lift (Webb 2006). When the body is appressed to the substratum, with the marginal fins lowered to be in contact with the bottom, flow over the eyed side results in lift because of a pressure reduction relative to the stagnation pressure (the pressure in the absence of flow) on the blind side. As flow increases, pressure on the blind side can be reduced by beating of the posterior parts of the dorsal and anal fins, resulting in a substantial, posteriorly directed flow under the body. Because pressure is reduced as flow increases, this behavior helps to equalize pressure between the eyed and blind sides, thus reducing the amount of lift and providing increased station-holding ability (Arnold and Weihs 1978; Webb 1989). Flatfishes also can bury into a soft substratum (Brainerd et al. 1997), a behavior likely important in crypsis as well as for reducing the risk of downstream displacement from currents. The Hogchoker (Achiridae, Trinectes maculatus), a small flatfish that occurs in coastal waters and rivers along the Atlantic and Gulf coasts (Burgess 1980), typically is found over nonvegetated, mud or sandy areas in rivers and illustrates this pattern (Ross 2001).

Finally, many fishes are gait generalists, using MPF and BCF gaits, both with welldeveloped red and white muscles. Generalists, such as Largemouth Bass (Micropterus salmoides), typically have good hovering and stationholding gaits, as well as fast-start, sprint, burstand-coast, and cruising gaits (Webb 1994). In addition to Black Basses (genus Micropterus), generalists include many groups that are represented in North American fresh water, such as salmonids, cyprinids, and ictalurids (Webb 1984). Burst-and-coast swimming can occur in a variety of gaits besides generalists.

EVOLUTIONARY TRENDS IN FORM AND FUNCTION

Anguilliform locomotion, which is used by jawless fishes such as lampreys and more derived freshwater eels, does not require as much resistance to compression as other types of BCF locomotion. Consequently, the evolution of nonanguilliform BCF swimming occurred in concert with selection for a vertebral column capable of resisting longitudinal compression by opposing muscles. The vertebral column consists of a series of rigid blocks separated by joints so that bending is permitted but not compression. High flexibility that is important to anguilliform locomotion becomes a hindrance in subcarangiform and carangiform locomotion, and one way to reduce flexibility is to decrease the number of vertebrae. For instance, American Eels have more than 100 vertebrae, and lampreys (Petromyzontidae) normally have between 46 and 74 myomeres (retaining a notochord and lacking ossified vertebral elements) (Scott and Crossman 1973; Ross 2001). In contrast, the number of vertebrae has been reduced to no more than 24 in more derived teleosts (Gosline 1971).

Placoderms, an extinct group of early fishes that are the sister group to all other gnathostomes (Figure 7.5), dominated the benthic environment of Devonian seas, with some groups likely entering freshwater habitats. These were generally heavily armored fishes with strongly dorso-ventrally compressed bodies and long tails (Moy-Thomas and Miles 1971; Janvier 1997b). Placoderms likely used BCF locomotion, but because heavy body armor would enhance station holding but detract from acceleration, a fast-start gait was likely absent (Webb 1994). The Acanthodii, another now extinct group of Paleozoic fishes that lived from the late Ordovician to early Permian periods, were mostly pelagic (Nelson 2006). Although initially marine, by the Devonian they primarily occurred in freshwater habitats. Fossil Acanthodii have been recovered from Laurasian shale deposits, including some in what is now eastern North America (Bardack and Zangerl 1968). Because they were primarily pelagic and lacked heavy armor, they likely were the first fishes to possess a fast-start gait (Webb 1994). In general, the performance of Paleozoic fishes was probably not high compared to modern fishes. Paleozoic fishes lacked vertebral strengthening and median fins, and appendages generally lacked mobility so that they had more of a role in trim than in propulsion (Webb 1994). An important aspect in the evolution of increasing maneuverability in fishes has been the reduction in the size of the fin base, thus increasing the extent of possible fin movement (Webb 2006).

Beginning in the Carboniferous, major advances in fish locomotion and control occurred within the Actinopterygii, the ray-finned fishes (Figure 7.5). MPF gaits likely arose when fins, which were initially used for trim, acquired independent movement. Over time, actinopterygian gaits improved with increases in strength of axial and appendicular skeletons and increased flexibility of paired fins (Webb 1994). Caudal fins of many teleosts are capable of extensive changes in shape, providing fine control over forces generated during swimming (Lauder and Tytell 2006), and such changes are possible through the evolution of the caudal skeleton and associated musculature.

FIGURE 7.7. Patterns in the evolution of caudal fin shape and control in the ray-finned fishes (Actinopterygii). The heavy white line indicates the approximate position of the hypochordal longitudinalis muscles (HL), which are involved in stiffening the upper lobe of the caudal fin. Based on Lauder (1989).

The primary support of the teleostean caudal fin is the terminal portion of the upturned vertebral column. Extending below and behind the upturned vertebral column are flat bony plates (the hypural bones) that support the fin rays (Box 7.1). The development of caudal fin locomotion resulted in most of the thrust being concentrated on one or several vertebrae at the base of the tail, and the evolution of nonanguilliform BCF swimming has been accompanied by a reduction in the number of vertebrae in the upturned part of the caudal fin to a maximum of two, and fusion of vertebrae within the caudal fin to increase resistance to compression (Gosline 1971, 1997). In the Chondrostei, a group of early ray-finned fishes that includes living forms such as sturgeon and paddlefish, the notochord and vertebral elements extend well into the upper fin lobe, and fin rays and other supporting elements are connected to the ventral side of the vertebral column (i.e., hypaxial in position) (Figure 7.7). Because the caudal fin lacks intrinsic muscles, the ability to change fin shape or stiffness is limited only to effects of trunk musculature on the caudal fin (Lauder 1989). In gars (Lepisosteiformes), the caudal fin is still supported only by hypaxial elements, but there are muscles within the caudal fin, including the hypochordal longitudinalis muscle (HL), that allow the beginning of independent control of caudal fin rays (Figure 7.7). The HL originates on caudal vertebrae, or on the ventral hypural bones in teleosts, and extends posteriorly to insert on the caudal fin rays. It functions first in stiffening the upper caudal lobe so that it is less inclined to the horizontal during swimming (Lauder 1989).

BOX 7.1 • Teleostean Caudal Skeleton

Major advances in fish locomotion are linked to the evolution of the caudal fin. The following figure illustrates key elements of a teleostean caudal skeleton as shown by a mesolarval River Carpsucker (Carpiodes carpio) from the Rio Grande. The hypural bones are the broad plates that provide support for the fin rays. The upturned portion of the vertebral column is generally greater in larval fishes, becoming reduced and differentiated during development into preural (proximal) and ural (distal) centra.

Caudal fin morphology of a larval River Carpsucker (Carpiodes carpio). Photo courtesy of Trevor Krabbenhoft.

Bowfin (Amiiformes) also have only hypaxial fin support but show more development of intrinsic caudal musculature, including the first occurrence, in the upper part of the caudal fin, of interradialis muscles, which go between adjacent caudal rays. The HL is increasingly differentiated and developed. Because of the development of intrinsic caudal musculature, the amiiform lineage shows the first ability to modulate caudal fin shape (Lauder 1989).

The development of the homocercal caudal fin occurs in teleosts (Figure 7.7). For the first time, the caudal fin is supported by epaxial as well as hypaxial elements. The epaxial support is provided by elongated neural arches, termed uroneurals, from the caudal vertebrae. In addition, there is a full complement of intrinsic caudal musculature and the HL originates from ventral hypural bones rather than the caudal vertebrae (Lauder 1989). These changes allow for much greater control over fin shape and stiffness compared to less derived groups. In modern teleosts such as Lepomis, tail thrust depends to great extent on caudal muscle activity, particularly the HL. At low to moderate constant swimming speeds, the HL is not contracted so that the upper lobe of the caudal fin provides less thrust than the lower lobe. However, during rapid acceleration, the HL is active and serves to stiffen the upper caudal lobe so that the caudal fin generates much more symmetrical thrust (Lauder 1989).

In lower teleosts, the pectoral fins are low on the body, the pelvic fins are posteroventral, and the pectoral fin base tends to be more horizontal (Figure 7.4B). In the more derived spiny-rayed teleosts (the Acanthomorpha) (Figure 7.5), the pelvic fins are anterior and ventrolateral rather than posterior, and the pectoral fins are located higher on the body, with the fin base shifted to a more vertical position (Figs. 7.4A, 7.6A) (Rosen 1982). The increased body depth and expansion of the median fins in certain groups, such as sunfishes, results in the placement of MPF propulsors around the center of mass. Such placement likely provides increased stability for hovering, slow swimming, and maneuvering (Webb 2006). However, the pattern in lower teleosts provides increased resistance to rolling, with the ventral position of pectoral and pelvic fins allowing increased stability through contact with the bottom. For instance, in a study of rolling stability in Creek Chub (Semotilus atromaculatus), Largemouth Bass, and Bluegill, Creek Chub showed significantly greater resistance to rolling compared to the two centrarchids (Eidietis et al. 2003). Creek Chub were particularly adept at gaining support from the ground using parr posturing, as described in the previous section. Furthermore, recent studies of pectoral fin function in salmonids have dispelled the idea that the more ventrally placed pectoral fin position of lower teleosts limits its functional importance. The salmonid pectoral fin shows a wide range of kinematic function in hovering, swimming, and braking. In slow-speed turning using pectoral fins, salmonids achieved the same angular velocities as Bluegill (a higher teleost) performing the same turning motion (Drucker and Lauder 2003). Thus changes in fin position and body morphology within the teleosts should be viewed not as one approach being superior to another, but instead as different solutions to the common problems of survival and reproduction (Rosen 1982). Certainly, the lower teleosts such as the Ostariophysi (Figure 7.5) are among the most speciose of all fish groups with close to 8,000 recognized species (Nelson 2006).

NATURAL SELECTION, PHENOTYPIC PLASTICITY, BODY FORM, AND FUNCTION

North American freshwater fishes exhibit a wide diversity of body forms, with variation in forms partitioned among orders and families, but also among and within species. As I have shown in previous sections of this chapter, ecological functions of varying morphologies, such as caudal fin shape and body shape, can be related to function, although some relationships have been more thoroughly studied than others. For instance, minnows having more elongate dorsal fins, wide caudal fin spans, and streamlined bodies showed greater swimming speed, as tested in a laboratory swim tunnel, and minnows with greater swimming speeds occur in habitats having higher current velocities (Leavy and Bonner 2009).

In general, fishes with shorter, deeper bodies tend to be associated with structurally complex habitats compared to those with longer, shallower bodies. A number of freshwater fish groups show polymorphisms in body form that are likely related to specific habitats (Baker et al. 2005). Examples include the endemic fish species of Lake Waccamaw, sunfishes of northeastern lakes, and sticklebacks from lakes of British Columbia and Alaska, with all the examples representing divergences since the middle to late Pleistocene.

Lake Waccamaw

Lake Waccamaw, North Carolina, a large (3,618 ha) Carolina bay lake of late Pleistocene origin, has a known fish fauna of at least 41 species, including three endemic forms (Shute et al. 1981; Menhinick 1991), with all colonization of the lake from adjoining streams (Hubbs and Raney 1946). The lake is generally clear and offers only sparse cover and a clean sand bottom (Krabbenhoft et al. 2009). The three endemic fishes in the lake, Waccamaw Silverside (Menidia extensa; Atheriniformes), Waccamaw Killifish (Fundulus waccamensis; Cyprinodontiformes), and Waccamaw Darter (Etheostoma perlongum; Perciformes) are sister species to the stream-dwelling Inland Silverside (Menidia beryllina), Banded Killifish (Fundulus diaphanus), and Tessellated Darter (Etheostoma olmstedi), respectively (Hubbs and Raney 1946; Krabbenhoft et al. 2009). Although representing three highly divergent lineages, the three endemic fish species have evolved similar shape differences from their stream ancestors, all having longer and shallower caudal peduncles and more slender bodies in comparison with their ancestral stream forms; the Waccamaw Silverside and Waccamaw Killifish also have significantly shorter and more slender heads than their stream ancestors (Figure 7.8). The local streams show much greater physical heterogeneity compared to the lake. Lake forms are likely adapted to a less complex habitat and more sustained swimming in open water in contrast to stream forms; even the benthic darter shows morphological shifts toward greater use of the water column (Krabbenhoft et al. 2009; Krabbenhoft 2010, pers. comm.).

Sticklebacks

As the Cordilleran ice sheet began its retreat from northwestern North America over the last 12,000–20,000 years (see Chapter 3), with the concomitant rebound of the land, newly formed lakes were independently colonized by the marine anadromous form of Threespine Stickleback (Gasterosteus aculeatus). This has provided a replicated system for studying stickleback evolution (McPhail 1994; S. A. Foster et al. 2003; Foster and Baker 2004). Arising from one or perhaps two original founder populations (Taylor and McPhail 1999; Wund et al. 2008), Threespine Sticklebacks found in lakes of British Columbia and Alaska often show distinct body forms associated with different lake habitats. The divergent body shapes are most likely related to variation in selection between highly structured, benthic habitats versus open water (limnetic) habitats. Benthic forms tend to be larger bodied, have fewer and shorter gill rakers, wider mouth gapes, and greater body depths in contrast to limnetic forms, which also tend to have larger, protruding eyes (Bentzen and McPhail 1984; Schluter and McPhail 1992; Rundle et al. 2000; Baker et al. 2005). Allopatric benthic and limnetic ecotypes occur in thousands of lakes, where developmental plasticity at least plays a role in shaping morphology; sympatric benthic and limnetic species pairs have evolved in at least six lakes; and partially differentiated benthic and limnetic ecotypes are known from two lakes (Schluter 2001; Baker et al. 2005; Wund et al. 2008).

FIGURE 7.8. Outlines of body shapes of the three endemic species from Lake Waccamaw, North Carolina (gray), and their closest stream relatives (black). Shapes are drawn from deformation grids of mean body shape for the six taxa, with differences in shape magnified three times. The diagonal lines (white = stream, black = lake) indicate the approximate posterior margin of the gill cover. Note the more dorsally oriented head in the Waccamaw Darter compared to the Tessellated Darter. Based on Krabbenhoft et al. (2009).

Sunfishes

A similar pattern of benthic and pelagic forms is apparent with sunfishes (Lepomis spp.), with colonization of the lakes likely occurring following the beginning retreat of the Laurentide ice sheet approximately 17,000 years ago (Chapter 3). In the Adirondack region of the northeastern United States, Bluegill and Pumpkinseed (Lepomis gibbosus) often occur together in lakes, with Bluegill occupying the water column and feeding primarily on planktonic prey, and Pumpkinseed occupying the benthic, littoral zone and feeding mostly on benthic invertebrates (Robinson et al. 1993, 2000). However, in lakes that were never colonized by Bluegill, the normal bottom-associated Pumpkinseed may also have a planktonic form that occurs in the water column and feeds primarily on planktonic prey such as cladocerans. Morphological divergence between the open- and shallow-water forms varies among lakes but is generally well developed (Robinson et al. 2000). Morphologically, the planktonic form of Pumpkinseed tends to approach the morphology of typical Bluegill, with shallower and longer bodies and longer pectoral fins. Robinson et al. (1993) suggest that the most likely explanation for the development of the planktonic form of Pumpkinseed is the exploitation of planktonic prey made available by the absence of Bluegill, although Robinson et al. (2000) showed that diffuse competition from other planktivorous fishes can also reduce the extent of morphological divergence.

Trade-Offs in Form and Function

Throughout this chapter the diversity of body forms in North American freshwater fishes, related to habitat use and behavior, suggests that some shapes are more efficient (e.g., in locomotion, predator avoidance, or prey capture) in some habitats compared to others. The examples of changes in body form of fishes in structurally simple versus complex habitats indicate that natural selection favors certain forms over others in particular habitats. However, there are few studies that have directly measured trade-offs in efficiency of different body forms in one type of habitat versus another. Indeed, well-designed studies of trade-offs of specializations in different habitats on any animal group have been relatively rare (Futuyma and Moreno 1988).

Trade-offs in foraging of sympatric benthic and limnetic “species” of Threespine Sticklebacks were investigated by Bentzen and McPhail (1984), who used laboratory-reared progeny of benthic and limnetic forms from Enos Lake, Vancouver Island, British Columbia. Benthic fish consumed larger prey relative to body size and were twice as efficient in foraging on a detrital substratum compared to limnetic fish. Both male and female benthic forms averaged 0.44 prey per feeding strike in contrast to male limnetic fish who averaged only 0.2 prey per strike. (Female limnetic forms would not feed on benthic prey.) In contrast, limnetic fish showed greater success in capturing small zooplankton, perhaps because of their larger, more protruding eyes. At the end of a five-day experiment where sticklebacks were introduced into mesh enclosures suspended in the lake and allowed access to a natural array of plankton, limnetic fish averaged 103 prey in the stomach, in contrast to approximately two prey in stomachs of benthic fish.

In a similar study, Robinson (2000) used benthic and limnetic forms of Threespine Stickleback to first ask whether morphological differences had a genetic basis and then to ask if the two forms differed in foraging efficiency between benthic and limnetic habitats. Fish were taken from limnetic and from shallow water, benthic habitats in Cranby Lake, Texada Island, British Columbia, which contains one of two known populations polytypic for body shape. The morphologies differ along a continuum from benthic to limnetic shapes (Baker et al. 2005). To determine if there was a genetic basis for the polymorphism, Robinson (2000) used laboratory-reared progeny from 30 crosses of benthic × benthic and 30 crosses of limnetic × limnetic morphotypes. Progeny of the benthic and limnetic fish retained their shape differences, even after being reared in a common environment, thus indicating at least a partial genetic basis for the differences. Second, laboratory-reared progeny were tested for feeding rate (number of prey consumed per unit time) and feeding efficiency (number of bites per prey item). The hypothesis that there would be trade-offs in foraging relative to morphology was supported. Limnetic forms were superior to benthic forms in the capture of planktonic prey, and benthic forms were superior to limnetic forms in capturing benthic prey—even though the shape differences were only on the order of 4–10%. When feeding on plankton, on average, benthic forms were less efficient, taking approximately twice the number of bites per prey item and 0.8 fewer prey per minute compared to limnetic forms. When feeding on benthic prey, limnetic forms had lowered efficiency, averaging 1.4 more bites per prey and 0.22 fewer prey per minute than benthic forms. The studies by Bentzen and McPhail (1984) and Robinson (2000) clearly illustrate that relatively minor shape differences, as well as other morphological correlates such as gill-raker length and number, mouth size, and eye shape, are significantly related to functional correlates—in these cases feeding rates and efficiencies.

Another, rather stringent test of functional trade-offs in Pumpkinseed focused on fish from Paradox Lake, New York. Fish in this lake are unimodal in the distribution of body shape, and shape differences are so subtle that comparison of the extremes of the distribution are difficult to separate by eye, instead requiring a multivariate shape analysis (Robinson et al. 1996; Robinson and Wilson 1996). Body shapes of fish from the pelagic habitat are more fusiform compared to those from the benthic habitat. Pumpkinseed were collected from pelagic and benthic habitats, and for each habitat, 10 fish were selected from the extremes of the morphological shape distribution (i.e., for the pelagic habitat, the 10 least fusiform and the 10 most fusiform body shapes; Robinson et al. 1996). For fish from each habitat, Robinson et al. determined body condition, measured as the percentage of lipids in muscle tissue. They also determined the relative growth rate, expressed as the standard length at a given age, with age determined from scale annuli. In fish from the pelagic habitat, those most extreme for a limnetic (fusiform) shape had higher condition factors and faster growth than those least extreme in limnetic shape. Fish from the benthic habitat did not show significant relationships in growth or condition factor nor in the multivariate index of body shape. In spite of this, the results are important because they illustrate functional differences in a case where measured morphological differences are very slight (Robinson et al. 1996).

These examples of trade-offs in function from different body forms illustrate how apparently minor changes in shape can have functional consequences. However, the linkage of form with function can be challenging because such relationships often are not linear, they can be dependent upon a particular context, and different structures sometimes can have equivalent functions. Because a large suite of morphological characters might govern a particular function, the choice of what is chosen to measure becomes critical to the outcome (Koehl 1996). For instance, in the previous example with Pumpkinseed, the studied relationship of body shape of pelagic fish to apparent foraging success might also be impacted by such things as gill-raker shape and morphology, head dimensions, and eye position, which were not studied.

DOES MORPHOLOGY PREDICT ECOLOGY?

This section takes a somewhat different approach to the relationship between morphology and function and considers whether ecological functions such as habitat and food use can be accurately predicted from morphology. To be able to do so would allow basic ecological questions of overlap in resource use, niche breadth, and species packing in assemblages (see Chapter 11) to be answered by studies of fish shapes, such as from museum specimens, rather than primarily through actual field or laboratory observations. The latter approach rapidly becomes logistically challenging as the number of potential resource dimensions and the number of species increase. In addition, short-term studies of resource use can be misleading because of natural variation, whereas morphology tends to average out such short-term environmental variation and, consequently, might be more useful as an indicator of overall ecological patterns (Ricklefs and Travis 1980; Douglas 1987). The underlying premise, then, is that morphological “space” can be directly mapped onto ecological “space,” where morphological and ecological spaces represent multidimensional hyperspace as determined via multivariate statistical analyses (see Chapter 4) (Strauss 1987). If there were not a strong relationship between morphology and ecology, then fish assemblages might not be structured or constrained by physical adaptations but perhaps by other processes (Douglas and Matthews 1992).

For a long time, morphological studies have been used to infer ecological relationships of fishes as well as other organisms (e.g., fishes: Keast and Webb 1966, Gatz 1979a, b; lizards: Collette 1961; birds: Van Valen 1965; bats: Findley 1976, with Findley [1976] being one of the earliest studies to use a multivariate, morphometric approach). The history of ecomorphological studies shows a mix of approaches and assumptions, with studies attempting to verify morphological and ecological relationships (generally rare) interspersed with those assuming strong relationships between morphology and ecology, and then using morphological studies to address ecological questions (generally more common).

Tests of the Ecomorphological Hypothesis

Studies of relationships between morphology and ecology show a progression of approaches, from providing only verbal descriptions of morphological and ecological relationships (e.g., Keast and Webb 1966), to using basic correlation analysis and limited use of factor analysis to summarize morphological differences (Gatz 1979b), to using sophisticated multivariate analyses capable of statistically testing the relationships between morphology and ecology (Douglas and Matthews 1992). In a series of papers, Gatz (1979a, 1979b, 1981) provided the first studies of freshwater fishes in which the relationships between morphology and ecology were tested statistically. He used 56 morphological features that included superficial body shape, pigmentation, mouth form, fin form and placement, gill-raker number and shape, swimbladder length and volume, digestive tract morphology, and brain structure. Ecological characters focused on broad-based dietary categories determined from stomach content analysis, with only limited information on habitat. Using information from the literature, for each morphological character he developed assumed functional predictions. The predictions were then tested against other morphological data, or with the ecological data. For instance, orientation of the mouth was assumed to indicate the location, relative to the fish, where feeding occurred. Hence a ventral mouth should indicate primarily benthic prey in the gut. Tests against prey categories known to occur on the surface versus those near or on the bottom showed general support for the morphological prediction. Using this approach, Gatz found that dietary hypotheses of species based on morphological features, even among morphologically similar species, were supported 90% of the time by actual dietary studies. Although this work provided important information, the prey categories were quite broad and ecological features were not, with a few exceptions, tested directly against morphological features but instead against hypotheses of function developed from the morphological characters.

Building on the studies by Gatz, Felley (1984) developed two suites of morphological characters that were then tested against detailed habitat information on a group of 43 minnow species from locations primarily in Oklahoma. The first morphological character group (a priori characters) was assumed to be related to habitat use, and characters were chosen primarily based on Gatz (1979a). The second group (a posteriori characters) was determined by statistically testing associations between morphology and habitat use by factor analysis and then choosing morphological measurements that were significantly related. These were then tested against habitat use in a different group of fishes than those used to identify what morphological features were important; however, the two groups differed in generic composition rather than being a randomly chosen group. No characters in the a priori group successfully predicted habitat use, illustrating the challenges in making generalizations of habitat use from one group of species to another. Among the a posteriori characters, the location of food source and the position in the water column were successfully predicted from morphology. Benthic species had morphologies associated with greater maneuverability, such as deeper bodies, dorsal fins near the center of mass, and longer fins, in contrast to more open-water species—an outcome similar to the previously mentioned studies on benthic and limnetic forms of fishes. The nature of the food source, measured as the relative use of organic detritus, also was predicted successfully from morphology. Species using larger amounts of detritus had longer intestines, greater amounts of pigmentation in the body cavity, and reduced size of the cerebellum. Overall, the Felley’s study, with both successes and failures in the prediction of habitat and food use from morphology, suggests caution in choosing certain aspects of an organism’s morphology and trying to relate it to function. As Felley (1984) cautioned, “We must consider the whole of a species’ morphology to properly identify its place in its ecosystem.”

One of the first studies to use multivariate analyses to rigorously test the association between morphology and ecology focused on an assemblage of southeastern fishes (Douglas and Matthews 1992). The study was based on 65 collections at 11 sites in the Roanoke River drainage of Virginia along a gradient from headwaters to fifth order streams. The analysis included most major groups of North American freshwater fishes—minnows, darters, suckers, madtom catfishes, and sculpins. Ecological data included food habits, based on analysis of stomach contents to major taxonomic levels, and microhabitat use (current speed, substratum size, and water depth). The shape data were based on 34 morphological characters and were analyzed in three different ways, resulting in three shape matrices. For each data set (food, habitat, and morphology), species were grouped by unweighted pairgroup with arithmetic mean (UPGMA) cluster analysis based on a measure of dissimilarity (i.e., a distance measure) (see Chapter 4), and the matrices were compared using the Mantel test. The Mantel test determines whether one matrix is correlated with another by testing against the null hypothesis of randomly arranged elements in the second matrix. The randomization is repeated many times and if the observed value of the correlation of the two matrices falls within the cluster of correlations based on the randomization, the two matrices are not related; if the observed value falls above or below the correlations based on randomization, then the two matrices are either positively or negatively correlated (Manly 1986).

The results showed that only the food use, morphological shape, and fish taxonomy matrices were significantly correlated. Although food use and body shape covaried, beyond the family level, taxonomy explained most of the variation in both the shape and ecology data sets. Consequently, the relationship between body shape and food use was completely overshadowed by the relationship between body shape and taxonomy. Because of the strong presence of phylogenetic history in structuring resource use of fish assemblages, morphological data may not be very useful in describing the structure of an assemblage any further than is already described by taxonomic data. Douglas and Matthews concluded “that taxonomy and trophic ecology are inextricably intertwined, and that resource use in the assemblage reflects [a] phylogenetic artifact.” To control for this, they repeated their analyses for a single family, the Cyprinidae. In this case, body shape was not significantly related to food use but was significantly related to habitat use, suggesting a greater importance of resource separation of freshwater fishes along habitat compared to food dimensions (Douglas and Matthews 1992). Freshwater fishes certainly differ along habitat dimensions; however, in a review of resource partitioning in fishes, Ross (1986) showed that both marine and freshwater fishes tended to separate more along food than habitat resources, and freshwater fish assemblages separated about equally along trophic and habitat dimensions. More likely, the difference in the relationship of morphology to food use in cyprinids was an artifact of the level of prey identification (see Chapter 11).

In a more limited study (eight cyprinids, three centrarchids, and four percids), Wood and Bain (1995) showed that body shape, expressed in multivariate space, was significantly correlated with habitat use (also expressed in multivariate space) in minnows and darters but not sunfishes. The lack of an association between sunfish morphology and habitat use was perhaps because of the few sunfish species, the choice of morphological measurements, or high variability in one of the morphological measurements.

Studies Assuming Validity of the Ecomorphological Hypothesis

Working on the assumption of a close relationship between morphology and ecology, Strauss (1987) used morphological analyses to compare species packing among seven fish assemblages from North and South America. He found that the separation of species in morphological space was not related to the number of species in an assemblage, so that species differences in morphology (and assumed ecology) remained similar across a broad range of fish assemblages. At the same time, Douglas (1987) found that minnow species were more tightly packed in morphological space in contrast to species of sunfishes. He also showed that as communities become more diverse, the amount of morphological space increases, suggesting that increased species packing is not because of smaller morphological distance between species. However, the increase in morphological space was essentially because of the increase in taxonomic diversity—morphological similarity was highly correlated with taxonomic similarity, which is not surprising given that taxonomies are largely based on morphology. Because of the confounding factor of phylogenetic relationships, Douglas (1987) concluded that comparisons of morphological space are more meaningful within than among major lineages.

On a much broader geographical scale, Winemiller (1991) compared packing in morphological space (based on 30 morphological characters) among 10 lowland fish assemblages from five major geographical regions, including North, Central, and South America and Africa. Key research questions included whether fish assemblages in faunistically richer areas showed greater morphological diversification and whether species in such areas showed greater or lesser similarity in morphology in contrast to assemblages from high latitude, species-poor areas. Tropical fish assemblages showed greater morphological diversity, which, assuming a close association between morphology and ecology, suggests greater niche diversity in such areas. However, species packing, inferred from morphological distance among species, did not differ among the 10 assemblages.

SUMMARY

Fishes primarily use their body and caudal fins (BCF) or median and paired fins (MPF) in locomotion, in some cases switching from MPF to BCF as the demand for speed relative to maneuverability increases. Within BCF locomotion, freshwater fishes vary in the amount of the propulsive wavelength that is generated within the body, ranging from anguilliform to carangiform as the caudal fin is increasingly involved and body undulations are reduced. However, almost all fishes use anguilliform locomotion during the larval period. In addition to providing propulsion, fins are used in a variety of ways to control stability, position in the water column, turning, and station holding. A major trend in the evolution of fishes has been the increased control of fin shape and function, especially with the caudal fin. Higher and lower teleosts differ in fin number and position, but differences generally are related to equally successful, albeit different, ways of dealing with environmental challenges. Fin positions, body shapes, and mode of swimming are generally related to water currents, habitat complexity, and prey use. Fishes in structurally complex habitats tend to have larger fins and deeper bodies in contrast to fishes in open-water habitats, and even slight differences in body form provide habitat-specific advantages in feeding. Although morphology is strongly related to ecological function, direct predictions of ecological function from morphology have proven difficult to make. In part this reflects the multiple roles that structures might have and the diverse ways that adaptation has responded to selective pressures within habitats.

SUPPLEMENTAL READING

Blake, R. W. 2004. Review paper: Fish functional design and swimming performance. Journal of Fish Biology 65:1193–222. An overview of functional morphology and fish locomotion.

Douglas, M. E., and W. J. Matthews. 1992. Does morphology predict ecology? Hypothesis testing within a freshwater stream fish assemblage. Oikos 65:213–24. One of the first multivariate analyses of the relationship between ecology and morphology.

Lauder, G. V., and E. D. Tytell. 2006. Hydrodynamics of undulatory propulsion, 425–68. In Fish Biomechanics. Fish Physiology. Vol. 23. R. E. Shadwick, and G. V. Lauder (eds.). Elsevier Academic Press, San Diego, California. Explains the relationship of water currents and thrust generated from undulatory locomotion in fishes.

Webb, P. W. 2006. Stability and maneuverability, 281–332. In Fish Biomechanics. Fish Physiology. Vol. 23. R. E. Shadwick, and G. V. Lauder (eds.). Elsevier Academic Press, San Diego, California. Explains the relationship between the degree of maneuverability in fishes and their hydrodynamic stability.

WEB SOURCES

The Tree of Life Web Project. http://tolweb.org.